diff options
Diffstat (limited to 'target/linux/realtek/files/drivers/net/wireless/rtl8192cd/change.txt')
-rw-r--r-- | target/linux/realtek/files/drivers/net/wireless/rtl8192cd/change.txt | 13333 |
1 files changed, 13333 insertions, 0 deletions
diff --git a/target/linux/realtek/files/drivers/net/wireless/rtl8192cd/change.txt b/target/linux/realtek/files/drivers/net/wireless/rtl8192cd/change.txt new file mode 100644 index 000000000..797d6ee7f --- /dev/null +++ b/target/linux/realtek/files/drivers/net/wireless/rtl8192cd/change.txt @@ -0,0 +1,13333 @@ +Change History +============== + +* DATE/WHOM/TYPE: +- 2012/10/02, button, feature refine + +* DIGEST: +- Refine for PSP IOT + +* TEST METHOD: +- Configure DUT to TKIP+AES, Let Sony PSP do site survery, +- check probility of DUT showing in PSP's list + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_headers.h + +* DESCRIPTION: +- Exclude HT and WMM IE of probe resonse which respond to 11b STA + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/26, jimmylin, feature refine + +* DIGEST: +- Disable PCIe clock shift adjustment + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/18, jimmylin, feature refine + +* DIGEST: +- Enable TX 2-path by default + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_hw.c + +* DESCRIPTION: +- If CCK B path has no calibration, use CCK A path and difference of MCS to + get CCK B path. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/14, weikaichuang, feature refine + +* DIGEST: +- Update 92d fw v.36 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +-/92d_data/rtl8192dfw_n.bin + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/14, chuangsw, feature refine + +* DIGEST: +- Refine 88E edca for wifi logo wmm test. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c +- OUTSRC/odm.c +- 8192cd_osdep.c +- change.txt +- 8192cd_cfg.h +- Hal8192CDMOutSrc.c +- 8192cd.h + + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/14, button, feature refine + +* DIGEST: +- Update MAC, PHY parameters of 8188E + +* TEST METHOD: +- Not verify yet. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- + +* DESCRIPTION: +- Sync RTL8188E_MAC_PHY_ForMPChip_v035 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/14, chuangsw, feature refine + +* DIGEST: +- Refine 88E tx early mode. + +* TEST METHOD: +- Disable tx early mode in 20M mode. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_sme.c +- OUTSRC/odm.c + +* DESCRIPTION: +- Turn on tx early mode in 20M mode may degrade TP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/14, YangQuan, feature refine + +* DIGEST: +- refine for smart repeater CHECK_AP_VXD_TIMEOUT. + +* TEST METHOD: +- enable smart repeater, STA connect to DUT AP interface, vxd client don't + connect to another AP +- STA ping DUT, verify packet lose or response time is better than orignal. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_sme.c + +* DESCRIPTION: +- enlarge CHECK_AP_VXD_TIMEOUT from 10s to 60s +- modify ss_req_ongoing from 3 to 1 in check_vxd_ap_timer + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/12, button, feature refine + +* DIGEST: +- Update MCU firmware + +* TEST METHOD: +- Not verify yet. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- + +* DESCRIPTION: +- Update MCU firmware. + 8192C/8188C fw ver. 86, UMC A-cut fw ver.80 + 8192D fw ver. 35 +- Hide PCIe power saving option + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/11, button, feature refine + +* DIGEST: +- Do not aggregate during RA try rate state + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Follow BB luke's suggestion + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/04, chuangsw, feature refine + +* DIGEST: +- Fix sdk 3.2 92C EVM not good than sdk 2.5.3 issue. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/09/04, jimmylin, feature refine + +* DIGEST: +- Add mib_misc and mib_wsc into mib_all + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_proc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/31, chuangsw, feature refine + +* DIGEST: +- Fix compile error of previous commition. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/31, chuangsw, feature refine + +* DIGEST: +- Fix 88e 2.4G frequency shift to 2.2G issue. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_osdep.c +- 8192cd_host.c + +* DESCRIPTION: +- 1. Add pcie clock auto scan in 97D. +- 2. Refine pcie parameter. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/31, chuangsw, feature refine + +* DIGEST: +- Improve 88E TP. + +* TEST METHOD: +- 8188E TP is less than 8188C. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_sme.c +- 8192cd_osdep.c +- 8192cd_cfg.h +- Hal8192CDMOutSrc.c +- 8192cd.h +- change.txt + +* DESCRIPTION: +- 1. Enable tx early mode for 8188E. +- 2. Adjust RQPN. +- 3. 88E "wifi_specific" must be "1" for logo test. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/31, jimmylin, bug fix + +* DIGEST: +- Fix CCK power not restricted PA in mp mode when using externel + +* TEST METHOD: +- In mp mode, CCK power is not restricted by hp_cck_pwr_max when using + externel PA. + +* ROOT CAUSE: +- The checking code is missing. + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/29, jimmylin, bug fix + +* DIGEST: +- Fix 20/40 MHz coexistence not working on 8188E + +* TEST METHOD: +- + +* ROOT CAUSE: +- Coding error when using 8188E because lacking of parentheses. Therefore + when legacy OBSS or legacy sta connectted, driver will not become 20MHz + mode. + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- Hal8192CDMOutSrc.c +- HalDMOutSrc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/28, weikaichuang, bug fixed + +* DIGEST: +- Fixed client mode unnecessarily reconnect with AP + +* TEST METHOD: +- If AP continuous change HT info "STA BW", the original code will + consider the AP has reinit and we tend to reconnect + +* ROOT CAUSE: +- Some APs will change HT info "STA BW" continuously, we should only + consider HT capability "BW" to check whether the AP has change BW or not + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/28, weikaichuang, bug fixed + +* DIGEST: +- Fixed WDS ping consecutive fail when reinit + +* TEST METHOD: +- WDS1 ping WDS2, then reinit WDS1, the ping will consecutive fail + +* ROOT CAUSE: +- The sequence number already receive by WDS2 will be dropped. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- If tx and rx idle for 5 sec, flush cache to 0xffff + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/23, Plus, feature refine + +* DIGEST: +- Fixed AP mode under VeriWave UDP testing Through-Put unstable issue + +* TEST METHOD: +- VeriWave UDP test ,packet size:1518,during time:10secs + 40Mbps,5G/2G,AMPDU enabled. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- During the veriwave testing,keep the TXOP turn on. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/20, jimmylin, bug fix + +* DIGEST: +- Fix the place to record sequence number for non-QoS data + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/15, jimmylin, bug fix + +* DIGEST: +- Fix 5G channel list without DFS in NCC domain + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- Remove 56, 60, and 64 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/14, weikaichuang, feature refined + +* DIGEST: +- refined auto channel selection and add disable_ch1213 mib + +* TEST METHOD: +- ch1 and ch11 put ap, check the result of auto channel is in middle channel + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_ioctl.c +- ieee802_mib.h + +* DESCRIPTION: +- mib disable_ch1213=1, autochannel won't select ch12 and ch13. +- the score calculated the number of AP plus the number of rx pkts and check + if fa over threshold or not + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/08/08, jimmylin, bug fix + +* DIGEST: +- Fix wrong bandwidth and guard interval info of station in fixed rate for + 8188E + +* TEST METHOD: +- Set system to be fixed rate. When showing bandwidth and guard interval + info of station, it will always be 20M mode and long GI for 8188E. + +* ROOT CAUSE: +- In 8188E, check_txrate_by_reg() will not be called. Therefore we need give + intial value to pstat->ht_current_tx_info in assign_tx_rate(). + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/07/31, davidhsu, feature refine + +* DIGEST: +- Add two regulation domain for global (14) and world-wide (15) + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c 8192cd_utils.c 8192cd.h wifi.h + +* DESCRIPTION: +- When regulation domain is set to global (14) or world-wide (15), the driver + will use passive scan when channel is switched to 12, 13, 14. +- If any hidden AP is found in these channels, it will do active scan again. +- When driver is set to AP mode and enable auto-channel under global and + world-wide domain, it will exclude the channel 12, 13 and 14. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/07/5, davidhsu, bug fix + +* DIGEST: +- Fix the issue that broadcast packet may be discarded by wifi client + during group key rekey + +* TEST METHOD: +- Enable WPA mixed mode and group key rekey +- Start a wifi client (ip: 192.168.1.1) to link to AP +- Issue the following command in PC console continuously as: + ping 192.168.1.1 -n 1 + arp -d +- Some PING may fail during group key rekey + +* ROOT CAUSE: +- When wlan driver set TKIP group key, it will always reset to TXPN to 0. + Thus, it will cause TXPN number is not matched with the number carried + in 2-1 message (RSC). So, client will drop the Rx broadcast packet + which PN is less than RSC. + +* MODIFIED FILES: +- 8192cd_security.c + +* DESCRIPTION: +- Do not reset TXPN when setting TKIP group key. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/07/02, jimmylin, refine + +* DIGEST: +- Refine code of LED control + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_led.c +- 8192cd_osdep.c +- 8192cd_rx.c +- 8192cd_tx.c +- 8192d_reg.h + +* DESCRIPTION: +- Refine code of LED control to be more orgnized. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/22, chuangsw, feature fine-tune + +* DIGEST: +- Add 96D pcie auto scan clock mechnism. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c +- 8192cd_hosti.c +- boards\rtl819xD\config.in + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/21, jimmylin, bug fix + +* DIGEST: +- Support copy_mib for repeater mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_ioctl.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/19, jimmylin, bug fix + +* DIGEST: +- Fix client mode DHCP getting/forwarding issue + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_br_ext.c +- 8192cd_headers.h + +* DESCRIPTION: +- When receiving DHCP Offer/Ack, transfer to broadcase packet + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/17, kevin, bug fixed + +* DIGEST: +- Change the timing of creating wds table back to the origianl one + +* TEST METHOD: + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- SD9 had changed the timing of creating wds table and found it causes + that two stations can communicate with different encryptions. Role + back to the original version. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/13, button, new feature + +* DIGEST: +- Extend GPIO control API to 8192C/ 8192D + +* TEST METHOD: + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_proc.c +- 8192cd_led.c +- linux-2.6.30/drivers/char/rtl_gpio.c + +* DESCRIPTION: +- Usage is same as 8188E before + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/13, jimmylin, bug fix + +* DIGEST: +- Fix sequence number check error when receiving QoS and non-QoS data packets + simultaneously + +* TEST METHOD: +- Set security mode to WPA-AES. NDSi will fail to connect to Nintendo test + server frequently. + +* ROOT CAUSE: +- NDSi can support QoS, but when WPA handshake NDSi will send EAPOL packets + as non-QoS data. We will treat these packets as QoS data with TID 0. After + that when receiving QoS data with the same sequence numbers and with TID 0, + these packets will be dropped. +- We should seperate sequence numbers between non-QoS data and QoS data with + TID 0. + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/07, jerryko, bug fix. + +* DIGEST: +- improve 1x1 throughput with intel driver v15.1.1.1 + +* TEST METHOD: +- intel 6300 use driver v15.1.1.1 +- use 96e+88e test + +* ROOT CAUSE: +- use small edca backoff values to get better throughput + +* MODIFIED FILES: +- OUTSRC/odm.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/06, jerryko, bug fix. + +* DIGEST: +- fix intel v15.1.1.1 iot wep mode + +* TEST METHOD: +- intel 6300 use driver v15.1.1.1 +- chariot UL or Bi-direction + +* ROOT CAUSE: +- when intel no ht cap, it still enable CF-End + +* MODIFIED FILES: +- OUTSRC/odm.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/05, jerryko, bug fix. + +* DIGEST: +- fix intel v15.1.1.1 iot issue for OUTSRC + +* TEST METHOD: +- intel 6300 use driver v15.1.1.1 +- chariot UL or Bi-direction + +* ROOT CAUSE: +- Client driver waits more time to tx. Then it cause throughput + to decrease. + +* MODIFIED FILES: +- OUTSRC/odm.c + +* DESCRIPTION: +- 1.adjust AP backoff time with intel client. +- 2.Send CF-End to inform client to use air. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/05, jerryko, bug fix. + +* DIGEST: +- fix intel v15.1.1.1 iot issue + +* TEST METHOD: +- intel 6300 use driver v15.1.1.1 +- chariot UL or Bi-direction + +* ROOT CAUSE: +- Client driver waits more time to tx. Then it cause throughput + to decrease. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- 1.adjust AP backoff time with intel client. +- 2.Send CF-End to inform client to use air. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/06/01, pluswang, feature refine. + +* DIGEST: +- macclone and nat25 conexist refine + +* TEST METHOD: +- An AP setting under router mode , Enable DHCP server +- Ulinker(enable DHCP client mode) let it connect to AP +- sometimes DUT can not get IP from DHCP server + +* ROOT CAUSE: +- hold macclone feature util DUT(br0) got IP and built a NAT25 entry successfully. + +* MODIFIED FILES: +- 8192cd_br_ext.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/28, jerryko, bug fix + +* DIGEST: +- RX AMSDU issue in 88E platform + +* TEST METHOD: +- use clients send amsdu +- ping large packets, amsdu packets receive fail + +* ROOT CAUSE: +- adjust MAC RX fifo boundary setting in 88E + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- adjust 0x116 rx fifo boundary register from 0x13ff to 0x25ff in 88E + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/25, jimmylin, feature refine + +* DIGEST: +- Unify 8188E GPIO control methods + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_led.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/23, jimmylin, bug fix + +* DIGEST: +- Don't drop IPv6 packets coming from guest port when guest_access is + enabled + +* TEST METHOD: +- In IPv6 passthrough case, traffic from guest port can't reach WAN side if + guest_access is enabled. + +* ROOT CAUSE: +- In IPv6 passthrough, the DA of packet will be the MAC of PC on WAN side. + The DA is not the MAC of DUT. In this case, we can't drop these packets. + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- In IPv6 passthrough case, the filter function should be implemented in + bridge module. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/23, jimmylin, feature add + +* DIGEST: +- Add proc API to support 8188E GPIO control + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.h +- 8192cd_proc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/22, chuangsw, feature refine. + +* DIGEST: +- Fix wifi logo 4.2.23-T06 in 92D DMDP mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/22, pluswang, feature refine. + +* DIGEST: +- macclone and nat25 conexist refine + +* TEST METHOD: +- A AP setting under bridge mode , don't enable DHCP server +- Device under test setting as STA mode ,enable DHCP client mode, let it connect to AP +- A DHCP server behide the AP (by wired) +- DUT got IP from DHCP server but can't ping the IP at internet +- after the patch it should be can ping the IP at internet + +* ROOT CAUSE: +- if the packets run macclone path then don't enter nat25 handle both tx path and rx path + +* MODIFIED FILES: +- 8192cd_br_ext.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/18, michael, refine + +* DIGEST: +- modify for 8198B watchdog support + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192d_hw.c +- 8192cd_osdep.c +- OUTSRC/rtl8192c/HalDMOutSrc8192C_AP.c +- Hal8192CDMOutSrc.c +- HalDMOutSrc.c + +* DESCRIPTION: +- The watchdog spec of 8198B is different from 8196C/8198/819XD + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/17, chuangsw, feature refine. + +* DIGEST: +- Fix wifi logo 4.2.25-T06 in 92D DMDP mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- Since BE will be very passitive when VI exist in 92D DMDP mode, +- this will cause the BE traffic too small while traffic forward +- from one sta to another sta. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Since the max throughput of this item is 20Mbps, so if total tp is <=20 + and vi exist, we may not change be_edca to be passitive. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/17, chuangsw, feature refine. + +* DIGEST: +- Refine 88E XCAP control setting. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/17, jimmylin, code rearrange + +* DIGEST: +- Remove redundant code about non-MERGE_FW + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_proc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/16, chuangsw, feature refine. + +* DIGEST: +- We should not modify the default value of REG_AFE_XTAL_CTRL in 88E. +- Enable 88E XCAP control setting by mp. + +* TEST METHOD: +- When do 88E MP test in low temp, the tx will stop. + +* ROOT CAUSE: +- REG_AFE_XTAL_CTRL (0x24) will be modified in wlan driver and this value + should use the default value or it will cause tx fail in low temp. + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/14, weikaichuang, sync sd9 patch + +* DIGEST: +- Merge some compiling issue for sd9 + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd_rx.c +- 8192cd_osdep.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/11, chuangsw, feature refine + +* DIGEST: +- Wifi logo patch for 92D 1x1 concurrent mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- Pages may not enough to pass the wifi wmm test in DMDP mode. + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_osdep.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Since this patch may degrade TX throughput from 90->70, + We should only enable when "wifi_specific" = 1. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/10, button, bug fix + +* DIGEST: +- 8197D bi-direction throughput issue + +* TEST METHOD: +- Run chariot high performance script, uplink 3 pairs, downlink 3pairs +- + +* ROOT CAUSE: +- swq_aggnum kepps at 1, not increase. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_sme.c +- 8192cd_ioctl.c +- OUTSRC/odm.c + +* DESCRIPTION: +- If only one sta, set swq_aggnum = 4, modify swq threshold when up_time is +- odd or even. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/09, jimmylin, bug fix + +* DIGEST: +- Fix RTS and CTStoSelf rates not 11Mbps in ERP protection mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- RTS and CTStoSelf rates are controlled by register in 88R/92C/92D +- If using CTStoSelf, the rate selection is wrong + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd_util.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/9, davidhsu, new feature + +* DIGEST: +- Add multiple profile support in client mode + +* TEST METHOD: +- Enable "Multiple AP profile Support" in menuconfig, and rebuild image +- After image is uploaded, set client mode, and profile by using: + iwpriv wlan0 set_mib ap_profile_add=ssid,sec_type,auth_type,key... + iwpriv wlan0 set_mib ap_profile_enable=1 +- Then, enable wlan driver. It will search APs configured in profiles. +- You may issue "cat /proc/wlan0/mib_ap_profile" to check current profile + setting. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h + 8192cd_cfg.h + ieee802_mib.h + 8192cd_ioctl.c + 8192cd_sme.c + 8192cd_proc.c + Kconfig + +* DESCRIPTION: +- Add a new define flag (SUPPORT_MULTI_PROFILE) for this new feature. +- When the interface is configured as client mode (including virtual + client) and profile is set and enabled, it will start scanning AP + by referring profile start from first to last entry looply till to + any AP in profile is found. +- Besides, when roaming, it will also look for AP by referring profile. +- When repeater mode is enabled, root AP will search vxd connected AP in + every 10 secs. In each time finding, it only look for one profile entry + at a time. If you set 5 profiles in vxd interface, it may take 40~50sec + for finding the 5th profile AP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/08, chuangsw, feature refine + +* DIGEST: +- Do not do aggregation while low tp + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- This file is loss at last commit. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/08, davidhsu, bug fix + +* DIGEST: +- Fix smart repeater mode vxd-client connection issue with hidden AP + +* TEST METHOD: +- Disable broadcast SSID in remote AP, and then set different channel in root + AP of repeater mode. Then, virtaul client of repeater can't find the + detinate AP to connect. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, chuangsw, feature refine + +* DIGEST: +- Do not do aggregation while low tp + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_ioctl.c +- 8192cd_tx.c + +* DESCRIPTION: +- add a mib "low_tp_no_aggr" to enable the feature +- when one sta'tp is < 2 Mbps, it will disable ampdu for that sta. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, jimmylin, bug fix + +* DIGEST: +- Fix system crash if struct aid_obj can not be allocated + +* TEST METHOD: +- + +* ROOT CAUSE: +- In alloc_stainfo(), if struct aid_obj can not be allocated, will jump to + no_free_memory to free allocated memory, but forgot to check if the pointer + is null or not. + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, kevin, bug fixed + +* DIGEST: +- Fix Osk mp test fail issue + +* TEST METHOD: +- + +* ROOT CAUSE: +- Continuously enter MP background mode would use all the Tx buffer + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, kevin, add feature for SD9 + +* DIGEST: +- Fix WPS fail when in power save mode for SD9 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_util.c +- 8192cd_osdep.c +- 8192cd_ioctl.c +- 8192cd_hw.c +- 8192cd_headers.h +- 8192cd_cfg.h +- 8192cd.h + +* DESCRIPTION: +- Before this patch, when in power save mode, SD9 cannot get WPS status. +- Modify that when WPS LED is in progress, stop entering power save mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, button, bug fix + +* DIGEST: +- Do IQK after changing band from 2.4G to 5G or 5 to 2.4G + +* TEST METHOD: +- Configure 92D to client mode, SMSP, ABGN, click site survey via web, +- Sniffer probe request sent by DUT in 5G band. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Before this patch, probe request can't be sniffered, all BSS info is +- collected by passive scan. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, kevin, feature refine + +* DIGEST: +- Upon receive mDNS multicast packet, do m2u to all STA + +* TEST METHOD: +- + +* ROOT CAUSE: +- In Mac os, once the STA does not send an IGMP join packet, AP does not + unicast the packet. + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/04, kevin, feature refine + +* DIGEST: +- Multiple patches form SD9 to refine features + +* TEST METHOD: + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd_util.c +- 8192cd_osdep.c +- 8192cd_ioctl.c +- 8192cd_sme.c +- 8192cd_psk.c +- 8192cd_proc.c +- Hal8192CDMOutsrc.c + +* DESCRIPTION: +- INTERFERENCE_CONTROL Refine +- RTS_INIT_RATE +- FINE_TUNE_JIFFIES +- FINE_TUNE_DATA_TYPE_USAGE +- FIXED_LE0_LED1_state_change_when_using_LED2_as_WIFI_LED +- REFINE_PROC_MSG +- Fixed to possibly use erroneous macid in H2C command when STA_EXT + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, chuangsw, bug fix + +* DIGEST: +- Fix 88E MP wrong setting when change bandwidth + +* TEST METHOD: +- 1. Use mp command "iwpriv wlan0 mp_bandwidth 40M=0" +- 2. Than "irf 0 18" can see BIT(11)=1 +- 3. "iwpriv wlan0 mp_bandwidth 40M=1" +- 4. "irf 0 18" ==> BIT(11) still is 1, this is wrong! + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_mp.c +- 8192cd_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, kevin, add RTL8672 support + +* DIGEST: +- Update pcie reset procedure for RTL8672 + +* TEST METHOD: + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_util.c +- 8192cd_tx.c +- 8192cd_rx.c +- 8192cd_osdep.c +- 8192cd_ioctl.c +- 8192cd_cfg.h +- 8192cd_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, jerryko, feature refine + +* DIGEST: +- Software queue refine + +* TEST METHOD: +- use 2/3/4 stations, chariot throughput test + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_ioctl.c +- 8192cd_proc.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- Hal8192CDMOutSrc.c +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h + +* DESCRIPTION: +- 1.combine HF fix swq flow +- 2.decide queue len by each station +- 3.queue len increase or decrease by queue timeout times. + more timeout decrease station's queue len.less timeout increase len. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, button, bug fix + +* DIGEST: +- Fix bug of cannot entering PCIe power saving state + +* TEST METHOD: +- + +* ROOT CAUSE: +- rtl8192cd_bcnProc do not process DMA ok event since Ver.9162 + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, kevin, bug fix + +* DIGEST: +- Recover HIMR when leaving PCIE power save mode + +* TEST METHOD: + +* ROOT CAUSE: +- After leave PCIE power safe, the original code does not recover HIMR + +* MODIFIED FILES: +- 8192cd_hc.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/03, jerryko, bug fix + +* DIGEST: +- change addba resp timeout 0 + +* TEST METHOD: +- use intel station +- chariot 10 pair dl to ul throughput test + +* ROOT CAUSE: +- after addba timeout, intel station delba, so ul throughput is not good + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/02, kevin, bug fix + +* DIGEST: +- Change the flow of dequeue_frame() to avoid crash dump + +* TEST METHOD: + +* ROOT CAUSE: +- Since dequeue_frame() checks if list is empty or not, and then save the + interupt. Once an interupt comes before it tries to get the packet, it + gets an empty list which causes the system crash. + +* MODIFIED FILES: +- 8192cd_util.h + +* DESCRIPTION: +- SAVE_INT_AND_CLI before check if the list is empty. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/02, kevin, bug fix + +* DIGEST: +- Fix system crash due to DMA cache sync when receiving a null data (size=0) + in MIPS system. + +* TEST METHOD: + +* ROOT CAUSE: +- In MIPS system, DMA cache sync checks the size of payload, if the size is + 0, the system considers sync is redundent and report crash. + +* MODIFIED FILES: +- 8192cd_util.h + +* DESCRIPTION: +- Check if size is zero befire DMA cache sync, if size is zero, return. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/02, jimmylin, bug fix + +* DIGEST: +- Fix smart repeater connection issue when 2nd side band differs with + target AP in 40M mode + +* TEST METHOD: +- Set root-AP of smart repeater in 40M mode with 2nd side band lower. Set + target AP in 40M mode with 2nd side band upper. Then smart repeater can't + connect to target AP. + +* ROOT CAUSE: +- When root-AP scanned target AP, driver only remembers the channel, not 2nd + side band. driver should also remember the 2nd side band, then the setting + can be the same with target AP. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/02, Kevin, feature refine + +* DIGEST: +- Add a condition to check if pstate is NULL or not in WDS mode + +* ROOT CAUSE: + +* TEST METHOD: + +* Description: +- In WDS mode, add a condition to check whether pstate is NULL or not when de-frame. +- Change creat_wds_table to the time before drv_state|=DRV_STATE_OPEN + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd_osdep.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/05/02, HF, bug fix + +* DIGEST: +- sync bug fix from zcom for pure WDS TP low + +* ROOT CAUSE: +- Pure WDS is configured as client mode but Beacon functions +- degrade the TP + +* TEST METHOD: + +* Description: +- DO NOT enable beacon function when pure WDS + +* MODIFIED FILES: +- 8192cd_hw.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/24, button, feature refine + +* DIGEST: +- Refine ODM EDCA turbo function + +* TEST METHOD: +- Compare ODM ver. EDCA parameters with original trunk ver. + +* ROOT CAUSE: + +* MODIFIED FILES: +- OUTSRC/odm.c + +* DESCRIPTION: +- Refine ODM EDCA turbo function + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/24, button, bug fix + +* DIGEST: +- Change PCIe power state to L0 at the time Vxd open. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/24, weikaichuang, update RF register + +* DIGEST: +- Update Tx gain table to get better output power linearity + +* TEST METHOD: +- Tested by RF team in channel 13 and compare the results between two version of register + +* ROOT CAUSE: +- +* MODIFIED FILES: +- data/radio_a_1T_n.txt + +* DESCRIPTION: +- Update RF register (Tx gain table) which gain better evm stablility in high channel number and better output power linearity + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPe: +- 2012/04/24, weikaichuang, bug fix + +* DIGEST: +- patch from SD9 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_headers.h +- 8192cd_mp.c +- 8192cd_osdep.c +- 8192cd_rx.c +- 8192cd_tx.c + +* DESCRIPTION: +- 1. VAP free RX queue when turn off the interface. +- 2. compile error under linux 2.6.32 +- 3. RTL867X_VLAN_mapping patch +- 4. Add 2-TX Mechanism in Low 11n Data rate(MCS0 ~MCS6) to increase performance. +- 5. Modify the timing of XTAL_BSEL under the defined flag "defined(CONFIG_RTL8672)" + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/23, jimmylin, bug fix + +* DIGEST: +- Fix not check ERP stations for 8188E + +* TEST METHOD: +- + +* ROOT CAUSE: +- In 8188E chip verification, ERP station check is disabled and not enabled + again. Fix it for 8188E. + +* MODIFIED FILES: +- 8192cd_ioctl.c +- 8192cd_net80211.c +- 8192cd_security.c +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/23, jimmylin, bug fix + +* DIGEST: +- Fix RTS rate under B sta + G sta + N sta mixed connection + +* TEST METHOD: +- Let B sta, G sta, and N sta connect to AP. Before sending data to G sta + and N sta, RTS should be sent first. RTS rate is 24Mbps, but should be CCK + rates. + +* ROOT CAUSE: +- When choosing RTS rate, we only consider N protection, forget to consider + ERP protection. Add ERP protection check when choosing RTS rate. + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/18, chuangsw, bug fix + +* DIGEST: +- 1. Fix MP Single Tone can not stop issue. +- 2. Fix MP power tracking get wrong index in CCK swing table. +- 3. Fix MP power tracking not work when current thermal (0x42) =0. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/17, jimmylin, feature refine + +* DIGEST: +- Modify timer unit of reorder check to avoid overflow +- Do reorder check according to address 1 + +* TEST METHOD: +- Test UL TP with 3 stations. Sometimes console will show "RC timer overflow" + +* ROOT CAUSE: +- If the time unit is microsecond, jiffies times 10000 may be overflow. Now + all time unit is adjusted to the same as jiffies. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_osdep.c +- 8192cd_rx.c + +* DESCRIPTION: +- Moreover, the condition to do reorder check is modified to address 1 no + matter about AP mode or client mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/16, button, feature refine + +* DIGEST: +- Modify ODM driverof 8188E + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- HalDMOutSrc.c +- OUTSRC/odm.c +- OUTSRC/rtl8188e/Hal8188ERateAdaptive.c + +* DESCRIPTION: +- Correct short GI setting +- Sync Rate adaptive parameters with trunk original ver. +- Validate mib dig_enable in ODM driver + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/16, PlusWang, feature refine + +* DIGEST: +- Let NAT25 and MacAddrClone feature can be use concurrent + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/13, button, new feature + +* DIGEST: +- Add ODM support of 8188E + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Makefile +- Kconfig +- 8188e_hw.c +- ieee802_mib.h +- Hal8192CDMOutSrc.c +- 8192d_hw.c +- 8192cd_hw.c .h +- 8192cd_util.h .c +- 8192cd_ioctl.c +- 8192cd_headers.h +- 8192cd_debug.h +- 8192cd_cfg.h +- 8192cd_reg.h +- 8192cd_tx.c +- 8192cd_rx.c +- 8192cd_sme.c +- 8192cd_osdep.c +- 1x_md5c.h +- 8192cd_proc.c + +* NEW FILES: +- HalDMOutSrc.c +- typedef.h +- odm_inc.h +- OUTSRC/HalPhyRf.c .h +- OUTSRC/odm.c .h +- OUTSRC/odm_debug.c .h +- OUTSRC/odm_HWConfig.c .h +- OUTSRC/odm_interface.c .h +- OUTSRC/odm_precomp.h +- OUTSRC/odm_reg.h +- OUTSRC/odm_types.h +- OUTSRC/HalHWImg8188E_BB.c .h +- OUTSRC/HalHWImg8188E_MAC.c .h +- OUTSRC/HalHWImg8188E_RF.c .h +- OUTSRC/rtl8188e/Hal8188ERateAdaptive.c .h +- OUTSRC/HalPhyRf_8188e.c .h +- OUTSRC/odm_RTL8188E.c .h + +* DESCRIPTION: +- Add an "ODM driver" option in menuconfig when 8188E is enabled. +- Default: disable, use original structure. +- HalDMOutSrc.c: Common DM functions +- Hal8192CDMOutSrc.c: 8192C/8188RE/8192D DM functions +- OUTSRC folder: ODM driver sync with MP,CE,ADSL platform + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/12, chuangsw, bug fix + +* DIGEST: +- Refine 88E 63 STAs connection fail in ecos. + +* TEST METHOD: +- Veriwave test + +* ROOT CAUSE: +- sta queue buffer is not enough. + +* MODIFIED FILES: +- 8188e_util.c + +* DESCRIPTION: +- sta dz mgt queue should be independent of other APSD queues. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/11, PlusWang, feature refine + +* DIGEST: +- Refine AFE Crystal Control circuit setting value + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/11, jimmylin, bug fix + +* DIGEST: +- Fix force_stop_wlan_hw() to reset both mac0 and mac1 for 92D concurrent + +* TEST METHOD: +- When system core dump, force_stop_wlan_hw() will be invoked to stop + hardware. With 92D concurrent mode, several times of core dump will cause + 92D Rx DMA hang. + +* ROOT CAUSE: +- Because of define flag, only mac0 of 92D concurrent will do + rtl8192cd_stop_hw(). Add CONFIG_RTL_92D_DMDP to enable mac1 to do + rtl8192cd_stop_hw(). + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/10, EricYeh, bug fix + +* DIGEST: +- Refine 8188E Tx Power Tracking function + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Do NOT do 88E IQK during 88E tx power tracking + beacause 88E IQK only works on channel 7 +- Save IQk results into 'IQKMatrixRegSetting' for tx power tracking + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/10, EricYeh, feature refine + +* DIGEST: +- Update RF parameters from suggestions of RF team (Anchin) + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- data_88e/PHY_REG_1T_88E.txt +- data_88e/radio_a_1T_88E.txt +- 8192cd_mp.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/05, victorlo, feature refine + +* DIGEST: +- Update RTL8188E phy register files to v25 + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* ADDED FILES: +- data_88e/MAC_REG_88E_TC.txt +- data_88e/PHY_REG_1T_88E_TC.txt +- data_88e/radio_a_1T_88E_TC.txt + +* MODIFIED FILES: +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_proc.c +- data_88e/MAC_REG_88E.txt +- data_88e/PHY_REG_1T_88E.txt +- data_88e/radio_a_1T_88E.txt + +* DESCRIPTION: +- MAC_REG_88E_TC.txt, PHY_REG_1T_88E_TC.txt and radio_a_1T_88E_TC.txt in + folder "data_88e" will stay at v20 to support testchip as a verified + version. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/03, jerryko, feature refine + +* DIGEST: +- Refine 88E tx early mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd_hw.c +- 8192cd.h + +* DESCRIPTION: +- reference http://dtdinfo1.realtek.com.tw:8080/~sub1/mantis-cn/view.php?id=12405 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/02, jerryko, feature refine + +* DIGEST: +- Support 88E tx early mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd_sme.c +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd.h + +* DESCRIPTION: +- reference http://dtdinfo1.realtek.com.tw:8080/~sub1/mantis-cn/view.php?id=12405 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/04/02, chuangsw, bug fix + +* DIGEST: +- 88E 63 STAs connection fail in ecos. + +* TEST METHOD: +- Veriwave test + +* ROOT CAUSE: +- sta queue buffer is not enough. + +* MODIFIED FILES: +- 8188e_util.c + +* DESCRIPTION: +- sta dz mgt queue should be independent of other APSD queues. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/28, chuangsw, feature refine + +* DIGEST: +- Refine throughput at initial time. + +* TEST METHOD: +- Test DL throughput with intel 6300. +- It may not get the highest TP after few seconds. + +* ROOT CAUSE: +- Some packets may not get in bridge shortcut at initial time. + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- Modify tx packet threshold to get in bridge shortcut. (ori:10 , now:1) +- We only need one packet to go throughput bridge to learn. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/21, button, bug fix + +* DIGEST: +- 88E Rx throughpout drop 20 Mbps ( 90 -> 70 Mbps) + +* TEST METHOD: +- power on or s/w re-init, run chariot uplink + +* ROOT CAUSE: +- IQK result is wrong + +* MODIFIED FILES: +- 8188e_hw.c + +* DESCRIPTION: +- use default value when calibrated index is unresonable. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/20, EricYeh, Bug Fix + +* DIGEST: +- Fix issue: "reg_dump all" will cause system reboot for 8192C + +* TEST METHOD: +- Enable printk function, try command "iwpriv wlan0 reg_dump all" + for 92C demo board, system will auto reboot + +* ROOT CAUSE: +- Command "reg_dump all" needs lots time to finish, which will cause + WATCHDOG reset if interrupt disabled. + +* MODIFIED FILES: +- 8192cd_ioctl.c + +* DESCRIPTION: +- Enable interrupt before command "reg_dump all" finished + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/20, EricYeh, feature refine + +* DIGEST: +- Fine tune thermal value range for 8188E + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- Modify MAX thermal value from 0x28 to 0x32 for 8188E + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/19, jimmylin, bug fix + +* DIGEST: +- Fix auto channel under smart repeater mode + +* TEST METHOD: +- Set auto channel in root AP under smart repeater mode, it will make vxd + client not connection. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/14, jerryko, feature refined + +* DIGEST: +- rl_recover_timer only init/del by root interface + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- priv->pshare->rl_recover_timer decides hw tx retry count and only initial/delete + by root interface. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/13, EricYeh, new feature + +* DIGEST: +- Add RTL8188E Tx Power Tracking function + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_mp.c +- 8192cd_sme.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Porting Tx Power Tracking function from WiFi client team. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/13, victorlo, bug fix + +* DIGEST: +- Use correct basic rate in client mode for connection packets + +* TEST METHOD: +- let client mode connects to an AP with only OFDM basic rates, + client mode uses CCK rate for connection packets tx. + +* ROOT CAUSE: +- During connection period, client mode has still not yet recognized + the connecting AP with all info including rates, will use its own + rate set to send connection packets. + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- Before connection success, there is no station info referenced in client + mode for APs. Since client mode have to use correct basic rates complying + connecting AP for sending management packets, driver should find rates + base on beacon and probe content from the AP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/05, victorlo, feature refine + +* DIGEST: +- Sync wlan driver after KLD 96E+88E (Kernel Linux 2.4) porting + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.h +- 8192cd_osdep.c +- ieee802_mib.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/01, jimmylin, bug fix + +* DIGEST: +- Fix bug of calculating tx pause timeout + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Put minuend and subtrahend in wrong positions in TSF_DIFF() + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/03/01, hf, refine + +* DIGEST: +- Refine WISP mode Passthrough when using smart repeater mode + +* ROOT CAUSE +- WISP using Vxd as client mode +- the driver code default handling root as client mode + +* TEST METHOD: +- Set WISP mode and Vxd enabled +- play VLC at APUT LAN and DUT LAN +- check if VLC can play + +* Description +- if WISP enabled, let the passthrough packets transmit from +- Vxd interface + +* MODIFIED FILES: +- 8192cd_tx.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/29, victorlo, feature refine + +* DIGEST: +- Refine for Intel IOT base on CET results + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c +- RateAdaptive.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/29, chuangsw, feature fine tune + +* DIGEST: +- Refine 88E RQPN setting to pass wifi logo. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/24, EricYeh, feature fine-tune + +* DIGEST: +- Refine 8188E RF parameters & Add 96E auto scan clock mechnism. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/24, jerryko, bug fix + +* DIGEST: +- Fix broadcast storm drop packets to cause rc timeout + +* TEST METHOD: +- use benchbee android app to test throughput +- compare throughput between bcm and rtk + +* ROOT CAUSE: +- Because benchbee sends many arp request pkts and cause broadcast storm. +- Then wlan driver drop broadcast pkts before reorder check pkts. +- So tcp ack packets from clients queue in rc queue and wait timeout to low + throughput. + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- Move the place of drop broadcast pkts in wlan driver. +- Driver will drop broadcast pkts in rtl_netif_rx and before + forward to wlan interface. +- It still let broadcast pkts into rc queue and + avoid rc timeout. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/24, victorlo, New Feature + +* DIGEST: +- Add supports for both RTL8188E test and MP chip + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_proc.c +- 8192cd_util.h +- data_88e/PHY_REG_PG_88E.txt + +* DESCRIPTION: +- Define flag "SUPPORT_RTL8188E_TC" for test chip related portions +- Enable power by rate feature for MP chip + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/23, chuangsw, feature refine + +* DIGEST: +- Refine MACID pause mechanism + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd.h + +* DESCRIPTION: +- Set a timer to avoid STA pause too long. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/23, button, feature refine + +* DIGEST: +- Refine DIG mechanism + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- Don't set initial gain to lower bound after new STA join BSS +- Adjust initial gain upper bound according to minumum RSSI +- Cancel DIG on/off threshold (RSSI:40,35) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/23, EricYeh, feature fine-tune + +* DIGEST: +- Refine 8188E IQK & Gain Table from suggestion of RF. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd_hw.c +- data_88e/radio_a_1T_88T.txt + +* DESCRIPTION: +- +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/23, qinjunjie, feature fine-tune + +* DIGEST: +- support flood some special multicast mac addresss to all sta, no matter it + sends igmp join or not + +* TEST METHOD: +- STA1 disable igmp(IGMPLevel=0),PC2 at ethernet run intel upnp tool to send + m-search packet, to check STA1 can receive m-search packet or not + +* ROOT CAUSE: +- When STA disable igmp, it won't send igmp report for upnp(m-search), and + won't receive multicast address format packet + +* MODIFIED FILES: +- 8192cd_osdep.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_cfg.h +- 8192cd_ioctl.c +- 8192cd.h +- 8192cd_proc.c +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/22, EricYeh, Feature refine + +* DIGEST: +- Refine 8188E IQK source codes + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8188e_reg.h +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/21, victorlo, Feature refine + +* DIGEST: +- Access Rate Adaptive relative functions for autorate only + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd_sme.c + +* DESCRIPTION: +- Check if driver is using autorate, then access RA functions; otherwise, + no access RA for better cpu utilizations. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/21, victorlo, Feature refine + +* DIGEST: +- Refine setting for rssi_dump debug print + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- By signin rssi_dump with different number as period setting, driver prints + rssi info under user request. e.g. set rssi_dump as 1, prints every second; + set rssi_dump as 3, prints every 3 seconds. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/20, button, feature fine tune + +* DIGEST: +- Improve throughput under weak signal condition + +* TEST METHOD: +- Let 2 wireless clients connec to 2 98+92c demoboard (DUT) respestively +- Move clients to a location where Tx rate of AP is aobut MCS3. +- Run throughput of one DUT pair(downlink).1 minute later, run the other pair. +- The throughput of second pair will be lower than the first pair. + +* ROOT CAUSE: +- EDCA turbo mode is not enabled + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- when aggregation is enabled, turn on TXOP, don't check throughput + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/20, button, feature fine tune + +* DIGEST: +- Improve 92D AP v.s client mode throughput + +* TEST METHOD: +- configure 2 92d DUT as AP and client mode, run chariot 3D+3U + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- rollback transceiver timing of AP mode to old value + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/20, jimmylin, feature finetune + +* DIGEST: +- Disable Rx short GI in band 4 according to baseband team + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/20, hf, refine + +* DIGEST: +- Refine transmit CTS_TO_SELF according to IC SPEC + +* TEST METHOD: +- Enable protection mode +- connect two client to DUT, one is in B mode, one is in G mode +- chariot before LAN PC and G mode client +- sniffer the packet to see if CTS_TO_SELF transmitted before TCP packet + +* ROOT CAUSE +- A single BIT(11) is used control if transmit CTS_TO_SELF +- according to new spec + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION +- if cts2self mib is true +- 1)Set CTS_EN bit when protection needed +- 2)NOT set RTS_EN any longer according to new spec + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/20, victorlo, bug fix + +* DIGEST: +- Refine RTL8188E NAV protection time setting + +* TEST METHOD: +- Downlink test with Intel 6300, throughput becomes around 82Mbps but not + ~90Mbps, by 3 pairs of high performance script on chariot. + +* ROOT CAUSE: +- NAV protection time has been set as a shorter time mistakenly by old IOT + function. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Set NAV protection time as h/w default to enhance throughput + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/17, jimmylin, bug fix + +* DIGEST: +- Fix bandwidth becoming from 40M to 20M after start_clnt_lookup() +- Fix tx stopping after site survey between DFS channels + +* TEST METHOD: +- Start AP in 40M mode and enable smart repeater mode. After site survey, + AP will become 20M mode. +- After site survey between DFS channels, WPS connection will fail. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/17, joey, code refine + +* DIGEST: +- move the called function "update_hw_l2table" from Bridge module to here + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- to avoid hacking the Linux kernel or the other kernel. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/17, victorlo, bug fix + +* DIGEST: +- Refine RTL8188E tx report packet drop counting + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c + +* DESCRIPTION: +- Tx report shows only 1 byte for tx drop caused by tx fail, but driver read + 2 bytes for counting mistakenly. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/17, EricYeh, new feature + +* DIGEST: +- Add RTL8188E IQK function + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8188e_hw.c +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- Add compile flag '_TMP_88E_IQK' to package IQK patches. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/16, victorlo, feature refine + +* DIGEST: +- Enable 63 clients connection support on RTL8188E + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_proc.c +- 8192cd_rx.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/16, EricYeh, new feature + +* DIGEST: +- Import Tx Power Tracking Tables from TXT files. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- data/REG_TXPWR_TRK.txt +- data/REG_TXPWR_TRK_hp.txt +- data_92d/REG_TXPWR_TRK_n_92d.txt +- data_92d/REG_TXPWR_TRK_n_92d_hp.txt +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- Hal8188EPwrSeq.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/16, victorlo, feature refine + +* DIGEST: +- Update rssi of clients for RTL8188E Rate Adaptive evaluation + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/15, jerryko, feature refine + +* DIGEST: +- enable 88E NO-Link function for DETECT_STA_EXISTANCE + +* TEST METHOD: +- (i)use broadcom STA to test chariot +- (ii)unplug STA when running chariot +- (iii)sniffer check if output pkt stop when set NO-Link function + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd_osdep.c +- 8192cd_ioctl.c +- 8188e_hw.c +- Hal8192CDMOutSrc.c +- 8192cd_cfg.h +- 8192cd.h + +* DESCRIPTION: +- (i)When 88E driver detects STA leaving,set NO-Link register to clean STA + output traffic. +- (ii)two conditions are STA leaving. + (a)continue pkt drop over max_pkt_fail(mib default 50) + (b)3 secs pkt drop over min_pkt_fail(mib default 30) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/14, PlusWang, feature refine + +* DIGEST: +- when reassembly wsc_ie at probe_req , recount it's length +* TEST METHOD: +- n/a +* ROOT CAUSE: +- n/a +* MODIFIED FILES: +- 8192cd_sme.c +* DESCRIPTION: +- when probe_req included multi parts wsc_ie + after reassembly wsc_ie reassign length of this TLV + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/15, jimmylin, feature refine + +* DIGEST: +- Modify 8188ER LED setting according to new circuit + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/14, EricYeh, feature refine + +* DIGEST: +- Add new AGC Table for 92D High Power PA. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- data_92d/AGC_TAB_n_92d_hp.txt +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- Add new AGC tables for 92D High Power PA according to the suggestion + from RF team. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/13, victorlo, feature refine + +* DIGEST: +- Enable RTL8188E Turbo mode and refine TP for Intel clients + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- Enable IOT Turbo mode. +- Enable h/w RTS for Intel clients in close and middle range to improve TP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/13, victorlo, feature refine + +* DIGEST: +- Update RTL8188E PHY register files to v20 + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- data_88e/MAC_REG_88E.txt + +* DESCRIPTION: +- Enable pre tx security circuit for using WEP or TKIP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/13, EricYeh, feature refine + +* DIGEST: +- Refine power tracking for High Power PA. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- + +* DESCRIPTION: +- Modify power tracking tables for High Power PA according to the suggestion + from RF team. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/13, davidhsu, new feature + +* DIGEST: +- Add smart repeater mode + +* TEST METHOD: +- Enable repeater mode and set root as AP and vxd as client +- Set AP channel is differnet with the AP, which vxd client want to connect +- Open the wlan0 and wlan0-vxd driver +- Then, it will auto switch AP channel the one as the vxd AP used + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- Revision: 9238 + 8192cd_cfg.h + 8192cd.h + 8192cd_headers.h + 8192cd_osdep.c + 8192cd_sme.c + +* DESCRIPTION: +- Add a new define flag (SMART_REPEATER_MODE) for this new feature. When + repeater mode is configured, it will automitically define this new flag. +- When it is defined and root interface is set to AP, it will start a timer (in + 10sec periodically) to check if vxd client has associated with AP + successfully. +- If vxd client did not connect to AP, it will issue site survey scan in root + interface. If any vxd destinated AP is found by matching SSID, it will auto + switch the current channel to that one. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/13, hf, bug fixed + +* DIGEST: +- fix assoc expired client always be active when client disconnect by plug out + +* ROOT CAUSE +- when client is plug out, DUT take client as associated +- and count down to expire. when the client is reused +- and issue probe request, STA's TX Rx count will change +- then DUT think the client is still associated +* TEST METHOD: + +* Description +- not to count the proble rsq by set station to NULL +- even set he HW TX report to mac id 0 by set station to NULL + +* MODIFIED FILES: +- 8192cd_tx.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/10, victorlo, Bug fix + +* DIGEST: +- Refine RTL8188E per station RA init flow + +* TEST METHOD: +- Set AP as TKIP only, after client connected, chariot can hardly run + throughput test. + +* ROOT CAUSE: +- Since TKIP (and WEP) mode, AP will force non-11N connection with clients, + driver will set the h/w RateID as B/G only. However driver was running + wrong steps by clearing the initial setting of tx rate without a new value + set instead, will finally cause tx rate forced as CCK 1Mbps. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Refine per station RA init flow, by reallocated RateID init step to the + front of rate mask setting, make RA mechanism generate the correct tx rate. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/10, jimmylin, bug fix + +* DIGEST: +- In client mode, check channel range for 40M mode and make sure to fit + region domain + +* TEST METHOD: +- Set region domain of AP to ETSI. Set AP to 40M mode. Set primary channel to + 8 and secondary channel to 12. Set region domain of client to FCC. Client + should connect in 20M mode because in FCC channel 12 is not allowed to use. + But client connected in 40M mode. + +* ROOT CAUSE: +- In client mode, we didn't check if the secondary channel is legal in the + current region domain. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/09, victorlo, feature refine + +* DIGEST: +- Update RTL8188E PHY register files to v19 + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c +- data_88e/AGC_TAB_1T_88E.txt +- data_88e/MAC_REG_88E.txt +- data_88e/PHY_REG_1T_88E.txt +- data_88e/PHY_REG_MP_88E.txt +- data_88e/radio_a_1T_88E.txt + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/09, victorlo, add new IC support + +* DIGEST: +- Add RTL8188E support + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* ADD FILES: +- 8188e_hw.c +- 8188e_reg.h +- data_88e/AGC_TAB_1T_88E.txt +- data_88e/MAC_REG_88E.txt +- data_88e/PHY_REG_1T_88E.txt +- data_88e/PHY_REG_MP_88E.txt +- data_88e/PHY_REG_PG_88E.txt +- data_88e/radio_a_1T_88E.txt +- Hal8188EPwrSeq.c +- Hal8188EPwrSeq.h +- HalPwrSeqCmd.c +- HalPwrSeqCmd.h +- RateAdaptive.c + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_debug.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_ioctl.c +- 8192cd_mp.c +- 8192cd_net80211.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_rx.c +- 8192cd_security.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- 8192cd_util.h +- 8192d_hw.c +- Hal8192CDMOutSrc.c +- wapi_wai.c + +* DESCRIPTION: +- n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/08, chuangsw, Bug fix + +* DIGEST: +- Fix low TP when client mode 5G with Broadcom AP in bi-direction. + +* TEST METHOD: +- Run chariot bi-direction when client mode 5G with Broadcom AP. +- The TP will be only 110 Mbps. + +* ROOT CAUSE: +- Wrong setting when client mode 5G with Broadcom AP. +- We set aggregation break time(0x51a) to 0x0f when client mode 5G with + Broadcom AP. +- It should only write 1 byte (0x51a) but we write 2 bytes. +- So 0x51b(slot time) is wrong. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Now the TP when client mode 5G with Broadcom AP will be 145 Mbps. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/02/06, jimmylin, feature fine tune + +* DIGEST: +- Set use_ext_lna to 0 by default for 92D + +* TEST METHOD: +- + +* ROOT CAUSE: +- According to RF engineers, mib use_ext_lna is suggested to be 0 by default + for 92D. + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +- 2012/02/01, huxin, Bug fix + +* DIGEST: +- Do m2u for MDNS packets + +* TEST METHOD: + +* ROOT CAUSE +- some device not able to join multicast group 224.0.0.251 need to receive +- MDNS packets. + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +- 2012/01/31, hf, new feature + +* DIGEST: +- Set to nomember protection mode when there is AP in HT-MIXED mode on same + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- Place an AP near DUT,make sure their channel is the same, both using 11N +- connect a BG client to the AP +- check beacon from DUT by sniffer to see if OM bit is 1 + +* Description +- check other AP's beacon, if found HT_MIXED mode +- then set DUT to nomember protection mode + +* MODIFIED FILES: +- 8192cd_sme.c +- ieee802_mib.h +- 8192cd_ioctl.c +- 8192cd.h +- 8192cd_proc.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/30, EricYeh, New Feature + +* DIGEST: +- Add Power By Rate tables for 92D High Power + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- /data_92d/PHY_REG_PG_92d_hp.txt +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- Add txt file 'PHY_REG_PG_92d_hp.txt' to be Power By Rate tables for 92D HP. +- In reload_txpwr_pg() function, add power by rate for all conditions. + (Because High Power is the condition NOT to add 'Power By Rate' before) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/19, button, Bug fix + +* DIGEST: +- Fix probe response content not sync in power saving state + +* TEST METHOD: +- Do WPS auto lock down test, input wrong AP's PIN code, then check +- WSC I.E. of beacon and probe rsp + +* ROOT CAUSE: +- WSC daemon modify probe rsp in L1 state, driver did not update content in +- Tx packet buffer. + +* MODIFIED FILES: +- 8192cd_ioctl.c + +* DESCRIPTION: +- Exit power saving state before modifying I.E. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/18, button, Feature refine + +* DIGEST: +- 96D wakeup pin setting + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- 96D's wake up pin is JTAG_TMS(GPIO A4), modify request_irq_for_wakeup_pin() + and gpio_wakeup_isr() +- Disable power saving feature of 8196E. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/17, hf, refine + +* DIGEST: +- Use skb_copy instead of skb_clone for BT stress testing + +* ROOT CAUSE +- Send to other station using skb_clone when rx a Mcast or Bcast packet, +- due to the sleep station, the packets may be queued in dz_queue then return +- but rx process going on, if some changes made to the packet in rx, +- there is a risk to break the system when the packet is released from +- dz_queue. + +* TEST METHOD: +- BT stress testing + +* Description +- Use Skb_copy always. + +* MODIFIED FILES: +- 8192cd_rx.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/17, jerryko, Feature refine + +* DIGEST: +- update SD9 patchs for RTL8672 platform + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd_osdep.c +- 8192cd_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/11, victorlo, Feature refine + +* DIGEST: +- Set PVB in beacon for data frame only + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Though wlan driver does not release buffered management packets while + receiving ps-poll, also management packets buffering was designed to take + care an IOT issue with one of Broadcom cards, so decided to set PVB for + data frame only to comply the original idea. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/10, EricYeh, bug fix + +* DIGEST: +- Fix 92C ctx+power tracking in MP mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- Do LCK during ctx will cause No Tx Power. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Do NOT do LCK & IQK during MP ctx test. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/10, EricYeh, bug fix + +* DIGEST: +- 92D intPA power tracking Fix + +* TEST METHOD: +- + +* ROOT CAUSE: +- During 92D 5G power tracking, registers to control Tx Power Index (ex: e00) + may be increased by Tx_Power_Level_DPK, and such registers is 4 bytes size. + But in previous source codes, + only the lowest byte of these registers will be changed. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_mp.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Increase Tx_Power_Level_DPK for all 4 bytes for related registers. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/10, victorlo, Feature refine + +* DIGEST: +- Refine setting condition of NAV/DUR for tx packets + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- Remove DA checking for whether it is a multicast/boardcast address, but + instead check the Address 1 of Wifi packets for decision on setting NAV/DUR + for tx. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/10, chuangsw, bug fix + +* DIGEST: +- Fix repeater mode as client, vxd use WPA2-PSK, the wireless client will +- connect to vxd ap fail in 4-2. + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- This patch should only turn on in client mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/10, button, feature refine + +* DIGEST: +- Finetune ATIM window, EIFS + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- ATIM window: 5 => 1ms. now, high queue data can be release outside + ATIMWND, don't need to reserved so much time. +- EIFS: 314 => 64 us. follow MAC team's suggestion + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/09, jimmylin, New feature + +* DIGEST: +- Add Tx no Ack function + +* TEST METHOD: +- Set mib "txnoack" to 1. Send unicast packets and observe by sniffer to see + if our MAC is waiting for Ack packet. Ps. aggregation should be disabled. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_tx.c +- ieee802_mib.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/09, chuangsw, Bug fix + +* DIGEST: +- Fix repeater mode as client, vxd use WPA2-PSK, the wireless client will +- connect to vxd ap fail in 4-2. + +* TEST METHOD: +- Set DUT to repeater mode and root is client, vxd use WPA2-PSK. +- DUT connect to AP. +- Use a wireless client to connect to vxd AP, then it will fail. + +* ROOT CAUSE: +- The PMK of vxd will be wrong after F() is interrupted. +- In the begging, root interface will run up first, and then vxd. +- When vxd is under open(), the root interface(client mode) had finish open() +- and receive assoc-rsp from ap, so it will call "enable_vxd_ap". +- Since the first time open() of vxd did not finish yet and when root client +- call "enable_vxd_ap", it will call open() again, so it may something wrong. + +* MODIFIED FILES: +- 8192cd_osdep.c +- 8192cd_psk.c + +* DESCRIPTION: +- When vxd open(), we will check if the root interface is in the link state, +- if not, it will return directly. +- Add CLI protect in derivePSK(). + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/09, victorlo, Feature Refine + +* DIGEST: +- Refine retry limit setting for different amount of Intel clients + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_ioctl.c +- 8192cd_sme.c +- 8192cd_tx.c + +* DESCRIPTION: +- Refine retry limit for IOT mode of Intel clients, to avoid multiple intel +clients block tx path with their own retry limits, and finally cause chain +reaction of low tx performance. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/09, button, Bug fixed + +* DIGEST: +- fix bug of MCAST2UI_REFINE + +* TEST METHOD: +- Plus found in Sercomm SDK, DUT connected by several phone core dump easily. + +* ROOT CAUSE: +- After DUT convert broadcast or multicast to unicast, If unicast destination +- station is sleeping, DUT will buffer this unicast packet, then take it back +- as broadcast later, and convert it to unicast again. + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c + +* DESCRIPTION: +- Add rtl8192cd_start_xmit_noM2U to void calling __rtl8192cd_start_xmit +- recursively. +- Enable MCAST2UI_REFINE + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2012/01/09, button, Bug fixed + +* DIGEST: +- Change to sw beacon mode before entering power saving state. + +* TEST METHOD: +- Jerry set a DUT's SSID to RealKungfu, put it in office for long run test with +- SDK1.4, and found tx DMA stop, hang up reset happened once a day. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Glayrainx suggests change to sw beacon before entering L1, and change it back +- to at the time exiting L1. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/06, jerryko, Feature Refine + +* DIGEST: +- RTL 97D cpu enqueue wlan tx packets when tx/rx tp over 55Mbps with txop 0x5e + +* TEST METHOD: +- (1)use 97D 10/100 demo board test with intel6300 +- (2)use 97D giga qa board test with intel6300 + +* ROOT CAUSE: +- 97D cpu use txop 0x5e, then the throughput is lower than txop 0x2e. +- When txop 0x5e, wlan tx use 0x5e txop time. But the number of packets are + not enough, the throughput is low. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_cfg.h +- 8192cd_util.c + +* DESCRIPTION: +- When the total throughput over upper bound(tx 55M and rx 55M), wlan tx start + to enqueue data to improve better throughput with txop 0x5e.When the total + throughput under lower bound(tx 35M or rx 35M), wlan tx don't enqueue data. +- enlarge skb number from 300 to 480 in RTL 97D + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/05, plus, Bug fixed + +* DIGEST: +-under STA mode ,correct deauth packet A2 address. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- under client mode , deauth packet the A2 address should be myself hwaddr, + but now it be filled BSSID + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2012/01/05, jerryko, Feature Refine + +* DIGEST: +- remove kernel defined for bridge shortcut get_eth_cached_dev function + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- To avoid wlan driver to merge to kernel 2.4, get_eth_cached_dev function + miss to enter bridge shortcut function. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/03, chuangsw, Feature Refine + +* DIGEST: +- Modify 5g available channels for RUSSIAN region domain. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- The available channels for RUSSIAN region domain in 5g should be + {36,40,44,48} but not {34,38,42,46}. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/02, victorlo, Feature Refine + +* DIGEST: +- Refine IQK triggering flow + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_hw.c +- 8192cd_sme.c +- 8192d_hw.c + +* DESCRIPTION: +- Refine IQK triggering flow to let driver do IQK once only on the same + channel, to prevent redundance and failure happens after success IQK + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/01, EricYeh, Feature Refine + +* DIGEST: +- Modify RF-00 register for High Power in MP mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/01, EricYeh, Feature Refine + +* DIGEST: +- Update 92D PHY parameters from mantis-7960 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- data_92d/PHY_REG_n.txt +- data_92d/PHY_REG_n_92d_hp.txt + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2012/01/01, EricYeh, Bug Fix + +* DIGEST: +- Fix MP ctx error + +* TEST METHOD: +- Use MP Tool to test 'Continuous Tx', No Tx detected in Power Meter machine. + +* ROOT CAUSE: +- Maybe setting of registers error. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_mp.c +- 8192cd_osdep.c +- 8192cd_tx.c + +* DESCRIPTION: +- Apply MP CTX patches from SD9 (default is NOT applied before) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/30, hf, refine + +* DIGEST: +- Fix IOT issue with Ralink when client mode or vxd mode + +* ROOT CAUSE +- coding error, break too early +- if VXD enabled and do peformance test between VXD and AP +- especially when AP is a ralink solution, +- the (tx_2s_avg + rx_2s_avg) is always 49. never reach 50 +- so the IOT_engine not started. + +* TEST METHOD: +- Do chariot test between VXD and AP(client mode) + +* Description +- change (tx_2s_avg + rx_2s_avg) Threshold to 45 when +- client mode and connected to Ralink AP + +* MODIFIED FILES: +- 8192cd_sme.c +- Hal8192CDMOutSrc.