- 28 May, 2015 2 commits
-
-
Luciano Coelho authored
The UMAC API supports multiple scan schedules, but now we use only a single one. Change the comment to make this clear and avoid confusion. Signed-off-by:
Luciano Coelho <luciano.coelho@intel.com> Reviewed-by:
David Spinadel <david.spinadel@intel.com> Signed-off-by:
Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Nicholas Krause authored
This removes the use of the two deprecated calls to the macro PTR_RET in iwl_mvm_get_regdomain and replaces them both to PTR_ERR_OR_ZERO. Signed-off-by:
Nicholas Krause <xerofoify@gmail.com> [Commit message editing] Signed-off-by:
-
- 26 May, 2015 3 commits
-
-
Eliad Peller authored
commit b112889c ("iwlwifi: mvm: add Aux ROC request/response flow") added aux ROC flow in addition to the existing ROC flow. While doing it, it moved the ROC reference release to a common work item, which is being called for both the ROC and aux ROC flows. This resulted in invalid reference accounting, as no reference was taken in case of aux ROC, while a reference was released on completion. Fix it by adding a reference for the aux ROC as well, and release only the relevant references on completion (according to the set bits). While at it, convert cancel_work_sync() to flush_work(), in order to make sure the references are being cleaned properly. Fixes: b112889c ("iwlwifi: mvm: add Aux ROC request/response flow") Signed-off-by:
Eliad Peller <eliadx.peller@intel.com> Reviewed-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Luciano Coelho authored
We should not call the iwl_pcie_set_pwr() functions in the suspend/resume flows for family 8000, because the register used is locked in devices from this family. Doing this causes an NMI protection error (RT_NMI_INTERRUPT_PREG_PROTECTION). To fix this, skip those calls if the device family is IWL_DEVICE_FAMILY_8000. Signed-off-by:
-
Emmanuel Grumbach authored
Signed-off-by:
-
- 21 May, 2015 6 commits
-
-
Eliad Peller authored
qos_seq points (to a struct) inside the command response data. Make sure to free the response only after qos_seq is not needed anymore. Reported-by:
Heng Luo <heng.luo@intel.com> Signed-off-by:
-
Luciano Coelho authored
If the device is reset during suspend with net-detect enabled, we leave the net-detect information dangling and this causes the next suspend to fail with a warning: [21795.351010] WARNING: at /root/iwlwifi/iwlwifi-stack-dev/drivers/net/wireless/iwlwifi/mvm/d3.c:989 __iwl_mvm_suspend.isra.6+0x2be/0x460 [iwlmvm]() [21795.353253] Modules linked in: iwlmvm(O) iwlwifi(O) mac80211(O) cfg80211(O) compat(O) [...] [21795.366168] CPU: 1 PID: 3645 Comm: bash Tainted: G O 3.10.29-dev #1 [21795.368785] Hardware name: Dell Inc. Latitude E6430/0CPWYR, BIOS A09 12/13/2012 [21795.371441] f8ec6748 f8ec6748 e51f3ce8 c168aa62 e51f3d10 c103a824 c1871238 f8ec6748 [21795.374228] 000003dd f8eb982e f8eb982e 00000000 c3408ed4 c41edbbc e51f3d20 c103a862 [21795.377006] 00000009 00000000 e51f3da8 f8eb982e c41ee3dc 00000004 e7970000 e51f3d74 [21795.379792] Call Trace: [21795.382461] [<c168aa62>] dump_stack+0x16/0x18 [21795.385133] [<c103a824>] warn_slowpath_common+0x64/0x80 [21795.387803] [<f8eb982e>] ? __iwl_mvm_suspend.isra.6+0x2be/0x460 [iwlmvm] [21795.390485] [<f8eb982e>] ? __iwl_mvm_suspend.isra.6+0x2be/0x460 [iwlmvm] [21795.393124] [<c103a862>] warn_slowpath_null+0x22/0x30 [21795.395787] [<f8eb982e>] __iwl_mvm_suspend.isra.6+0x2be/0x460 [iwlmvm] [21795.398464] [<f8eb9d7c>] iwl_mvm_suspend+0xec/0x140 [iwlmvm] [21795.401127] [<c104be11>] ? del_timer_sync+0xa1/0xc0 [21795.403800] [<f8d4107e>] __ieee80211_suspend+0x1de/0xff0 [mac80211] [21795.406459] [<c168e43d>] ? mutex_lock_nested+0x25d/0x350 [21795.409084] [<c1586b64>] ? rtnl_lock+0x14/0x20 [21795.411685] [<f8cf0076>] ieee80211_suspend+0x16/0x20 [mac80211] [21795.414318] [<f8c4e014>] wiphy_suspend+0x74/0x710 [cfg80211] [21795.416916] [<c141e612>] __device_suspend+0x1e2/0x220 [21795.419521] [<f8c4dfa0>] ? addresses_show+0xa0/0xa0 [cfg80211] [21795.422097] [<c141f997>] dpm_suspend+0x67/0x210 [21795.424661] [<c141fd6f>] dpm_suspend_start+0x4f/0x60 [21795.427219] [<c108d8e0>] suspend_devices_and_enter+0x60/0x480 [21795.429768] [<c168646a>] ? printk+0x4d/0x4f [21795.432295] [<c108de76>] pm_suspend+0x176/0x210 [21795.434830] [<c108ca5d>] state_store+0x5d/0xb0 [21795.437410] [<c108ca00>] ? wakeup_count_show+0x50/0x50 [21795.439961] [<c13208db>] kobj_attr_store+0x1b/0x30 [21795.442514] [<c11e3a4b>] sysfs_write_file+0xab/0x100 [21795.445088] [<c11e39a0>] ? sysfs_poll+0xa0/0xa0 [21795.447659] [<c1179655>] vfs_write+0xa5/0x1c0 [21795.450212] [<c1179af7>] SyS_write+0x57/0xa0 [21795.452699] [<c1699ec1>] sysenter_do_call+0x12/0x32 [21795.455146] ---[ end trace faf5321baba2bfdb ]--- To fix this, call the iwl_mvm_free_nd() function in case of any error during resume. Additionally, rename the "out_unlock" label to err to make it clearer that it's only called in error conditions. Cc: stable@vger.kernel.org [3.19+] Signed-off-by:
-
Luciano Coelho authored
The __iwl_mvm_resume() function always returns 1, which causes mac80211 to do a reconfig with IEEE80211_RECONFIG_TYPE_RESTART. This type of reconfig calls iwl_mvm_restart_complete(), where we unref the IWL_MVM_REF_UCODE_DOWN, so we should always take the reference in this case. This prevents this kind of warning from happening: [40026.103025] WARNING: at /root/iwlwifi/iwlwifi-stack-dev/drivers/net/wireless/iwlwifi/mvm/mac80211.c:236 iwl_mvm_unref+0xc9/0xd0 [iwlmvm]() [40026.105145] Modules linked in: iwlmvm(O) iwlwifi(O) mac80211(O) cfg80211(O) compat(O) ctr ccm arc4 autofs4 snd_hda_codec_hdmi snd_hda_codec_idt joydev coretemp kvm_intel kvm aesni_intel ablk_helper cryptd lrw aes_i586 snd_hda_intel xts snd_hda_codec gf128mul snd_hwdep snd_pcm snd_seq_midi dell_wmi snd_rawmidi sparse_keymap snd_seq_midi_event snd_seq uvcvideo dell_laptop videobuf2_core dcdbas microcode videodev psmouse snd_timer videobuf2_vmalloc videobuf2_memops serio_raw snd_seq_device btusb i915 snd bluetooth lpc_ich drm_kms_helper soundcore snd_page_alloc drm i2c_algo_bit wmi parport_pc ppdev video binfmt_misc rpcsec_gss_krb5 nfsd mac_hid nfs_acl nfsv4 auth_rpcgss nfs fscache lockd sunrpc msdos lp parport sdhci_pci sdhci ahci libahci e1000e mmc_core ptp pps_core [last unloaded: compat] [40026.117640] CPU: 2 PID: 3827 Comm: bash Tainted: G W O 3.10.29-dev #1 [40026.120216] Hardware name: Dell Inc. Latitude E6430/0CPWYR, BIOS A09 12/13/2012 [40026.122815] f8effd18 f8effd18 e740fd18 c168aa62 e740fd40 c103a824 c1871238 f8effd18 [40026.125527] 000000ec f8ec79c9 f8ec79c9 d5d29ba4 d5d2a20c 00000000 e740fd50 c103a862 [40026.128209] 00000009 00000000 e740fd7c f8ec79c9 f1c591c4 00000400 00000000 f8efb490 [40026.130886] Call Trace: [40026.133506] [<c168aa62>] dump_stack+0x16/0x18 [40026.136115] [<c103a824>] warn_slowpath_common+0x64/0x80 [40026.138727] [<f8ec79c9>] ? iwl_mvm_unref+0xc9/0xd0 [iwlmvm] [40026.141319] [<f8ec79c9>] ? iwl_mvm_unref+0xc9/0xd0 [iwlmvm] [40026.143881] [<c103a862>] warn_slowpath_null+0x22/0x30 [40026.146453] [<f8ec79c9>] iwl_mvm_unref+0xc9/0xd0 [iwlmvm] [40026.149030] [<f8ec7a4d>] iwl_mvm_mac_reconfig_complete+0x7d/0x210 [iwlmvm] [40026.151645] [<f8b74b20>] ? ftrace_raw_event_drv_reconfig_complete+0xc0/0xe0 [mac80211] [40026.154291] [<f8b6769e>] ieee80211_reconfig+0x28e/0x2620 [mac80211] [40026.156920] [<c10ef0ea>] ? ring_buffer_unlock_commit+0xba/0x100 [40026.159585] [<f8b4a04d>] ieee80211_resume+0x6d/0x80 [mac80211] [40026.162206] [<f8a79722>] wiphy_resume+0x72/0x260 [cfg80211] [40026.164799] [<c141e2e7>] ? device_resume+0x57/0x150 [40026.167425] [<f8a796b0>] ? wiphy_suspend+0x710/0x710 [cfg80211] [40026.170075] [<c141e26e>] dpm_run_callback+0x2e/0x50 [40026.172695] [<c141e321>] device_resume+0x91/0x150 [40026.175334] [<c141f636>] dpm_resume+0xf6/0x200 [40026.177922] [<c141f920>] dpm_resume_end+0x10/0x20 [40026.180489] [<c108d9f7>] suspend_devices_and_enter+0x177/0x480 [40026.183037] [<c168646a>] ? printk+0x4d/0x4f [40026.185559] [<c108de76>] pm_suspend+0x176/0x210 [40026.188065] [<c108ca5d>] state_store+0x5d/0xb0 [40026.190581] [<c108ca00>] ? wakeup_count_show+0x50/0x50 [40026.193052] [<c13208db>] kobj_attr_store+0x1b/0x30 [40026.195608] [<c11e3a4b>] sysfs_write_file+0xab/0x100 [40026.198055] [<c11e39a0>] ? sysfs_poll+0xa0/0xa0 [40026.200469] [<c1179655>] vfs_write+0xa5/0x1c0 [40026.202893] [<c1179af7>] SyS_write+0x57/0xa0 [40026.205245] [<c1699ec1>] sysenter_do_call+0x12/0x32 [40026.207619] ---[ end trace db1d5a72a0381b0a ]--- Signed-off-by:
EliadX Peller <eliad@wizery.com> Reviewed-by:
-
Emmanuel Grumbach authored
There are buses that can't handle ASYNC command without copying them. Duplicate the host command instead. Signed-off-by: -
Liad Kaufman authored
Take the MAC address from the OTP even if one is present in the NVM, if that MAC address happens to be a reserved one. Signed-off-by:
Liad Kaufman <liad.kaufman@intel.com> Signed-off-by:
-
Haim Dreyfuss authored
fw_status is the only pointer pointing to a block of memory allocated above and should be freed after use. Note: this come from Klockwork static analyzer. Cc: stable@vger.kernel.org [3.19+] Fixes: 2021a89d ("iwlwifi: mvm: treat netdetect wake up separately") Signed-off-by:
Haim Dreyfuss <haim.dreyfuss@intel.com> Signed-off-by:
-
- 15 May, 2015 3 commits
-
-
Emmanuel Grumbach authored
A few triggers have status = MLME_SUCCESS and they are still interesting. E.g. if we want to collect data upon deauth, the status will be MLME_SUCCESS. Fix that. Fixes: d42f5350 ("iwlwifi: mvm: add trigger for firmware dump upon MLME failures") Signed-off-by:
-
Avri Altman authored
Disabling the clocks is a standard procedure while stopping the device. On family 8000 however, disabling the bus master DMA clock increases the NIC's power consumption. To fix this, skip this call if the device family is IWL_DEVICE_FAMILY_8000. Signed-off-by:
Avri Altman <avri.altman@intel.com> Reviewed-by:
-
Emmanuel Grumbach authored
3165 really needs to load iwlwifi-7265D-13.ucode. This device is supported starting from -13.ucode, update the MIN and OK defines accordingly. While at it, add 3165 to the device list in the Kconfig file. Signed-off-by:
-
- 10 May, 2015 1 commit
-
-
Emmanuel Grumbach authored
