summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
| | * | | | Cafstaging Release 1.0.0.233AnjaneeDevi Kapparapu2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Cafstaging Release 1.0.0.233 Change-Id: I9b28251e3a773f8f954bacbbcf4ebeff4e88aa95 CRs-Fixed: 688141
| | * | | | qca-cld:Fix for roaming failure with pmf requiredLeela Venkata Kiran Kumar Reddy Chirala2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Roaming fails with pmf Ap due to capabilities mismatch, as the initial connection pmf capabilities are not saved in the connected profile.save the pmf capabilities in the connected profile so that they can be used to prepare filter while roaming. CRs-fixed: 754455 Change-Id: I507e026328e7c707da70e0fa2e00fd1e485a8e34
| | * | | | qcacld: HDD: Increase threshold value for tcp delayed ack triggerRavi Joshi2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Provide fix for low tput numbers in BT-CoEx. Higher tcp delayed ack value seem to cause issues in BT-CoEx case reducing the overall tput. The increase of tput threshold from 4000 to 10000 will keep the tcp delayed ack value at 1 for BT-CoEx and other low tput conditions like HT20, VHT20, HT40 and VHT40. VHT80 mode is unaffected by this change. CRs-Fixed: 754237 Change-Id: I2000175f9159759d58eb1dbf9a1fcff1e86f1200
| | * | | | qcacld: Add cfg.ini support to enable/disable MAC address spoofingSrinivas Girigowda2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add support to enable/disable MAC address spoofing. By default feature is enabled; To disable set "gEnableMacAddrSpoof" to 0 Change-Id: I0235404949ac6c6ad3cffb5071d792497b03d37f CRs-Fixed: 753471
| | * | | | qca-cld: BMI: Revise SSR for USB interface in X86Chien-Ming Chen2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Revise USB SSR to dispatch an udev KOBJ_OFFLINE event. Upper framework could handle SSR by unloading/loading driver with this event type. There are files (/etc/udev/rules.d/71-qcacld.rules, /etc/qcacld-restart.sh) which handle this reloading by default. Change-Id: I8899ccd6abf01adc07effc7c443833b736de9ece CRs-Fixed: 747301
| | * | | | qcacld: Fix pktlogmod exit for the second timeChien-Ming Chen2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix kernel crash from remove_proc_entry() due to continuous calling of pktlogmod_exit(). Both PKTLOG_PROC_DIR and pl_dev->pl_info don't exist, we must prohibit this redundance. Variable usb_sc->hdd_removed isn't synchronized. To avoid race condition, new atomic variable - hif_usb_hdd_remove is added to protect. Change-Id: Ie75930b72ef56e930c75690a9205994c3ab55bee CRs-Fixed: 751681
| | * | | | qcacld-new: Move adf creiation/destroy into HIF for USB and SDIOMingcheng Zhu2014-11-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Current adf creation is inside of hdd_wlan_startup. And this implementation has problem in handling hdd_wlan_startup error condition where the socket buffer cannot be unmapped. The issue is addressed in this gerrit. Change-Id: Iac902edc229f12375c32e5ed6dd81b9dc8b8bd16 CRs-Fixed: 752044
| * | | | | Merge "Release 1.0.0.232 QCACLD WLAN Drive"Linux Build Service Account2014-11-11
| |\ \ \ \ \
| | * | | | | Release 1.0.0.232 QCACLD WLAN DriveAnjaneedevi Kapparapu2014-11-11
| | |\| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Merge remote-tracking branch 'origin/caf/caf-wlan/master' into HEAD * origin/caf/caf-wlan/master: Cafstaging Release 1.0.0.232 qcacld-new: Fix HIF memory leak when HIFStart returns error qcacld: UMAC: Fixing memory leak in pe/mac open qcacld: Fix memory leak in htt_rx.c::htt_rx_deinit() qcacld: Remove async_task processing in Rx path for optimization qcacld: Add extra debug information for scan request and results Change-Id: Ibf5bf90ef2949f179a1c50e2f16a051b6a73e091
| | | * | | | Cafstaging Release 1.0.0.232AnjaneeDevi Kapparapu2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Cafstaging Release 1.0.0.232 Change-Id: I92f53c71eedc16f69c87207893453314bc56ab95 CRs-Fixed: 688141
| | | * | | | qcacld-new: Fix HIF memory leak when HIFStart returns errorMingcheng Zhu2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix the memory leak when the HIFStart returns error. Also a typo in HIF_CE_completion_state memory allocation is fixed. Change-Id: Ibe26f4c314711ad7e5e2ace387ff1f9fb3e9eec7 CRs-Fixed: 753596
