Skip to content

K
Kernel

openeuler / Kernel
1 年多 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 06 9月, 2012 3 次提交
  3. 22 8月, 2012 1 次提交
  5. 14 8月, 2012 2 次提交
    • B
      ath5k: fix spin_lock_irqsave/spin_lock_bh nesting in mesh · 7dd6753f
      由 Bob Copeland 提交于 
      Lockdep found an inconsistent lock state when joining a mesh with
ath5k.  The problem is that ath5k takes the lock for its beacon state,
ah->block, with spin_lock_irqsave(), while mesh internally takes the
sync_offset_lock with spin_lock_bh() in mesh_sync_offset_adjust_tbtt(),
which in turn is called under ah->block.

This could deadlock if the beacon tasklet was run on the processor
that held the beacon lock during the do_softirq() in spin_unlock_bh().

We probably shouldn't hold the lock around the callbacks, but the
easiest fix is to switch to spin_lock_bh for ah->block: it doesn't
need interrupts disabled anyway as the data in question is only accessed
in softirq or process context.

Fixes the following lockdep warning:

[  446.892304] WARNING: at kernel/softirq.c:159 _local_bh_enable_ip+0x38/0xa6()
[  446.892306] Hardware name: MacBook1,1
[  446.892309] Modules linked in: tcp_lp fuse sunrpc cpufreq_ondemand acpi_cpufreq mperf ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 nf_conntrack_ipv4 ip6table_filter nf_defrag_ipv4 xt_state nf_conntrack ip6_tables ext2 arc4 btusb bluetooth snd_hda_codec_idt snd_hda_intel carl9170 snd_hda_codec coretemp joydev ath5k snd_hwdep snd_seq isight_firmware ath snd_seq_device snd_pcm applesmc appletouch mac80211 input_polldev snd_timer microcode cfg80211 snd lpc_ich pcspkr i2c_i801 mfd_core soundcore rfkill snd_page_alloc sky2 tpm_infineon virtio_net kvm_intel kvm i915 drm_kms_helper drm i2c_algo_bit i2c_core video
[  446.892385] Pid: 1892, comm: iw Not tainted 3.6.0-rc1-wl+ #296
[  446.892387] Call Trace:
[  446.892394]  [<c0432958>] warn_slowpath_common+0x7c/0x91
[  446.892398]  [<c04399d7>] ? _local_bh_enable_ip+0x38/0xa6
[  446.892403]  [<c04399d7>] ? _local_bh_enable_ip+0x38/0xa6
[  446.892459]  [<f7f9ae3b>] ? mesh_sync_offset_adjust_tbtt+0x95/0x99 [mac80211]
[  446.892464]  [<c043298f>] warn_slowpath_null+0x22/0x24
[  446.892468]  [<c04399d7>] _local_bh_enable_ip+0x38/0xa6
[  446.892473]  [<c0439a52>] local_bh_enable_ip+0xd/0xf
[  446.892479]  [<c088004f>] _raw_spin_unlock_bh+0x34/0x37
[  446.892527]  [<f7f9ae3b>] mesh_sync_offset_adjust_tbtt+0x95/0x99 [mac80211]
[  446.892569]  [<f7f7650f>] ieee80211_beacon_get_tim+0x28f/0x4e0 [mac80211]
[  446.892575]  [<c047ceeb>] ? trace_hardirqs_on_caller+0x10e/0x13f
[  446.892591]  [<f7fdc541>] ath5k_beacon_update+0x40/0x26b [ath5k]
[  446.892597]  [<c047ad67>] ? lock_acquired+0x1f5/0x21e
[  446.892612]  [<f7fdf9fb>] ? ath5k_bss_info_changed+0x167/0x1b2 [ath5k]
[  446.892617]  [<c087f9ea>] ? _raw_spin_lock_irqsave+0x78/0x82
[  446.892632]  [<f7fdf9fb>] ? ath5k_bss_info_changed+0x167/0x1b2 [ath5k]
[  446.892647]  [<f7fdfa09>] ath5k_bss_info_changed+0x175/0x1b2 [ath5k]
[  446.892651]  [<c0479dd4>] ? lock_is_held+0x73/0x7b
[  446.892662]  [<c0458fd5>] ? __might_sleep+0xa7/0x17a
[  446.892698]  [<f7f5d8f7>] ieee80211_bss_info_change_notify+0x1ed/0x21a [mac80211]
[  446.892703]  [<c0449875>] ? queue_work+0x24/0x32
[  446.892718]  [<f7fdf894>] ? ath5k_configure_filter+0x163/0x163 [ath5k]
[  446.892766]  [<f7f95fa4>] ieee80211_start_mesh+0xb9/0xbd [mac80211]
[  446.892806]  [<f7f6e610>] ieee80211_join_mesh+0x10c/0x116 [mac80211]