There are devices that forbid MIMO by the mean of the NVM. Detect thoses devices and forbid MIMO otherwise the firmware would crash. STBC is still allowed on these devices. Signed-off-by:
-
- 03 May, 2015 1 commit
-
-
David Spinadel authored
Fix a copy paste bug that didn't copy wildcard SSIDs to scan requests. This bug causes scan with only wildcard SSID to be passive, and scans with more than one SSID to send only the direct probes. Fixes: 2a28ac14c518 ("iwlwifi: mvm: add common scan params to thw iwl_mvm_scan_params struct") Signed-off-by:
-
- 29 Apr, 2015 24 commits
-
-
Luciano Coelho authored
Some more tidbits left over from the legacy scan removal. Signed-off-by:
-
Chaya Rachel Ivgi authored
The thermal throttling parameters were constant and hardcoded, not allowing changes for different NIC families. Change this so that the values are part of the NIC family configuration and are not constant (so they can be changed dynamically in the future). Signed-off-by:
Chaya Rachel Ivgi <chaya.rachel.ivgi@intel.com> Reviewed-by:
-
Johannes Berg authored
When a firmware restart is done, don't try to reprogram the keys to new slots but rather just keep the old key index, while skipping keys that weren't programmed before. Not only does this restore the state more faithfully, but it will also allow using the HW key index for internal purposes as an array index. Signed-off-by:
-
Luciano Coelho authored
The generic scan command functions are now irrelevant, since both sched and regular scans are in the same code. So rename this functions to dwell and isolate all the dwell-related setting to them. Keeping the dwell code separate makes it easier to compare the LMAC and UMAC versions. Signed-off-by:
-
Luciano Coelho authored
Having an LMAC counterpart for the existing UMAC flags function makes things more consistent and easy to compare and spot the differences. The flags are the same, but are in different bits, so unfortunately we can't use a single function for both APIs. Signed-off-by:
-
Luciano Coelho authored
Clean the code a little by moving all the flags into the right function. Additionally, rename the function, since "common" is now irrelevant. Signed-off-by:
-
David Spinadel authored
Signed-off-by:
-
Emmanuel Grumbach authored
When we use an external buffer, it is allocated from the t DRAM and can be as big as 64MB. This buffer is huge and might not be needed for the specific issue being chased. Especially if lots of dumps are going to be created. Allow to limit the size of the buffer in the configuration. Signed-off-by: -
Eliad Peller authored
qos_seq points (to a struct) inside the command response data. Make sure to free the response only after qos_seq is not needed anymore. type=bugfix bug=not-tracked fixes=I78c07110b59dcd389207bd2b168b0a2760839fe0 Change-Id: I619ce2c17e064dc98c7be9abc1e23175fdc8fb9a Reported-by:
-
Luciano Coelho authored
With just a few differences left in the UMAC scan functions now, we can merge them into one, taking care of the small difference according to the total number of iterations required. Signed-off-by:
-
Luciano Coelho authored
In theory, LMAC scans can handle a 32-bit delay, but since waiting for over 18 hours to start the scan is a bit silly and to keep it aligned with UMAC scans (which only support 16-bit delays), trim it down to 16-bits. This makes the LMAC vs. UMAC and the UMAC reg scan vs. UMAC sched scan code more similar. Signed-off-by:
-
Luciano Coelho authored