| | | * | | | qcacld: UMAC: Fixing memory leak in pe/mac openKrishna Kumaar Natarajan2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently we leak memory in peOpen if for some reason peOpen returns failure. Memory is freed only in peClose. This change set fix all the memory leak in peOpen and macOpen. Change-Id: I8f500ae69c71de16fc670a00a1f80ed14addeab8 CRs-Fixed: 753755
| | | * | | | qcacld: Fix memory leak in htt_rx.c::htt_rx_deinit()Orhan K Akyildiz2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes two memory leaks: 1) A success path memory leak which happens every time wlan driver is unloaded. This is because htt_rx_hash_deinit function used to miss deallocation of [htt_]pdev->hash_table. 2) In function htt_rx_hash_deinit(), there is also a memory leak in failure path. It happens if the hash_table allocation is successful, but allocation of one of the subsequent hash bucket entries fails. In this case, previously allocated hash-bucket entries are deallocated properly, but the hash_table was not de-allocated, causing the leak. CRs-Fixed: 753700 Change-Id: I4cd7ec0e671fca2e360eacd8fa0b4445c9bd8806
| | | * | | | qcacld: Remove async_task processing in Rx path for optimizationChandrasekaran, Manishekar2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Multiple context swicthes are invovled, when async_task is used in the Rx path and this increases the processing delay and reduces the throughput. This change directly invokes HIF read/write without involving the async task. Change-Id: Ib821e1affa463d71d92f82aa42415636d471745f CRs-Fixed: 750890
| | | * | | | qcacld: Add extra debug information for scan request and resultsDeepak Dhamdhere2014-11-11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Added debug logs to find out why SME scan request does not start fresh scan. Also added debug logs to find out reason for the "Found 0 BSS" message. This is a temporary change. We will change log levels after debugging is complete. Change-Id: I5d645739f556162e7d7d86174764c95946b8d2ad CRs-Fixed: 752920
| * | | | | | Merge "Release 1.0.0.231 QCACLD WLAN Drive"Linux Build Service Account2014-11-11
| |\| | | | |
| | * | | | | Release 1.0.0.231 QCACLD WLAN DriveAnjaneedevi Kapparapu2014-11-10
| | |\| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Merge remote-tracking branch 'origin/caf/caf-wlan/master' into HEAD * origin/caf/caf-wlan/master: Cafstaging Release 1.0.0.231 qca_cld: Update APIs for 3.14 kernel qcacld: update fw common interface files qcacld: Remove the unnecessary code from HIF suspend/resume qcacld: reset protection bits at regular interval qcacld: MSM8994: Fix correct feature set for get_concurrency_matrix qcacld HDD: Block scan during DFS operation Change-Id: Icf50b856f7f3e5a78cbcb99aa8341b20d437c468
| | | * | | | Cafstaging Release 1.0.0.231AnjaneeDevi Kapparapu2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Cafstaging Release 1.0.0.231 Change-Id: I9d6e6ac6aa300de73321f05935a4955ed94e7bc1 CRs-Fixed: 688141
| | | * | | | qca_cld: Update APIs for 3.14 kernelYue Ma2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Some CFG/NL80211 and kernel APIs etc. are changed during kernel update to 3.14 version. Update the corresponding APIs in host driver to fix compilation errors. Change-Id: Idc692e83307ff3f3b881d4e1027ba9d8286e3cbe CRs-fixed: 745059
| | | * | | | qcacld: update fw common interface filessyeh2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This change set update the fw common interface files associated with the CL 1132271 Change-Id: I4f0fb97911beed62ba0c157d54f5946595a7f5f5 CRs-Fixed: 750310
| | | * | | | qcacld: Remove the unnecessary code from HIF suspend/resumeRandy Chiu2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove unnecessary scan status check in hif resume callback and redundant vos check in hif suspend/resume callback for USB interface. Change-Id: I145bba58561191e57c8faf8a7ddfbb48777186e0 CRs-Fixed: 749509
| | | * | | | qcacld: reset protection bits at regular intervalNaveen Rawat2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently protection bits in beacon template are updated, when a legacy AP beacon heard by SAP. If the AP goes away, those bit cannot be reset, hence this gerrit reset them at regular interval. If legacy APs are still present, their beacons will again update protection bit, but at least we can ensure that currect status of protection bits in beacon template. Change-Id: I8a29c74e2ddf700a4ea7e41804d031f990458610 CRs-Fixed: 747886
