- 29 Sep, 2015 34 commits
-
-
Zhaoyang Liu authored
This patch add support for usb multi endpoints resync. Once multi channel event is received from firmware, update usb_mc_setp flag to block TX data until setup is over. And available data endpoint will be attached to BSS interface. Signed-off-by:
Zhaoyang Liu <liuzy@marvell.com> Signed-off-by:
Cathy Luo <cluo@marvell.com> Signed-off-by:
Amitkumar Karwar <akarwar@marvell.com> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org>
-
Zhaoyang Liu authored
This patch add support for USB interface to TX data with different endpoint. And previous TX information are saved in new designed structure. Signed-off-by:
-
Zhaoyang Liu authored
This patch add multi channel event process for USB multi tx data endpoints. Driver receives firmware multi channel event only in case that new connection is setup or interface is disconnect. Different BSS interface need update used usb endpoint. Signed-off-by:
-
Zhaoyang Liu authored
This patch move data_sent flag update to usb.c file. >From now all data_sent update cases only happened in specific file: sdio.c, usb.c, pcie.c. Outside ot these files, it is only allowed to check the value of data_ent. Signed-off-by:
-
Zhaoyang Liu authored
This patch fix missing tx data_sent flag update for usb interface. Except USB interface, data_sent flag has been updated in specific file such as sdio.c and pcie.c. So only USB interface type need check when TX data completed. Signed-off-by:
-
Amitkumar Karwar authored
We have preallocated buffer pool for Tx data. During high data traffic, all buffers are submitted to USB and driver needs to wait until one of the buffers get available for next Tx packet. "data: -ENOSR is returned" errors is expected in this case. Let's lower the priority of this message. Signed-off-by:
-
Zhaoyang Liu authored
This patch fixes observed issues while updating counter for number of paused data packets in wmm queue when trying to delete packet or delete station entry. Signed-off-by:
-
Amitkumar Karwar authored
Preallocated PCIe buffer is being reused for all PCIe interface events. Physical address of the buffer is shared with firmware so that it can perform DMA on it. As event length is specified in the header, there should not be a problem if the buffer gets overwritten. We will save some cycles by avoiding memset everytime while submitting the buffer to firmware. Fixes: 2728cecd(mwifiex: corrections in PCIe event skb) Signed-off-by:
-
Aniket Nagarnaik authored
This NULL pointer dereference is observed during suspend resume stress test. All pending commands are cancelled when system goes into suspend state. There a corner case in which host may receive response for last scan command after this and try to trigger extra active scan for hidden SSIDs. The issue is fixed by adding a NULL check to skip that extra scan. Fixes: 2375fa2b (mwifiex: fix unable to connect hidden SSID..) Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by:
Aniket Nagarnaik <aniketn@marvell.com> Signed-off-by:
-
Aniket Nagarnaik authored
Current TDLS implementation always includes ht/vht information in tdls setup confirm frame which causes teardown by legacy peer station after TDLS handshake. We will inclue ht/vht capabilities in tdls setup confirm frame only if peer station supports it to fix this problem. Signed-off-by:
-
Xinming Hu authored
cfg80211_inform_bss would return null in some cases, such as memory allocation failure. This patch adds sanity check for this case, to avoid possible issues when above corner case is hit. Signed-off-by:
Xinming Hu <huxm@marvell.com> Signed-off-by:
-
Xinming Hu authored
64k Tx and Rx buffers are allocated during driver initialization for SDIO level data aggregations. When host is under memory pressure situation, kzalloc() request for 64k may fail. We will try allocating 32k buffers and disable our rx single port aggreagation feature in this situation. If the allocation still fails, we will disable our sdio multport aggregation feature as well. In this way, we will transmit and receive packets one by one, thus reduce the demand for big memory. Signed-off-by:
-
Xinming Hu authored
This patch includes below changes 1) Check if multiport aggregation feature flag while doing rx length validation check. 2) Reset buffer size variables when buffers are freed. Signed-off-by:
-
Zhaoyang Liu authored
This patch fixes the issue of delay time in A-MSDU tx packet. In generated new A-MSDU packet, the time stamp is initialized to zero. Choose the time of first MSDU packet as aggregated packet's time. Signed-off-by:
-
Xinming Hu authored
Interface counters should be incremented in add_virtual_intf() and decremented in del_virtual_intf() function. This patch corrects it. Signed-off-by:
-
Guodong Xu authored
Align fields reg_ch_conf_last and reg_ch_conf_pending of struct wl1271{} to 64bit. Without this, on 64bit ARM, wlcore_set_pending_regdomain_ch() fails at the point it calls set_bit(ch_bit_idx, (long*)wl->reg_ch_conf_pending); Here is the error message while doing iw wlan0 scan or connect: [ 10.666857] wlcore: IRQ work [ 10.670046] wlcore: intr: 0x40 (fw_rx_counter = 1, drv_rx_counter = 0, tx_results_counter = 0) [ 10.678697] wlcore: WL1271_ACX_INTR_DATA [ 10.682810] Unhandled fault: alignment fault (0x96000021) at 0xffffffc037a817f4 [ 10.690139] Internal error: : 96000021 [#1] PREEMPT SMP [ 10.695366] Modules linked in: [ 10.698437] CPU: 3 PID: 894 Comm: irq/60-wl18xx Tainted: G W 4.2.0-rc6-linaro-hikey #2 [ 10.707501] Hardware name: HiKey Development Board (DT) [ 10.712733] task: ffffffc03a9d1680 ti: ffffffc039e18000 task.ti: ffffffc039e18000 [ 10.720239] PC is at set_bit+0x14/0x30 [ 10.724002] LR is at wlcore_set_pending_regdomain_ch+0x40/0x4c Signed-off-by:Guodong Xu <guodong.xu@linaro.org> Signed-off-by:
-
Jan Kaisrlik authored
The patch adds support for "outside the context of a BSS"(OCB) mode to ath9k driver and extends debugfs files by OCB ralated information. This patch was tested on AR9380-AL1A cards. Signed-off-by:
Jan Kaisrlik <kaisrja1@fel.cvut.cz> Cc: Michal Sojka <sojkam1@fel.cvut.cz> Signed-off-by:
-
Hiroaki KAWAI authored
Fix rssi calculation error which was introduced in otus to ar9170 porting. Signed-off-by:
Hiroaki KAWAI <hiroaki.kawai@gmail.com> Acked-by:
Christian Lamparter <chunkeey@googlemail.com> Signed-off-by:
-
Geoff Levand authored
Fixes the output of 'modinfo --field firmware'. Signed-off-by:
Geoff Levand <geoff@infradead.org> Signed-off-by:
-
Miaoqing Pan authored
This patch fix https://lists.openwrt.org/pipermail/openwrt-devel/ 2015-August/034979.html. As the peak detect calibration is set incorrectly. Signed-off-by:
Miaoqing Pan <miaoqing@qca.qualcomm.com> Signed-off-by:
-
Joe Perches authored
The mwifiex_dbg macro has two tests that could be consolidated into a function reducing overall object size ~10KB (~4%). So convert the macro into a function. $ size drivers/net/wireless/mwifiex/built-in.o* (x86-64 defconfig) text data bss dec hex filename 233102 8628 4809 246539 3c30b drivers/net/wireless/mwifiex/built-in.o.new 243949 8628 4809 257386 3ed6a drivers/net/wireless/mwifiex/built-in.o.old Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Hante Meuleman authored
When PCIE type devices are being FW reloaded without being properly reset then the device ends up in a locked state, requiring the device to be completely powered down. This patch adds a reset through watchdog at the moment the device (cores) has been recognized. This will solve warm reboot issues. Cc: Rafal Milecki <zajec5@gmail.com> Reviewed-by:
Arend Van Spriel <arend@broadcom.com> Reviewed-by:
Franky (Zhenhui) Lin <frankyl@broadcom.com> Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Signed-off-by:
Hante Meuleman <meuleman@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
In case of error during brcmf_bus_start() the network interfaces were freed using free_netdev(). However, the interfaces may have additional memory allocated which is not freed. The netdev has destructor set to brcmf_cfg80211_free_netdev() which frees the additional memory if allocated and call free_netdev(). The brcmf_net_detach() either calls brcmf_cfg80211_free_netdev() directly or uses unregister_netdev() when struct net_device::reg_state indicates the netdev was registered. Reported-by:
Daniel (Deognyoun) Kim <dekim@broadcom.com> Reviewed-by:
-
Arend van Spriel authored
Instead of passing ifidx and drvr just pass struct brcmf_if pointer which holds both parameters. Reviewed-by:
-
Arend van Spriel authored
The brcmf_fws_txstatus_suppressed() function prototype specifies an ifidx parameter which is not used within the function implementation. Reviewed-by:
-
Arend van Spriel authored
Both PCIe and SDIO devices have the possibility to log the firmware console output in kernel log. For PCIe it is logged when PCIE debug level is enabled. For SDIO it is logged when user specifies a non-zero console interval through debugfs. This patch tries to make it a bit more consistent. The firmware console output is only logged when FWCON debug level is enabled. Reviewed-by:
Pontus Fuchs <pontusf@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
Because the P2P Device interface in firmware uses the same interface index as the primary interface we use the bsscfg index as index in the struct brcmf_pub::iflist. However, in the data path we get the interface index and not the bsscfg index. So we need a mapping of interface index to bsscfg index, which can be determined upon handle adding the interface. Reviewed-by:
-
Arend van Spriel authored
Most call sites of brcmf_txfinalize already have struct brcmf_if instance so pass that to brcmf_txfinalize() as the function needs it anyway. Reviewed-by:
-
Arend van Spriel authored
The knowledge on how to map the interface index to a struct brcmf_if instance is in brcmf_get_ifp() so use that function when only the interface index is known instead of accessing brcmf_pub::iflist directly. Reviewed-by:
-
Arend van Spriel authored
The p2pdev interface is setup in firmware resulting in a interface event. This event has role and no-if flag. When role is p2p client and no-if flag is set it indicates that this is the p2pdev interface. This info is used in handling the event and adding interface in the driver. Reviewed-by:
-
Arend van Spriel authored
In brcmf_bus_start() the function brcmf_cfg80211_attach() is called which may fail. If this happens we should not call brcmf_cfg80211_detach() in the failure path as it will result in NULL pointer dereference: brcmf_fweh_activate_events: Set event_msgs error (-5) brcmf_bus_start: failed: -5 brcmf_sdio_firmware_callback: dongle is not responding BUG: unable to handle kernel NULL pointer dereference at 0000000000000068 IP: [<ffffffff811e8f08>] kernfs_find_ns+0x18/0xd0 PGD 0 Oops: 0000 [#1] SMP Modules linked in: brcmfmac(O) brcmutil(O) cfg80211 auth_rpcgss CPU: 1 PID: 45 Comm: kworker/1:1 Tainted: G O Hardware name: Dell Inc. Latitude E6410/07XJP9, BIOS A07 02/15/2011 Workqueue: events request_firmware_work_func task: ffff880036c09ac0 ti: ffff880036dd4000 task.ti: ffff880036dd4000 RIP: 0010:[<ffffffff811e8f08>] [<ffffffff811e8f08>] kernfs_find_ns+0x18/0xd0 RSP: 0018:ffff880036dd7a28 EFLAGS: 00010246 RAX: ffff880036c09ac0 RBX: 0000000000000000 RCX: 000000007fffffff RDX: 0000000000000000 RSI: ffffffff816578b9 RDI: 0000000000000000 RBP: ffff880036dd7a48 R08: 0000000000000000 R09: ffff880036c0b340 R10: 00000000000002ec R11: ffff880036dd7b08 R12: ffffffff816578b9 R13: 0000000000000000 R14: ffffffff816578b9 R15: ffff8800c6c87000 FS: 0000000000000000(0000) GS:ffff88012bc40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000068 CR3: 0000000001a0b000 CR4: 00000000000006e0 Stack: 0000000000000000 ffffffff816578b9 0000000000000000 ffff8800c0d003c8 ffff880036dd7a78 ffffffff811e8ff5 0000000ffffffff1 ffffffff81a9b060 ffff8800c789f880 ffff8800c0d00000 ffff880036dd7a98 ffffffff811ebe0d Call Trace: [<ffffffff811e8ff5>] kernfs_find_and_get_ns+0x35/0x60 [<ffffffff811ebe0d>] sysfs_unmerge_group+0x1d/0x60 [<ffffffff81404ef2>] dpm_sysfs_remove+0x22/0x60 [<ffffffff813f9db9>] device_del+0x49/0x240 [<ffffffff815da768>] rfkill_unregister+0x58/0xc0 [<ffffffffa06bd91b>] wiphy_unregister+0xab/0x2f0 [cfg80211] [<ffffffffa0742fe3>] brcmf_cfg80211_detach+0x23/0x50 [brcmfmac] [<ffffffffa074d986>] brcmf_detach+0x86/0xe0 [brcmfmac] [<ffffffffa0757de8>] brcmf_sdio_remove+0x48/0x120 [brcmfmac] [<ffffffffa0758ed9>] brcmf_sdiod_remove+0x29/0xd0 [brcmfmac] [<ffffffffa0759031>] brcmf_ops_sdio_remove+0xb1/0x110 [brcmfmac] [<ffffffffa001c267>] sdio_bus_remove+0x37/0x100 [mmc_core] [<ffffffff813fe026>] __device_release_driver+0x96/0x130 [<ffffffff813fe0e3>] device_release_driver+0x23/0x30 [<ffffffffa0754bc8>] brcmf_sdio_firmware_callback+0x2a8/0x5d0 [brcmfmac] [<ffffffffa074deaf>] brcmf_fw_request_nvram_done+0x15f/0x5e0 [brcmfmac] [<ffffffff8140142f>] ? devres_add+0x3f/0x50 [<ffffffff810642b5>] ? usermodehelper_read_unlock+0x15/0x20 [<ffffffff81400000>] ? platform_match+0x70/0xa0 [<ffffffff8140f400>] request_firmware_work_func+0x30/0x60 [<ffffffff8106828c>] process_one_work+0x14c/0x3d0 [<ffffffff8106862a>] worker_thread+0x11a/0x450 [<ffffffff81068510>] ? process_one_work+0x3d0/0x3d0 [<ffffffff8106d692>] kthread+0xd2/0xf0 [<ffffffff8106d5c0>] ? kthread_create_on_node+0x180/0x180 [<ffffffff815ed35f>] ret_from_fork+0x3f/0x70 [<ffffffff8106d5c0>] ? kthread_create_on_node+0x180/0x180 Code: e9 40 fe ff ff 48 89 d8 eb 87 66 0f 1f 84 00 00 00 00 00 66 66 66 66 90 55 48 89 e5 41 56 49 89 f6 41 55 49 89 d5 31 d2 41 54 53 <0f> b7 47 68 48 8b 5f 48 66 c1 e8 05 83 e0 01 4d 85 ed 0f b6 c8 RIP [<ffffffff811e8f08>] kernfs_find_ns+0x18/0xd0 RSP <ffff880036dd7a28> CR2: 0000000000000068 ---[ end trace 87d6ec0d3fe46740 ]--- Reported-by:
-
Arend van Spriel authored
Just pass the interface to be removed, ie. the struct brcmf_if instance. Reviewed-by:
-
Arend van Spriel authored
Avoid spreading the ifidx in the driver, but have it return the struct brcmf_if instance. Reviewed-by:
-
Arend van Spriel authored
In rx path the firmware provide an interface index which is used to map to a struct brcmf_if instance. However, this involves some trick that is done in two places. This is changed by having driver core providing brcmf_get_ifp() function. Reviewed-by:
-
- 26 Sep, 2015 1 commit
-
-
Kalle Valo authored
Merge tag 'iwlwifi-next-for-kalle-2015-09-21' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next * some debugfs improvements; * fix signedness in beacon statistics; * deinline some functions to reduce size when device tracing is enabled; * filter beacons out in AP mode when no stations are associated; * deprecate firmwares version -12; * fix a runtime PM vs. legacy suspend race; * one-liner fix for a ToF bug; * clean-ups in the rx code; * small debugging improvement; * fix WoWLAN with new firmware versions;
-
- 21 Sep, 2015 5 commits
-
-
Eliad Peller authored
In order to verify d0i3 flow, add debug print to indicate d0i3 exit was completed (right after tx was re-enabled), along with the wakeup reasons. Signed-off-by:
Eliad Peller <eliadx.peller@intel.com> Signed-off-by:
Luca Coelho <luciano.coelho@intel.com>
-
Eliad Peller authored
Recent fw version added assert to make sure wowlan configuration is configured only when a station is connected. Change the driver behavior to pass this configuration only if we indeed have ap station id (i.e. connected). Signed-off-by:
-
Johannes Berg authored
The RX API is currently mixed up into the general fw-api.h file, but we're going to need to extend it significantly in the future, so move it to its own file. Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Johannes Berg authored
Remove some unused values from the RX API; these were used with older firmware API that didn't have the RX energy API, support for which was removed a long time ago. Signed-off-by:
-
Johannes Berg authored
Since the driver can never configure the data here, this field will always be reported as 0 by the firmware. Even if this was not the case, however, it wouldn't matter since the extra data would be added beyond the end of the phy_info structure we use in the driver, so wouldn't harm anything in this code either. Signed-off-by:
-