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/29, jimmylin, new feature + +* DIGEST: +- Add new tables for new region domain - CN + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_sme.c +- 8192cd_util.c +- wifi.h + +* DESCRIPTION: +- Also fix the error of channel array of 2.4G for NCC + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/29, Button, bug fix + +* DIGEST: +- Fix 92d system hang during power on test + +* TEST METHOD: +- Build 92D DMDP image, enable PCIe power saving. +- Run power on endurance test by auto test tool, sometimes console cannot +- type or show any message after 4 days. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Switch to function 1 at ehe end of init_pcie_power_saving(). + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/29, EricYeh, Bug Fix (MAJOR) + +* DIGEST: +- Fix 92D IQK failed in MP mode. + +* TEST METHOD: +- + +* ROOT CAUSE: +- In 92D IQK, 'iqk_2g_done' 'iqk_5g_done' are used to judge the continue of + performing IQK, but this part is NOT suitable in MP mode. + +* MODIFIED FILES: +- 8192d_hw.c + +* DESCRIPTION: +- Always do IQK in MP mode regardless of 'iqk_2g_done'&'iqk_5g_done' + +----------------------------------------------------------------------------- + ++:Add vxd shortcut +* DATE/WHOM/TYPE: +- 2011/12/28, hf, refine + +* DIGEST: +- Add Vxd shortcut + +* ROOT CAUSE +- N/A +* TEST METHOD: + +* Description +- for vxd client, the shortcut can improve TP +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_rx.c +- 8192cd_osdep.c +- 8192cd_cfg.h + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/28, jerryko, feature refined + +* DIGEST: +- Interference resist(define INTERFERENCE_CONTROL)(patch from sd9) + +* TEST METHOD: +- strong non-ap interference signal on 2.4G channel +- use wifi card to connect ap + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_util.c +- 8192cd_sme.c +- 8192cd_ioctl.c +- 8192cd_hw.c +- 8192cd_headers.h +- 8192cd_hw.h +- 8192cd.h +- 8192cd_cfg.h + +* DESCRIPTION: +- 1.add baseband notch filter dynamic method to filter interference +- 2.enable dig by default and adjust dig flow + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/28, jerryko, feature refined + +* DIGEST: +- Separate four ac VO/VI/BE/BK stream(define WMM_VIBE_PRI)(patch from sd9) + +* TEST METHOD: +- use 4 pair downlink chariot (VO/VI/BE/BK) +- check VI/BE throughput + +* ROOT CAUSE: +- When VO stream run, VI EDCA=BE EDCA.So throughput can't separate VI and BE. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_tx.c +- 8192cd_sme.c +- 8192cd_hw.h +- 8192cd.h +- 8192cd_cfg.h + +* DESCRIPTION: +- When VO stream run, check if BE stream exist. If BE stream exist, VI EDCA>BE + EDCA. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/26, EricYeh, feature fine tune + +* DIGEST: +- 92D High Power 2.4G support + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- /dara_92d/PHY_REG_n_92d_hp.txt +- /dara_92d/radio_a_n_92d_hp.txt +- /dara_92d/radio_b_n_92d_hp.txt +- 8192cd_headers.h +- 8192cd_hw.c +- wifi.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/21, EricYeh, bug fix + +* DIGEST: +- CCK Shaping Filter for 2.4G channel-14 & other channels during MP. + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION: +- Because values of registers a22 - a29 shall be different between + channel-14 & others, so, when changing channel during mp mode, + call set_CCK_swing_index() to re-assign the correct values. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/16, jerryko, bug fix + +* DIGEST: +- add TBTT time protection to avoid TBTT hold > dot11BeaconPeriod + +* TEST METHOD: +- + +* ROOT CAUSE: +- AP single SSID, min beacon is 20ms. But TBTT hold is 32ms. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- When AP single SSID, beacon period <= 40ms, TBTT hold set 0x1df (15ms). + AP beacon period > 40ms, TBTT hold set 0x400 (32ms). + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/14, Eric Yeh + +* DIGEST: +- Merge MP patches from SD9 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_mp.c +- 8192cd_osdep.c +- 8192cd_tx.c + +* DESCRIPTION: +- Use 'CONFIG_RTL8672' to pack these patches. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/14, jerryko, bug fixi + +* DIGEST: +- avoid compiler error when mp mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/13, Button, feature refined + +* DIGEST: +- Merge init flow patch from sd9 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_hw.c +- 8192c_reg.h +- 8192cd_osdep.c +- 8192cd_headers.h +- data/rtl8192cfwua.bin +- data/rtl8192cfwn.bin +- data_92d/rtl8192dfw_n.bin + +* DESCRIPTION: +- Add MCU unlock for some corner case. +- 110511_8192cd_load_firmware_patch.zip, 110531_8192cd_load_firmware_patch.zip + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/13, jerryko, feature refined + +* DIGEST: +- if low tp, disable txop(define LOW_TP_TXOP)(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c +- 8192cd_ioctl.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/13, jerryko, feature refined + +* DIGEST: +- RTS init rate adjust update patch.(default no enable)(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/09, hf, bug fix + +* DIGEST: +- Fix repeater mode as client, associate repeatly + +* ROOT CAUSE +- AP_BW should be init as -1 for vxd also + +* TEST METHOD: + +* Description + +* MODIFIED FILES: + 8192cd_sme.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/01, jerryko, feature refined + +* DIGEST: +- Intel 6205 speed test not stable(patch from sd9) + +* TEST METHOD: +- speed test throughput with Intel 6205 + +* ROOT CAUSE: +- Intel 6205 enter sleep mode + +* MODIFIED FILES: +- 8192cd.h +- Hal8192CDMOutSrc.c +- 8192cd_proc.c +- 8192cd_ioctl.c +- 8192cd_sme.c + +* DESCRIPTION: +- 1.add Intel 6205 OUI +- 2.adjust edca cwmax (define LOW_TP_TXOP) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/08, jerryko, bug fix + +* DIGEST: +- Fix AP 11n only with WPA2PSK, non-11n client connect issue + +* TEST METHOD: +- 1.AP set 11n only with WPA2PSK +- 2.use 11abg client to connect AP +- 3.sometimes client can ping to AP success + +* ROOT CAUSE: +- From sniffer, AP send assoc response with status=18.Then AP still send 1-4 key + shaking with client. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- modify to check non-11n client code.Check non-11n client first, then send + event. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/12/02, joeylin, New feature +* DIGEST +- add TCP SYN DoS filter from WLAN interface +* TEST METHOD +- issue "xdos.exe [br0 ip] 0-65535" from STA +* ROOT CAUSE +* MODIFIED FILES +- /trunk/RLXLinux/rlxlinux-v2.2/users/boa/Makefile +- /trunk/RLXLinux/rlxlinux-v2.2/users/boa/system/set_bridge.c +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/net/rtl/features/rtl_features.c +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/net/rtl/features/rtl_ps_hooks.c +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/Kconfig +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd.h +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_sme.c +- /trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_ioctl.c +* DESCRIPTIO +- this feature is disabled by default +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/01, jerryko, new feature + +* DIGEST: +- RTS init rate adjust base on the tx rate of each client.(default no enable)(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_hw.h +- 8192cd_hw.c +- 8192cd_sme.c +- 8192cd_ioctl.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/12/01, jerryko, new feature + +* DIGEST: +- When autoChannel,vap up after root scan finish in RTL8672 platform.(patch from + sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_osdep.c +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/30, victorlo, bug fix [fatal] + +* DIGEST: +- AutoChannelSelection is not working in 5G Band when DFS is not built + +* TEST METHOD: +- select auto channel in 5G Band with verison without DFS built, + channel will remain at ch36. + +* ROOT CAUSE: +- Since there is a wrong comparison sentence to stop site survey for + AutoChannelSelection, site survey becomes infinite loop and no response + for the channel selection when DFS is not built. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- AutoChannelSelection is not working in 5G Band when DFS is not built, + caused by a wrong comparison sentence. When DFS is built, + AutoChannelSelection is still working though channel number can fit + the case and will not lead to infinite loop. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/30, jerryko, feature refined + +* DIGEST: +- update missing patch and fix compiler error in RTL8672 platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_header.h +- 8192cd_ioctl.c +- 8192cd_mp.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_psk.c +- 8192cd_rx.c +- 8192cd_sme.c +- 8192cd_tx.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/30, jerryko, feature refined + +* DIGEST: +- modify config and add new rtl8672_port.c for RTL8672 platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h +- Makefile +- rtl8672_port.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/29, button, new feature + +* DIGEST: +- Antenna diversity for 92D dual band concurrent mode + +* TEST METHOD: +- Check antenna switch pin by josephchang + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_hw.c +- 8192cd_proc.c +- 8192cd_sme.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Enable steps: + 1. Enable antenna diversity in menuconfig + 2. Define CONFIG_ANT_SWITCH, SW_ANT_SWITCH, HW_ANT_SWITCH in 8192cd_cfg.h +- Antenna switch control: + WLAN0: 5G, GPIO4/5 (SW), WLAN1: 2.4G, ANT_SEL_P/N (HW) + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/11/28, davidhsu, feature refine + +* DIGEST: +- Refine some codes for client mode + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- Rev: 8122 + 8192cd_hw.c, 8192d_hw.c, 8192cd.h, 8292cd_tx.c, 8192cd_sme.c + +* DESCRIPTION: +- In rtl8192cd_init_hw_PCI(), call clnt_ss_check_band() before + PHY_IQCalibrate() when SMSP is used. This is to make sure the + mib phyBandSelect and channel is sync before IQK. +- During ss_timer(), call IQK when band is switched. +- Modify IQK_92D_5G_n(), IQK_92D_2G() to not allow do IQK twice +- In start_clnt_lookup(), set is_40m_bw = 0 because it may set bandwidth + to 40M after site survery end. +- In rtl8192cd_tx_clientMode(), do not call mac_clone_handle_frame() + to do mac clone when wps is enabled. +- In OnBeaconClnt_Bss(), modify the check bandwidth algorithm that + it will cause miss checking in some cases. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/25, EricYeh, Bug Fix + +* DIGEST: +- 92C+92D nfjrom, 2.4G can ONLY TX at channel 1 + +* TEST METHOD: +- Use MP TOOL do open>close>open, 2.4G channel switch failed. + +* ROOT CAUSE: +- Because 92C set to 5G after second lauch from MP Tool. + (First launch from MP TOOL is OK) + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- In 8192cd_open(), check if (IC_TYPE == 92C/88RE) && (in MP mode), + config band setting as [BGN][2.4G][CH_1]. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/22, button, feature refine + +* DIGEST: +- Update RSSI, signal quality in /proc/wlan0/mib_bssdesc + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- DNI request this feature under client mode + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/22, button, bug fix + +* DIGEST: +- Manual EDCA doesn't apply to beacon I.E. + +* TEST METHOD: +- Manual assign EDCA parameters by mib, then check beacon of DUT + +* ROOT CAUSE: +- code merge mistake (ver.2557) + +* MODIFIED FILES: +- 8192d_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/22, jerryko, feature refined + +* DIGEST: +- power tracking sync with SD9 patch + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- 8192cd_osdep.c + +* Description +- remove priv->pshare->ThermalValue=0 in rtl8192cd_stop_sw() function. + clean pshare->ThermalValue in rtl8192cd_init_sw() function for root + interface. + +--------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/18, Brian, New feature + +* DIGEST: +- Enable accounting WIFI client to drop connection + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd.h +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_cfg.h +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_ioctl.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_proc.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_security.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_sme.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/ieee802_mib.h + +* Description +- ieee802_mib.h updated, VERSION is updated to 10 +- Set act_enabled through flash MIB MIB_ACCOUNT_ENABLED. + Whan accounting is enabled, there are two conditions make WIFI client dropped. + a.WIFI client keep connect exceed miniutes defined by acct_timeout_period + b.The summation of TX and RX is lower than acct_timeout_pthroughput (will be refined) + c.Both acct_timeout_period and acct_timeout_pthroughput are zero turns off accounting. + +--------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/18, EricYeh, bug fix + +* DIGEST: +- Fix MP 92D Power Tracking + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- 8192cd_mp/c + +* Description +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/16, EricYeh, New Feature + +* DIGEST: +- Dynamically switch LNA for 97D High Power MP + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_mp.c +- 8192cd_rx.c + +* Description +- Add tag 'MP_SWITCH_LNA' to package this patch (Default is turn off) +- The algorithm is, when signal level of received packets > Threshold, + turn LNA OFF, otherwise turn LNA ON. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/16, EricYeh, refine + +* DIGEST: +- Refine DPK procedure of Normal Mode + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_sme.c + +* Description +- Remove DPK from swchnl() + > Site Survey has no need to do DPK. + > DPK shall be performed AFTER IQK finished. +- Do DPK at rtl8192cd_init_hw_PCI() + > In AP mode, just do DPK once at the beginning. +- Do DPK at OnAssocRsp() + > In Client mode, do DPK when connected to an AP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/16, EricYeh, refine + +* DIGEST: +- SYNC new DPK function from driver of WiFi station team. + +* ROOT CAUSE +- N/A + +* TEST METHOD: +- N/A + +* MODIFIED FILES: +- 8192d_hw.c + +* Description +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/16, joeylin, refine + +* DIGEST: +- for 8196D/8197D: modify GPIO wake up pins for power saving feature. + +* ROOT CAUSE +- GPIO pins used for 8196D/8197D are different. + +* TEST METHOD: +- after the modification, STA can connect to 2G/5G AP with power saving enabled. + +* MODIFIED FILES: +- 8192cd_hw.c + +* Description +- modify request_irq_for_wakeup_pin() and gpio_wakeup_isr() of 8192cd_hw.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/10, victorlo, new feature + +* DIGEST: +- Add WMM-PS client mode support for logo test + +* TEST METHOD: +- Enable WMM-PS in client mode by mib, and also set different power-save + state by mib for each TID(VO/VI/BE/BK), client mode will announce its + PS capability when it's connecting to a WMM-PS AP. +- Dynamically change power management mode in client mode by mib, packet + with power management bit set will be delivered after then, and WMM-PS + test can be started. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_headers.h +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_rx.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- ieee802_mib.h +- wifi.h + +* DESCRIPTION: +- WMM-PS client mode support for logo test only, since our solution will not + change to power-save mode automatically while in-use. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/09, jerryko, bug fix + +* DIGEST: +- Fix receive not correct large size mp packets, system not stable when + RX_BUFFER_GATHER enable(patch from sd9) + +* TEST METHOD: +- APUT mp mode +- run packet size > RX BUFFER SIZE +- APUT system not stable + +* ROOT CAUSE: +- When receive not correct AMSDU, reuse=1 and pointer to last fragment. But + first fragment not reused. + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- When receive not correct AMSDU, reuse=1. Then change to reuse=0 and + free first fragment. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/11/04, jason, Enhance the PCIe Clock swing amplitude + +* DIGEST: +- Enhance the PCIe Clock swing amplitude and clock signal quality + +* TEST METHOD: +- Run long-time test to verify the PCIe bus error status and the status is very well. + +* ROOT CAUSE: +- Hardware layout issue. + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- Follow the RD center suggestion. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/10/31, button, bug fix + +* DIGEST: +- Fix buf of configuration read error after changing band setting + +* TEST METHOD: +- Build 96c+92d DMDP firmware, configure to 2.4G/5G Concurrent Mode, +- change to selective Mode, apply change, then change back to concurrent +- mode. Read chip id by command: ew b8b0100c 0, dw b8b10000, ew b8b0100c 1 +- dw b8b10000, Configuration read of function 1 will fail. + +* ROOT CAUSE: +- MAC_PHY_CR setting is wrong. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Follow pisa's suggestion. set MAC_PHY_CR to 0xfc/0xf1/0xf3 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/27, button, bug fix + +* DIGEST: +- Fix tx hang up when using WSend flood multicast packets through ethernet + +* TEST METHOD: +- let a client connect to DUT, flood multicast packets from LAN PC by tool +- "WSend", check the field "check_cnt_tx" of "proc/wlan0/stats" + +* ROOT CAUSE: +- DMA level of mgt queue and high queue are the same(low), when a lot of +- traffic generated from high queue, tx packet buffer will be exhausted, and +- packets in mgt queue will be pending for more than 1 min, then triger tx +- hang up reset. + +* MODIFIED FILES: +- 8192cd_osdep.c +- 8192cd_sme.c +- 8192c_reg.h +- 8192cd_tx.c +- 8192cd.h + +* DESCRIPTION: +- when beacon Tx ok, and did put packets into high queue, set HIQ_NO_LMT_EN=1, +- change it back to 0 at the time next beacon early interrupt occurs. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/27, Button, Bug fix + +* DIGEST: +- fix bug of connecting to 92d fail at Link state L1 + +* TEST METHOD: +- Enable PCIe power saving on 8198+8192c+8192d +- Let 92d enter L1 state, use a notebook connect to 92d AP + +* ROOT CAUSE: +- rx packet overflow, such that fw can not parsing packet any more +- set rx filter before entering L1, relieve fw loading + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_osdep.c + +* DESCRIPTION: +- 1. do not enable ASPM of NIC +- 2. set rx filter(accept auth, probe request only) +- 3. wait till tx/rx packet buffer empty then enter L1 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/26, chuangsw, bug fix + +* DIGEST: +- Fix 92D client mode coredump issue when enable WIFI_11N_2040_COEXIST. + +* TEST METHOD: +- Use 92d client mode (2.4 + 5G) and enable 20/40 coexist function. +- Connect to a 20/40 coexist ap. +- It will crash in 5 mininute. + +* ROOT CAUSE: +- We will parse DSSET_IE to get the channel information in beacon + and this should be in bg ap. +- Then we set timeout value correspond to the channel in DSSET_IE. +- If we get a channel number > 14 (in 5G), it will overlap the structure + bg_ap_timeout_ch[14] so that lead to crash. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Ensure channel number is <= 14 when we parse DSSET_IE. + +---------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/26, chuangsw, bug fix + +* DIGEST: +- Fix wifi logo 5.2.35 T05 fail. + +* TEST METHOD: +- Use client mode to assoc with 20/40 MHz capable AP. +- Start ping to AP (ping <AP IP> -t -l 10000). +- Use another 11n 20MHz only STA to connect to AP. +- Then AP will go "20 MHz protection mode" (MODE_2). +- Check if there is RTS before the ping req. If yes, then fail. + +* ROOT CAUSE: +- We should only enable ht_protection mode when detect MODE_1 or MODE_3. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Correct the condition when enable ht_protection. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/26, chuangsw, Feature Refine + +* DIGEST: +- For wifi logo 5.2.31 requirement in testplan 2.0.12. + +* TEST METHOD: +- Follow wifi logo testplan 2.0.12 5.2.31 +- We will fail on ckecking packet's prioirty. + +* ROOT CAUSE: +- Original we just change EDCA para when AP's ACM bit is set. +- But this will fail since new testplan may check packet's priority. + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_tx.c + +* DESCRIPTION: +- change packet's priority when detect AP's ACM bit on. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/10/26, button, bug fix + +* DIGEST: +- Fix client mode connect to Cisco 541N AP fail reported by Pegatron + +* TEST METHOD: +- Configure security of AP541N to WPA2 AES. Let a DUT(92c client) connect +- to 541N, reboot AP, DUT may connect to AP fail. + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_security.c + +* DESCRIPTION: +- After our client associate to AP, it will wait AP's EAPOL 4-1, if wait 4-1 +- timeout, change to roaming state. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/24, davidhsu, Feature Refine + +* DIGEST: +- Add Tx early mode for 8192d driver + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: + Revision: 7718 +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_ioctl.c +- 8192cd_proc.c +- 8192cd_sme.c +- 8192cd_util.c +- 8192cd_tx.c + +* DESCRIPTION: +- When 'CONFIG_RTL_TX_EARLY_MODE_SUPPORT' is enabled from + menuconfig, it will define TX_EARLY_MODE in 8192cd_cfg.h. +- A mib 'em_enable' is defined to control Tx eary mode is enabled + or not. It is enabled is default. +- When 'em_enable' is enabled, it will detect if Tx throughput is > 48M + and last for 3 sec. If yes, it will put Tx packet into sw queue. +- When sw queued packet number is > 30 or timeout, it will dequeue + it and put it into tx descriptor. +- In Tx decriptor, it needs fill 8 byte more data for eary mode. It + includes the length for each subsequent pending packet. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/21, EricYeh, Feature Refine + +* DIGEST: +- Use compile tag '_DEBUG_RTL8192CD_' to pack debug command 'reg_dump tx' + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_ioctl.c + +* DESCRIPTION: +- Use compile tag '_DEBUG_RTL8192CD_' to pack debug command 'reg_dump tx' + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/20, button, bug fix + +* DIGEST: +- Fix client mode connect to Cisco AP fail reported by Pegatron + +* TEST METHOD: +- Let a DUT(92c client) connect to Cisco AP, power off and power on AP, + DUT may connect to AP fail and do nothing. + +* ROOT CAUSE: +- In Pegatron SDK, there's a timer checking DUT status periodic, + if no link, it will trigger a site survery request from upper layer. + At the time receiving ss req. If DUT is in authentication phrase, + after doing site survy, it will do nothing, won't try to connect to + original AP + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- After doing site survery, if DUT is in authentication phrase, change to + roaming state. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/20, EricYeh, Feature Refine + +* DIGEST: +- Add new Gain Table for 92D-intPA & 96D-intPA-USB + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- /data_92d/radio_a_intPA_GM_new1.txt +- /data_92d/radio_a_intPA_GM_new.txt +- /data_92d/radio_a_intPA_new.txt +- /data_92d/radio_b_intPA_GM_new1.txt +- /data_92d/radio_b_intPA_GM_new.txt +- /data_92d/radio_b_intPA_new.txt +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- Add two compile options in 8192cd_cfg.h + 'RTL8192D_INT_PA_GAIN_TABLE_NEW' & 'RTL8192D_INT_PA_GAIN_TABLE_NEW1' +- 'RTL8192D_INT_PA_GAIN_TABLE_NEW' if for intPA (both Non-USB & USB power) + If 'RTL8192D_INT_PA_GAIN_TABLE_NEW' defined, + >> Decrease the Tx Power Gap between indices 39 & 40. (original is 2dBm) +- 'RTL8192D_INT_PA_GAIN_TABLE_NEW1' if for USB power only. + If 'RTL8192D_INT_PA_GAIN_TABLE_NEW1' defined, + >> Decrease the Tx Power Gap between indices 39 & 40. (original is 2dBm) + >> Incresde Tx Power level + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/19, chuangsw, feature refine + +* DIGEST: +- Refine client mode dhcp renew patch. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_be_ext.c +- 8192cd_tx.c +- linux-2.6.30\net\bridge\br_input.c +- linux-2.6.30\net\bridge\pocket_filter.c +- linux-2.6.30\net\bridge\Makefile +- linux-2.6.30\drivers\net\rtl819x\rtl_nic.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/19, Plus, refine + +* DIGEST: +- WPS client mode feature refine + +* TEST METHOD: +under client mode that already associated to some AP +1.start WPS and 2.immediately stop WPS, +3.after WPS be cancelled client must recovery it's assoc to orignal AP + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_ioctl.c - revision :7792 + +* DESCRIPTION: +-n/a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/18, Plus, bugfix + +* DIGEST: +- under client mode resolved IOT issue with MarvellAP(WPS1 test bed) + +* TEST METHOD: +- run wps1.0 test plan 5.1.1 will fail +- under client mode do wps with Marvell AP (5G/channel36/WPA+TKIP ) will fail + +* ROOT CAUSE: +- after WPS handshake done MarvellAP should sent deauth to STA but it haven't + so it's sta_info is not on correct state + +* MODIFIED FILES: +- 8192cd_osdep.c - revision :1.127.2.4 + +* DESCRIPTION: +- when STA run close() function it send deauth but not disassoc , + let Marvell's sta_info correct and process 4-way handshake correct + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/18, jerryko, feature refine + +* DIGEST: +- 1.wifi_specific change default value to 2;CHECK_CRYPTO disable + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_ioctl.c +- 8192cd_sme.c + +* DESCRIPTION +- 1.some customers use original wifi_specific default value=0 to fail wifi + logo test. So change to 2. +- 2.CHECK_CRYPTO will cause tx hang in customer test. So CHECK_CRYPTO disable. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/17, button, feature refine + +* DIGEST: +- Refine auto channel selection + +* TEST METHOD: +- Tested by BB tesm Wilson + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h -revision: 7771 +- 8192cd_sme.c -revision: 7771 + +* DESCRIPTION +- Scan channel with 20M/40M BW respectively, select clear channel by + estimated traffic + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/17, chuangsw, bug fix/feature refine + +* DIGEST: +- 1. When client mode choose 2.4+5G, it will disable cck rate +- even connect with a "b" only AP. +- 2. Refine rx gain tracking table. + +* TEST METHOD: +- Set client mode to 2.4+5G and connect to b only AP. +- From sniffer we will see client mode always use 1M to send packet. + +* ROOT CAUSE: +- When client mode choose 2.4+5G, it's OPMODE is ABGN. +- In add_RATid(), we will disable cck rate if OPMODE & A. +- In this case, OPMODE is ABGN so cck rate is disabled and bitmap is wrong. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c - revision : 1.7 + +* DESCRIPTION: +- 1. In AP mode, we can disable cck rate by checking network type include A or + not. +- In client mode, we should not only check network type but also check + phyband. +- 2. Refine rx gain tracking table by BB Jenyu's suggestion. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/10/17, jerryko, bug fix + +* DIGEST: +- fix 92c high power board can't tx beacon (0x810 rollback to original value) + +* TEST METHOD: +- use 92c high power board, and the board can't tx beacon + +* ROOT CAUSE: +- bbp team offer new 0x810 reg value will cause high power board issue + +* MODIFIED FILES: +- data/PHY_REG_1T_n_hp.txt +- data/PHY_REG_2T_n_hp.txt + +* DESCRIPTION: +- bbp offer new 0x810 reg value check in at version 7584 svn trunk. + The value used to solve IOT issue when client mode with broadcom 5g ap. + But the value cause 92c high power can't tx beacon and 92d high power + can tx beacon in 97d+92c+92d (high power)platform. Therefore, in 92c high + power board 0x810 register value rollback to original value 0x10005388. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/14, jerryko, bug fix + +* DIGEST: +- fix rx reorder to cause tp drop in client mode.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- In client mode, broadcast packets and unicast packets use different seq number. + But rx reorder method in client mode don't separate broadcast and unicast. + If broadcast seq larger than win start+win size, win start will move to + broadcast seq-win size. Then unicast seq rx smaller than win start, the + packet will be dropped. + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- In client mode, rx broadcast packets don't check reorder. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/07, EricYeh, bug fix + +* DIGEST: +- Fix NDSi Client Connection Issue with 92D. + +* TEST METHOD: +- Use NDSi to connect our AP with connection setting nember 1 2 3, + it will indicate the test result FAIL with error code 051300-2. + (Connection setting 4 5 6 is for advanced, + and these connections can NOT duplicate this issue.) + +* ROOT CAUSE: +- Beacause when NDSi send AUTH frame with rate=2M, + 92D still response ACK with rate=1M, + this behavior will cause many Retries from NDSi. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Open Response Rate 2M for 92D chip. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/10/04, chuangsw, bug fix + +* DIGEST: +- Devices under our client mode AP will get DHCP renew fail. + +* TEST METHOD: +- Set the AP's DHCP lease time to 1 min. +- Set our AP to client mode and be DHCP client. +- Connect client mode AP with a PC. +- After 1 min, PC will DISCOVER AP again. + +* ROOT CAUSE: +- When helf time of DHCP lease time, PC will send DHCPREQ (ciaddr is its IP). +- When AP receive this DHCPREQ, it will send DHCPACK in unicast since ciaddr + is not zero. BTW, the da of DHCPACK is PC's MAC addr. +- But AP's assoc list did not have PC's mac addr so it will drop DHCPACK. +- PC didn't receive DHCPACK so finally it will timeout and send DHCPDISCOVER + after lease time. + +* MODIFIED FILES: +- linux-2.6.30\net\bridge\br_input.c +- linux-2.6.30\net\bridge\pocket_filter.c +- linux-2.6.30\net\bridge\Makefile +- linux-2.6.30\drivers\net\rtl819x\rtl_nic.c +- linux-2.6.30\drivers\net\wireless\rtl8192cd\8192cd_br_ext.c +- linux-2.6.30\drivers\net\wireless\rtl8192cd\change.txt + +* DESCRIPTION: +- Modify the DHCPREQ from PC as follow: +- 1. Add option "DHCP_REQUEST_IP" in DHCPREQ if it's ciaddr is not empty. +- 2. Set the ciaddr to zero. +- Since we will enable DHCP broadcast flag and ciaddr is empty, the AP + will send DHCPACK in broadcast so that PC will receive. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/10/03, victorlo, new feature + +* DIGEST: +- Add and refine 3 tables for clients and connections statistics + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_psk.h +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- Add station info (rx rate, auth mode, enc mode) reporting to webpage +- Add 2 tables for connection statistics, including connects and rejects +- http://dtdinfo1.realtek.com.tw:8080/~sub1/mantis-cn/view.php?id=11222 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/30, EricYeh, Feature Refine + +* DIGEST: +- Fix MP error: Unable to open WLAN0 + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/30, hf, refine + +* DIGEST: +- make 1.5.5 toolchain happy when wapi enabled + +* ROOT CAUSE + +* TEST METHOD: + +* Description + +* MODIFIED FILES: +- 8192cd_tx.c +- wapiCyrpo.h + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/29, jimmylin, feature add + +* DIGEST: +- Add LED type 15 for 8192C (Link, Tx/Rx data) + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/29, hf, refine + +* DIGEST: +- Refine for WAPI client aggregation and tx rate + +* ROOT CAUSE +- when WAPI enabled, legcy check mechanism will take the client as +- legcy client and mask the mcs rate + +* TEST METHOD: + +* Description +- set mask_mcs_rate = 0 when wapi used + +* MODIFIED FILES: +- 8192cd_sme.c + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/29, hf, bug fix + +* DIGEST: +- Fix DemoBoard(96c+92c) as WAPI client can not connect to Demo AP + +* ROOT CAUSE +- WAPI AE Challenge is changed when RecvUnicastKeyAggrementRequest +- sometimes we using the Next Challenge (for unicast key update) to +- generate the Response which will be discarded by AP due to AE Challenge +- no consistent + +* TEST METHOD: +- Build WAPI image +- one as WAPI client, one as WAPI AP +- Connect client to AP + +* Description +- Add waiAEChallangeNext to record the AE Challenge for key update + +* MODIFIED FILES: +- 8192cd_util.h +- wapi_wai.c +- wapi_wai.h + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/28, jerryko, feature refined + +* DIGEST: +- IOT_EDCA_switch function support new TXOP/CW values for 96D/97D platform + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- For performance issue, use lower TXOP/CW in 96D/97D platform + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/28, chuangsw, bug fix + +* DIGEST: +- Fix 92D 5G rssi drop 10% by rx_gain_tracking_92D. + +* TEST METHOD: +- 1. Use a client to connect to our 92d AP in 5G. +- 2. Run chariot and dump rssi, it will drop 10% about 30s. + +* ROOT CAUSE: +- When current ther > ther_mib, it will get wrong index mapping of rx gain + table, so RF Reg0x3c will write an unpredictable value. + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- Restrict the condition to trigger rx_gain_tracking_92D only when current + ther < ther_mib. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/28, jerryko, new feature + +* DIGEST: +- client mode support txop enlarge dynamically and update mac/bbp parameters + +* TEST METHOD: +- 1.use 96c+92d client mode and test netgear 3700v3 and netgear 3800 +- 2.use chariot +- 3.when enable txop enlarge, the tp will increase 5~8Mbps + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- data/PHY_REG_1T_n_hp.txt +- data/PHY_REG_2T_n_hp.txt +- data/PHY_REG_1T_n.txt +- data/PHY_REG_2T_n.txt +- data/PHY_REG_1T.txt +- data/PHY_REG_2T.txt +- data_92d/MACPHY_REG.txt +- data_92d/PHY_REG_n.txt +- 8192cd_sme.