| | | * | | | qcacld: MSM8994: Fix correct feature set for get_concurrency_matrixSrinivas Girigowda2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently we support only STA and P2P concurrency on MSM8994. Hence remove other unsupported concurrency feature set from the get_concurrency_matrix API. Change-Id: I17c71896cc8ca2723d44e4c9b717d0d2955ea87f CRs-Fixed: 752651
| | | * | | | qcacld HDD: Block scan during DFS operationManikandan Mohan2014-11-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | DFS Sta-AP mode not supported. Also in AP-AP mode, DFS is allowed in SCC mode only. But when Sec AP starts, we observe OBSS scan which results in phy error resulting in FAKE RADAR event. Fixed this issue by blocking all scan during DFS Master mode operation. Both wext and nl scan will be aborted and null scan result is sent. Change-Id: Ibecf38a57c0043d88056f73084feb7309e544480 CRs-fixed: 748080
| * | | | | | Merge "Release 1.0.0.230 QCACLD WLAN Drive"Linux Build Service Account2014-11-11
| |\| | | | |
| | * | | | | Release 1.0.0.230 QCACLD WLAN DriveAnjaneedevi Kapparapu2014-11-08
| | |\| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Merge remote-tracking branch 'origin/caf/caf-wlan/master' into HEAD * origin/caf/caf-wlan/master: Cafstaging Release 1.0.0.230 wlan: qca_cld: WAR to reject bounced back IPv6 Multicast (NS) Frames qcacld-new: Avoid allocate large memory at runtime. qcacld-new: move adf_ctx creation and initialization from HDD to HIF qcacld-new: fix the adf_ctx creation/initialization issue qcacld-2.0: Clean up IsMemoryDebugSupportEnabled CFG item. qcacld:wma: Add reason code for station kickout event Change-Id: I8caf646b3b867a3cd898b4ba52b78473eedbc53c
| | | * | | | Cafstaging Release 1.0.0.230AnjaneeDevi Kapparapu2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Cafstaging Release 1.0.0.230 Change-Id: I56af17c0d0b72430e86d5464659c5398a236accb CRs-Fixed: 688141
| | | * | | | wlan: qca_cld: WAR to reject bounced back IPv6 Multicast (NS) FramesDebashis Dutt2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Description: This is a workaround for IPv6 Multicast (Neighbor Solicitation) frames being not rejected by the WLAN driver when these frames are being bounced back from the AP. Currently, this is just a workaround, till the point we completely root-cause this issue (figure out how this works for the partial-offload case). Change-Id: If5d48ed903f484805e7f4be9df52582e50bd6ce5 CRs-Fixed: 748105
| | | * | | | qcacld-new: Avoid allocate large memory at runtime.Mingcheng Zhu2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In HIF probing, HIF allocates a very large memory for its complete_state link list. This cause stability issues when the system memory is fragmented. The implementation changed to allocate a list of small memory blocks to construct the free complete_state link list. Change-Id: Idb0f68aa836621892ed42cbfd427665f60f89a36 CRs-Fixed: 752935
| | | * | | | qcacld-new: move adf_ctx creation and initialization from HDD to HIFMingcheng Zhu2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is a problem with existing implementation: if hdd_wlan_startup returns error HIF cannot unmap and free nbufs because the adf_ctx is freed by hdd_wlan_startup. To address this problem adf_ctx creation and initialization needs to be within hif_pci_prob. Change-Id: If2fdb2212a41cf6f2f282f5544baa10a49fed8cf Crs-Fixed: 752044
| | | * | | | qcacld-new: fix the adf_ctx creation/initialization issueMingcheng Zhu2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Current implementation is to create and initialize adf_ctx inside hdd_wlan_startup function. This implemplementation has an issue that HIF cannot unmap and free socket buffers if hdd_wlan_startup failed. The fix is to move the adf_ctx craetion and initialization to hif_pci_prob. The adf_ctx deinitialization is also moved from HDD to HIF alyer. vos_mem_init function is also moved to hdd_driver_init so that HIF can use VOSS memory debug feature. Change-Id: I204ef0527949bf2de4d57b5953b66439306f7187 CRs-Fixed: 752044