[  446.892834]  [<f7a96b90>] __cfg80211_join_mesh+0x176/0x1b3 [cfg80211]
[  446.892855]  [<f7a96c1c>] cfg80211_join_mesh+0x4f/0x6a [cfg80211]
[  446.892875]  [<f7a89891>] nl80211_join_mesh+0x1de/0x1ed [cfg80211]
[  446.892908]  [<f7a8db99>] ? nl80211_set_wiphy+0x4cf/0x4cf [cfg80211]
[  446.892919]  [<c07cfa36>] genl_rcv_msg+0x1d5/0x1f3
[  446.892940]  [<c07cf861>] ? genl_rcv+0x25/0x25
[  446.892946]  [<c07cf009>] netlink_rcv_skb+0x37/0x78
[  446.892950]  [<c07cf85a>] genl_rcv+0x1e/0x25
[  446.892955]  [<c07cebf3>] netlink_unicast+0xc3/0x12d
[  446.892959]  [<c07cee46>] netlink_sendmsg+0x1e9/0x213
[  446.892966]  [<c079f282>] sock_sendmsg+0x79/0x96
[  446.892972]  [<c04eb90d>] ? might_fault+0x9d/0xa3
[  446.892978]  [<c07a81d8>] ? copy_from_user+0x8/0xa
[  446.892983]  [<c07a852c>] ? verify_iovec+0x43/0x77
[  446.892987]  [<c079f4d8>] __sys_sendmsg+0x180/0x215
[  446.892993]  [<c045f107>] ? sched_clock_cpu+0x134/0x144
[  446.892997]  [<c047992f>] ? trace_hardirqs_off+0xb/0xd
[  446.893002]  [<c047bf88>] ? __lock_acquire+0x46b/0xb6e
[  446.893006]  [<c047992f>] ? trace_hardirqs_off+0xb/0xd
[  446.893010]  [<c045f149>] ? local_clock+0x32/0x49
[  446.893015]  [<c0479ec1>] ? lock_release_holdtime.part.9+0x4b/0x51
[  446.893020]  [<c0479dd4>] ? lock_is_held+0x73/0x7b
[  446.893025]  [<c050d127>] ? fcheck_files+0x97/0xcd
[  446.893029]  [<c050d4df>] ? fget_light+0x2d/0x81
[  446.893034]  [<c07a01f3>] sys_sendmsg+0x3b/0x52
[  446.893038]  [<c07a07b4>] sys_socketcall+0x238/0x2a2
[  446.893044]  [<c0885edf>] sysenter_do_call+0x12/0x38
[  446.893047] ---[ end trace a9af5998f929270f ]---
[  447.627222]
[  447.627232] =================================
[  447.627237] [ INFO: inconsistent lock state ]
[  447.627244] 3.6.0-rc1-wl+ #296 Tainted: G        W
[  447.627248] ---------------------------------
[  447.627253] inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
[  447.627260] swapper/0/0 [HC0[0]:SC1[1]:HE1:SE0] takes:
[  447.627264]  (&(&ah->block)->rlock){+.?...}, at: [<f7fdd2d1>] ath5k_tasklet_beacon+0x91/0xa7 [ath5k]
[  447.627299] {SOFTIRQ-ON-W} state was registered at:
[  447.627304]   [<c047cdbf>] mark_held_locks+0x59/0x77
[  447.627316]   [<c047ceeb>] trace_hardirqs_on_caller+0x10e/0x13f
[  447.627324]   [<c047cf27>] trace_hardirqs_on+0xb/0xd
[  447.627332]   [<c0439a3d>] _local_bh_enable_ip+0x9e/0xa6
[  447.627342]   [<c0439a52>] local_bh_enable_ip+0xd/0xf
[  447.627349]   [<c088004f>] _raw_spin_unlock_bh+0x34/0x37
[  447.627359]   [<f7f9ae3b>] mesh_sync_offset_adjust_tbtt+0x95/0x99 [mac80211]
[  447.627451]   [<f7f7650f>] ieee80211_beacon_get_tim+0x28f/0x4e0 [mac80211]
[  447.627526]   [<f7fdc541>] ath5k_beacon_update+0x40/0x26b [ath5k]
[  447.627547]   [<f7fdfa09>] ath5k_bss_info_changed+0x175/0x1b2 [ath5k]
[  447.627569]   [<f7f5d8f7>] ieee80211_bss_info_change_notify+0x1ed/0x21a [mac80211]
[  447.627628]   [<f7f95fa4>] ieee80211_start_mesh+0xb9/0xbd [mac80211]
[  447.627712]   [<f7f6e610>] ieee80211_join_mesh+0x10c/0x116 [mac80211]
[  447.627782]   [<f7a96b90>] __cfg80211_join_mesh+0x176/0x1b3 [cfg80211]
[  447.627816]   [<f7a96c1c>] cfg80211_join_mesh+0x4f/0x6a [cfg80211]
[  447.627845]   [<f7a89891>] nl80211_join_mesh+0x1de/0x1ed [cfg80211]
[  447.627872]   [<c07cfa36>] genl_rcv_msg+0x1d5/0x1f3
[  447.627881]   [<c07cf009>] netlink_rcv_skb+0x37/0x78
[  447.627891]   [<c07cf85a>] genl_rcv+0x1e/0x25
[  447.627898]   [<c07cebf3>] netlink_unicast+0xc3/0x12d