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- decide txop enlarge by checking tx aggregation ability using ADDBA_ready variables +- update mac register by mac team suggestion + (i)0x63c~0x63f value change to 0x0c0c0808 + (ii)0x51a change to 0x0f only in client mode and Broadcom AP and 5g band +- update bbp register by bbp team suggestion + (i)0x810 value change to 0x10000330 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/09/26, jimmylin, Bug fix [FATAL] + +* DIGEST +- Fix system crash when client mode received disassociation request + +* TEST METHOD +- Let client mode connect to our turnkey AP. Apply change on turnkey AP will + cause client mode crash and reboot. + +* ROOT CAUSE +- We do rtl8192cd_rx_isr() in clnt_ss_check_band() will cause infinite + recursive calls in rtl8192cd_rx_isr(). + +* MODIFIED FILES +- 8192cd_sme.c - revision: 7569 + +* DESCRIPTION +- Adding rtl8192cd_rx_isr() in clnt_ss_check_band() is in order to fix client + mode connection failed with 5G AP. The previous issue is: + [2011/09/07, EricYeh, Bug Fix, DIGEST: Fix 92D(SMSP) Client Mode connection + issue.] +- I removed the adding rtl8192cd_rx_isr() in clnt_ss_check_band(). Instead, + I add a channel check in collect_bss_info() to filter a beacon without + channel information but the current channel is in 2.4G because beacon in + 2.4G should carry channel information. Then we will get correct channel + information and connection will success. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/09/26, chuangsw, Bug fix + +* DIGEST +- Fix HD7 DHCP cannot get IP. + +* TEST METHOD +- Use HD7 to connect to our AP, sometimes it may get IP fail. + +* ROOT CAUSE +- HD7 ask AP to reply DHCP OFFER/ACK in broadcast. +- If AP send DHCP OFFER/ACK when HD7 go to sleep, it may loss. + +* MODIFIED FILES +- 8192cd.h +- 8192cd_hw.c +- 8192cd_ioctl.c +- 8192cd_tx.c + +* DESCRIPTION +- Add a mib "bcast_to_dzq" to turn on/off the patch as follow +- 1. Put all of the broadcast packets to dzqueue +- 2. Enlarge ATIMWND(0x55a) from 0x05 to 0x0a + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/09/26, Button, Bug fix + +* DIGEST +- Fix client mode connect to 2.4G AP fail issue + +* TEST METHOD +- Set DUT to client mode, band 2.4G+5G (ABGN), and try co connect to an 2.4G +- AP + +* ROOT CAUSE +- Connect fail becasue channel list do not contain 2.4G, it will not do site + survery in 2.4G channel + +* MODIFIED FILES +- rtl8192cd/8192cd_util.c + +* DESCRIPTION +- Sync channel list condition with kernel 2.4 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/26, jerryko, feature refine + +* DIGEST +- Check phyband and channel match or not in 8192cd_open function + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- To avoid phyband and channel mismatch to cause wlan not stable. When wlan + up, driver check phyband and channel. If phyband and channel mismatch, + driver will print error message and wlan interface down. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/22, button, feature refine + +* DIGEST +- Add range check to avoid getting wrong tx rate + +* TEST METHOD: +- Netami reports turn on rssi_dump, tx rate show as below: +- [1] 61% tx 0s rx 24 (ss 61 59)(FA 6)(CCA 12)(DIG 0x38)(TP 0,0) +- [1] 64% tx 0s rx 24 (ss 64 62)(FA 13)(CCA 20)(DIG 0x38)(TP 0,0) + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Hal8192CDMOutSrc.c - revision: 7515 + +* DESCRIPTION: +- Array size of dot11_rate_table is 12, but access index may large than 12 if +- firmware reports wrong rate index, add range check to avoid access wrong +- rate. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/22, jimmylin, feature add + +* DIGEST: +- Add Tx 2-path for 1T MCS, OFDM, and CCK +- Add Tx beamforming for 92C + +* TEST METHOD: +- Test in channel emmulator by baseband team + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- revision: 7514 +- 8192cd_hw.c +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192d_hw.c +- Hal8192CDMOutSrc.c +- ieee802_mib.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/22, EricYeh, Feature Refine + +* DIGEST: +- Tx Power Debug Command support "ADD_TX_POWER_BY_CMD" + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_ioctl.c + +* DESCRIPTION: +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/21, jerryko, bug fix + +* DIGEST: +- In beq/bkq/viq/voq timer function, fix priv=NULL pointer + +* TEST METHOD: +- If get priv member in timer function, it maybe cause kernel crash + +* ROOT CAUSE: +- Because priv=(struct rtl8192cd_priv *)skb->dev->priv, but skb=null. Then + priv point unknown address + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- Get skb and check skb not null. Then priv point skb->dev->priv. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/21, button, feature refine + +* DIGEST: +- Sync EDCCA setting for both 92C and 92D + +* TEST METHOD: +- Compare beacon tx ok couner in open space + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_hw.c +- Hal8192CDMOutSrc.c + +* DESCRIPTION: +- do not change EDDCA setting of MAC, just control in BB register. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/19, jerryko, bug fix + +* DIGEST: +- When 92d client mode only, fix max_fw_macid number check to avoid fw ra not + work + +* TEST METHOD: +- 1.choose client only and auto rate +- 2.connect AP and run chariot +- 3.ib 485 and ob 485 rate_idx +- 4.ib 485 and check 485 value change or not + +* ROOT CAUSE: +- 92d fw will check max_fw_macid field. When 92d client mode only, + max_fw_macid will set 1. Fw only control macid0 and can't control macid1 + rate. The TP sometimes drop to low during a period time when AP rx not good. + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- check max_fw_macid with total_sta_number(NUM_STAT)+1. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/16, jimmylin, bug fix [MINOR] + +* DIGEST: +- Fix wrong parentheses in rtl8192cd_tx_clientMode() to get NAT2.5 shortcut + work + +* TEST METHOD: +- + +* ROOT CAUSE: +- The position of parentheses are wrong in rtl8192cd_tx_clientMode() + therefore TX flow of client mode can not go into shortcut of NAT2.5. + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 7433 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/16, button, new feature + +* DIGEST: +- RTL8198 +92C+92D PCIE Power Saving feature. + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - revision : 7428 +- 8192cd_hw.c - revision : 7428 +- Kconfig - revision : 7428 + +* DESCRIPTION: +- 8198 disable this feature at default, beacause wake up pin of mini PCI + doesn't connect to 8198 GPIO + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/15, button, feature refine + +* DIGEST: +- Move dynamic mechanism to Hal8192CDMOutSrc.c + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - revision : 7417 +- 8192cd_headers.h - revision : 7417 +- 8192cd_sme.c - revision : 7417 +- 8192cd_util.c - revision : 7417 +- 8192d_hw.c - revision : 7417 +- 8192cd_hw.c - revision : 7417 +- Makefile - revision : 7417 +- Hal8192CDMOutSrc.c - revision : 7417 + +* DESCRIPTION: +- Move dynamic mechanism to Hal8192CDMOutSrc.c according to BB tesm's + suggestion + +----------------------------------------------------------------------------- + +- 2011/09/14, EricYeh, Feature Refine [MAJOR] + +* DIGEST: +- 92D EXT HIGH Power Fix - Add support of 92D extPA + HIGH POWER + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_ioctl.c +- 8192cd_mp.c +- data_92d/radio_a_n_92d_hp.txt +- data_92d/radio_b_n_92d_hp.txt +- Kconfig + +* DESCRIPTION: +- Do NOT apply settings of registers and mibs of 92C to 92D +- Set limit of HIGH Power Index = 63 +- Modify Gain Tables for 92D High Power +- Make 92D can also apply HIGH_POWER_EXT_PA in menuconfig +- Add support of data_92d\radio_b_n_92d_hp.txt + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/14, EricYeh, Bug Fix [MAJOR] + +* DIGEST: +- Fix 92D(SMSP) Client Mode connection issue. + +* TEST METHOD: +- Try to connect an AP in channel 36, + there may be probability to be "connection timeout" + +* ROOT CAUSE: +- WiFi Driver listen Beacons to verify the channel of target AP. + But if the Beacons of target AP NOT contain channel information, + WiFi Dirver will use the current channel to connect. + But, this Beacon may be received from the previous channel setting. + (Ex, set channel 36 + > start change to channel 1 + > RF receive Beacons(of Target AP) from channel 36 + > change to channel 1 done + > Handle Beacons(of Target AP)received from channel 36 before + > Consider Target AP is in channel 1) + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- In condition of "92D Client Mode - SMSP", + Before switching Band (2.4G <-> 5G), + >> stop receiving management frames and try to clear Rx queue. + After switching Band done, + >> restore receiving management frames. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/13, jerryko, new feature + +* DIGEST: +- pstat tx_pkts only calculate data and no M2U in RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8274 + +* MODIFIED FILES: +- 8192cd_tx.c +- 8192cd_util.h +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/09, victorlo, feature refine + +* DIGEST: +- Add dynamic EDCA switching for BE to comply Wifi logo test + +* TEST METHOD: +- WiFi approved 11n test plan 4.2.22 T06 + +* ROOT CAUSE: +- Sometimes test case failed caused by our BE tx stream depressing VI rx + stream from testbed, with unknown reason. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_ioctl.c +- 8192cd_rx.c +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- With dynamic EDCA switching for BE, by wlan mib, can help customers pass + logo test more smoothly once if they hit the same issue. Default disabled. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/09, chuangsw, bug fix [MAJOR] + +* DIGEST: +- Fix client mode connection issue when co-exist is enabled in AP + +* TEST METHOD: +- Config DIR-615's BW to auto 20/40 mode. +- Config our AP to client mode and associate with DIR-615 in open space. +- Ping <AP> and we will find lot of ping loss. + +* ROOT CAUSE: +- DIR-615 will auto enable the 20/40 coexist future. +- In r7084's modification, we check the AP's BW based on the + _HTCAP_SUPPORT_CH_WDTH_ bit in HT Capibility IE. +- In this case, _HTCAP_SUPPORT_CH_WDTH_ = 1, so we think that AP is 40M, +- But AP is 20M since it goes to 20/40 coexist mode, so _HTIE_STA_CH_WDTH_=0 +- After assoc, our CurrentBW=0 (20M) but we think AP'BW=1 so we will go + ReConn in OnBeaconClnt_Bss(). + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_sme.c + +* DESCRIPTION: +- Not only check _HTCAP_SUPPORT_CH_WDTH_ but also _HTIE_STA_CH_WDTH_ + to decide AP's BW. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/09, Button, Bug Fix [MAJOR] + +* DIGEST: +- Probe response content error under 92D DMDP mode. + +* TEST METHOD: +- Let 96c+92d enter power saving mode, sniifer the probe response sent +- by firmware, the last part may be a lot of zero when WPS is enabled. + +* ROOT CAUSE: +- when we activate fw offlaod, we have to put beacon and probe response to +- that space reserved for bacon, in worse case, we need 8 pages, but packet + buffer +- reserved for beacon queue in DMDP is 4 page only. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Refine tx packet buffer under DMDP mode. +- move 3 pages from public queue to beacon queue. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/08, jerryko, new feature + +* DIGEST: +- Add USE_TXQUEUE flag in RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8276 + +* MODIFIED FILES: +- ieee802_mib.h +- 8192cd_util.c +- 8192cd_tx.c +- 8192cd_proc.c +- 8192cd_osdep.c +- 8192cd_ioctl.c +- 8192cd_util.h +- 8192cd_cfg.h +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/08, jerryko, new feature + +* DIGEST: +- Add LOW_TP_TXOP flag in RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_util.c +- 8192cd_ioctl.c +- 8192cd_hw.c +- 8192cd_cfg.h +- 8192cd.h + +* DESCRIPTION: +- Support dynamically adjust TXOP in low throughput feature + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/08, jerryko, new feature + +* DIGEST: +- Add ETHWAN Config.(default no enable)(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8277 + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/07, button, new feature + +* DIGEST: +- Enable RTL8196C+92D PCIE Power Saving feature. + +* TEST METHOD: +- 1. Configured "Enable PCIE power saving support" in menuconfig. +- 2. Check PCIe link status in proc/wlan0/stats +- 3. When AP is in L1 mode, it will wake up when a client attempts to connect + it. +- 4. Repeat the test in selective / concurrent mode. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - revision : 7319 +- 8192cd_headers.h - revision : 7319 +- 8192cd_hw.c - revision : 7319 +- 8192cd_tx.c - revision : 7319 +- 8192cd_rx.c - revision : 7319 +- 8192cd_sme.c - revision : 7319 +- 8192cd_osdep.c - revision : 7319 +- 8192d_reg.h - revision : 7319 +- Kconfig - revision : 7319 +- data_92d/rtl8192dfw_n.bin - revision : 7319 +* DESCRIPTION: +- Stop power saving command: iwpriv wlan0 stopps +- Power consumption +- 2.4G 2x2: L0: 0.25 mA, L1: 0.18 mA +- 5G 2x2: L0: 0.365 mA, L1: 0.313 mA +- 1x1 concurrent: L0: 0.256 mA, L1: 0.214 mA + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, feature refine + +* DIGEST: +- Modify USE_RLX_BSP using different kernel in RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8273 + +* MODIFIED FILES: +- 8192cd_cfg.h +- 8192cd_mp.c +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, feature refine + +* DIGEST: +- BR_SHORTCUT/RX_BUFFER_GATHER enable/disable by menuconfig in + RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8268 +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8269 + +* MODIFIED FILES: +- 8192cd_cfg.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, feature refine + +* DIGEST: +- update pcie reset procedure in RTL8672(CONFIG_RTL8672) platform.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8264 + +* MODIFIED FILES: +- 8192cd_osdep.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, bug fix + +* DIGEST: +- TP not stable when DWA-140 near with AP.(patch from sd9) + +* TEST METHOD: +- short distance, DWA-140 client with our AP. And the TP not stable + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8263 + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_hw.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, feature refine + +* DIGEST: +- should_forbid_Nmode function add alg==_NO_PRIVACY_ check.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8279 + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, bug fix + +* DIGEST: +- somtimes M2U funcion not work when multiple sta connect and disconnect AP.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8270 + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, feature refine + +* DIGEST: +- define macro SKB_MAC_HEADER/SKB_IP_HEADER to get mac/ip header in different + kernel version.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8272 + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, bug fix + +* DIGEST: +- Get_subframe function:skb copy len err when nextskb->len<rest_len.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8266 + +* MODIFIED FILES: +- 8192cd_rx.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, bug fix + +* DIGEST: +- LED GPIO State change when wlan down/up.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8265 + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/06, jerryko, bug fix + +* DIGEST: +- tx power tracking. ThermalValue not correct.(patch from sd9) + +* TEST METHOD: +- + +* ROOT CAUSE: +- Please reference http://cnsd6-bts.realtek.com/view.php?id=8260 + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/09/01, chuangsw, bug fix [MINOR] + +* DIGEST: +- Ignore basic rate checking. + +* TEST METHOD: +- Some clients cannot connect to our AP. + +* ROOT CAUSE: +- Client will get ASSOC_FAIL after check_basic_rate(). +- E.g. AP's basic rate set is {1M, 2M, 5.5M, 11M} + Client's basic rate set is {1M} + Client's basic rate set is not match of AP's basic rate set. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Originally check_basic_rate is for "bg" wifi logo test, and now we have + check that 11n wifi testplan does not have negative basic rate set test. +- It may cause some IOT issue if client's basic rate set is not perfect match + with AP's basic rate set. +- So now we remove check_basic_rate(). + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/30, button, bug fix + +* DIGEST +- Re-init Peer interface for 92D dual-mac dual-phy mode + +* TEST METHOD: +- In sercomm SDK, config 92d as DMDP, wlan0: client mode, wlan1: AP mode. + Let wlan0 coonnect to an encryption AP several times, input correct and + wrong key, there are 3 kinds of abnormal: + 1. wlan0 keeps Tx the same packet + 2. no interrupt in Wlan0 except beacon error + 3. Rx ok, but no Tx Packet in air + +* ROOT CAUSE: +- In DMDP mode, if anoter MAC is working, we do not close power, reset BB, MAC + MAC state machine error + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 7133 +- 8192cd_sme.c +- 8192cd_ioctl.c +- 8192cd_headers.h +- 8192cd.h + +* DESCRIPTION: +- When close one interface, if the MIB "peerReinit" is set, we also reinit + antoter interface + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/30, chuangsw, feature fine tune + +* DIGEST: +- Update PHY_REG 0xc8c + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- data/PHY_REG_1T_n_hp.txt +- data/PHY_REG_1T_n.txt +- data/PHY_REG_1T.txt +- data/PHY_REG_2T_n_hp.txt +- data/PHY_REG_2T_n.txt +- data/PHY_REG_2T.txt +- data_92d/PHY_REG_n.txt + +* DESCRIPTION: +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/26, chuangsw, feature fine tune + +* DIGEST: +- Check AP's BW capability in beacon for client mode + +* TEST METHOD: +- Let client mode connect to DIR-665 in 40MHz mode and ping OK. Switch AP's +bandwidth from 40MHz mode to 20MHz mode. Then ping will fail. +- Make client mode reconnect to AP manually, then ping OK. + +* ROOT CAUSE: +- When we switch DIR-665's bandwidth, it will not disconnect its clients. +And the re-initialization time of DIR-665 is too short for our client mode +to detect to do roaming. So our client mode will remain connection with +DIR-665 while it changes bandwidth, so traffic will fail. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Add checking AP's BW for client mode. AP's HT capability should +not change after it starts. If the BW changed, it means AP may +change some configurations and re-initialize. Then we should do +roaming. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/25, jimmylin, feature fine tune + +* DIGEST: +- Fine tune hangup reset flow for 92D dual-mac dual-phy + +* TEST METHOD: +- In 92D dual-mac dual-phy mode, if one interface needs to do silent reset + and does silent reset by itself only, the behaviors of both two interfaces + will become abnormal. + +* ROOT CAUSE: +- When 92D dual-mac dual-phy does silent reset, both interfaces need to do + silent reset. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 7057 + +* DESCRIPTION: +- In check_hangup(), if one interface needs to do silent reset, it will check + the other interface is up or not. If the other interface is up, will reset + the other interface, too. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/08/25, EricYeh, Feature Refine + +* DIGEST: +- Fix errors of setting OFDM Tx Power. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_ioctl.c + +* DESCRIPTION +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/08/25, EricYeh, Feature Refine + +* DIGEST: +- Redefine the setting of Tx Power according to the discussion. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- PHY_REG_PG_CE.txt +- PHY_REG_PG_FCC.txt +- radio_a_n_92d_hp.txt +- radio_b_n_92d_hp.txt +- TXPWR_LMT.txt +- TXPWR_LMT_CE.txt +- TXPWR_LMT_FCC.txt +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_ioctl.c +- 8192cd_mib.c +- 8192cd_mp.c +- 8192cd_proc.c +- 8192cd_sme.c +- Kconfig + +* DESCRIPTION +- Redefine the procedure about 'Power by Rate'. +- Redefine the procedure about 'Band Edge Limit'. +- Redefine the procedure about 'USB Power'. +- Add 'radio_a_n_92d_hp.txt' & 'radio_b_n_92d_hp.txt' for 92D EXT High PA. +- Add debug command 'iwpriv wlan%d reg_dump tx' + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/24, button, new feature + +* DIGEST: +- Power tracking of internal PA 92d + +* TEST METHOD: +- Verify by RF team Binson + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 7040 +- 8192cd_hw.h - revision: 7040 +- 8192d_hw.c - revision: 7040 + +* DESCRIPTION +- porting form NIC driver + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/19, victorlo, feature refine + +* DIGEST: +- Add maximum tx desc limit for both bands on 92D 1x1 concurrent mode + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_cfg.h +- 8192cd_hw.c +- 8192cd_mp.c +- 8192cd_osdep.c +- 8192cd_proc.c +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_tx.h + +* DESCRIPTION: +- When device is working as 92D 1x1 concurrent mode, and transmission + is not good in one of the bands, would cause lots of ethernet buffers + queuing in tx desc ring of this band. That will affect performance of + the one another band if the amount of tx desc is large. By modifing the + maximum number of available tx desc preventing the effect on the one + another. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/19, jimmylin, bug fix + +* DIGEST: +- Do re-order for multicast/broadcast packets from clients + +* TEST METHOD: +- Do ping test and send multicast from client to AP at the same time. Ping + delay time will vary because we do re-order for unicast but not for + multicast. + +* ROOT CAUSE: +- We didn't do re-order for multicast/broadcast. But packets from client are + actually unicast in 802.11 protocol. So we should do re-order for multicast + and broadcast. + +* BUG LEVEL +- Minor + +* MODIFIED FILES: +- 8192cd_rx.c - revision: 6998 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/08/09, EricYeh, Bug Fix + +* DIGEST: +- SYNC procedure of switching channel for MP and Noraml Driver. + +* TEST METHOD: +- Set the same power index of channel 149, + than comapred the tx power measured from MP-NFJROM and Normal Driver, + expected result is almost the same. (the diff shall be less than 1 dBm) + +* ROOT CAUSE: +- The behavior of switching channel of MP mode is not synchronize with + Normal Driver. + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION +- Synchronize the setting RF-REG 0xB when sitching channel in MP mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/08/08, button, feature refine + +* DIGEST: +- Add Japan 5G band (safety) spec support + +* TEST METHOD: +- Set different power level to make sure AP will stop tx (packet + and beacon) when there is a continuous signal over the specified + power level. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 6911 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/04, jimmylin, bug fix + +* DIGEST: +- Check AP's HT capability in beacon for client mode + +* TEST METHOD: +- Let client mode connect to DIR-665 in 40MHz mode and ping OK. Switch AP's + bandwidth from 40MHz mode to 20MHz mode. Then ping will fail. +- Make client mode reconnect to AP manually, then ping OK. + +* ROOT CAUSE: +- When we switch DIR-665's bandwidth, it will not disconnect its clients. + And the re-initialization time of DIR-665 is too short for our client mode + to detect to do roaming. So our client mode will remain connection with + DIR-665 while it changes bandwidth, so traffic will fail. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6888 + +* DESCRIPTION: +- Add checking AP's HT capability for client mode. AP's HT capability should + not change after it starts. If the HT capability changed, it means AP may + change some configurations and re-initialize. Then we should do roaming. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/01, EricYeh, feature refine + +* DIGEST: +- Add a menuconfig config to support 40/25 MHz clock + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_osdep.c +- Kconfig + +* DESCRIPTION: +- Use 'CONFIG_PHY_EAT_40MHZ' to select Clock Source. + Defined - 40MHz, Otherwise - 25MHz. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/01, jerryko, bug fix + +* DIGEST: +- 1.update update_beacon function and change RQPN to avoid beacon same SN and + shorten (20ms) interval beacon when mbssid +- 2.change TBTT hold time to avoid many beacon lost to let WNA1100 auth again + in noise environment test + +* TEST METHOD: +- 1.same SN and shorten interval beacon + (a)enable mbssid and use heavy traffic + (b)sniffer monitor beacon delta time and see same SN and shorten interval + beacon +- 2.WNA1100 auth + (a)In noise test lab and use 3 sta , Intel 6300/WNA1100/3100v2 to connect + AP ch6 + (b)STA 6300 use BT/offload/FTP upload/FTP download, WNA1100 and 3100v2 use + FTP upload/FTP download + (c)use other 5 AP in ch1/ch3/ch6/ch9/ch11 and tx multicast data len 1360 + Bytes and tx rate is 1Mbps and TP use 700Kbps + (d)WNA1100 will auth again after one or two hours running + +* ROOT CAUSE: +- 1.Beacon and BE mapping to high queue. When heavy traffic running, high + queue no page, the dma will not move beacon data to mac. Then up to TBTT time, + mac use old beacon data output. +- 2.When multicast start, AP beacon sometimes are blocked. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_sme.c +- 8192cd_hw.c +- 8192cd_osdep.c + +* DESCRIPTION: +- 1.High queue only used for beacon (4 pages). BE/VO/Mgmt move to low + queue.After change that, sniffer no fould same SN beacon or shorten interval + beacon. +- 2.Enlarge TBTT hold time. By mac Scott suggestion, single ssid change to + 0x400 (32ms) and mBssid change to 0x1df (15ms). When total ssid > 5, beacon + interval change to 200ms to avoid TBTT hold time > each Bssid beacon delta + time. After change that, single SSID AP test 24 hours and WNA1100 no auth + again in noise test lab. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/01, jerryko, fine tune + +* DIGEST: +- change retry count to 0x30 when client mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- client mode use more retry to reduce packet lost in open environment + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/08/01, jerryko, bug fix + +* DIGEST: +- Fix wifi client mode lost packet issue + +* TEST METHOD: +- 1.AP open and client connect +- 2.AP tx broadcast packet +- 3.AP tx unicast packet and seq same as broadcast packet +- 4.client will drop unicast packet + +* ROOT CAUSE: +- client mode record unicast cache seq number same as broadcast cache + +* MODIFIED FILES: +- 8192cd_rx.c +- 8192cd.h + +* DESCRIPTION: +- client mode add one variable to save broadcast/multicast packet seq + number + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/29, chuangsw, bug fix + +* DIGEST: +- Fix client mode site survey for duplicate BSSID with different channel issue. + +* TEST METHOD: +- + +* ROOT CAUSE: +- In collect_bss_info(), when we find the MACADDR that already in bss list, + we'll only consider the situation of the same channel. +- So when we reveive a packet in different channel from the same MACADDR, we + will also create a bss entry for it. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- When we reveive a packet in different channel from the same MACADDR, we will + compare it with the original entry and update the channel and rssi if the + newer packet's rssi is greater than the original entry. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/25, chuangsw, bug fix + +* DIGEST: +- Fix wifi logo client mode 5.2.17/5.2.18 + +* TEST METHOD: +- + +* ROOT CAUSE: +- Client didn't send MIC Failure Report when it receive the bad MIC frame. + +* MODIFIED FILES: +- 8192cd_security.c + +* DESCRIPTION: +- +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/25, jimmylin, fine tune + +* DIGEST: +- Check disable_brsc per SSID + +* TEST METHOD: +- + +* ROOT CAUSE: +- If there is an application that some SSIDs need to disable brsc but some + not, we need to check disable_brsc per SSID + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 6785 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/25, jimmylin, fine tune + +* DIGEST: +- Handle unknown type packets in NAT2.5 + +* TEST METHOD: +- + +* ROOT CAUSE: +- Originally, in NAT2.5, if we receive an unknown type packet from ethernet, + we will drop it. Now we will remember the source mac, convert it to our + mac, and send to air. Then when receiving an unknown type packet from air, + we will convert the dest mac as what we remembered and then send it to + ethernet. So we can handle unknown type packets, but only for one PC. + +* MODIFIED FILES: +- 8192cd.h - revision: 6781 +- 8192cd_sme.c - revision: 6781 +- 8192cd_proc.c - revision: 6781 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/16, EricYeh, feature refine + +* DIGEST: +- Improve production pass rate of 6195B from the suggestion of Anchi. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_mp.c + +* DESCRIPTION: +- When 6195B perform mp Tx, and select channel 4 or 12, + set PhyReg 0x26 with value 0x4f000. (0x4f200 is the previous value) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/14, jimmylin, bug fix + +* DIGEST: +- Fix wrong beacon period variable used in updateTSF() + +* TEST METHOD: +- + +* ROOT CAUSE: +- Should use beacon period got from AP and check if it is 0 + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6712 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/14, chuangsw, bug fix + +* DIGEST: +- Fix wifi logo 4.2.28 Spatial Multiplexing Power Save Operation + +* TEST METHOD: + +* ROOT CAUSE: +- Since we choose highTP station only depend on the percentage of the total + Tx+Tx counts, so if there is only one station, it will be the highTP + station. +- If we found highTP station, we'll set the txop to 0x5e so that we'll + not send RTS per frame. +- The test plan says that if we send the packet with the rate greater + than MCS7, we should send RTS/CTS before each frame. + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- Modify the TXOP enlarge machanism : + If the throughput of the highTP station is greater than 10 Mbps, + we will turn on the txop. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/13, EricYeh, feature refine + +* DIGEST: +- Add new gain tables for USB power. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- radio_a_intPA_GM.txt (Add) +- radio_b_intPA_GM.txt (Add) +- 8192cd_headers.h +- 8192cd_hw.c + +* DESCRIPTION: +- Add two files 'radio_a_intPA_GM.txt' & 'radio_b_intPA_GM.txt' to store + gain table of USB power. +- When WiFi Driver fits these conditions: '92D'+'Internal PA'+'USB Power' + driver will utilize the files described above to set gain tables. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/13, button, feature refine + +* DIGEST: +- Extend STA cache for tx desc break field to 4 AC queues + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd. - revision: 6701 +- 8192cd_tx.c - revision: 6671 + +* DESCRIPTION: +- Different TID packet do not aggregate together. If each priority only used +- by 1 STA, we doesn't need to set break field in txdesc. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/12, jerryko, bug fix + +* DIGEST: +- client mode LAN PC can't get dhcp ip when wep enable (enable flag USE_WEP_DEFAULT_KEY) + +* TEST METHOD: +- 1.AP set wep, shared-key auth +- 2.After client connect AP,LAN PC can't get dhcp + +* ROOT CAUSE: +- Dhcp reply packet drop by client mode. client receive dhcp reply decrypt by + hw. client use UseSwCrypto=1 and check wep decrypt error. Then client drop packet. + +* MODIFIED FILES: +- 8192cd_util.c + +* DESCRIPTION: +- After client connet ap, packet decrypt by hw and don't use sw decrypt checking. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/07, button, bug fix + +* DIGEST: +- HTC HD7 can't connect to 88RE AP + +* TEST METHOD: +- Let a STA connect to DUT and play streaming video, +- then using HD7 connect to DUT. + +* ROOT CAUSE: +- HD7 is critical in timing. Our SIFS is too long such that sometimes +- it can't receive our DHCP offer. After retry several times, it finally +- timeout and get IP fail. + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 6670 +- data/MACPHY_REG_92C.txt - revision: 6670 +- 8192cd_hw.c - revision: 6674 +* DESCRIPTION: +- reduce SIFS delay (2us) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/06, jimmylin, feature fine tune + +* DIGEST: +- Don't send beacon when func_off of vap is 1 + +* TEST METHOD: +- + +* ROOT CAUSE: +- Need to set register to prevent sending beacon while func_off is 1 for vap. + It can be done dynamically. + +* MODIFIED FILES: +- 8192cd.h - revision: 6659 +- 8192cd_sme.c - revision: 6659 + +* DESCRIPTION: +- For root interface, we cannot prevent sending beacon dynamically. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/04, chuangsw, feature refine + +* DIGEST +- Enable RTS when HT MIMO Dynamic power save is set and rate > MCS7. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- Rollback the modification during fixing intel power save issue. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/07/01, chuangsw, feature refine + +* DIGEST +- Refine 88RE/92CE PHY parameter. + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- data/PHY_REG_1T_n_hp.txt +- data/radio_a_2T_n_hp.txt +- data/radio_b_2T_n_hp.txt + +* DESCRIPTION: +- Adjust Tx gain table to avoid output power nonlinearity from index 39 to 40. +- Adjust Tx gain table to protect PA from high input power. (Tx index > 57) +- Fix 88RE high-power dead zone issue during throughput test. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/24, victorlo, feature refine + +* DIGEST +- Refine rssi calculation, and DIG upper bound for using ext LNA + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_rx.c + + +* DESCRIPTION: +- Refine rssi calculation with a new and more precise formula for both + typical and LNA-using devices of RTL8191c series. +- Refine DIG upper bound for using ext LNA to improve performance at noisy + surroundings. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/20, chris, bug fix + +* DIGEST +- Pause TX when one interface is down during concurrent mode + +* TEST METHOD: +- 1. Config as DMDP + 2. ifconfig wlan1 down + 3. Use sniffer to check if wlan1 still TX beacon + +* ROOT CAUSE: +- When stop hw, if another MAC is working, the previous code only + clean IMR without pausing TX. + +* MODIFIED FILES: +- 8192cd_hw.c + + +* DESCRIPTION: +- To stop hw, now we add TX pause and clean RCR even the other MAC is + working. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/17, PlusWang, feature add +* DIGEST: + - synchronized WIFI-DIRECT(WIFI-P2P) feature from branch-2.4 + +* TEST METHOD: +* ROOT CAUSE: +* DESCRIPTION: +*ADD FILE: + 8192cd_p2p.h + 8192cd_p2p.c +* MODIFIED FILES: + Kconfig + Makefile + wifi.h + 8192cd_cfg.h + 8192cd_debug.h + 8192cd.h + 8192cd_header.h + 8192cd_osdep.c + 8192cd_sme.c + 8192cd_tx.c + 8192cd_security.c + 8192cd_security.h + 8192cd_util.c + 8192cd_hw.c + + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/15, chuangsw, feature refine + +* DIGEST: +- Enable RTL8198+92c PCIE Power Saving feature. + +* TEST METHOD: +- 1. Use RTL8198 AP configured "Enable PCIE power saving support"(in + menuconfig). +- 2. Power on RTL8198 AP and don't let any client connect to it, it will go + to L1 power saving mode. +- 3. When RTL8198 AP is in L1 power saving mode, it will not wake up + when a client attempt to connect to it. + +* ROOT CAUSE: +- Default IRQ setting is wrong. + +* MODIFIED FILES: +- boards/rtl8198/bsp/irq.c +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_osdep.c + +* DESCRIPTION: +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/13, EricYeh, feature refine + +* DIGEST: +- Do NOT perform APK. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- APK could calibrate with wrong curve. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Do NOT call function PHY_APCalibrate() during PCI Init. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/13, chris, feature enhancement + +* DIGEST +- add a 5 min DFS check timer + +* TEST METHOD: +- When running Chariot, the AP will switch channel under following conditions: + 40M mode -- When TP > 45M, two radar pulses happen in 5 min. + 20M mode -- When TP > 20M, two radar pulses happen in 5 min. + +* ROOT CAUSE: +- To prevent false alarm of DFS when running Chariot. + +* MODIFIED FILES: +- 8192cd.h +- 8192cd_ioctl.c +- 8192cd_osdep.c +- 8192cd_sme.c +- 8192cd_util.h + + +* DESCRIPTION: +- When running Chariot, we switch channel only when the radar is detected + twice in 5 min. Otherwise, the radar pulse could be a false alarm. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/09, Button, feature refine + +* DIGEST: +- Improve rx sensitivity of 6195B by Anchi + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- Update RF parameters for for 6195B. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/08, chris, bug fix + +* DIGEST: +- Fix DMDP MAC-1 should take power index offset from path-B + +* TEST METHOD: +- Check power index of 92D DMDP MAC-1 HT20/HT_2S/OFDM rates. + The driver should apply indexes from path-B rather than path-A. + +* ROOT CAUSE: +- DMDP MAC-1 RF should take the index and setting from path-B. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- For offsets of HT20/HT_2S/OFDM, DMDP MAC-1 RF should take the offsets from path-B. +- For Power-by-rate, DMDP MAC-1 should take the offsets from path-B. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/02, ericyeh, feature refine + +* DIGEST: +- Do NOT perform APK when in MP mode. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- APK may not choose the correct value, + which will cause Mask and EVM be worse. + +* MODIFIED FILES: +- 8192cd_hw.c + +* DESCRIPTION: +- If mib "mp_specific" has non-zero value, + Do NOT try to do APK during the period of initiation of WiFi driver. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/06/01, chuangsw, feature refine + +* DIGEST: +- Refine EDCA BE parameter for ralink client. + +* TEST METHOD: +- Use 98+92d and ralink-2870 client to run the channel emulator. +- Set the chariot script to UL. +- Default EDCA BE parameter (0x508) is 0x005e642b and throughput is 140 Mbps. +- Write the 0x508 register to 0x005e4319, the throughput will up to 160 Mbps. + +* ROOT CAUSE: +- Original EDCA BE parameter is weakness than ralink. + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_util.c + +* DESCRIPTION: +- For ralink client, and if it is the highest TP station. +- Reduce the EDCA BE parameter (CW and AIFS) so that we can send data aggressively. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/05/27, hf ,bug fix + +* DIGEST +- Free skb in dz_queue every 5 seconds when station is idle + +* TEST METHOD +- + +* ROOT CAUSE +- dz_queue is freed at 5th-8th second after station becomes idle +- in original code, if packets is queued after eight seconds +- there is no chance to free the packets util expire_to becomes zero + +* MODIFIED FILES +- 8192cd_sme.c +- 8192cd_util.c +- 8192cd.h + +* DESCRIPTION +- Add idle_count flag +- call free_sta_skb after the first 5 idle seconds +- then call free_sta_tx_skb every 5 idle seconds + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE +- 2011/05/27, hf ,bug fix + +* DIGEST +- Do NOT queue skb when station expired and release queued skb in assoc_expire +- when station expired + +* TEST METHOD +- Connecting a USB dongle to AP +- disconnect the dongle by pulling it out +- check skb_queue_len for station's dz_queue by cat /proc/sta_info + +* ROOT CAUSE +- the unicast packets sent to an expired and sleeped station is queued +- in the dz_queue without increasing TX counts,these packets are not freed +- since free_sta_skb is called only when pstat->expire_to is great than zero, +- which cause system lacked of NIC private buffer + +* MODIFIED FILES +- 8192cd_sme.c +- 8192cd_tx.c +- 8192cd_util.c +- 8192cd_headers.h + +* DESCRIPTION +- Do Not Queue packets if station expired +- free_sta_tx_skb in assoc_expire when station expired + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/27, chuangsw, feature refine + +* DIGEST: +- Refine NAV upper bound for 92c/92d. + +* TEST METHOD: +- Reference to 802.11n test plan 5.2.3 +- Config 92c AP to client mode and connet to an AP1 which is in B mode. +- Use an AP2 to generate 802.11 Null Data with large duration (25000 us), + da set to 00:01:02:03:04:05, and the channel is the same as AP1. +- Start TCP FILESNDL chariot script (92c client mode -> AP1). +- When 92c client mode AP receive a 802.11n Null Data with large duration, + it should update its NAV but not. + +* ROOT CAUSE: +- Default NAV upper bound (0x652) setting is 0x20, unit is 128 microseconds, + e.g. 32*128 = 4096 microsecond. +- So when we receive a packet which duratin is greater than 4096, we + will not update NAV. + +* MODIFIED FILES: +- data/MACPHY_REG_92C.txt +- data_92d/MACPHY_REG.txt + +* DESCRIPTION: +- Enlarge the default NAV upper bound(0x652) setting to 0xC8 so + that it can tolorate 25000 microseconds' duration. +- 0xC8 = 200d, 200 * 128 = 25600 > 25000 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/27, jerryko, bug fix + +* DIGEST: +- Add Wifi tx packet length check(skb->len) to avoid wrong tx length to stop wifi + tx dma + +* TEST METHOD: +- (1)use netami-v2.4 branch code base + platform is 96CT+92CE +- (2)AP boot, intel client connect and ping ok +- (3)AP console reboot +- (4)intel client connect and ping fail +- (5)wait a moment, wifi driver check txhalt and reload wifi driver + +* ROOT CAUSE: +- When AP boot, upper layer send small length(1 or 5 bytes) packets to wifi + driver. If wifi clients already connect before sending small length packets, + wifi driver will calculate wrong tx length and kick packet to wifi tx dma. + It will cause tx dma halt. + +* MODIFIED FILES: +- 8192cd_tx.c + +* DESCRIPTION: +- If wifi transmit data packets length < 15 , drop it. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/25, victorlo, bug fix + +* DIGEST: +- Refine tx rate adaptive bitmap setting for short GI + +* TEST METHOD: +- Use 92Ce station set as 20/40M auto mode connects 20M mode AP, + AP will not transmit shortGI rate in autorate. + +* ROOT CAUSE: +- When station connects 20M mode AP with bandwidth setting as 20/40M auto, + AP will mistakenly discard short GI as an usable tx rate for that station. + Eventually AP will only transmit MCS15(2T) or MCS7(1T) as peak rate for it. + +* MODIFIED FILES: +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- Refine tx rate adaptive bitmap setting for short GI, to have a correct + setting for shortGI enable with different bandwidth settings of stations. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/23, victorlo, bug fix + +* DIGEST: +- Excluded DS IE checking for beacon statistic in client mode at 5G band connection + +* TEST METHOD: +- When client mode connected a 5G AP(Netgear WNCE3001), and do no data traffic + afterwards, reconnection happens every 5~6 seconds. + +* ROOT CAUSE: +- Since client mode cannot find out DS IE in beacon, by which we can check the + channel setting of the AP, will make a wrong statistic count of beacon to be none. + Carrying DS IE is mandatory in 2.4G band setting only, but driver will check + wherever the band is, caused a zero count and start roaming every 5~6 + seconds. + +* MODIFIED FILES: +- 8192cd_sme.c + +* DESCRIPTION: +- Excluded DS IE checking for beacon statistic in client mode at 5G band + connection, since this IE is mandatory for 2.4G band only. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/23, EricYeh, feature refine + +* DIGEST: +- Improve EVM & rx sensitivity of UMC 88re 6195B. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c +- 8192cd_mp.c + +* DESCRIPTION: +- Update RF parameters for UMC 88re 6195B. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/23, jimmylin, bug fix + +* DIGEST: +- Fix 92D TP issue with Intel + +* TEST METHOD: +- Config 92D to 2x2 mode. Test throughput with Intel client either in 2.4G + or 5G mode. The Tx rate will be not stable and TP will get lower than + before. + +* ROOT CAUSE: +- In order to fix power saving mode with Intel, we don't sent RTS to Intel. + But in 92D 2x2 mode, sending RTS is necessary for higher TP. Therefore + we will send RTS to Intel for 92D and high TP condition. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6383 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/23, davidhsu, feature refine + +* DIGEST: +- Enable PRIV_STA_BUF in dualband + +* TEST METHOD: +- N/A + + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_osdep.c + 8192cd_util.c + 8192cd_cfg.h + +* DESCRIPTION: +- Enable PRIV_STA_BUF flag in dualband. When CONCURRENT_MODE + and PRIV_STA_BUF are defined, it will use pre-allocate array buffer + for driver buffer pool and sta info under dualband. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/19, chris, feature refine + +* DIGEST: +- Halt init ONLY for part number check fail + +* TEST METHOD: +- Part Number check for illegal PG 92D IC will cause load firmware fail. + Then, the initialization of WiFi dirver will be halted. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c - rev. 6337 + +* DESCRIPTION: +- We would like to halt the initialization of 92D only for "part number fail". + After firmware load fail, if (0x1c5==0xe0), the reason is caused by part number + fail. The init. is therefore halted. Otherwise, we continue the open procedure. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/05/19, jimmylin, bug fix + +* DIGEST: +- Fix IOT issue with AR5007 + +* TEST METHOD: +- Connect AR5007 to AP mode. If AR5007 starts entering and exiting power + saving mode, packets from AR5007 will be dropped sometimes. + +* ROOT CAUSE: +- When AR5007 exits power saving mode with null data with power management + bit of 0, it will clear the sequence number to 0. The sequence numbers of + following packets will start from 1. Then the packets will be dropped by + AP beacuse of tuple cache checking. + +* MODIFIED FILES: +- 8192cd_rx.c - revision: 6336 + +* DESCRIPTION: +- If receiving null packet with sequence number of 0, clearing the tuple + cache for that client. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/19, jimmylin, bug fix + +* DIGEST: +- Fix 92D DMDP mode TP issue with Intel + +* TEST METHOD: +- Config 92D to DualMAC DualPHY mode. Test throughput with Intel client + either in 2.4G or 5G mode. The Tx rate will be not stable and TP will get + lower than before. + +* ROOT CAUSE: +- In order to fix power saving mode with Intel, we don't sent RTS to Intel. + But in 92D DMDP mode, sending RTS is necessary for higher TP. Therefore + we will send RTS to Intel only in 92D DMDP mode and high TP condition. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6329 +- 8192cd_tx.c - revision: 6329 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/16, victorlo, feature refine + +* DIGEST: +- Refine rssi_dump and FA statistic combination + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 6271 +- 8192cd_hw.c - revision: 6271 +- 8192cd_sme.c - revision: 6271 + +* DESCRIPTION: +- Refine rssi_dump by replacing SNR and SQ with FA, CCA and DIG info + for future debug helpfulness +- Add 2 more reporting registers into FA statistic combination, + in BB team request + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/12, jimmylin, bug fix + +* DIGEST: +- Fix missing of skb refill handler in tx_stuck_fix() + +* TEST METHOD: +- Refer to 69.3 mantis #382 + +* ROOT CAUSE: +- Should handle skb refill in tx_stuck_fix() + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6230 +- 8192cd_rx.c - revision: 6230 + +* DESCRIPTION: +- Refer to 69.3 mantis #382 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/10, chris, bug fix + +* DIGEST: +- Fix endian issue for USB power setting +- Add USB power support for MP + +* TEST METHOD: +- Check if txpower add PHY_REG_PG value for MCS4 and MCS12. + +* ROOT CAUSE: +- In driver code, index of priv->pshare->phw->MCSTxAgcOffset_A[x] does + not map to MCS y. Ex: priv->pshare->phw->MCSTxAgcOffset_A[7] = value for MCS4. + If one add txpower by the index, the value will be wrong. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 6192 & 6206 +- 8192cd_mp.c - revision: 6206 + +* DESCRIPTION: +- Driver read PHYREG_PG.txt and store the value in register. However, driver + read the register directly and does not care the endian problem. This causes + the index of MCSTxAgcOffset_A does not map to the real MCS index. Driver must + carefully use the offsets. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/03, victorlo, bug fix + +* DIGEST: +- Refine tx power control method to avoid damaging (high power) PA + +* TEST METHOD: +- High power AP with smaller odd power index in one or both + RF Path A and B, in power index for MCS8 to MCS15, + less than 0x10. +- Use an 11n station connects the AP in middle range, ping long + packet success +- Move station to the AP, let AP enters tx power control mode to + degrade tx power for the station +- One or both external PA would be damaged shortly. + +* ROOT CAUSE: +- When AP detected connected station in a short range by rssi, + driver will dynamically modify tx desc by an negative power + index for degrading tx power. +- Since signin power index have to be an even number, + driver masks the last bit of power index after 2's compliment. +- By using the mask of last bit, will mistakenly changed the + degraging power index to decrease more, and overflow the final + power index as a very large positive one and damage the PA. + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 6134 + +* DESCRIPTION: +- driver masks the power index by the last bit before 2's compliment + +----------------------------------------------------------------------------- + +- 2011/05/03, chuangsw, feature refine + +* DIGEST: +- update 92c firmware version 69.0. +- This version can support 63 clients. + +* TEST METHOD: +- AP use the new firmware to run up the channel emulator, client is Intel 6300 +- chariot : DL, 20M, +- the throughput is better than original firmware at the attenuation 76~80 dB. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- N/A + +* DESCRIPTION +- New firmware use new rate adaptive algorithm and it seems better. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/03, button, feature refine + +* DIGEST: +- refine wlan init, stop flow by filen + +* TEST METHOD: +- run reboot, reinit script by tera term macro. +- recover wlan after exceptions happened, such as watch dog reset, core dump, +- access null pointer, divided by zero. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 6125 +- 8192cd_headers.h - revision: 6125 +- 8192cd_osdep.c - revision: 6125 + +* DESCRIPTION: +- SD9 found 92c init fail in reboot long run test. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/03, jerryko, feature refine + +* DIGEST: +- missing release cache pstat pointer when fix MIC error + +* TEST METHOD: +- when use 92D DMDP, the cache hwaddr still exist after clients disassociate. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_util.c - revision: 6111 +- 8192cd.h - revision: 6111 +- wapiCrypto.c - revision: 6111 +- wapi_wai.c - revision: 6111 + +* DESCRIPTION: +- in release_stainfo function, missing clean stainfo_cache pointer. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/03, button, feature refine + +* DIGEST: +- refine for antenna diversity + +* TEST METHOD: +- run chariot with DWA-160 A2, DL, 6 pairs. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 6104 + +* DESCRIPTION: +- do not clear clear tx short-cut every sec. just update it. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/02, jerryko, bug fix + +* DIGEST: +- MIC error happen + +* TEST METHOD: +- Use customer board and config 11g mode and wpa-pak with tkip. +- Use two wifi clients to connect the AP and use chariot 1 DL+1 UL + each wifi clients. And mic error happen randomly. + +* ROOT CAUSE: +- When driver found mic error packets, print packets information + da/sa/len/tid/key. And found the unicast key use wrong key. +- In tkip_rx_mic function, get_stainfo get wrong pstat. + get_stainfo function memcmp (cache hwaddr)=hwaddr, but cache pstat is other + client pointer. + +* MODIFIED FILES: +- 8192cd_util.c - revision: 6090 + +* DESCRIPTION: +- 1.don't memcpy cache hwaddr, just save cache pstat pointer. +- 2.memcmp compare (cache hwaddr pointer) with hwaddr. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/05/02, button, bug fix + +* DIGEST: +- avoid overflow in function mod64 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 6088 + +* DESCRIPTION: +- maximum of divisor= 2^20, maximum of residue= (2^20)-1, 12 bit left shift +- will never overflow. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/27, chrisyeh, code refine + +* DIGEST: +- refine 92D internal PA to prevent burning PA +- remove " 92D INTERNAL PA" item from menuconfig because using the same + code base + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- revision: 6047 +- 8192cd_cfg.h +- 8192cd_hw.c +- 8192cd_ioctl.c +- 8192d_reg.h +- Kconfig + +* DESCRIPTION: +- Driver take MIB values "trsw_pape_CC" and "trsw_pape_C9" to + config as internal PA. If the values are not identified, alway + init as external PA to prevent burning PA. + +- Legal internal PA setting + trsw_pape_C9/trsw_pape_CC = AA/A0 + trsw_pape_C9/trsw_pape_CC = AA/AF + trsw_pape_C9/trsw_pape_CC = 00/A0 + + External PA + + trsw_pape_C9/trsw_pape_CC = 00/FF and others + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/27, chrisyeh, bug fix + +* DIGEST: +- fix eFuse write offset > 128 byte for 8192D +- add 92d XCAP/TRSW_CTRL(C9)/PAPE_CTRL(CC) efuse write + +* TEST METHOD: +- PG offset > 128 value + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- revision: 6047 + 8192cd_hw.c + 8192cd_hw.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/04/26, jimmylin, feature add + +* DIGEST: +- Add eFuse read/write functions for 8192D + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- revision: 6032 + 8192cd_hw.c + 8192cd_hw.h + 8192cd_ioctl.c + 8192d_hw.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/25, davidhsu, bug fixed + +* DIGEST: +- Fix the issue of Tx MIC error in client mode TKIP + +* TEST METHOD: +- Config DUT as client mode in TKIP encryption +- Connect to a AP +- Ping from a Ethernet PC to AP. Ping will fail after a while. + +* ROOT CAUSE: +- When MCAST2UI_REFINE is defined, it will skip the checking for DA in + get_tx_sc_index() routine. However, it will cause problem in client mode. + +* MODIFIED FILES: +- revision: 6027 + 8192cd_tx.c + +* DESCRIPTION: +- Modify function get_tx_sc_index() that it will skip DA check only + in AP mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/19, chuangsw, feature refine + +* DIGEST: +- 98+ 92c/92d DL performance issue + +* TEST METHOD: +- 1. build svn trunk image (r5944) +- 2. run chariot from AP to client (Intel 6300) +- 3. DL performance will be 140~160, avg 145 Mbps. + +* ROOT CAUSE: +- 1. "highTP_found_pstat" will be reset in rtl8192cd_expire_timer() every second but + the highTP_found_pstat is choosen every two seconds. +- 2. In choose_IOT_main_sta(), tx_2s_avg and rx_2s_avg may be overflow so it + will not found the highTP_found_pstat. + +* MODIFIED FILES: +- revision: 5961 +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- 1. reset highTP_found_pstat every two seconds. +- 2. avoid tx_2s_avg/rx_2s_avg overflow. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/04/14, button, bug fix + +* DIGEST: +- endian issue + +* TEST METHOD: +- 1) ZTE found a client connect to 92c in BGN, AES mode, open rssi dump, +- 92c rx rate is MCS15, but tx rate is limited to 54 Mbps. +- 2) tx power by rate error + +* ROOT CAUSE: + +* MODIFIED FILES: +- revision: 5901 +- 8192cd_hw.c +- 8192cd_sme.c + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/11, chrisyeh, feature add + +* DIGEST: +- Add band edge Tx power check for 8192d + +* TEST METHOD: +- use MP tool to verify if the power-by-rate strategy is limited by band edge power. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: + revision : 5746 +- 8192cd_headers.h +- 8192cd_ioctl.c +- 8192cd_mp.c +- 8192cd_cfg.h +- 8192cd.h +- 8192cd_util.c +- 8192cd_hw.c +- data_92d/TXPWR_LMT.txt (Add) +- 8192cd_hw.h +- 8192cd_proc.c +- ieee802_mib.h + +* DESCRIPTION: +1. Must set a target_pwr for MP target, default: 13dBm +2. Driver will check TXPWR_LMT.txt according to the current channel. Then, + it caculates the difference between txpwr_lmt & target_pwr. +3. Power-by-rate index is bound by the difference. + +Default: OFF +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/04/08, button, bug fix + +* DIGEST: +- Avoid skb leak + +* TEST METHOD: +- Enable DELAY_REFILL_RX_BUF, set NUM_RX_DESC larger than REAL_MAX_SKB, +- down/up wlan several times, and check skb_free_num in proc + +* ROOT CAUSE: +- At wlan stop phase, if cur_rx_refill doesn't equal to cur_rx, skb will +- refill to rx ring, such that memory leak. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 5839. + +* DESCRIPTION: +- Reset cur_rx_refill, cur_rx before free skb. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/04/08, button, bug fix + +* DIGEST: +- MP single tone for high power boards + +* TEST METHOD: +- +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_mp.c revision: 5821 + +* DESCRIPTION: +- Enable PAPE by software during single tone test + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/04/07, chrisyeh, bug fix + +* DIGEST: +- fix 92d DMDP: ONLY wlan1 open alone - part 2 + +* TEST METHOD: +- DMDP mode, wlan0 & wlan1 down. Then, wlan0 up, it will load firmware fail. + +* ROOT CAUSE: +- BB ad/da 0x888 should be set in MAC0 + +* MODIFIED FILES: + revision : 5811 +- 8192d_hw.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/31, chrisyeh, bug fix + +* DIGEST: +- fix 92d DMDP: ONLY wlan1 open alone + +* TEST METHOD: +- DMDP mode, wlan0 & wlan1 down. Then, wlan0 up, it will load firmware fail. + +* ROOT CAUSE: +- Typo causes wlan1 always cannot load fw. + +* MODIFIED FILES: + revision : 5727 +- 8192d_hw.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/31, chuangsw, bug fix + +* DIGEST: +- fix wifi logo client mode MIC error test fail + +* TEST METHOD: +- DUT : AP in client mode +- use an AP to send MIC error packet to DUT, DUT should response a MIC Failure +Report but not. + +* ROOT CAUSE: +- Client will not send MIC Failure Report if it receive a MIC error packet + from AP's ethernet addr, we should get the AP's addr by BSSID but not sa. + +* MODIFIED FILES: + revision : 5708 +- 8192cd_psk.h +- 8192cd_psk.c +- 8192cd_security.c + +* DESCRIPTION: + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/03/25, chrisyeh, feature add + +* DIGEST: +- Tx Beamforming for 92D + +* TEST METHOD: +- use emulator to test if performance is enhanced + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: + revision: 5619 & 5620 + - 8192cd_headers.h + - 8192cd_ioctl.c + - 8192cd_proc.c + - 8192cd_sme.c + - 8192cd_tx.c + - 8192d_hw.c + - ieee802_mib.h + +* DESCRIPTION: +- BB will follow tx desc to enable TX_BF + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/25, jerryko, bug fix + +* DIGEST: +- fix beacon err too many in 8198+92c platform + +* TEST METHOD: +- open environment, cat "/proc/wlan0/stats" display beacon err is almost the same + as beacon ok + +* ROOT CAUSE: +- In 8198 platform, mac EIFS register (0x642) use 0x4000 value. + It means when hw recv crc error, then it stops tx 16ms(0x4000). The time is too + long and will effect beacon tx. So use default value(0x13A) and the value is + the same as 8196c+92c platform. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 5604 + +* DESCRIPTION: +- After using 0x13A, the beacon err decrease a lot. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/24, chris, feature add + +* DIGEST: +- Add EFUSE thermal value write function + +* TEST METHOD: +- iwpriv wlanX efuse_set HW_11N_EFUSE=## +- iwpriv wlanX efuse_get HW_11N_EFUSE + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 5591 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/23, chuangsw, feature refine + +* DIGEST: +- AP mode wifi logo fail 4.2.43-T05 + +* TEST METHOD: +- Use 2T STA to connect to our AP +- AP ping STA (ping <STA> -n 100 -l 1000) +- Check if the PING-REQ at data rate thar >= MCS 13 + +* ROOT CAUSE: +- The init RA register is set to MCS7, and it will update to more than MCS7 + untill it receives heavy data flow. + +* MODIFIED FILES: + revision: 5561 +- 8192cd.h +- 8192cd_sme.c +- 8192cd_util.c + +* DESCRIPTION: +- We will the the init rate to MCS15 if the STA is 2T and RSSI > 55 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/23, button, bug fix + +* DIGEST: +- Fix reboot failed in 96c+92c + +* TEST METHOD: +- run a term macro, execute ifocnifg wlan0 down; reboot + +* ROOT CAUSE: +- We disable clock for saving power(50mW). When clock is disabled, it needs +- to do PCIe reset procedure to establish the link, demo board v120 may +- reset fail in 50 times reboot test. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 5554 + +* DESCRIPTION: +- Do not turn off PCIe clock in all condition + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/22, chuangsw, bug fix + +* DIGEST: +- Client mode wifi logo MIC error test fail (5.2.17) + +* TEST METHOD: +- Setup a AP in WPA+TKIP mode +- Setup another AP in client mode and connect to above AP +- AP issue a mic error packet to the client +- Client response incorrect MIC Failure Report (not a EAPOL-key frame format + in sniffer) +- AP issue the second mic error packet in 60 seconds +- Client did not disassoc with AP + +* ROOT CAUSE: +- The MIC Failure Report that client send is not in EAPOL-key frame format. +- Client did not disassoc itself when it receive two mic error packets. + +* MODIFIED FILES: +- New Revision: 5525 + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/03/20, david, bug fix + +* DIGEST: +- Fix the issue that AP client may go to a lock state during connect to AP + with wrong security. + +* TEST METHOD: +- Setup a AP in WEP mode +- Config DUT as AP client in WPA mode and with the same SSID of above AP +- Power on the DUT, and then check the status of web page. You will see DUT + will always stay in the state of "waiting for key" + +* ROOT CAUSE: +- When AP client connect to AP, it wait for 4-1 of 4-way handshaking. + However, it did not start timer to detect if lost 4-1 packet. +- So, I modify that it will indicate a ASSOC_IND event to psk module + in client mode when WPA/WPA2 enabled. +- In psk module, it will start a timer to wait for 5 sec for 4-1. If timeout, + it will disconnect AP and scanning again. + +* MODIFIED FILES: +- New Revision: 5445 +- 8192cd.h +- 8192cd_headers.h +- 8192cd_ioctl.c +- 8192cd_psk.c +- 8192cd_psk.h +- 8192cd_security.c +- 8192cd_sme.c + + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/18, button, new feature + +* DIGEST: +- 1x1, 2x2 hardware antenna diversity + +* TEST METHOD: +- run lukelee's test plan in channel emulator + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- rtl8192cd/8192cd.h - revision: 5440 +- rtl8192cd/8192cd_cfg.h - revision: 5440 +- rtl8192cd/8192cd_headers.h - revision: 5440 +- rtl8192cd/8192cd_hw.c - revision: 5440 +- rtl8192cd/8192cd_ioctl.c - revision: 5440 +- rtl8192cd/8192cd_osdep.c - revision: 5440 +- rtl8192cd/8192cd_proc.c - revision: 5440 +- rtl8192cd/8192cd_rx.c - revision: 5440 +- rtl8192cd/8192cd_rx.h - revision: 5440 +- rtl8192cd/8192cd_sme.c - revision: 5440 +- rtl8192cd/8192cd_tx.c - revision: 5440 +- rtl8192cd/8192cd_util.c - revision: 5440 +- rtl8192cd/Kconfig - revision: 5440 + +* DESCRIPTION +- enable hardware rx diversity, and assign tx path by collecting rx result +- from rx phy status,. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/03/16, hf, bug fix + +* DIGEST: +- When using iwlist scan, we can not get results when command issued + +* TEST METHOD: +- build image with Wireless Tools v29 support +- iwlist wlan scan + +* ROOT CAUSE: +- iwlist will can rtl_siwscan then rtl_giwscan +- in rtl_siwscan just started sitesurvey +- Do Not waiting sitesurvey done + +* MODIFIED FILES: +- New Revision: 5393 +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd.h +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_comapi.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_osdep.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_sme.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/Kconfig +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/Makefile +* DESCRIPTION: +- No longer bind RTL_COMAPI_WLTOOLS with RTL_COMPAI_CFGFILE +- add a wait_queue to wait sitesurvey done +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/16, chrisyeh, Feature refine + +* DIGEST: +- Modify Client mode Site Survey Rule + +* TEST METHOD: +- config as client, site survey will follow UI band setting. Only configured as + A+B+G+N, client will do full-band SS. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- New Revision: 5362 +- 8192cd_sme.c +- 8192cd.h + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/16, chrisyeh, Bug fix + +* DIGEST: +- Fix client site survey issue (RF18 r/w error) + +* TEST METHOD: +- config as client, set band A+B+G+N, do SS more than 10 times. The scanned AP + will get fewer and fewer. + +* ROOT CAUSE: +- RF18 read/write error. + +* MODIFIED FILES: +- New Revision: 5362 +- 8192cd_hw.c +- 8192cd_mp.c +- 8192d_hw.c + +* DESCRIPTION: +- Pre-store RF18 when update band mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/15, hf, Bug fix + +* DIGEST: +- Add key ID ioctl for iwpriv when WDS using wep security + +* TEST METHOD: +- + +* ROOT CAUSE: +- T&W need to using 2~4 wep key when WDS using WEP + +* MODIFIED FILES: +- New Revision: 5363 +- 8192cd_osdep.c +- 8192cd_ioctl.c + +* DESCRIPTION: +- Add ioctl for key index setting + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/16, victorlo, add feature + +* DIGEST: +- Periodically update TXOP from Beacon for Wifi 11n Client mode test + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 5365 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/15, YangQuan, Bug fix + +* DIGEST: +- bug fix for SSID error in beacon frame after WPS when broadcast SSID is + disabled + +* TEST METHOD: +- + +* ROOT CAUSE: +- we should set the SSID value only, not the type/length/value in IE + +* MODIFIED FILES: +- 8192cd_ioctl.c - revision: 5344 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/03/13, victorlo, refine + +* DIGEST: +- Refine Client mode EDCA setting flow when ACM bit is set by AP + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 5292 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/13, victorlo, refine + +* DIGEST: +- Refine IOT mode checking for Wifi 11n Client logo test + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 5290 +- 8192cd_util.c - revision: 5290 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/13, victorlo, add feature + +* DIGEST: +- Support Client mode 11n protection mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 5287 +- 8192cd_sme.c - revision: 5287 +- 8192cd_tx.c - revision: 5287 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/11, jimmylin, bug fix + +* DIGEST: +- Move rx statistics in validate_mpdu() to avoid using invalid skb + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_rx.c - revision: 5274 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/11, hf, fine-tune + +* DIGEST: +- refine for TP and 2.4g 5g concurrent ipert testing + +* TEST METHOD: +- + +* ROOT CAUSE: +- active tx_recycle by raise tx tasklet during TX benefit the TP + +* MODIFIED FILES: +- New Revision: 5255 +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_rx.c +- trunk/RLXLinux/rlxlinux-v2.2/linux-2.6.30/drivers/net/wireless/rtl8192cd/8192cd_tx.c +* DESCRIPTION: +- during Tx when there are 1/2 tx ring is used, then raise tx tasklet +- in order to do tx_recycle just after rx_tasklet + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +* DATE/WHOM/TYPE: +- 2011/03/10, victorlo, add feature + +* DIGEST: +- Support Wifi 11n Client logo 20/40 Coexist detection + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 5238 +- 8192cd_proc.c - revision: 5238 +- 8192cd_sme.c - revision: 5238 +- 8192cd_tx.c - revision: 5238 +- 8192cd_util.c - revision: 5238 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/09, button, bug fix + +* DIGEST: +- rollback for 92d DMDP mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 5234 + +* DESCRIPTION: +- 92d MIMO TR mode depends on MIB value, can't check chip ID only init_one +- phase only. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/07, Qjj/jimmylin, bug fix + +* DIGEST: +- Update TIM bit and send high queue packet if there are packets queued in + high queue + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 5155 + +* DESCRIPTION: +- We only update TIM bit if there are packets queued in software queue. But + some packets may be queued in high queue due to short DTIM window. So we + need to update TIM bit if there are packets queued in high queue and send + them out to prevent high queue stop. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/03/07, victorlo, bug fix + +* DIGEST: +- Add crypto mode checking to avoid s/w encrypted packet use tx shortcut + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 5153 + +* DESCRIPTION: +- s/w encrypted packet needs 3 desc for every single transmit, + but there are only 2 desc signin thru tx shortcut, and will + cause tx hangup by desc content mismatched. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/07, chuangsw, bug fix + +* DIGEST: +- If AP's hardware tx queue still has data when client go to power save mode, it will + send RTS packet. +- Since the client is sleeping, it will not response with CTS, so RTS will go to retry + limit soon and drop the packet. + + +* TEST METHOD: +- Enable Intel client's power save mode (set "power management" to lowest). +- Run up chariot (one pair) + +* ROOT CAUSE: +- RTS retry too fast + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 5142 +- 8192cd_rx.c - revision: 5142 + +* DESCRIPTION: +- Do not sent RTS for Intel client +- Use data retry fot Intel client + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/03/07, jimmylin, bug fix + +* DIGEST: +- Add CLI protection for rtl8192cd_firetx() for Rx tasklet case +- Don't clear ISR in rtl8192cd_rx_dsr() for return value check of interrupt + handler in kernel 2.6 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 5140 +- 8192cd_rx.c - revision: 5140 + +* DESCRIPTION: +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/07, EricYeh, new feature + +* DIGEST: +- Support Wpa_Supplicant. + +* TEST METHOD: +- Remove daemons 'auth' 'iapp' 'wsc daemon', run daemon 'wpa_supplicant' + to control WiFi module as a WiFi station, check the basic functions + (like scan results, associate to appointed WiFi AP... etc) + can be performed succeddfully or not? + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 5120 (svn) +- 8192cd_cfg.h - revision: 5120 (svn) +- 8192cd_comapi.c - revision: 5120 (svn) +- 8192cd_headers.h - revision: 5120 (svn) +- 8192cd_ioctl.c - revision: 5120 (svn) +- 8192cd_net80211.c - revision: 5120 (svn) +- 8192cd_osdep.c - revision: 5120 (svn) +- 8192cd_security.c - revision: 5120 (svn) +- 8192cd_sme.c - revision: 5120 (svn) +- 8192cd_util.c - revision: 5120 (svn) + +* DESCRIPTION +- Add compile tag 'WIFI_WPAS' to figure out which modifications made for + supporting Wpa_Supplicant. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/04, YangQuan, bug fix + +* DIGEST: +- Bug fix for WDS ping fail + +* TEST METHOD: +- ping between two WDS mode AP + +* ROOT CAUSE: +- wdsDev is not initialization in __rtl8192cd_start_xmit_out() for WDS traffic + +* MODIFIED FILES: +- 8192cd_rx.c New Revision: 5091 + +* DESCRIPTION +- init wdsDev in __rtl8192cd_start_xmit_out() for WDS traffic + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/04, hf, bug fix + +* DIGEST: +- Fine Tune for VAP TP + +* TEST METHOD: +- run chariot high performance script, when VAP enabled + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_rx.c New Revision: 5089 + +* DESCRIPTION +- Clear IMR when in RX tasklet + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/03, button, bug fix + +* DIGEST: +- DWA-160 A2 IOT + +* TEST METHOD: +- run chariot high performance script, 3 pairs rx in 5G 20M mode, AES + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd.tx.c - revision: 5078 + +* DESCRIPTION +- enlarge AMPDU density from 5 to 7 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/03/01, button, new feature + +* DIGEST: +- UMC B-cut 88c(6195B) support + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: + +- 8192cd.osdep.c - revision: 5058 +- 8192cd_proc.c - revision: 5058 +- 8192cd_hw.c - revision: 5058 +- 8192cd_util.h - revision: 5058 +- 8192cd_headers.h - revision: 5058 + +* DESCRIPTION +- modify radio parameters, IQK, power mode + +------------------------------------------------------------------------ + +* DATE/WHOM/TYPE: +- 2011/02/25, chris, bug fix + +* DIGEST: +- slave mode DFS support & channel map for DFS + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h - rev 5040 +- 8192cd_cfg.h - rev 5040 +- 8192cd_headers.h - rev 5040 +- 8192cd_hw.c - rev 5040 +- 8192cd_osdep.c - rev 5040 +- 8192cd_sme.c - rev 5040 +- 8192cd_util.c - rev 5040 +- Kconfig - rev 5040 + + +* DESCRIPTION +- No DFS channel if DFS is not defined. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2011/02/24, chris, bug fix + +* DIGEST: +- fix RF reg 0x28 write fail during auto-channel switching + +* TEST METHOD: +- Config 92d demo board to 2G SMSP +- Use "iwpriv wlan0 autoch" to trigger auto-channel switching +- Check if Tx/Rx count changes + +* ROOT CAUSE: +- RF Reg seems cannot be read/write too frequently. Because 92D will + reload many RF registers, there could be read or write error. + +* MODIFIED FILES: + +- 8192cd.h - revision: 5011 +- 8192cd_hw.c - revision: 5011 +- 8192cd_util.h - revision: 5011 +- 8192d_hw.c - revision: 5011 + +* DESCRIPTION +- Pre-store RF reg 0x18 & 0x28 in share_info to prevent frequent r/w RF reg. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/23, button, feature refine + +* DIGEST: +- refine dynamic noise control + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: + +- 8192cd.h - revision: 5006 +- 8192cd_ioctl.c - revision: 5006 +- 8192cd_hw.c - revision: 5006 + +* DESCRIPTION +- add a mib to enable or disable noise control + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/23, victorlo, refine + +* DIGEST: +- Refine ADDBA action frames and status handling + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: revision 4999 +- 8192cd.h +- 8192cd_sme.c +- 8192cd_tx.c + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/21, button, feature refine + +* DIGEST: +- dynamic noise control + +* TEST METHOD: +- 2 demo boards set bandwidth to 20MHz, channel 1 and 11 respectively. +- run chariot throughput script, see if it the patch improves the performance. + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 4974 +- 8192cd.h - revision: 4974 +- 8192cd_proc.c - revision: 4974 +- 8192cd_hw.c - revision: 4974 + +* DESCRIPTION +- without ext. LNA: enlarge initial gain dynamically +- with ext. LNA: change T/R switch to Tx mode to suppress noise + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2011/02/21, chuangsw, bug fix + +* DIGEST: +- refine intel power saving problem + +* TEST METHOD: +- turn on intel's power saving mode (set "power management" to lowest) and + connect to our AP (ex. 98+92c) +- run up the chariot test (one pair, DL) then we can see a big cut at about +every one minute. +- ps. the newer intel wifi module (ex. Intel 6300) will go to power saving + mode more positive + +* ROOT CAUSE: +- we cannot stop the HW tx queue in time when intel client is going to +power saving mode. + +* MODIFIED FILES: +- 8192cd.h - revision 4972 +- 8192cd_sme.c - revision 4972 + +* DESCRIPTION +- we set the "TX_PAUSE" register to stop the transmission of tx queue when + receive power management packet that indicate it will go to sleep. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/18, chris, bug fix + +* DIGEST: +- fix 92D SMSP ping fail due to new IQK-R03 bug + +* TEST METHOD: +- Config 92D as SMSP, client can link-up but ping failed + +* ROOT CAUSE: +- In IQK-R03, path-B IQK will overwrite the result of path-A IQK (0xe30). + +* MODIFIED FILES: +- 8192d_hw.c - rev. 4951 + +* DESCRIPTION +- In IQK-R03, path-B IQK will overwrite the result of path-A IQK (0xe30) with + default value. The wave-form cannot be demodulized by the clients. +- Fix it by moving the section "fill IQK path-A result" after "IQK path B". + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/16, davidhsu, bug fix + +* DIGEST: +- Fix the issue that it will not use s/w crypto engine in WEP when h/w + crypto engine is full (max 28) + +* TEST METHOD: +- Use veriWave to test max client capacity under WEP encryption. + It can only reach 28 in max. Can't reach 31. + +* ROOT CAUSE: +- When driver do s/w crypto check in transmitting/receiving packet, + it has bug in WEP mode. So, it will think it used h/w cryptio even + the associated client is above h/w max (28). + +* MODIFIED FILES: +- 8192cd_util.c - revision: 4929 + +* DESCRIPTION +- Modify UseSwCrypto() that it will check "pstat->keyInCam" flag for + each STA in WEP mode if USE_WEP_DEFAULT_KEY is not defined. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/10, jimmylin, refine + +* DIGEST: +- If there is traffic > 10Kbps, don't send disassoc req of MIC error + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_security.c - revision: 4879 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/10, jimmylin, bug fix + +* DIGEST: +- Check da before sending disassoc req and deauth req + +* TEST METHOD: +- + +* ROOT CAUSE: +- After 1.2d, we will receive strange frame randomly. If we receive a bad + association req with multicast sa, we will send deauth req to that + multicast address. It will cause other stations disconnected. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 4879 + +* DESCRIPTION +- At this moment, we add da checking at issue_deauth() and issue_disassoc() + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/09, davidhsu, feature refine + +* DIGEST: +- Reduce pre-allocate rx buffer size in 92d only + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 4868 +- 8192cd_util.c - revision: 4868 +- 8192cd_cfg.h - revision: 4868 +- 8192cd_header.h - revision: 4868 + +* DESCRIPTION +- When 92d-ony and dual-mac & dual-phy are defined, it will define + a new flag DUALBAND_ONLY +- When DUALBAND_ONLY is defined, the skb pool number will be + declared as half size of original one +- When driver is opened, it will check if current configuration is + DUALMAC_DUALPHY. If yes, it will call a new added function + split_pool() to check if current skb pool size is over max size. + If yes, it will put the extra buffer into the pool of the other interface. +- If it is not DUALMAC_DUALPHY mode, it will call a new added function + merge_pool() to check if any skb buffer existed in the pool of the other + inteface. If yes, it will move these buffers into the pool of current + interface. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/09, chuangsw, feature refine + +* DIGEST: +- avoid power index too small + +* TEST METHOD: +- when we use the mp tool set the power index to 1, it will burn the pa. + +* ROOT CAUSE: +- all power index will be reduced 16, so if the power index is 1 or some small + values, it will be very big after reduced 16 (negative number is trancformed + into unsigned int). + +* MODIFIED FILES: +- 8192cd.h - revision: 4864 +- 8192cd_hw.c - revision: 4864 +- 8192cd_tx.c - revision: 4864 + +* DESCRIPTION +- 1. choose the minimum power index (min_pwr_idex) +- 2. if (min_pwr_idex < 16) + the power index will be reduced "min_pwr_idex" + else + the power index will be reduced 16 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/09, button, bug fix + +* DIGEST: +- fix bug of leave sta detect + +* TEST METHOD: +- netgear BT test + +* ROOT CAUSE: +- once a sta connected to vap fall into in leave state, will never change +- to normal state. + +* MODIFIED FILES: +- 8192cd_rx.c - revision: 4849 + +* DESCRIPTION +- when receving packet from leaving sta, change state to normal in +- rx_shortcut, process_datafrme, rtl8192cd_rx_procNullPkt. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/02/08, jerryko, feature refine + +* DIGEST: +- modify SW_TX_QUEUE flow to improve performance in 98+92c+92d platform + +* TEST METHOD: +- veri-wave test 31 clients to measure performance + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h -revision: 4845 +- 8192cd_proc.c -revision: 4845 +- 8192cd_sme.c -revision: 4845 +- 8192cd_util.c -revision: 4845 +- 8192cd_tx.c -revision: 4845 + +* DESCRIPTION +- modify + 1.record mac or tid when client number over one + 2.dynamic enable sw queue when client number over one + 3.use function call instead of repeated parts + 4.tx out priority first when only one client + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/26, button, bug fix + +* DIGEST: +- avoid external PA burn out + +* TEST METHOD: + +* ROOT CAUSE: + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 4810 + +* DESCRIPTION +- disable adding extra power by rate of high power boards + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/19, EricYeh, new feature + +* DIGEST: +- Porting WiFi driver into X86 PC with linux OS, + for purpose of supporting Hostapd under the environment of + little endian CPU and Non-Realtek platform. + +* TEST METHOD: +- Make WiFi driver under X86 PC with linux OS. +- Insert the generated WiFi module. +- Install Hostapd. +- Run Hostapd to verify AP function can work or not. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 4749 (svn) +- 8192cd_cfg.h - revision: 4749 (svn) +- 8192cd_comapi.c - revision: 4749 (svn) +- 8192cd_hw.c - revision: 4749 (svn) +- 8192cd_ioctl.c - revision: 4749 (svn) +- 8192cd_mp.c - revision: 4749 (svn) +- 8192cd_net80211.c - revision: 4749 (svn) +- 8192cd_osdep.c - revision: 4749 (svn) +- 8192cd_proc.c - revision: 4749 (svn) +- 8192cd_psk.c - revision: 4749 (svn) +- 8192cd_psk.h - revision: 4749 (svn) +- 8192cd_psk_hapd.c - revision: 4749 (svn) +- 8192cd_rx.c - revision: 4749 (svn) +- 8192cd_security.c - revision: 4749 (svn) +- 8192cd_security.h - revision: 4749 (svn) +- 8192cd_sme.c - revision: 4749 (svn) +- 8192cd_tkip.c - revision: 4749 (svn) +- 8192cd_tx.c - revision: 4749 (svn) +- 8192cd_util.c - revision: 4749 (svn) +- 8192cd_util.h - revision: 4749 (svn) +- wifi.h - revision: 4749 (svn) + +* DESCRIPTION +- Add below compile tags for this patch. +- 'NOT_RTK_BSP' + Avoid utlizing Realtek platform specified items + and fit other platforms. +- 'NETDEV_NO_PRIV' + If linux version >= 2.6.29 and Non-Realtek platform, + struct "net_device" does NOT have element named "priv", + so use this tag to solve the issue. +- 'SMP_SYNC' + Because the previous protection is designed for uni-processor, + and such design will incur dead-lock when CPU supporting SMP, + so add this tag for modifications about SMP. +- 'PACK_STRUCTURE' + No padding in some structures for specified platform. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/19 , YangQuan + +* DIGEST: +- sync code from kernel 2.4 branch 1.2 at 20110118 for the following issue +- INCLUDE_WPS +- A4_STA +- SW_ANT_SWITCH +- igmp, icmpv6 +- _SLINUX + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- + +* DESCRIPTION +- there are some differences in mlcst2unicst() +- please see 1.2 branch change.txt for more details about the new feature + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/17 , victorlo, bug fix + +* DIGEST: +- Refine LCK + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 4723 + +* DESCRIPTION +- Refine LCK + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/13, jason, new feature + +* DIGEST: +- Added rtl8196cs support. + +* TEST METHOD: + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h -revision: 4720 +- 8192cd_hw.c -revision: 4720 +- 8192cd_osdep.c -revision: 4720 + +* DESCRIPTION +- Added rtl8196CS support. + +----------------------------------------------------------------------------- + + +- 2011/01/15, chrisyeh, rollback + +* DIGEST: +- rollback firmware FW v10.2 +- disable TX_REPORT for 92D + +* TEST METHOD: +- Run Chariot with sleep mode enabled Intel client. BB will hangup. + +* ROOT CAUSE: +- fw v14 may cause CCK hangup in 5G band + +* MODIFIED FILES: +- 8192cd_cfg.h -revision: 4709 +- data_92d/rtl8192dfw_n.bin -revision: 4709 + +* DESCRIPTION +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/14, victorlo, code merge + +* DIGEST: +- Rollback commited ECOS code from source + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: revision 4707 +- 1x_kmsm_aes.c +- 1x_kmsm_hmac.c +- 8192cd.h +- 8192cd_aes.c +- 8192cd_dmem.c +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_ioctl.c +- 8192cd_mib.c +- 8192cd_osdep.c +- 8192cd_psk.c +- 8192cd_psk.h +- 8192cd_rx.c +- 8192cd_security.c +- 8192cd_security.h +- 8192cd_sme.c +- 8192cd_tkip.c +- 8192cd_tx.c +- 8192cd_util.c +- ieee802_mib.h +- wifi.h + +* DESCRIPTION +- Rollback commited ECOS code from source + +----------------------------------------------------------------------------- +- 2011/01/14, chrisyeh, feature add + +* DIGEST: +- update firmware 8192D_8051FW_V0014_release +- enable TX_REPORT for 92D + +* TEST METHOD: + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h -revision: 4702 +- data_92d/rtl8192dfw_n.bin -revision: 4702 + +* DESCRIPTION +- enable 92D TX_REPORT from V0014 + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/14, victorlo, code merge + +* DIGEST: +- Merge ECOS from CVS trunk + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: revision 4686 +- 1x_kmsm_aes.c +- 1x_kmsm_hmac.c +- 8192cd.h +- 8192cd_aes.c +- 8192cd_dmem.c +- 8192cd_cfg.h +- 8192cd_headers.h +- 8192cd_hw.c +- 8192cd_hw.h +- 8192cd_ioctl.c +- 8192cd_mib.c +- 8192cd_osdep.c +- 8192cd_psk.c +- 8192cd_psk.h +- 8192cd_rx.c +- 8192cd_security.c +- 8192cd_security.h +- 8192cd_sme.c +- 8192cd_tkip.c +- 8192cd_tx.c +- 8192cd_util.c +- ieee802_mib.h +- wifi.h + +* DESCRIPTION +- Merge ECOS from CVS trunk + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/13, jerryko, feature refine + +* DIGEST: +- 1.add define SW_TX_QUEUE to improve performance when multiple clients +- 2.add define MCAST2UI_REFINE to improve vlc traffic + +* TEST METHOD: +- 1.veri-wave test 31 clients to measure performance +- 2.wmm 4.2.20 and 4.2.21 test +- 3.vlc test ethernet to wlan +- 4.vlc test wlan to wlan + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h -revision: 4682 +- 8192cd_cfg.h -revision: 4682 +- 8192cd_ioctl.c -revision: 4682 +- 8192cd_proc.c -revision: 4682 +- 8192cd_rx.c -revision: 4682 +- 8192cd_sme.c -revision: 4682 +- 8192cd_util.c -revision: 4682 +- 8192cd_tx.c -revision: 4682 + +* DESCRIPTION +- 1.When driver detect transmit DA mac or tid switch times over + high_threshold, sw queue will enable and start to queue incoming packets. + When queue packets over limit or timeout, driver will send all packets in + queue. When driver detect transmit DA mac or tid switch times lower than + low_threshold , sw queue will disable and dequeue all packets in queue. +- 2.MCAST2UI_REFINE use skb_clone instead of skb_copy when open and ccmp + security for wlan tx multicast to unicast. +- 3.For improve wlan to wlan vlc traffic, wlan rx check if multicast data or + not. And multicast data will send ethernet tx and wlan tx directly. + If multicast management will send to upper layer. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/13, davidhsu, bug fix + +* DIGEST: +- Fix 92d diver hang up issue when receiving fragment packets but not AMSDU + +* TEST METHOD: +- Config 96c/98 + 92d only, enable 1x1 dualband concurrent +- Let a DWA-130 client to connect to AP in 2.4G band (This client will send + packet in AMSDU) +- Let a 5G client connect to AP in 5G band +- Run BT in these two clients +- AP will hang up (console hang) after running a preiod time + +* ROOT CAUSE: +- 5G mac will report fragment packets but not in AMSDU. Thus, it will + cause driver looply in parsing Rx packet. + +* MODIFIED FILES: +- 8192cd_rx.c revision: 4663 + +* DESCRIPTION +- Discard fragment packets if it is not AMSDU + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/12, hf, debug refine + +* DIGEST: +- show dz queue len in proc + +* TEST METHOD: + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_proc.c -revision: 4657 + +* DESCRIPTION +- show dz queue len in proc + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/11, button, feature refine + +* DIGEST: +- Refine auto channel selection + +* TEST METHOD: + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_header.h -revision: 4343 +- 8192cd_cfg.h -revision: 4343 +- 8192cd.h -revision: 4343 +- 8192cd_hw.c -revision: 4343 +- 8192cd_sme.c -revision: 4343 +- 8192cd_ioctl.c -revision: 4643 + +* DESCRIPTION +- select channl by BSS number, false alarm and CCA count + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/11, chuangsw, bug fix + +* DIGEST: +- fix wifi logo problem (4.2.4 & 4.2.21-T06) + +* TEST METHOD: +- 4.2.4 : we can duplicate the problem by ping big packet. + ex: ping xxx -l 1500 + +- 4.2.21-T06 : setup as the wifi logo test discription. + we can duplicate it by turning on the aggregation of + the AP when doing the WMM chariot test. + +* ROOT CAUSE: +- 4.2.4 : the MAX_SKB_BUF(2256) is not enough, so it will be drop + in reassemble() (at 8192cd_rx.c). + error message: + "RX DROP: over rx buf size after reassemble". + +- 4.2.21-T06 : the root cause is not found + +* MODIFIED FILES: +- 8192cd_cfg.h -revision: 4641 +- 8192cd_tx.c -revision: 4641 + +* DESCRIPTION +- 8192cd_cfg.h : Ajust the "MAX_SKB_BUF" size 2256 to 2260. +- 8192cd_tx.c : delete the is_intel_sta cheching before set_desc(TX_AggEn) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/10, chrisyeh, bug fix + +* DIGEST: +- No full-band SS for AP auto-channel + +* TEST METHOD: +- Select auto-channel and see if the channel is at the right band. + +* ROOT CAUSE: +- Do full-band SS for Client, AP Site survey, but no auto-channel. + +* MODIFIED FILES: +- 8192cd_sme.c -revision: 4625 + +* DESCRIPTION +- Switch to full-band if auto_channel != 1 (that means not AP auto-channel) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/10, jimmylin, feature refine + +* DIGEST: +- Add cli protection in some functions for association list + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_security.c - revision 1.18 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/10, button, feature refine + +* DIGEST: +- update for high power board + +* TEST METHOD: +- test in spectrum analyzer by Anchi + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c revision: 4587 + +* DESCRIPTION +- modify PCIe clock from 100M to 102M to reduce harmonic in 2.4G high channel +- in order conform to FCC standard + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/01/07, victorlo, bug fix + +* DIGEST: +- Refine RSC signin of EAPOL 4-3 and 2-1 packet with a correct PN value + +* TEST METHOD: +- Set Enc mode as WPA2/AES. +- Use Broadcom station (e.g. iphone) associates AP, DHCP and ARP req + from ethernet success. Ping success from ethernet. +- After client associated, generate lots of broadcast packets from AP + (e.g. ARP attack) +- Stop broadcast packets when "dot11EncryptKey.dot11TXPN48.val48" + greater than 0xffff, which is displayed in proc/wlan0/mib_gkeytbl +- Reconnect Broadcom station again, DHCP and ARP req fail. Ping fails + +* ROOT CAUSE: +- Wrong PN value of group key has been mistakenly signed in EAPOL 4 way + (or 2 way) handshake when greater than 0xffff, cause associating + station accepted a wrong unreasonable large PN value as the initial + sequence number of broadcast packet from AP. +- After 4 way (or 2 way) handshake, broadcast packet with correct PN + received by client will be dropped. + +* MODIFIED FILES: +- 8192cd_psk.c - revision:4582 + +* DESCRIPTION +- Endian issue of signin packet value by memcpy, + refine RSC signin of EAPOL 4-3 and 2-1 packet with a correct PN value + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/7, chrisyeh, bug fix + +* DIGEST: +- handle sleep mode management frame +- fine tune dynamic noise control + +* TEST METHOD: +- Broadcom client cannot aggregate packet when running chariot + +* ROOT CAUSE: +- Broadcom client send ADDBAReq when sleeping + +* MODIFIED FILES: +- 8192cd_cfg.h -revision: 4567 +- 8192cd_headers.h -revision: 4567 +- 8192cd_hw.c -revision: 4567 +- 8192cd_sme.c -revision: 4567 +- 8192cd_osdep.c -revision: 4567 +- 8192cd_tx.c -revision: 4567 + +* DESCRIPTION +- We queue management frame if client is sleeping + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/7, chrisyeh, bug fix + +* DIGEST: +- fix 92D load firmware fail & reinit fail + +* TEST METHOD: +- Long run test of reinit. + +* ROOT CAUSE: +- Move Load_92D_Firmware later than PHY init +- Remove MAC I/O off before Firmware Self-reset + +* MODIFIED FILES: +- 8192cd_hw.c -revision: 4561 +- 8192cd_osdep.c -revision: 4561 +- 8192d_hw.c -revision: 4561 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/3, button, bug fix + +* DIGEST: +- fix 92d tx EVM drop 5 dB + +* TEST METHOD: +- build 92c + 92d nfjrom, test in IQ-view + +* ROOT CAUSE: +- 92d power depends on 92c circuit in v630 board + +* MODIFIED FILES: +- 8192cd_hw.c -revision: 4526 + +* DESCRIPTION +- 92c power always in PWM mode + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2011/1/3, button, bug fix + +* DIGEST: +- console hang after long run sw re-init + +* TEST METHOD: + +* ROOT CAUSE: +- CPU continues receive GPIO interrupt, because IRQ status is not zero. + +* MODIFIED FILES: +- 8192cd_hw.c -revision: 4523 +- 8192cd.h -revision: 4523 + +* DESCRIPTION +- request before WLAN open, ensure all IRQ is processed. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/24, chris, bug fix + +* DIGEST: +- Fix SW_LCK cause SiteSurvey fail +- Add compile flag 'SW_LCK_92D' for software LCK + +* TEST METHOD: +- Do Site Survey in SMSP mode. See if there are full-band APs. + +* ROOT CAUSE: +- Reload LCK value should follow the current channel setting, not mib value. + +* MODIFIED FILES: +- 8192cd_cfg.h - rev. 4422 +- 8192cd_headers.h - rev. 4422 +- 8192cd_hw.c - rev. 4422 +- 8192cd_mp.c - rev. 4422 +- 8192d_hw.c - rev. 4422 + +* DESCRIPTION +- In ReloadLCK function, channel is obtained by query RF register. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/20, jerryko, feature add + +* DIGEST: +- 92D support DFS feature + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - rev. 4332 +- 8192cd_cd.h - rev. 4332 +- 8192cd_hw.c - rev. 4332 +- 8192cd_ioctl.c - rev. 4332 +- 8192cd_proc.c - rev. 4332 +- 8192cd_sme.c - rev. 4332 +- 8192d_hw.c - rev. 4332 + +* DESCRIPTION +- support DFS feature. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/13, button, bug fix + +* DIGEST: +- Atheros, Broadcom can't go back to OFDM rate once it dropped to +- CCK rate in range testing. + +* TEST METHOD: +- use attenuator to emulate long range test, let client tx rate dtop to CCK by +- setting attenuation 50dB, run chariot uplink 10 sec and set attenuation to 0 dB, +- see if client's rate go back to OFDM. + +* ROOT CAUSE: +- Broadcom client is sensitive to CCK SIFS, we set it too large before. +- Atheros rate adaptive mechanism may reference ACK rate, we mask +- 5.5 11 Mbps ACK will let Atheros stay in CCK rate. + +* MODIFIED FILES: +- 8192cd_h - revision: 4274 +- data/MACPHY_REG_92C.txt - revision: 4274 + +* DESCRIPTION +- adjust CCK SIFS with 10 uS +- enable all CCK rate for ACK frame. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/9, chris, feature add + +* DIGEST: +- 92D Software LCK & TX power tracking + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_headers.h - rev. 4257 +- 8192cd_hw.c - rev. 4257 +- 8192cd_mp.c - rev. 4257 +- 8192cd_proc.c - rev. 4257 +- 8192cd_tx.c - rev. 4257 +- 8192d_hw.c - rev. 4257 + +* DESCRIPTION +- 92D LCK is implemented by software. LCK value should be loaded after switch channel each time. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/9, chris, code refine + +* DIGEST: +- Remove 92D TESTCHIP support. +- Refine 92D+92C dual card. +- Add PHY_REG_PG.txt for 92D. + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - rev. 4250 +- 8192cd_headers.h - rev. 4250 +- 8192cd_hw.c - rev. 4250 +- 8192cd_mp.c - rev. 4250 +- 8192cd_proc.c - rev. 4250 +- 8192cd_sme.c - rev. 4250 +- 8192d_hw.c - rev. 4250 + +- data/PHY_REG_MP_n.txt - removed +- data/PHY_REG_MP_n_92C.txt - added rev. 4251 +- data/PHY_REG_PG.txt - removed +- data/PHY_REG_PG_92C.txt - added rev. 4251 +- data_92d/AGC_TAB_2G.txt - removed +- data_92d/AGC_TAB_5G.txt - removed +- data_92d/PHY_REG.txt - removed +- data_92d/PHY_REG_MP.txt - removed +- data_92d/radio_a.txt - removed +- data_92d/radio_b.txt - removed +- data_92d/rtl8192dfw.bin - removed +- data_92d/rtl8192dfw_smcc.bin - removed + +* DESCRIPTION +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/7, chris, bug fix + +* DIGEST: +- 92D DMDP wlan1 tx power MUST read path-B values + +* TEST METHOD: +- N/A + +* ROOT CAUSE: +- Because DMDP wlan1 use path-B to Tx/Rx, we should use path-B values. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 4219 + +* DESCRIPTION +- N/A + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/6, chris, bug fix + +* DIGEST: +- Enable firmware self reset function +- Add 92D firmware header auto-recognize + +* TEST METHOD: +- Soft reboot many times and see if firmware load fail. + +* ROOT CAUSE: +- Firmware self reset function should be enabled. +- Add: Mask IMR when firmware in reset + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 4205 +- 8192d_hw.c - revision: 4205 + +* DESCRIPTION +- If recognize 92D f/w header 92DX, add 32 bytes in front of fw_start. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/12/3, chris, bug fix + +* DIGEST: +- FW: 8192D_FW_v10.2_1202_3(with header), fix intel IOT + +* TEST METHOD: +- Intel client will use low rate (<=MCS1) tx to DUT + +* ROOT CAUSE: +- Firmware has some timing issue, not clearified yet. + +* MODIFIED FILES: +- data_92d/rtl8192dfw_n.bin - rev. 4184 + +* DESCRIPTION +- N/A +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/29, jason, bug fix + +* DIGEST: +- Modified AP RX performance penalty for dual card 92D+92C in the rtl8198 platform. + +* TEST METHOD: +- Use rtl8198 DUT and build dual card 92D+92C support to test performance. + +* ROOT CAUSE: +- When enabling RX_GATER, the AP RX performance of the 92D is lower than 100Mbps in dual card platform (92D+92C). So roll back to the disable RX_GATER. + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 4138 +- 8192cd_util.c - revision: 4138 + +* DESCRIPTION +- Roll back to dsiable RX_GATER and modified the MAX_SKB_NUM to improve the performance. +- The bug is reported to the Mantis. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/11/26, chris, bug fix & feature add + +* DIGEST: +- 92D firmware update 8192D_8051FW_V0012_release +- fix wep/tkip RA fail +- check 92D PN by firmware + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h - revision: 4096 +- 8192cd_headers.h - revision: 4096 +- 8192cd_hw.c - revision: 4096 +- 8192cd_util.c - revision: 4096 +- 8192d_hw.c - revision: 4096 +- data_92d/rtl8192dfw_n.bin - revision: 4096 +- data_92d/rtl8192dfw_smcc_n.bin - removed! + +* DESCRIPTION +- firmware 8192D_8051FW_V0012_release needs driver claim maximum used macid for + rate adaptive. Then using legacy rate for WEP/TKIP can be fixed. +- 92D part number should be checked by firmware. If firmware loaded fail, driver + should not start. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/26, button, feature refine + +* DIGEST: +- refine for high power demo boards + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h revision:4086 +- 8192cd_headers.h revision:4086 +- 8192cd_hw.c revision:4086 +- 8192cd_ioctl.c revision:4086 +- 8192cd_mp.c revision:4086 +- data/PHY_REG_1T_n_hp.txt revision:4087 + +* DESCRIPTION +- add phy parameters for 88RE high power demo board +- add mib "hp_ofdm_max", "hp_cck_max" to limit maximum tx power in MP mode + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/24, jason, feature add + +* DIGEST: +- "Fixed the PCIe bios code dump issue if existing 92D+92C. + +* TEST METHOD: +- Put DUT side by side, see the dual wlan interface work normally. + +* ROOT CAUSE: +- Null pointer access in the PCIe bios state if the private data does not exist. + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 4049 +- 8192cd_osdep.c - revision: 4049 + +* DESCRIPTION +- The pointer is not initalization before the private data is not allocation. +- The private data is dynamically allocated in the PCIe bios state. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/11/23, chris, feature add + +* DIGEST: +- 92D Dynamic Noise Control + +* TEST METHOD: +- Put DUT side by side, see if TP will drop when all DUT traffic starts + +* ROOT CAUSE: +- TP will be affected by noise from adjacent 5G channels + +* MODIFIED FILES: +- 8192cd.h - revision: 4011 +- 8192cd_cfg.h - revision: 4011 +- 8192cd_headers.h - revision: 4011 +- 8192cd_hw.c - revision: 4011 +- 8192cd_osdep.c - revision: 4011 +- 8192cd_proc.c - revision: 4011 +- 8192cd_tx.c - revision: 4011 +- 8192cd_util.c - revision: 4011 + +* DESCRIPTION +- Avoid Rx noise +- Lower Tx power. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/17, button, feature refine + +* DIGEST: +- Support USB power + +* TEST METHOD: +- + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h - revision: 3915 +- 8192cd_cfg.h - revision: 3915 +- 8192cd_hw.c - revision: 3919 +- 8192cd_util.c - revision: 3919 +- Kconfig - revision: 3915 + +* DESCRIPTION +- Lower 2 spatial stream power index +- Mask MCS12, MCS11 of sta tx map + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/15, chrisyeh, feature add + +* DIGEST: +- 92D power tracking +- 92D crystal cap function +- EFUSE MAP is also refined + +* TEST METHOD: +- if there is thermal meter value, tx power index will change as the thermal value varies. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_headers.h - revision: 3835 +- 8192cd_hw.c - revision: 3835 +- 8192cd_hw.h - revision: 3835 +- 8192cd_ioctl.c - revision: 3835 +- 8192cd_mp.c - revision: 3835 +- 8192cd_osdep.c - revision: 3835 +- 8192cd_phyreg.h - revision: 3835 +- 8192cd_proc.c - revision: 3835 +- 8192cd_sme.c - revision: 3835 +- 8192d_hw.c - revision: 3835 +- 8192d_reg.h - revision: 3835 +- ieee802_mib.h - revision: 3835 + +* DESCRIPTION +- 92D power tracking related registers and procedures are different from 92C. +- 92D crystal cap function +- 92D efuse map length is double of 92C's. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/15, chrisyeh, bug fix + +* DIGEST: +- use skb_copy for 92D broadcast packet MIC error + +* TEST METHOD: +- Use WAP-WAP2-mixed encryption for dual band concurrent mode, then client will be disassociated due to MIC error. + +* ROOT CAUSE: +- Orignally, broadcast packet is only cloned. It seems like the data is damaged. + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 3835 +- 8192cd_rx.c - revision: 3835 + +* DESCRIPTION +- In rx and tx processes, use skb_copy instead of skb_clone. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/11/11, button, feature refine + +* DIGEST: +- ADD_TX_POWER_BY_CMD supports minus value + +* TEST METHOD: +- set power plus value by iwpriv command for calibrated and uncalibrated +- demo boards, and check the difference of power index (reg e00, e04... ) + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 3764 +- 8192cd.h - revision: 3764 +- 8192cd_hw.h - revision: 3764 +- 8192cd_ioctl.c- revision: 3764 + +* DESCRIPTION +- use example: iwpriv wlan0 set_mib txPowerPlus_cck_1=-2 +- available range:-128 ~ +126, +127 use default PHY_REG_PG.txt + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/11, button, feature fine tune + +* DIGEST: +- fine tune for high power demo board + +* TEST METHOD: +- channel emulator test + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd.h - revision: 3276 +- 8192cd_headers.h - revision: 3276 +- 8192cd_hw.c - revision: 3276 +- 8192cd_hw.h - revision: 3276 +- 8192cd_osdep.c - revision: 3276 +- 8192cd_proc.c - revision: 3276 +- 8192cd_sme.c - revision: 3276 +- 8192cd_util.c - revision: 3276 +- data/PHY_REG_2T_n_hp.txt - revision: 3276 +- data/PHY_REG_PG_hp.txt - revision: 3276 + +* DESCRIPTION +- update phy parameter +- modify rssi threshold of dynamic algorithm +- add tx power control, disable beacon AGC + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/10, victorlo, feature refine + +* DIGEST: +- Enhance Beacon tx + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 3720 + +* DESCRIPTION +- By MAC register refine, to enlarge TBTT hold time and shorter the beacon + IFS time, to enhance beacon tx opportunity + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/10, davidhsu, bug fix + +* DIGEST: +- Packet length is 0 when re-send wpa 4-ways message in client mode + +* TEST METHOD: +- Config in client mode with PSK enabled. +- It may have chance in failure of wpa 4-ways handshaking + +* ROOT CAUSE: +- EAPOL packet length did not be recorded in 1st message + +* MODIFIED FILES: +- 8192cd_psk.c - revision: 1.7 + +* DESCRIPTION +- Record the eapol packet length in 1st message + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/10, chris, bug fix + +* DIGEST: +- For beacon Tx after each bcn_err happens + +* TEST METHOD: +- In open space, sniffer beacon and should not see beacon interval larger than 200ms. + +* ROOT CAUSE: +- Beacon can be blocked for a long time if CCA keeping happening. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 3685 + +* DESCRIPTION +- Enable MAC Force beacon function: +- a) 0x524[13] = 1??b1 +- b) 0x55d[7:0] set to a value not equal to 8??hff, the number indicates that 92DE will + TX bcn regardless of the channel once the BCN failed number is larger than the register setting. +- c) 0x642[7:0] set the EIFS less than TBTT_PROHIBIT_TIME + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/09, button, bug fix + +* DIGEST: +- Auto channel side band issue + +* TEST METHOD: +- Configure 92d to 5G 2x2, bandwidth 40MHz, auto channel. +- check the channel number, and side band selected by our algorithm + +* ROOT CAUSE: +- we consider 2.4G only in sideband selection. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 3681 + +* DESCRIPTION +- select channel and sideband follow the document + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/09, button, bug fix + +* DIGEST: +- Tx rate error for legacy encryption STA + +* TEST METHOD: +- Configure DUT to AES/TKIP mix mode, and set tx rate to MCS 15. +- Let a wireless client connect to DUT in TKIP and ping DUT, +- sniffer the packets from DUT to STA. + +* ROOT CAUSE: +- If user set fix tx rate, we just assign data rate in tx descriptor, +- do not check the capability of target STA. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 3668 +- 8192cd_tx.c - revision: 3668 +- 8192cd_util.h - revision: 3668 + +* DESCRIPTION +- If the rate set of STA do not contain the fixed tx rate assign by user, +- we use FW rate adaptive instead. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/11/04, Jason, bug fixed + +* DIGEST: +- Fixed the rtl8192d rebooting issue about missing rtl8192d PCIe card. + +* TEST METHOD: +- Build 98+AP package and make to fail in initialize AP MIB. Reboot the DUT and make sure the 92d PCIe existing. + +* ROOT CAUSE: +- The rtl8192d PHY can not link the PCIe root complex in the rtl8198 platform. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 3648. (svn) + +* DESCRIPTION +- 1) Added the following sequence in stop hardware. + - Reset the Root Complex PHY and pReset. + - Initialize the MDIO reset. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/11/04, Jason, new feature + +* DIGEST: +- Added auto-recognition pcie card and bypass non-exist method in rtl8198 platform. + + +* TEST METHOD: +- Removed the 92D and run the image normally with 92C. +- And both (92D+92C) run normally in the rtl8198 platform. + +* ROOT CAUSE: +- The recognition of the PCIE does not have error exception handler. + + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 3588. (svn) + +* DESCRIPTION +- 1) Check PCIE Link status and decide the wlan interface number. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/11/04, Jason, new feature + +* DIGEST: +- 92C+92D Concurrent mode for dual PCIe cards in rtl8198 platform. + + +* TEST METHOD: +- Run each Single Mode chariot up to 160Mbps and Dual Mode chariot up to 26xMbps. + + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Kconfig - revision: 3565. (svn) +- 8192cd_osdep.c - revision: 3565. (svn) +- 8192cd_hw.c - revision: 3565. (svn) +- 8192cd_cfg.h - revision: 3565. (svn) +- 8192cd_headers.h - revision: 3565. (svn) +- 8192cd_proc.c - revision: 3565. (svn) +- Makefile - revision: 3565. (svn) +- data/AGC_TAB_n_92C.txt - revision: 3565. (NEW) +- data/MACPHY_REG_92C.txt - revision: 3565. (NEW) + +* DESCRIPTION +- 1) Added compile tag 'CONFIG_RTL_DUAL_PCIESLOT_BIWLAN_D' to support rtl8192c +rtl8192d dual card support in rtl8198 platform. +- 2) Change data/AGC_TAB_n to data/MACPHY_REG.txt to data/AGC_TAB_n_92C to data/MACPHY_REG_92C.txt to avoid the same files name in data and data_92d directory. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/10/21, Chris, new feature & bug fix + +* DIGEST: +- 92D Client mode full-band site survey. +- 92D MP-Chip Smart concurrent. +- 92D fine tune. + +* TEST METHOD: +- Full-band Site Survey: Set AP to 2x2 Client mode, the site survey result will show both 5G and 2G APs. +- Smart Concurrent: Build with "Smart concurrent support", if only one band is connected, the client uses 2T rates. + But clients on the other band can still discover the AP. When both bands are connected, the clients use 1T rates. +- Update PHY parameters 92D_MPChip_PHYforMP_101013 and Synthersizer setting. Add FW for smart concurrent + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Kconfig - revision: 3409 (svn) +- 8192cd_osdep.c - revision: 3409 (svn) +- 8192c_reg.h - revision: 3409 (svn) +- 8192cd_hw.c - revision: 3409 (svn) +- 8192d_reg.h - revision: 3409 (svn) +- 8192cd_hw.h - revision: 3409 (svn) +- 8192cd_sme.c - revision: 3409 (svn) +- 8192cd_tx.c - revision: 3409 (svn) +- 8192cd_cfg.h - revision: 3409 (svn) +- 8192cd_ioctl.c - revision: 3409 (svn) +- 8192d_hw.c - revision: 3409 (svn) +- 8192cd_mp.c - revision: 3409 (svn) +- 8192cd_util.c - revision: 3409 (svn) +- 8192cd_headers.h - revision: 3409 (svn) +- 8192cd.h - revision: 3409 (svn) +- data_92d/radio_a_n.txt - revision: 3409 (svn) +- data_92d/radio_b_n.txt - revision: 3409 (svn) +- data_92d/rtl8192dfw_smcc_n.bin - revision: 3409 (svn) (NEW) + + +* DESCRIPTION +- 1)When starting site survey, set network type to A/B/G/N to get all channel map. + When doing site survey, function "clnt_ss_check_band()" checks channel and switches to + the respective PHY_BAND. +- 2)Smart concurrent: open interface by 2x2_A0->1x1A0G1 then 1x1A0G1->2x2G1->1x1A0G1 to + record IQK results and related reg values. +- 3)For new PHY-parameters, 92D must load synthersizer setting before switching channel. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/10/15, ZhangJian, buf fix + +* DIGEST: +- Refine for hostapd. + +* TEST METHOD: +- Enable hostapd that is choosing "Realtek hostapd support" at kernel menuconfig + and selecting "openssl" & "hostapd in host controlled mode" at users menuconfig, + using 1) RTL8192su(winXP driver: 1084.36.708.2010, 2010-7-8), + 2) DWA140(winXP driver: 1.4.3.0, 2009-5-19), + 3) WN111(3.0.0.141, 2009-1-13), + test wpa2-psk, wps push button and wps pin OK. + +* ROOT CAUSE: +- 1) If rsn ie differs between beacon (or probe rsp) and 3/4 msg of 4-Way Handshake, + wlan client will not send out 4/4 msg of 4-Way Handshake. + +* MODIFIED FILES: +- 8192cd_net80211.c - revision: 3334 (svn) +- 8192cd_osdep.c - revision: 3334 (svn) +- 8192cd_psk_hapd.c - revision: 3334 (svn) +- Kconfig - revision: 3334 (svn) + +* DESCRIPTION +- 1) rsn ie constructed at wlan driver shoud be consistent with hostapd code + such as wmm_enabled. + 2) expired_time of /proc/wlan0/sta_info should be consistent with + ap_max_inactivity at hostapd conf file. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/09/15, EricYeh, new feature + +* DIGEST: +- Support requests(listed below) from TP-Link. + 1. WiFi AP works without daemons 'auth' 'iapp' 'iwcontrol' + 2. 802.1x Authentication (RADIUS) & IAPP are handled by 'Hostapd-0.5.9' + 3. PSK is handled by 'WiFi Driver' itself. + 4. WPS is handled by 'wscd'. + +* TEST METHOD: +- Remove daemons 'auth' 'iapp' 'iwcontrol', run 'wsc daemon' and modified + 'Hostapd-0.5.9', verify related functions (Security, WPS, WDS, IAPP...) + can work well or not. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 2973 (svn) +- 8192cd_cfg.h - revision: 2973 (svn) +- 8192cd_ioctl.c - revision: 2973 (svn) +- 8192cd_net80211.c - revision: 2973 (svn) +- 8192cd_osdep.c - revision: 2973 (svn) +- 8192cd_psk.c - revision: 2973 (svn) +- 8192cd_psk.h - revision: 2973 (svn) +- 8192cd_psk_hapd.c - revision: 2973 (svn) +- 8192cd_rx.c - revision: 2973 (svn) +- 8192cd_security.c - revision: 2973 (svn) +- 8192cd_sme.c - revision: 2973 (svn) +- 8192cd_util.c - revision: 2973 (svn) + +* DESCRIPTION +- Add compile tag 'HAPD_DRV_PSK_WPS' to figure out which modifications made + for supporting this request. +- Use tag HAPD_DRV_PSK_WPS to make driver can fully handle WPA-PSK funciton. +- Use tag HAPD_DRV_PSK_WPS to restore all EAP_BY_QUEUE function, in order to + make 'wsc daemon' can receive EAP packets and madatory events. +- To support this feature, the modified Hostapd-0.5.9 shall ONLY manage + 802.1x Authentication & IAPP. +- To support this feature, compile tag 'NO_IWCONTROL' shll be defined when + making 'wsc daemon'. +- To support this feature, compile tag 'WIFI_HAPD' shall also be defined. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/09/15, EricYeh, new feature + +* DIGEST: +- Support Hostapd. + +* TEST METHOD: +- Remove daemons 'auth' 'iapp' 'wsc daemon', run Hostapd as daemon, + check WiFi settings(ex: /proc/wlan0/mib_all) are corresponding to + imported hostpad configuration file, and verify related functions + (Security, WPS, WDS, IAPP...)can work well or not. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 1x_kmsm_hmac.c - revision: 2973 (svn) +- 1x_md5c.c - revision: 2973 (svn) +- 8192cd.h - revision: 2973 (svn) +- 8192cd_cfg.h - revision: 2973 (svn) +- 8192cd_comapi.c - revision: 2973 (svn) +- 8192cd_headers.h - revision: 2973 (svn) +- 8192cd_ioctl.c - revision: 2973 (svn) +- 8192cd_net80211.c - revision: 2973 (svn) +- 8192cd_net80211.h - revision: 2973 (svn) +- 8192cd_osdep.c - revision: 2973 (svn) +- 8192cd_proc.c - revision: 2973 (svn) +- 8192cd_psk.c - revision: 2973 (svn) +- 8192cd_psk.h - revision: 2973 (svn) +- 8192cd_psk_hapd.c - revision: 2973 (svn) +- 8192cd_rx.c - revision: 2973 (svn) +- 8192cd_security.c - revision: 2973 (svn) +- 8192cd_sme.c - revision: 2973 (svn) +- 8192cd_util.c - revision: 2973 (svn) + +* DESCRIPTION +- Add compile tag 'WIFI_HAPD' to figure out which modifications made for + supporting Hostapd. +- Hostapd will handle all EAP packets, so 'EAP_BY_QUEUE' & + 'INCLUDE_WPA_PSK' are undefined in 8192cd.h, which means EAP packets + can be received by Hostapd from l2 socket, and EAP parts of WPA-PSK + will not be managed in driver. +- In order to keep driver handle 802.11 management frames for WPA-PSK + (Hostapd only handle EAP packets), add new file '8192cd_psk_hapd.c' and + utilizing WIFI_HAPD to retain these PSK parts. + (Because INCLUDE_WPA_PSK will be undefined if define WIFI_HAPD) +- Add new file '8192cd_net80211.c' & '8192cd_net80211.h' to deal with + IO Control sent from driver interface of Hostapd. +- Add function 'rsn_init()' to generate rsn_ie for 802.1x(RADIUS). +- Add function 'event_indicate_hapd()' in 8192cd_util.c to report + events to Hostapd by netlink. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/29, button, buf fix + +* DIGEST: +- perofrmance drop after htc connect + +* TEST METHOD: +- connect to DUT and disconnect, observe proc/wlan0/sta_info, +- test fail if remapped_aid = 0 + +* ROOT CAUSE: +- remapped_aidarray won't clear when STA disconnect + +* MODIFIED FILES: +- 8192cd_cd.h - revision: 1.76 +- 8192cd_util.c - revision: 1.65 +- 8192cd_headers.h - revision: 1.42 +- 8192cd_sme.c - revision: 1.116 +- 8192cd_proc.h - revision: 1.45 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/29, button, bug fix + +* DIGEST: +- fix reconnection issue of vxd AP of repeater mode + +* TEST METHOD: +- set a AP as repeater mode. let root client connect to a AP, +- then run WPS PIN configuration of vxd AP and other STA. +- The STA cannot connet to VXD AP. + +* ROOT CAUSE: +- RCR CBSSID was set, HW filtered the packets to vxd AP. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.115 +- 8192cd_osdep.c - revision: 1.85 + +* DESCRIPTION +- CBSSID only set in AD-HOC mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/28, davidhsu, new feature + +* DIGEST: +- Add IPv6 pass-through in client + +* TEST METHOD: +- Setup two IPv6 PCs. One PC connected to AP Ether LAN, and otherother PC + connected to AP-client by Ethernet. +- Run VLC streaming between these two PC with IPv6 multicast address. + +* ROOT CAUSE: +- N/A + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 1.77 +- 8192cd_br_ext.h - revision: 1.2 +- 8192cd_br_ext.c - revision: 1.3 + +* DESCRIPTION +- Add a new define flag CL_IPV6_PASS for this feature. +- When CL_IPV6_PASS is defined: + * networkaddr of NAT25 DB will be enlarged from 11 to 17 in order to + store ipv6 address (16 bytes). + * When converting mac address from Ethernet mac to WiFi mac, beside to + replace dst mac addr, it needs check for network discovery packet to + update link-layer-address field to WiFi mac in ICMPv6 option field. + * When receiving a IPv6 packet from wlan and did not find a matched + IP address, it needs to check if dst IP is multicast (first byte is 0xff). + If yes, it need reverse dst MAC to multicast addr because our AP + may convert multicast to unicast address and then sent it to AP-client. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/28, button, buf fix + +* DIGEST: +- perofrmance drop after htc connect + +* TEST METHOD: +- Wireless client AR928X connect to DUT, and runs the throughput.. +- cell phone(htc tattoo) connects to DUT, +- AR928X's throughput would drop a lot, even the cell phone is not accessing data. +- After disconnect the cell phone and wait for many minutes, +- AR928X's throughput did not recover to the original performance. + +* ROOT CAUSE: +- "remapped_aidarray" do not reset after re-init, +- if macid is used before, pstat->remapped_aid will be 0, +- and maps to broadcast rate table. after htc (BG mode) connects, +- broadcast rate table will be modified. perofrmance drop is due to 11G tx rate, + +* MODIFIED FILES: +- 8192cd_cd.h - revision: 1.75 +- 8192cd_util.c - revision: 1.64 + +* DESCRIPTION +- move remapped_aidarray into pshare clear area + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/23, button, new feature + +* DIGEST: +- hardware WAPI support + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- wapi_wai.c - revision: 1.5 +- wapi_wai.h - revision: 1.4 +- wapiCrypto.c - revision: 1.4 +- wapiCrypto.c - revision: 1.2 +- 8192cd_tx.c - revision: 1.49 +- 8192cd_rx.c - revision: 1.31 +- 8192cd_cfg.h - revision: 1.76 +- 8192cd_osdep.c - revision: 1.84 +- 8192cd_hw.c - revision: 1.138 +- 8192cd_sme.c - revision: 1.114 +- 8192cd_util.c - revision: 1.63 +- 8192cd_util.h - revision: 1.13 +- 8192cd_ioctl.c - revision: 1.46 +- 8192cd_proc.c - revision: 1.44 +- 8192cd_security.h - revision: 1.4 +- 8192cd_security.c - revision: 1.11 +- ieee802_mib.h - revision: 1.18 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2010/07/23, button, buf fix + +* DIGEST: +- power tracking error after reinit + +* TEST METHOD: +- keyin "flash set HW_WLAN0_11N_THER 10" in console, then reboot, +- record swing index(0xc80), compare with the value after reinit via web. + +* ROOT CAUSE: +- "priv->pshare->ThermalValue" do not reset in normal driver + +* MODIFIED FILES: +- 8192cd_cd.h - revision: 1.73 +- 8192cd_mp.c - revision: 1.39 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/20, jimmylin, bug fix + +* DIGEST: +- Fix abnormal behavior of interface closing if WDS earlier than root + +* TEST METHOD: +- Enable WDS interfaces. While applying change, close WDS interfaces earlier + than root interface. Then opening root interface will cause abnormal + behavior. + +* ROOT CAUSE: +- Because closing WDS interface, we would change the drv_state of root + interface. Then while closing root interface, the normal procedure will + be skipped. After that, re-open root interface will cause problems. + +* MODIFIED FILES: +- 8192cd_osdpe.c - revision: 1.82 + +* DESCRIPTION +- While closing WDS interfaces, don't change the drv_state of root interface. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/15, BUTTON, bug fix + +* DIGEST: +- Modify for TX report + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.71, 1.72 +- 8192cd_util.c - revision: 1.61, 1.62 +- 8192cd_sme.c - revision: 1.112, 1.113 +- 8192cd_osdep.c - revision: 1.81 + +* DESCRIPTION +- add CLI protection for tx report + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/14, jimmylin, bug fix + +* DIGEST: +- Fix reconnection issue of root client of repeater mode + +* TEST METHOD: +- Set a AP as WPA-PSK. Set DUT as repeater mode while root is client mode and + vxd is AP mode. Let root-client connect to the AP. Then power off and then + power on the AP. The root-client can no longer connect to the AP. + +* ROOT CAUSE: +- After root-client rescaned and found the AP, driver will check if the vxd + is existing or not to decide to set RCR_CBSSID or not. But the vxd + interface is closed while root-client starts roaming. So we should use + netif_running() instead of IS_DRV_OPEN() to check if the vxd is opened. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.111 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/14, button, feature fine-tune + +* DIGEST: +- refine L1/L2 -> L0 procedure, check FW status. + +* TEST METHOD: +- set encryption type: WPA2, a STA associated and expired, + let L1 -> L0 procedurce invoke by Group key rekey timer. + (Group Key Rekey: 40 sec, STA expire: 17 s) + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.137 + +* DESCRIPTION +- confirm FW accept H2C commad, then turn on rx DMA. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/13, jimmylin, bug fix + +* DIGEST: +- Modify for MP h/w ctx + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_mp.c - revision: 1.29 + +* DESCRIPTION +- According to Annie/SD9, 0x820 and 0x828 can't be set while h/w ctx. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/02, jimmylin, bug fix + +* DIGEST: +- Unlock MAC register while rebooting + +* TEST METHOD: +- Load 1.2b image to platform and then reload v1.2 image. The wlan interface + can not work. + +* ROOT CAUSE: +- After wlan shutdown, the MAC register will be locked. So we need to unlock + MAC register. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.80 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/02, victorlo, bug fix + +* DIGEST: +- RX BAR IOT issue for Atheros NIC, where atheros BAR has bug + +* TEST METHOD: +- Chariot uplink test with high performance script, sometimes would cause + NIC to send BAR + +* ROOT CAUSE: +- Atheros NIC tx BAR has bug + +* MODIFIED FILES: +- data/MACPHY_REG.txt - revision: 1.12 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/07/01, jimmylin, bug fix + +* DIGEST: +- Fix not restricting HT capability in WPA/WPA2 mixed mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.110 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/29, jimmylin, bug fix + +* DIGEST: +- Fix client mode connection unstable issue + +* TEST METHOD: +- Set Netgear WGM124 as WPA-PSK/TKIP and use client mode to connect. + Sometimes the connection will fail. + +* ROOT CAUSE: +- Sometimes WGM124 will send many retries of association responses with + different sequence numbers to client after 4-way handshake. The state + machine of our client mode is not working well and will re-process these + association responses and will clear the unicast key which is obtained + from the previous 4-way handshake. Then the traffic between WGM124 and + client mode can not pass through. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.107 + +* DESCRIPTION +- After we receive the first association response, our state will become + association success. After association success, we will not re-process the + subsequent retry association response. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/28, jimmylin, new feature + +* DIGEST: +- Add MP CTX of single carrier +- Modify for external high power PA support + +* TEST METHOD: +- Use "iwpriv wlan0 mp_ctx background,scr" for single carrier + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.70 +- 8192cd_mp.c - revision: 1.28 +- 8192cd_phyreg.h - revision: 1.4 +- data/PHY_REG_2T_n_hp.txt - revision: 1.2 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/24, button, bug fix + +* DIGEST: +- NDSi can't link under WEP encryption. + +* TEST METHOD: +- + +* ROOT CAUSE: +- NDSi cannot receive AP's Ack with 1M date rate, and send many retry packets, +- then cannot finish all connection test in time. + +* MODIFIED FILES: +- data/MACPHY_REG.txt - revision: 1.11 + +* DESCRIPTION +- add 2Mbps into response rate set + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/23, button, feature fine-tune + +* DIGEST: +- e-phy reset once + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.78 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/23, button, feature fine-tune + +* DIGEST: +- improve tx EVM + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.132 + +* DESCRIPTION +- clk inverted, by Arthur + + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/23, jimmylin, new feature + +* DIGEST: +- Decrease the number of proc files + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.68 +- 8192cd_proc.c - revision: 1.41 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/23, button, new feature + +* DIGEST: +- tx report (ok/fail) + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- data/rtl8192cfwn.bin - revision: 1.7 +- 8192cd.h - revision: 1.67 +- 8192cd_sme.c - revision: 1.105 +- 8192cd_osdep.c - revision: 1.76 +- 8192cd_util.h - revision: 1.11 +- 8192cd_util.c - revision: 1.59 +- 8192cd_hw.c - revision: 1.131 +- 8192cd_headers.h - revision: 1.40 +- 8192cd_cfg.h - revision: 1.72 + +* DESCRIPTION +- request tx report every 5 sec, count in "tx_pkt", "tx_fail" field of sta_info + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/23, jimmylin, bug fix + +* DIGEST: +- Correct beacon statistics when MBSSID is enabled + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.66 +- 8192cd_osdep.c - revision: 1.75 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/21, jimmylin, new feature + +* DIGEST: +- Add external high power PA support + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- data/PHY_REG_2T_n_hp.txt - revision: 1.1 +- data/radio_a_2T_n_hp.txt - revision: 1.1 +- data/radio_b_2T_n_hp.txt - revision: 1.1 +- 8192cd_headers.h - revision: 1.39 +- 8192cd_hw.c - revision: 1.129 +- 8192cd_mp.c - revision: 1.27 +- 8192cd_proc.c - revision: 1.40 + +* DESCRIPTION +- Should turn on compiler flag of HIGH_POWER_EXT_PA and set mib of use_ext_pa + to 1 to support. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/18, victorlo, bug fix + +* DIGEST: +- Fix WPA/WPA2 group key rekey fail issue + +* TEST METHOD: +- Enable group key rekey with rekey period set as no less than 10 minutes. + After station has connected to AP, and let both AP & station do nothing. + After group key rekey done, AP with different value of key content in cam + and mib. New connecting stations will have different cam content with AP, + and all wireless stations cannot ping each other after then. + +* ROOT CAUSE: +- When AP switch to PCIE power save mode after all stations have expire + count as 0. At that mode, AP will drop all tx packets and also cannot set + any wlan nic registers. With conditions above, AP cannot update its own + cam content and send out group key rekey packets. But AP will update + its group key mib content. That makes AP update its mib content to all + stations while new or re-connections but its own cam info. Multicast + packets will finally cannot pass through the whole BSS correctly. + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 1.45 + +* DESCRIPTION +- Fix WPA/WPA2 group key rekey fail issue. When there is any legal + tx packets from AP to BSS, AP will relieve itself from PCIE power + save mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/18, jimmylin, bug fix + +* DIGEST: +- Fix wrong sequence of beacon with broadcast bit set in DTIM field and + queued broadcast packets + +* TEST METHOD: +- Let power save client connect to AP and send broadcast from AP to see if + the DTIM BC bit is set and the sequence of beacon and queued broadcast + packets + +* ROOT CAUSE: +- We used data queue to send queued broadcast packets, but the hardware + doesn't guarantee the sequence of beacon and data queue. In haredware + design, we should use high queue to send queued broadcast packets. Then + hardware MAC will send broadcast packets after beacon. + +* MODIFIED FILES: +- 8192cd.h - revision: 1.64 +- 8192cd_cfg.h - revision: 1.70 +- 8192cd_headers.h - revision: 1.38 +- 8192cd_hw.c - revision: 1.127 +- 8192cd_osdep.c - revision: 1.74 +- 8192cd_sme.c - revision: 1.102 +- 8192cd_tx.c - revision: 1.44 +- 8192cd_util.c - revision: 1.58 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/11, button, new feature + +* DIGEST: +- tx hangup reset + +* TEST METHOD: +- set TxBufferAddr in txdesc (Dword8) less than 1M +- let tx hangup occur, and see if the procedure could recover. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.126 +- 8192cd_osdec.c - revision: 1.73 +- 8192cd_cfg.h - revision: 1.68 +- 8192cd_sme.c - revision: 1.101 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/6/10, ericyeh, bug fixed + +* DIGEST: +- Fix some connection fail under WEP shared key (ex. WNDA3100-v1) + +* TEST METHOD: +- Set AP(8196C) to WEP shared key, + use WNDA3100-v1 to do authentication and association. + +* ROOT CAUSE: +- Test the WEP Shared key with Atheros 11n clients + (WNDA3100-v1, Driver v3.0.0.2), + Failed due to Challenge Text rejection. +- During Authentication, + WNDA3100-v1 will send several "Retry" shared key authentication frames, + and AP will do reaction with ALL these "Retry" frmaes(if received), + like assigning different "Challenge Texts" for these "Retry" frames. + Such behavior will cause problems when making comparison between + assigned and received challenge texts, + and also may cause WNDA3100-v1 NOT issue Assocation Request. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.99 + +* DESCRIPTION +- In mgt_handler(), + update pstat->tpcache_mgt when received unhandled Retry frames. +- In OnAuth() and OnAssocRsp(), + add "pstat->tpcache_mgt = GetTupleCache(pframe);" + after calling alloc_stainfo() & init_stainfo(), + because these two functions will reset the value of tpcache_mgt, + such behavior will cause error + when try to discard repeated frames in mgt_handler(). + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/10, keith, bug fix + +* DIGEST: +- Add restriction of WPA/WPA2 TKIP from N mode + +* TEST METHOD: +- 1.Enable wireless security with 802.1x and tkip only. +- 2.Client connect to AP. +- 3.The Probe rsp, Assoc req and Assoc rsp have n mode rate. + +* ROOT CAUSE: +- We don't forbid n mode in this condition. + +* MODIFIED FILES: +- rtl865x/AP/goahead-2.1.1/LINUX/flash.c, revision: 1.124 +- rtl865x/linux-2.4.18/drivers/net/rtl8192cd/8192cd_util.c, revision: 1.57 + +* DESCRIPTION [description about the change]: +- Enable 802.1x and tkip, client will connect with N mode. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/07, button, bug fix + +* DIGEST: +- tx stuck issue + +* TEST METHOD: +- wlan down / up + +* ROOT CAUSE: +- unknown + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.124 +- 8192cd_osdec.c - revision: 1.72 + +* DESCRIPTION +- sync card disable flow with NIC team, +- disable RF before reset BB, inform MCU before reset it. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/06/04, button, feature fine-tune + +* DIGEST: +- beacon loss issue reported from huawei + +* TEST METHOD: +- configure DUT to BG mode, fix rate: 2M bps, +- run chariot thoughput script ( LAN to WLAN, 3 pairs) +- click apply chage via webpage + +* ROOT CAUSE: +- ephy hangup + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.123 +- 8192cd_sme.c - revision: 1.95 + +* DESCRIPTION +- "tx_stuck_fix" doesn't work in huawei's configuration, +- finetune it to fit more condition. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/31, jimmylin, bug fix + +* DIGEST: +- Remove 4/26 changing AMPDU facility dynamically by Button + +* TEST METHOD: +- + +* ROOT CAUSE: +- This mechanism will cause Win7 WPS test fail and iPhone 3GS connection + issue + +* MODIFIED FILES: +- 8192cd.h - revision: 1.63 +- 8192cd_sme.c - revision: 1.94 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/24, jimmylin, bug fix + +* DIGEST: +- DTIM broadcast bit didn't be set if broadcast packets are queued + +* TEST METHOD: +- Let power save client connect to AP and send broadcast from AP to see if + the DTIM BC bit is set or not + +* ROOT CAUSE: +- The BC packets are not queued for VAP interfaces. The timing to send out + queued BC packets are not correct, either. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.121 +- 8192cd_osdep.c - revision: 1.71 +- 8192cd_tx.c - revision: 1.43 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/21, jimmylin, bug fix + +* DIGEST: +- Fix bug of handling QoS Null Data in WMM-PS + +* TEST METHOD: +- Use BlackBerry associated with WMM-PS. Ping to BlackBerry will get lost. + +* ROOT CAUSE: +- BlackBerry uses QoS Null Data as check frame to query queued packets in + WMM-PS mode. AP didn't send any packet out after receiving QoS Null Data. + +* MODIFIED FILES: +- 8192cd_rx.c - revision: 1.30 + +* DESCRIPTION +- After AP received QoS Null Data, we try to get pstat by AID, but there is + no AID field in QoS Null Data. After modifying to get pstat by sa, we can + get the correct pstat and we can send the queued packets out. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/20, button, bug fix + +* DIGEST: +- DTIM count error + +* TEST METHOD: +- look up TIM IE in beacon frame + +* ROOT CAUSE: +- download beacon twice + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.70 +- 8192cd_sme.c - revision: 1.93 +- 8192cd_hw.c - revision: 1.120 + +* DESCRIPTION +- download beacon twice before entering L1/L2, download once in general + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/19, jimmylin, feature fine-tune + +* DIGEST: +- Add mib to control t/r switch + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.119 +- 8192cd_ioctl.c - revision: 1.42 +- 8192cd_proc.c - revision: 1.38 +- ieee802_mib.h - revision: 1.17 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/18, jimmylin, feature fine-tune + +* DIGEST: +- Add TID-TxQ remapping + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 1.64 +- 8192cd_ioctl.c - revision: 1.40 +- 8192cd_osdep.c - revision: 1.68 +- 8192cd_proc.c - revision: 1.37 +- 8192cd_tx.c - revision: 1.41 +- ieee802_mib.h - revision: 1.16 + +* DESCRIPTION +- Add mib to customize TID and TxQ (Traffic class) mapping + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/18, button, bug fix + +* DIGEST: +- vap beacon content error + +* TEST METHOD: +- sniffer the beacon sent by vap + +* ROOT CAUSE: +- if PCIE_POWER_SAVING is defined, we need to downlaod beacon twice, +- download beacon and probe rsp at the first time, and then download real beacon. +- but VAP miss the second time. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.67 + +* DESCRIPTION +- VAP or VXD download beacon once (not include probe rsp) + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/07, button, feature fine-tune + +* DIGEST: +- add efuse support in menuconfig + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- Config.in - revision: 1.9 +- 8192cd_osdep.c - revision: 1.64 +- 8192cd_cfg.h - revision: 1.62 +- 8192cd_hw.c - revision: 1.112 +- AP/goahead-2.1.1/LINUX/Makefile - revision: 1.54 +- AP/goahead-2.1.1/LINUX/flash.c - revision: 1.120 + +* DESCRIPTION +- menuconfig: Network device support -> Wireless LAN -> Efuse Support +- if Efuse Support is selected, read MAC Addr and Tx Power in efuse instead of flash. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/07, brian, new feature + +* DIGEST: +- Enable user specify multicast rate used + +* TEST METHOD: +- Sniffer packets from LAN to WLAN in multicast traffic + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_tx.c - revision: 1.40 + +* DESCRIPTION +- when lowestMlcstRate = 0, means multicast rate use the highest rate in + in basic rate set ; others, it is the bitmap as used in fixedRate. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/07, jimmylin, feature fine-tune + +* DIGEST: +- Add GPIO8 LED2 lighting pattern 12 + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.61 +- 8192cd_sme.c - revision: 1.92 + +* DESCRIPTION +- Add GPIO8 LED2 lighting patter of one LED of enable_tx_rx_data. The + behavior is link LED0 enable_tx_rx_data (mode 7 pattern). LED_TYPE retains + as 12. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/07, jimmylin, bug fix + +* DIGEST: +- Fix WAPI crash issue from ZhaoBo + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.60 +- 8192cd_util.c - revision: 1.55 +- wapi_wai.c - revision: 1.4 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/06, brian, new feature + +* DIGEST: +- enable config EDCA parameters manually + +* TEST METHOD: +- enable then interface down-up +- verify by beacon and chariot throughput testing + +* ROOT CAUSE: +- +* MODIFIED FILES: +- 8192cd.h - revision: 1.59 +- 8192c_cfg.h - revision: 1.61 +- 8192cd_headers.h - revision: 1.34 +- 8192cd_hw.c - revision: 1.111 +- ieee802_mib.h - revision: 1.15 +- 8192cd_proc.c - revision: 1.36 +- 8192cd_ioctl.c - revision: 1.38 +- 8192cd_sme.c - revision: 1.91 +- 8192cd_util.c - revision: 1.54 +- ieee802_mib.h - revision: 1.15 + + +* DESCRIPTION +- default not enabled +- only take effect after interface down-up, reboot will be default + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/05/03, button, new feature + +* DIGEST: +- read tx power from efuse + +* TEST METHOD: +- parsing the efuse content of a known value 92ce + +* ROOT CAUSE: +- +* MODIFIED FILES: +- 8192cd.h - revision: 1.58 +- 8192c_reg.h - revision: 1.8 +- 8192cd_headers.h - revision: 1.33 +- 8192cd_hw.c - revision: 1.109 +- 8192cd_hw.h - revision: 1.12 +- ieee802_mib.h - revision: 1.13 +- 8192cd_mp.c - revision: 1.26 +- 8192cd_proc.c - revision: 1.35 +- 8192cd_ioctl.c - revision: 1.37 +- 8192cd_osdep.c - revision: 1.63 + + +* DESCRIPTION +- if autoload from efuse success, and the mib "use_efuse" is set, +- PG value is valid, then use the tx power index in efuse. + +----------------------------------------------------------------------------- + + +* DATE/WHOM/TYPE: +- 2010/04/26, victorlo, bug fix + +* DIGEST: +- Dynamic NAV protect length switch for IOT + +* TEST METHOD: +- Intel 5100 connect to AP with using AC power adapter, TP is ok. But chariot + TP test will drop a lot when AC power adapter of Intel 5100 is cut off + +* ROOT CAUSE: +- AP NAV protect length is too short for Intel 5100 switching to + MIMO power-save dynamic mode + +* MODIFIED FILES: +- 8192cd.h - revision: 1.56 +- 8192cd_headers.h - revision: 1.32 +- 8192cd_hw.c - revision: 1.107 +- 8192cd_sme.c - revision: 1.90 +- 8192cd_util.c - revision: 1.52 +- data/MACPHY_REG.txt - revision: 1.10 + +* DESCRIPTION +- Dynamic NAV protect length switch for Intel 5100 + 11n MIMO power-save IOT issue + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/26, Button, feature fine-tune + +* DIGEST: +- chaneg AMPDU facility dynamically + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.55 +- 8192cd_sme.c - revision: 1.89 + +* DESCRIPTION +- default ampdu facility is 32k, when receives assoc req from broadcom sta, +- change to 16k, and send dissasoc. 10 sec later, change back to 32k. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/23, victorlo, new feature + +* DIGEST: +- Add Japan 2.4G band (safety) spec support + +* TEST METHOD: +- Set different power level to make sure AP will stop tx (packet + and beacon) when there is a continuous signal over the specified + power level. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 1.54 +- 8192cd_headers.h - revision: 1.31 +- 8192cd_hw.c - revision: 1.106 +- 8192cd_hw.h - revision: 1.11 +- 8192cd_ioctl.c - revision: 1.36 +- 8192cd_proc.c - revision: 1.34 +- 8192cd_sme.c - revision: 1.88 +- 8192cd_util.c - revision: 1.51 + +* DESCRIPTION +- Add Japan 2.4G band (safety) spec support +- Power threshold can be set by mib "edcca_thd", 0 is default as + disabled, suggested value is 50. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/22, Button, bug fix + +* DIGEST: +- 20Mhz Reperater mode performance issue + +* TEST METHOD: +- 96b: pure AP +- 96c: repeater mode, + root: AP 20MHz BW, + vxd: client, connect to 96b AP +- chariot can't run under the condition: 96b tx 96c rx + +* ROOT CAUSE: +- Tx rate of 96b is 40M MCS rate. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.87 + +* DESCRIPTION +- vxd set BW support in probe req according to target AP and root AP's capability. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/22, victorlo, feature fine-tune + +* DIGEST: +- Update APK to version R07 + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.104 + +* DESCRIPTION +- Update APK to version R07, to bound calibrated output range + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/21, Button, feature bug fix + +* DIGEST: +- fix beacon stop issue in ad-hoc mode + +* TEST METHOD: +- 96b 96c connect via ad-hoc mode + +* ROOT CAUSE: +- hw may sync wrong tsf in the init flow or join IBSS procedure. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.84 1.85 +- 8192cd_hw.c - revision: 1.100 + +* DESCRIPTION +- check tsf every sec, correct it when sync wrong value. +- disable beacon function in MacInit + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/21, pluswang, merge code + +* DIGEST: +- merge MP new feature from 2.6 + design:caryLee + +* TEST METHOD: +* ROOT CAUSE: + +* MODIFIED FILES: +-1.51 8192cd.h +-1.48 8192cd_util.c +-1.30 8192cd_headers.h +-1.27 8192cd_rx.c +-1.22 8192cd_mp.c + +* DESCRIPTION + +manual: +iwpriv wlan0 mp_arx filter_init //filter init +iwpriv wlan0 mp_arx filter_SA=FF:FF:FF:FF:FF:FF //setting SA filter +iwpriv wlan0 mp_arx filter_DA= FF:FF:FF:FF:FF:FF //setting DA filter +iwpriv wlan0 mp_arx filter_BSSID= FF:FF:FF:FF:FF:FF //setting BSSID filter +iwpriv wlan0 mp_arx filter_clean // clean all filter +iwpriv wlan0 mp_arx filter_print // display all filter + +example: +1 init Filter: + iwpriv wlan0 mp_arx filter_init +2 add filter: + //eg setting DA and BSSID + iwpriv wlan0 mp_arx filter_DA=00:26:5A:08:65:ed + iwpriv wlan0 mp_arx filter_BSSID=da:85:05:95:53:b6 + + 3 start arx test: + iwpriv wlan0 mp_arx start + +4 stop arx test: + //will 1)stop arx test and 2)display result and 3)clean filter + iwpriv wlan0 mp_arx stop + +only those packets that meet DA and BSSID filter +will be received and census. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/21, jimmylin, bug fix + +* DIGEST: +- Add key id check for key-mapping key + +* TEST METHOD: +- Config 802.1x+WEP, the authentication is done but the traffic is not + passed. + +* ROOT CAUSE: +- In new CAM design, bit(8) of register SECCFG(0x680) should be set if the + key id table is filled. + +* MODIFIED FILES: +- 8192c_reg.h - revision: 1.7 +- 8192cd_hw.c - revision: 1.98 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/20, victorlo, new feature + +* DIGEST: +- Concurrent support up to 64 clients + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 1.52 +- 8192cd_cfg.h - revision: 1.57 +- 8192cd_proc.c - revision: 1.33 +- 8192cd_sme.c - revision: 1.83 +- 8192cd_tx.c - revision: 1.39 +- 8192cd_util.c - revision: 1.49 + +* DESCRIPTION +- Concurrent support up to 64 clients + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/20, Button, feature fine-tune + +* DIGEST: +- reduce receive AMPDU size form 32k to 16k + +* TEST METHOD: +- run chariot with broadcom WPC600N, BW 20 MHz, uplink + +* ROOT CAUSE: +- rx packet buffer size of 92ce is 10k bytes, if we declare rx AMPDU size=32k in HTCAP IE +- rx FIFO overflow (HIMR_RXFOVW) will occur frequently, and degrades the throughpt. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.81 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/09, Button, feature fine-tune + +* DIGEST: +- add a bit ctrl to forbid N mode + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.60 +- 8192cd_util.c - revision: 1.47 + +* DESCRIPTION +- if bit 3 of the MIB "lgyEncRstrct" is set, and enc type is WEP or TKIP, +- AP works under BG mode + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/09, Button, feature fine-tune + +* DIGEST: +- extend the waiting time of entering L1/L2 state + +* TEST METHOD: +- Microssoft DTM + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_cfg.h - revision: 1.54 +- 8192cd_ioctl.c - revision: 1.34 +- 8192cd_rx.c - revision: 1.25 +- 8192cd_security.c - revision: 1.10 + +* DESCRIPTION +- extend the waiting time from 3s to 10s in order to pass WCN wireless test + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/09, victorlo, new feature + +* DIGEST: +- Enable APK + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.97 + +* DESCRIPTION +- Enable APK(PA Analog Pre-distortion calibration) to enhance tx EVM + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/08, Button, feature fine-tune + +* DIGEST: +- Refine RF CCK shape filter setting + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.50 +- 8192cd_mp.c - revision: 1.21 +- 8192cd_hw.c - revision: 1.95 + +* DESCRIPTION +- adjust swing index according to power tracking effect + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/06, jimmylin, bug fix + +* DIGEST: +- When restrict Realtek client in N mode for WEP/TKIP, not only restrict + 8192S client + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.49 +- 8192cd_osdep.c - revision: 1.59 +- 8192cd_sme.c - revision: 1.81 +- 8192cd_util.c - revision: 1.46 + +* DESCRIPTION +- + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/07, button, bug fix + +* DIGEST: +- beacon frame includes beacon and probe rsp. + +* TEST METHOD: +- sniffer the beacon packet sent by 88c/92c test chip + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.58 +- 8192cd_sme.c - revision: 1.80 + +* DESCRIPTION +- download beacon once if power saving is not enable + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/06, jimmylin, feature fine-tune + +* DIGEST: +- Fine tune DIG according to Luke Lee + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.48 +- 8192cd_hw.c - revision: 1.94 +- 8192cd_ioctl.c - revision: 1.33 +- 8192cd_sme.c - revision: 1.79 + +* DESCRIPTION +- Luke Lee tested new DIG flow on ADSL platform and asked all AP driver to + sync the new DIG flow. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/04/01, button, but fix + +* DIGEST: +- fix L1 unstable problem + +* TEST METHOD: +- let an AP enter L1 state, then use a client associate to it. +- or start site survey via webpage when AP in L1 state, check the BSS list. + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.92 + + +* DESCRIPTION +- if disable printk function, the timing of entering L1 flow changes, +- it will cause PCIe link unstable, and change back to L0 fail. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/31, jimmylin, feature fine-tune + +* DIGEST: +- Restrict AMPDU packet number in ADDBA Req/Rsp for Broadcom in 8M package + +* TEST METHOD: +- Test uplink throughput with WPC600N and get bad performance + +* ROOT CAUSE: +- We doubt Broadcom send too large AMPDU to us and we can not handle it under + 8M package so there are some packets lost. Therefore the uplink throughput + is bad. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.78 + +* DESCRIPTION +- We restrict AMPDU packet number in ADDBA Req/Rsp for Broadcom in 8M package + to 32 to let Broadcom send less AMPDU to us. Then we can handle it smoothly + and get better performance. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/31, button, feature fine-tune + +* DIGEST: +- disable pcie clk in driver + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.91 +- 8192cd_osdep.c Rev. 1.55 1.56 + +* DESCRIPTION +- turn on / off pcie clk in rtl8192cd_open, rtl8192cd_close +- needn't patch set_init.c in all package. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/30, victorlo, feature fine-tune + +* DIGEST: +- Refine RF CCK shape filter setting + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- 92C formal chip CCK BB HW doesnot support 0.5dB/gain step, and + 1dB/gain setp is the minumum calibration scale supporting. CCK + power cal will have issue. + +* MODIFIED FILES: +- 8192cd_mp.c Rev. 1.20 + +* DESCRIPTION +- Refine RF CCK shape filter setting to achieve the 0.5dB gain step, + and has no effect to OFDM/MCS power. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/26, button, bug fix + +* DIGEST: +- hang up or core dump when disable or enable WLAN interface + +* TEST METHOD: +- disable or enable WLAN interface via webpage + +* ROOT CAUSE: +- if clk ever disable, we needs do pcie reset again + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.88 +- 8192cd_osdep.c Rev. 1.51 1.53 +- AP/goahead-2.1.1/LINUX/system/set_init.c Rev. 1.39 + +* DESCRIPTION +- when "PCIE_POWER_SAVING" is defined, enable pcie clk rtl8192cd_open of +- root interface, disable in "radio_off" func. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/26, victorlo, feature fine-tune + +* DIGEST: +- Refine GPIO8 LED2 lighting pattern + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h Rev. 1.46 +- 8192cd_osdep.c Rev. 1.52 +- 8192cd_rx.c Rev. 1.24 +- 8192cd_sme.c Rev. 1.76 +- 8192cd_tx.c Rev. 1.37 + +* DESCRIPTION +- Refine GPIO8 LED2 lighting pattern, changed from one LED enable_tx_rx mode + (mode 7 pattern) to one LED link_rx_tx mode (mode 3 pattern). LED_TYPE + retains as 11. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/25, button, new feature + +* DIGEST: +- power tracking for MP & normal driver + +* TEST METHOD: +- check siwng index(c80,c88) vary with thermel meter correclty + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.87 +- 8192cd_sme.c Rev. 1.74 +- 8192cd.h Rev. 1.45 +- 8192cd_headers.h Rev. 1.27 +- 8192cd_phyreg.h Rev. 1.2 +- 8192cd_mp.c Rev. 1.18 + +* DESCRIPTION +- porting form NIC driver + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/25, button, feature fine-tune + +* DIGEST: +- disable RTS for Intel client in long range + +* TEST METHOD: +- emulator setting: attenuation 24dB + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h Rev. 1.45 +- 8192cd_sme.c Rev. 1.74 +- 8192cd_tx.c Rev. 1.36 + +* DESCRIPTION +- RTS data date down to 1 Mbps, occupy channel resource, +- disable RTS when RSSI <30, to get better throughput. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/25, button, bug fix + +* DIGEST: +- system hang when leaving L2 + +* TEST METHOD: +- + +* ROOT CAUSE: +- while(1) loop + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.87 + +* DESCRIPTION +- auto-load fail card will set power switch to PFM mode when doing pcie reset, +- but auto-load success card set power mode according to efuse 6[5] +- assume power mode will change to PFM mdoe when doing pcie reset can't fit all situation. + +----------------------------------------------------------------------------- +* DATE/WHOM/TYPE: +- 2010/03/25, victorlo, feature fine-tune + +* DIGEST: +- Refine VI EDCA setting to comply WiFi approved-n logo test + +* TEST METHOD: +- WiFi test plan 4.2.20 T07 + +* ROOT CAUSE: +- AP's VI tx stream is too aggressive to VO stream from other devices + +* MODIFIED FILES: +- 8192cd.h Rev. 1.44 +- 8192cd_hw.c Rev. 1.86 +- 8192cd_sme.c Rev. 1.73 +- 8192cd_util.c Rev. 1.45 + +* DESCRIPTION +- AP always detects if there is any VO stream in the same BSS, and degrades + VI EDCA parameter setting when any is found. By temporary cancelling TXOP + in VI EDCA setting, AP will comply WiFi logo test. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/24, victorlo, bug fixed + +* DIGEST: +- Refine RRSR bitmap at GN mode +- Set RF reg depends on band in use for better tx EVM + +* TEST METHOD: +- WPC600N connects AP at N mode, chariot test with 3D3U high-TP script. Open + shielding room door when chariot test becomes stable, tx rate of WPC600N + will degrade to CCK. Tx rate of WPC600N will hardly resume after the door + has closed. + +* ROOT CAUSE: +- WPC600N (Broadcom) cannot easily receive AP's 802.11 Ack with rate 11M and + 5.5M. After its tx rate degraded to CCK, it can hardly recover its tx rate + to MCS. + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.84 + +* DESCRIPTION +- Excludes CCK 11M and 5.5M at GN mode to prevent Broadcom IOT +- Set RF reg depends on band in use for better tx EVM + reg 0x26 is 0x4f000 for G and N mode + reg 0x26 is 0xf400 for B mode + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/23, Button, feature fine-tune, bug fixed + +* DIGEST: +- refine enter, leaving L2 flow +- add a iwpriv cmd to stop power saving + +* TEST METHOD: +- let a client connect and disconnect to an 96c AP having power saving functions, +- after enter and leving L2 seveal times, AP won't send any packet except beacon. + +* ROOT CAUSE: +- PERST will let SPS enter PFM mode +- + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.83 +- 8192cd_ioctl.c Rev. 1.31 + +* DESCRIPTION +- PCIe power saving ON/OFF + on: setmig ps 119 + off: iwpriv wlan0 stopps + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/19, jimmylin, bug fixed + +* DIGEST: +- Fix management frame buffer leakage after several times of entering and + leaving L2 + +* TEST METHOD: +- Set both AP and AP-client to WPA2-PSK but different keys. Let AP-client try + to connect to AP repeatedly. Eventually, AP can't send any management frame. + +* ROOT CAUSE: +- After entering L2, Tx and Rx descriptor registers will be clear. After + leaving L2, Tx and Rx descriptor registers will be reset, and Tx and Rx + head and tail pointers will be clear to 0. But the buffers used in Tx and + Rx descriptors are not freed correctly. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.82 + +* DESCRIPTION +- Before entering L2, we should call rtl8192cd_tx_dsr() and + rtl8192cd_rx_isr() to free used buffers first after stoping DMA. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/17, Button, bug fixed + +* DIGEST: +- fix bug of repeater mode. +- repeater mode, client mode, ad-hoc mode test ok. + +* TEST METHOD: +- + +* ROOT CAUSE: +- NETYPE, BCN_CTRL setting error. + +* MODIFIED FILES: +- 8192cd_hw.c Rev. 1.80 +- 8192cd_sme.c Rev. 1.71 +- 8192c_reg.h Rev. 1.6 +- 8192cd_osdep.c Rev. 1.47 +- 8192cd_util.c Rev. 1.43 + +* DESCRIPTION +- repeater mode function doesn't work + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/17, jimmylin, feature fine-tune + +* DIGEST: +- Add 1R CCA feature + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.43 +- 8192cd_hw.c - revision: 1.79 +- 8192c_sme.c - revision: 1.70 +- 8192c_ioctl.c - revision: 1.29 +- 8192c_proc.c - revision: 1.27 + +* DESCRIPTION +- In some customers' boards, there is noise around one of the receiving + path. Using 1R CCA to reduce noise influence. There is a new mib of "1rcca" + added. When 1rcca is set to 1, only path A will be used to CCA. When 1rcca + is set to 2, only path B will be used to CCA. The default value is 0 and + means using both 2 pathes to CCA. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/3/12, Button, new feature + +* DIGEST: +- 8192ce formal chip support + +* TEST METHOD: +- + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.76 +- 8192cd_osdep.c - revision: 1.46 +- 8192cd.h - revision: 1.42 +- 8192cd_proc.c - revision: 1.26 +- 8192cd_headers.h - revision: 1.24 +- data/radio_b_2T_n.txt - revision: 1.1 +- data/radio_b_2T.txt - revision: 1.1 +- data/radio_a_2T_n.txt - revision: 1.1 +- data/radio_a_2T.txt - revision: 1.1 +- data/radio_a_1T_n.txt - revision: 1.1 +- data/radio_a_1T.txt - revision: 1.1 +- data/PHY_REG_2T_n.txt - revision: 1.1 +- data/PHY_REG_2T.txt - revision: 1.1 +- data/PHY_REG_1T_n.txt - revision: 1.1 +- data/PHY_REG_1T.txt - revision: 1.1 +- data/AGC_TAB_n.txt - revision: 1.1 + +* DESCRIPTION +- upgrade PHY parameter from Mantis, + 8192C_PCIE_PHYforMP_100113.zip + 8188C_PCIE_PHYforMP_100113.zip + 8192C_Formal_PCIE_PHYforMP_100311.zip + 8188C_Formal_PCIE_PHYforMP_100311_1.zip + modify radio_a_2T_n.txt radio_a_1T_n.txt 0x26 for better Tx EVM + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/10, victorlo, new feature + +* DIGEST: +- Add LED2 GPIO8 lighting mode + +* TEST METHOD: +- LED2 background on, when throughput keeps increasing, + frequency of LED on and off becomes faster + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 1.41 +- 8192cd_sme.c - revision: 1.67 + +* DESCRIPTION +- Add LED2 lighting by using GPIO8 as output pin + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/09, chris, add friendly API + +* DIGEST: +- Add text-based config file support +- Support iwconfig/iwlist + +* TEST METHOD: +- config file in /etc/Wireless/RTL8192CD.dat with format: + "interface"_"mib_command" (autotest supported) + e.g. wlan0_ssid=XXXX +- use iwconfig and iwlist + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_comapi.c - revision: 1.2 +- 8192cd_comapi.h - revision: 1.2 +- 8192cd_headers.h - revision: 1.21 +- 8192cd_ioctl.c - revision: 1.27 +- 8192cd_osdep.c - revision: 1.42 +- 8192cd_proc.c - revision: 1.24 +- 8192cd_security.h - revision: 1.2 +- Config.in - revision: 1.8 +- Makefile - revision: 1.7 + +* DESCRIPTION +- 8192cd_comapi.c - file parsing and call set_mib() +- 8192cd_comapi.h - file parsing +- 8192cd_ioctl.c - support iwconfig/iwlist +- 8192cd_osdep.c - load config file when open +- 8192cd_proc.c - for autotest +- 8192cd_security.h - wps related +- Config.in - new menuconfig item + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/09, victorlo, feature fine-tune + +* DIGEST: +- Refine IQK retry setting + +* TEST METHOD: +- n/a + +* ROOT CAUSE: +- Current IQK retry limit and relative settings would cause unreasonable + number of times of retry + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.73 + +* DESCRIPTION +- Modify IQK retry limit and way of retry counting + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/07, jimmylin, bug fixed + +* DIGEST: +- Fix client mode MC/BC decryption error issue + +* TEST METHOD: +- Set client mode and DHCP client. AP-client can't get IP. + +* ROOT CAUSE: +- Formal mask IC has changed the CAM search flow. But our CAM setting didn't + follow up. + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.71 +- 8192c_reg.h - revision: 1.5 + +* DESCRIPTION +- Modify CAM setting according to the latest spec. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/04, jimmylin, bug fixed + +* DIGEST: +- Fix incorrect HT IE of client mode if AP changed bandwidth + +* TEST METHOD: +- Set AP in 20M mode, then let AP-client connect to the AP. Then change + AP to 40M mode, and let AP-client reconnect to the AP automatically. + +* ROOT CAUSE: +- When applying change of bandwidth of the AP, AP will send disassociation + request to AP-client. AP-client will not re-scan and reconnect to the AP + directly. So AP-client doesn't know the AP has already changed the + bandwidth. + +* MODIFIED FILES: +- 8192cd_sme.c - revision: 1.65 + +* DESCRIPTION +- When AP-client received disassociation request, do re-scan channels. Then + use the new capability to connect to the AP. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/04, jimmylin, new feature + +* DIGEST: +- Add LC Calibration + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.64 + +* DESCRIPTION +- In the init sequence, we will do LCK after IQK. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/3/3, victorlo, bug fixed + +* DIGEST: +- Tx rate might degrade when time goes by. + +* TEST METHOD: +- Use test tool like VLC to do downlink continuously, will casue + AP tx initial rate become lower. + +* ROOT CAUSE: +- When it is now downlink continuously, and client may sometimes + change its mode to power-save mode. Then packets queued in AP + FIFO will retry until tx success, and will cause AP eventually + degrade initial tx rate to that client mistakenly. + +* MODIFIED FILES: +- 8192cd.h - revision: 1.39 +- 8192cd_headers.h - revision: 1.20 +- 8192cd_osdep.c - revision: 1.41 +- 8192cd_sme.c - revision: 1.64 +- 8192cd_util.c - revision: 1.39 +- rtl8192cfwn.bin - revision: 43 + +* DESCRIPTION +- By this bug fixed, AP will not do statistics of tx success to a + client when one changes its mode to power-save. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/3/3, jason, feature fine-tune + +* DIGEST: +- PCIe PHY updated by designer + +* TEST METHOD: +- LAN/WLAN Chariot test okay and throughput is the same as the previous version. + +* ROOT CAUSE: +- Run PCIE PHY test with rtl8196c with RD center. + +* MODIFIED FILES: +- 8192cd_osdep.c - revision: 1.40 + +* DESCRIPTION +- change PCIe PHY parameter for 88c normal chip + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/3/2, Button, new feature + +* DIGEST: +- PCIe L2 support + +* TEST METHOD: +- if there's no sta links to AP, PCIe link status(b8b00728) should be 0, +- use a wireless client connect to AP, link status b[0-4] should change to 0x11, +- AP must respond Auth in 100ms, use sniffer to verify it. + +* ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd_hw.c - revision: 1.63 +- 8192cd_ioctl.c - revision: 1.25 +- 8192cd_rx.c - revision: 1.20 + +* DESCRIPTION +- change PCIe link state to L2 under idle mode (no STA linls) for 88c normal chip + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/03/2, davidhsu, new feature + +* DIGEST: +- Add lazy WDS feature + +* TEST METHOD: +- Set a mac address in a wds entry of AP1 manually +- Enable lazy WDS (w/o need to set mac address of wds entry) in AP2 +- Then, two APs could commnucate in WDS channel + + * ROOT CAUSE: +- n/a + +* MODIFIED FILES: +- 8192cd.h - revision: 1.38 +- 8192cd_cfg.h - revision: 1.42 +- 8192cd_headers.h - revision: 1.19 +- 8192cd_osdep.c - revision: 1.38 +- 8192cd_proc.c - revision: 1.23 +- 8192cd_rx.c - revision: 1.19 +- 8192cd_security.c - revision: 1.6 +- 8192cd_sme.c - revision: 1.62 +- 8192cd_util.c - revision: 1.38 +- 8192cd_psk.c - revision: 1.5 + +* DESCRIPTION +- Add a new define flag (LAZY_WDS) for this feature, default is not defined +- When lazy wds is enabled (mib value of wds_enable is set to 2), it will + dynamically learn wds mac address by incoming packet. +- Also, after this wds entry is created dynamically, the expired timer will + count down its expired time, like a noraml client did. When expired, this + entry will be flushed. Currently the max entry is 8 (NUM_WDS). +- Please note, when using lazy wds, wds interface (e.g., wlan0-wds0...) + must be binded to bridge interface (br0) and up in advanced. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/02/24, PlusWang, feature fine-tune + +* DIGEST: +- Disable 802.11d feature under default case + +* TEST METHOD: +- + +* ROOT CAUSE: +- + +* MODIFIED FILES: +- 8192cd.h - revision: 1.40 + +* DESCRIPTION +- Disable define flag "DOT11D" + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/02/23, jimmylin, feature fine-tune + +* DIGEST: +- Process ADDBA/DELBA according to TID + +* TEST METHOD: +- Send packet with DSCP value of non-0, but we will send ADDBA of TID 0. + +* ROOT CAUSE: +- We didn't process ADDBA/DELBA according to TID. We have only one variable + of ADDBA status for all traffic classes. + +* MODIFIED FILES: +- 8192cd.h - revision: 1.37 +- 8192cd_sme.c - revision: 1.53 +- 8192cd_tx.c - revision: 1.32 + +* DESCRIPTION +- We add an array of variables for all traffic classes and then will process + ADDBA/DELBA according to TID. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/02/11, jimmylin, feature fine-tune + +* DIGEST: +- Fine tune IOT dynamic mechanism for vap interfaces + +* TEST METHOD: +- Test TP with client associated in VAP interfaces will get lower result + comparing with associating to root interface + +* ROOT CAUSE: +- Originally we pick the highest TP station associated in root interface + as the target to do IOT dynamic mechanism. + +* MODIFIED FILES: +- 8192cd.h - revision: 1.36 +- 8192cd_headers.h - revision: 1.18 +- 8192cd_sme.c - revision: 1.52 +- 8192cd_util.c - revision: 1.37 + +* DESCRIPTION +- Now we will take the stations associated in VAP interfaces into + consideration. + +----------------------------------------------------------------------------- + +* DATE/WHOM/TYPE: +- 2010/02/11, davidhsu, bug fixed + +* DIGEST: +- System hangup when issue mp ctx command + +* TEST METHOD: +- Issue "mp_ctx" command, then system will hang-up. + +* ROOT CAUSE: +- When "mp_ctx" is executed, driver will read any keyboard key + (by uart) to break the transmission. Before transmitting, it will + disable uart interrupt, and then restore uart IMR when breaking. + The problem is uart IMR definition is wrong in driver. and which + will cause interrupt dispatch routine can't break the loop when + uart interrupt is happened. + +* MODIFIED FILES: +- 8192cd_mp.c - revision: 1.15 + +* DESCRIPTION +- Do not use a absolute value to access system register. Instead, + it will refer the pre-defined constant, defined in asm/rtl865x/platform.h + to access the register. + +----------------------------------------------------------------------------- + + +** 2010-02-04 victorlo ** + +- Refine WDS, set WDS peers to be BG devices as initial + +- Refine Desc number for RTL8198 + +------------------------------------------------------------------------------ + +** 2010-02-02 jimmylin ** + +- Send AMPDU only when the peer responses to send BA + * Fix IOT with EeePC in WEP mode + +------------------------------------------------------------------------------ + +** 2010-02-01 jimmylin ** + +- Restrict TKIP in N mode for WiFi test + +------------------------------------------------------------------------------ + +** 2010-01-27 jimmylin ** + +- Modify Tx shortcut to support more than 1 ethernet mac + +------------------------------------------------------------------------------ + +** 2010-01-26 jimmylin ** + +- Modify AMPDU spacing for AES IOT + +- Fix wrong HT IE under auto channel + +------------------------------------------------------------------------------ + +** 2010-01-25 victorlo ** + +- Refine LED setting + +------------------------------------------------------------------------------ + +** 2010-01-22 jimmylin ** + +- Modify get_shortcut_dev for VAP interfaces + +- Rearrange code in iRam + +------------------------------------------------------------------------------ + +** 2010-01-20 jimmylin ** + +- Turn on Tx power addition per rate + +------------------------------------------------------------------------------ + +** 2010-01-19 jimmylin ** + +- Add Tx count limit for MP continueous Tx + +------------------------------------------------------------------------------ + +** 2010-01-12 jimmylin ** + +- Fix WDS fixed rate issue + +------------------------------------------------------------------------------ + +** 2010-01-07 jimmylin ** + +- Fix unaligned access while parsing IE + +- Fix temporary removing key and adding back flow + +------------------------------------------------------------------------------ + +** 2010-01-06 jimmylin ** + +- Enable client mode and repeater mode + +------------------------------------------------------------------------------ + +** 2009-12-31 victorlo ** + +- Display fw header info + +------------------------------------------------------------------------------ + +** 2009-12-31 jimmylin ** + +- Fix WDS no aggregation issue + +------------------------------------------------------------------------------ + +** 2009-12-30 jimmylin ** + +- Fix WDS mode VLC play failed issue + +- Fix AP-client mode VLC play failed issue + +------------------------------------------------------------------------------ + +** 2009-12-28 jimmylin ** + +- Modify for station rate infomation + +------------------------------------------------------------------------------ + +** 2009-12-23 jimmylin ** + +- Add more protection and disable RX_RDU interrupt for system stability + +------------------------------------------------------------------------------ + +** 2009-12-21 victorlo ** + +- Enable SGI tx for all interface + +------------------------------------------------------------------------------ + +** 2009-12-18 victorlo ** + +- Enable per station ampdu aggr size + +------------------------------------------------------------------------------ + +** 2009-12-17 victorlo ** + +- Enable two stages beacon pwr enhance for Intel IOT + +------------------------------------------------------------------------------ + +** 2009-12-15 jimmylin ** + +- Don't do beacon hang check when system just init + +------------------------------------------------------------------------------ + +** 2009-12-14 victorlo ** + +- Stop aggr tx when ADDBA decline received, for non-intel stations + +------------------------------------------------------------------------------ + +** 2009-12-11 victorlo ** + +- Enable IOT mode (for n stations of Realtek and Intel) + +------------------------------------------------------------------------------ + +** 2009-12-07 jimmylin ** + +- Merge modification for Kernel 2.6 + +------------------------------------------------------------------------------ + +** 2009-12-04 button ** + +- Add PCIe power saving support + +- Set RTS initial rate to 24M + +- Using H2C cmd after fw download + +- Modify LLT table setting + +------------------------------------------------------------------------------ + +** 2009-12-03 victorlo ** + +- Enable CCK CCA dynamic adjust + +------------------------------------------------------------------------------ + +** 2009-12-02 victorlo ** + +- Enable per rate tx power increment + +------------------------------------------------------------------------------ + +** 2009-12-02 jimmylin ** + +- Add vxd AP and VAP interfaces to process dzqueue + +- Sync with 8192SE driver + * Fix 0xA0A in 0x83 in MP mode + * Update dynamic CCK CCA mechanism + +------------------------------------------------------------------------------ + +** 2009-12-01 davidhsu ** + +- Enlarge WPS IE buffer size to 260 bytes (PROBEIELEN). + +------------------------------------------------------------------------------ + +** 2009-11-27 victorlo ** + +- Refine MP for ctx + * Size down DMA burst size to avoid mp packet ctx hang + * Refine RCR to prevent any rx packet while ctx + +------------------------------------------------------------------------------ + +** 2009-11-27 jimmylin ** + +- Sync with 8192SE driver + * Add more protection for critical session + * Add timer_pending check before delete timer + * Add check in time-out function to see if driver is opened + * Add more protection for critical session + * Fix MBSSID power saving issue for PS-Poll + +------------------------------------------------------------------------------ + +** 2009-11-26 jimmylin ** + +- Modify Tx power offset range from -8 to +7 + +- Add traffic and rate infomation for power saving + +------------------------------------------------------------------------------ + +** 2009-11-26 victorlo ** + +- Update IQK for 8192c, to version 091112 + +- Add Realtek OUI bit for 8188c and 8192c + +------------------------------------------------------------------------------ + +** 2009-11-24 jimmylin ** + +- Fix Tx fragment bug + +------------------------------------------------------------------------------ + +** 2009-11-20 jimmylin ** + +- First release + |