| | | * | | | qcacld-2.0: Clean up IsMemoryDebugSupportEnabled CFG item.Rajeev Kumar2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In ROME based targets SLUB debug build is by default made and tested so no need of CFG INI item to control SLUB debug feature as its already controlled by CONFIG_SLUB_DEBUG in Kbuild. This fix will allow us to use VOS memory debug capability even for those memory allocations which are done during HIF PCIe probe/registration. Change-Id: Id73e6c7ca5e8c13ca9509450a28db960baf9abc8 CRs-Fixed: 752858
| | | * | | | qcacld:wma: Add reason code for station kickout eventManikandaraja Venkatachalapathy2014-11-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Handle station kickout event that is embedded in wow wakeup event for the suspend resume scenario. Change-Id: I192e776091f7d2c43abc28d49693d96372d40edf CRs-Fixed: 751569
| * | | | | | Merge "Release 1.0.0.229 QCACLD WLAN Drive"Linux Build Service Account2014-11-11
| |\| | | | |
| | * | | | | Release 1.0.0.229 QCACLD WLAN DriveAnjaneedevi Kapparapu2014-11-07
| | |\| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Merge remote-tracking branch 'origin/caf/caf-wlan/master' into HEAD * origin/caf/caf-wlan/master: Cafstaging Release 1.0.0.229 qcacld: TDLS: Make tdls logging as per UMAC modules wlan: qcacld: Disable "gAllowDFSChannelRoam" by default in cfg file wlan: qcacld: Enable DFS channel roam in ini file qcacld: BMI: Support SSR for SDIO interface in X86 qcacld SAP: Set ACS Scan as active scan type qcacld HDD: Fix dynamic phymode change issues qcacld: hdd: add condition to check to acquire dfs wakelock qcacld: UMAC: Use the defined macros consistently across qcacld:SDIO Fix for Data Stall issue qcacld: Fix the handling of invalid input to SETBAND qcacld-2.0: Fix of wlan logger thread getting stuck qcacld: Reset the center frequency segment with fallback qca-cld:hdd: Set proper IE field for HT20/HT40 mode Change-Id: Ice1ad586e8ee8ad1a2e4dc8d9a30cf7a49a235a7
| | | * | | | Cafstaging Release 1.0.0.229AnjaneeDevi Kapparapu2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Cafstaging Release 1.0.0.229 Change-Id: Iefd5a6ca6bd131413ab873c4632861b23505e561 CRs-Fixed: 688141
| | | * | | | qcacld: TDLS: Make tdls logging as per UMAC modulesAtul Mittal2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | TDLS module uses different macros to enable/disable logging. Change the same to match other UMAC modules. This will enable the tdls logs to be routed to logger application. Change-Id: Ic406e74d3a7c93844a205b7f6cde89dc9bf5b62c CRs-Fixed: 691352
| | | * | | | wlan: qcacld: Disable "gAllowDFSChannelRoam" by default in cfg fileGirish Gowli2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In cfg file, change default value of "gAllowDFSChannelRoam" to CFG_ROAMING_DFS_CHANNEL_DISABLED Change-Id: I664f4143300ab1a5cb67470db47e61fad1e5a053 CRs-Fixed: 744321
| | | * | | | wlan: qcacld: Enable DFS channel roam in ini fileGirish Gowli2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Enaable DFS channel roam in ini file using the ini parameter gAllowDFSChannelRoam Change-Id: I46cdbd3717bd499f36e7c43ba6fc0a0a98ce401b CRs-Fixed: 744321
| | | * | | | qcacld: BMI: Support SSR for SDIO interface in X86syeh2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Leverage the udev framework to reload the wlan SDIO driver to support SSR for SDIO interface in x86. We modify the driver code to report a KOBJ_OFFLINE event when SSR is triggered. And the udev user-level application would parse this event and trigger the script we provide (/etc/udev/rules.d/71-qcacld.rules, /etc/qcacld-restart.sh) to reload the SDIO driver. Besides, we might set "gEnableFwSelfRecovery=1" in qcom_cfg.ini to enable this fw recovery mechanism in the driver. When firmware crash occurs, it will trigger fw recovery and then wlan sdio driver will be reloaded. Change-Id: I8ef00565729e89f79034943ac7a94a5bfb4ef44c CRs-Fixed: 750263
| | | * | | | qcacld SAP: Set ACS Scan as active scan typeManikandan Mohan2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ACS scan type is set active scan to reduce the ACS scan delay and thereby improving the SAP startup time. Further even though scan type is give as active, the scan offload done in FW takes care of passive scan for DFS channels. Change-Id: I8512526c363886ac50ca2a9952d47b9afd38fbcd CRs-fixed: 746027