The last remaining difference between the regular scan and scheduled scan flows for LMAC is the FW capabilities check for EBS scans. Merge these checks into a new function and then combine the LMAC scan functions into a single one. Signed-off-by:
-
Luciano Coelho authored
As another step towards combining the scan and sched scan functions, add parameters that tell the scan function how many iterations we want (i.e. 1 for normal scan, more for scheduled scan) and that set the full scan multiplier (only meaningful for LMAC). Signed-off-by:
-
Luciano Coelho authored
The probe request to be added to both LMAC and UMAC scan commands are identical, so move the generation out of the LMAC/UMAC-specific code. Signed-off-by:
-
Luciano Coelho authored
To make things clearer, rename the iwl_mvm_scan_calc_params() function to iwl_mvm_calc_dwell() and make it calculate and fill in only dwell-related parameters. Signed-off-by:
-
Luciano Coelho authored
Now that both scheduled scan and regular scan SSID populating functions do the same thing, they can be combined into a single function. Signed-off-by:
-
Luciano Coelho authored
The firmware inverts the order of the SSIDs sent out in probe requests (for some reason). For regular scans, we've been passing the SSIDs in the opposite order so they go out in the order we want. With scheduled scans, we were not doing that, so they were sent out in reverse order of priority. Fix that by using the reverse order when populating the SSIDs array for scheduled scans as well. Signed-off-by:
-
Luciano Coelho authored
The only difference in the ssid_bitmap between LMAC and UMAC scans is that in LMAC bits 1 to 20 are used, while in UMAC bits 0 to 19 are used (*sigh*). So we can combine the bitmap creation into a single function and simply shift left if LMAC is used. Signed-off-by:
-
Luciano Coelho authored
Many parameters are common for all scans. Instead of parsing the cfg80211 scan and sched scan requests differently in each flow, move the parsing outside of the API/scan-type specific functions. In this way, we only need to differentiate between scan types once. Signed-off-by:
-
Luciano Coelho authored
Similarly to the regular scan patch, a lot of the UMAC and LMAC sched scan code is also almost identical. Grab the low hanging fruits by combining the obvious parts. Signed-off-by:
-
Luciano Coelho authored
A lot of the UMAC and LMAC scan code is almost identical. Grab the low hanging fruits by combining the obvious parts. Signed-off-by:
-
Emmanuel Grumbach authored
The lists of the devices supported by either MVM or DVM firmwares was incomplete. Point to https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi#firmware instead of maintaining the lists. Signed-off-by:
-
Arik Nemtsov authored
Usually during HW recovery the state of all active interfaces is cleaned up during drv_start(). There's a special case where a HW restart is requested when an interface is going down. In this case the iface-iterator won't see this interface and we won't clean it. This has bad consequences once the interface is legitimately brought up again. Signed-off-by:
Arik Nemtsov <arikx.nemtsov@intel.com> Reviewed-by:
-
Luciano Coelho authored
Instead of repeating the same code in 4 different places, combine the comparisons into a new function. Additionally, this change fixes UMAC scans where the RRM IEs were not taken into consideration when calculating the IE length. Signed-off-by:
-