| | | * | | | qcacld HDD: Fix dynamic phymode change issuesManikandan Mohan2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 1. Fix 11AC VHT20 advertising secondary ch offset by disabling channel bonding for 11AC VHT20 2. Fix hostapd restart failure after dynamic enabling of 5g channel bonding support. Change-Id: I978bf6562f1a312fb416878a6c921eca068f8781 CRs-fixed: 752170
| | | * | | | qcacld: hdd: add condition to check to acquire dfs wakelockRyan Hsu2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When SAP operates in DFS channel, sap_dfs_ref_cnt is used to keep tracking the wakelock that has been acquired to prevent the system go into suspend. When start the SAP, it'll acquire the lock first, 0. acquire lock for old channel A (sap_dfs_ref_cnt++) And when radar comes in, the following logic will happen 1. acquire lock for new channel B (sap_dfs_ref_cnt++) 2. release lock for old channel A (sap_dfs_ref_cnt--) Also the bssState will be used as sainty checking, to decide if we've to acquire/release the lock or not. With above logic, the bssState is already BSS_START. So that the step#1 to acquire lock on new channel would never be able to done. And only the step#2 release lock on old channel will be done. This then allow the system goes into suspend state. Add the sap_df_ref_cnt as the combo condition to decide if we have to acquire the lock or not. Change-Id: Id2c565a2cd48c419d5245b276cabbf8279ab9c47 CRs-fixed: 751325
| | | * | | | qcacld: UMAC: Use the defined macros consistently acrossSrinivas Girigowda2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In the UMAC code, limSystemRole variable is accessed in three ways, 1. Directly accessing the variable psessionEntry->limSystemRole; 2. Using GET_LIM_SYSTEM_ROLE macro; and 3. By invoking the function limGetSystemRole(). With this change, use GET_LIM_SYSTEM_ROLE to access the variable consistently across. Similarly replace pMac->gDriverType with ANI_DRIVER_TYPE(pMac). Change-Id: I82c083afd96f7fa8514a2fcc5aa11cef349c4851 CRs-Fixed: 745779
| | | * | | | qcacld:SDIO Fix for Data Stall issueGanesh Babu Kumaravel2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is race where HTCSendDataPkt() thinks the tx queue is in access by referring TxProcessCount so it puts the data in tx queue and returns. But HTCTrySend() which gets called from HTCTxCompletionHandler() serviced the queue but not yet cleared the TxProcessCount. But unlocks the traget htc lock. This race results in a case where the tx packet is not processed at all and txrx scheduler is waiting for this credit to download other packets. So fix it by unlocking the htc lock after clearing TxProcessCount. CRs-Fixed: 751626 Change-Id: If00d9e8c4757313671c58aee3d9869553424fdd6
| | | * | | | qcacld: Fix the handling of invalid input to SETBANDSrinivas Girigowda2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The root cause, if the command is "SETBAND 100" the existing code reads only the 8th byte and ignores the following bytes and hence "SETBAND 100" is treated similar to "SETBAND 1" and proceeds without throwing error. The fix is to read the entire ASCII string and convert it to integer and let the code take care of error handling. Change-Id: Ibcd1a7a5c88e9b05c49d54ef3c7163be536c5f25 CRs-Fixed: 752065
| | | * | | | qcacld-2.0: Fix of wlan logger thread getting stuckRajeev Kumar2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | if cnss_diag is not registered then do not wake up logger thread also do not send any broadcast indication from WLAN driver to CNSS_DIAG. CNSS_DIAG to check WLAN interface status and register itself once it detects WLAN interface is up. Change-Id: I5e829b13b1608cb11290a30ec6dd73d5f66d6c9d CRs-fixed: 750700
| | | * | | | qcacld: Reset the center frequency segment with fallbackKiran Kumar Lokere2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Reset the center frequency segment when SAP fallbacks to 20/40Mhz mode from 80MHz. Change-Id: I72ee1ddf4e19eb27fcc29add41b42cf7f52195af CRs-Fixed: 750795
| | | * | | | qca-cld:hdd: Set proper IE field for HT20/HT40 modePeng Xu2014-11-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Based on hostapd.conf, INI and iwpriv setphymode settings, set proper HT Capabilities, OBSS scan parameter and HT operation IE field in beacon. Change-Id: I7ba312d0eb3bfc6cdd71408efb5df3eb28800715 CRs-fixed: 748530
| * | | | | | Merge "Release 1.0.0.228 QCACLD WLAN Drive"Linux Build Service Account2014-11-11
| |\| | | | |
generated by cgit v1.2.3 (git 2.25.1) at 2026-02-08 05:58:21 +0000