[  447.627907]   [<c07cee46>] netlink_sendmsg+0x1e9/0x213
[  447.627915]   [<c079f282>] sock_sendmsg+0x79/0x96
[  447.627926]   [<c079f4d8>] __sys_sendmsg+0x180/0x215
[  447.627934]   [<c07a01f3>] sys_sendmsg+0x3b/0x52
[  447.627941]   [<c07a07b4>] sys_socketcall+0x238/0x2a2
[  447.627949]   [<c0885edf>] sysenter_do_call+0x12/0x38
[  447.627959] irq event stamp: 1929200
[  447.627963] hardirqs last  enabled at (1929200): [<c043a0e9>] tasklet_hi_action+0x3e/0xbf
[  447.627972] hardirqs last disabled at (1929199): [<c043a0c0>] tasklet_hi_action+0x15/0xbf
[  447.627981] softirqs last  enabled at (1929196): [<c043999d>] _local_bh_enable+0x12/0x14
[  447.627989] softirqs last disabled at (1929197): [<c040443b>] do_softirq+0x63/0xb8
[  447.627999]
[  447.627999] other info that might help us debug this:
[  447.628004]  Possible unsafe locking scenario:
[  447.628004]
[  447.628009]        CPU0
[  447.628012]        ----
[  447.628016]   lock(&(&ah->block)->rlock);
[  447.628023]   <Interrupt>
[  447.628027]     lock(&(&ah->block)->rlock);
[  447.628034]
[  447.628034]  *** DEADLOCK ***
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      7dd6753f
    • L
      ath9k: fix decrypt_error initialization in ath_rx_tasklet() · e1352fde
      由 Lorenzo Bianconi 提交于 
      ath_rx_tasklet() calls ath9k_rx_skb_preprocess() and ath9k_rx_skb_postprocess()
in a loop over the received frames. The decrypt_error flag is
initialized to false
just outside ath_rx_tasklet() loop. ath9k_rx_accept(), called by
ath9k_rx_skb_preprocess(),
only sets decrypt_error to true and never to false.
Then ath_rx_tasklet() calls ath9k_rx_skb_postprocess() and passes
decrypt_error to it.
So, after a decryption error, in ath9k_rx_skb_postprocess(), we can
have a leftover value
from another processed frame. In that case, the frame will not be marked with
RX_FLAG_DECRYPTED even if it is decrypted correctly.
When using CCMP encryption this issue can lead to connection stuck
because of CCMP
PN corruption and a waste of CPU time since mac80211 tries to decrypt an already
deciphered frame with ieee80211_aes_ccm_decrypt.
Fix the issue initializing decrypt_error flag at the begging of the
ath_rx_tasklet() loop.
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      e1352fde
  7. 11 8月, 2012 2 次提交
  9. 03 8月, 2012 1 次提交
  11. 18 7月, 2012 27 次提交
  13. 13 7月, 2012 4 次提交
    • M
      ath9k: do not disable hardware while wow is enabled · 4a17a50d
      由 Mohammed Shafi Shajakhan 提交于 
      Hardware needs to be AWAKE and should maintain association
with the AP to process WoW triggers any time

Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: vadivel@qca.qualcomm.com
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      4a17a50d
    • M
      ath9k: Add WoW related mac80211 callbacks · b11e640a
      由 Mohammed Shafi Shajakhan 提交于 
      add suspend/resume/set_wakeup callbacks to the driver

*suspend

- bail out only if all the conditions for configuring WoW.
  is fine, currently multivif case is not handled
- check for associated state.
- map wow triggers from user space data.
- add deauth/disassoc pattern and user defined pattern,
  for the later a list is maintained.
- store the interrupt mask before suspend, enabled beacon
  miss interrupt for WoW.
- configure WoW in the hardware by calling ath9k_hw_wow_enable.

*resume

- restore the interrupts based on the interrupt mask
  stored before suspend.
- call ath9k_hw_wow_wakeup to configure/restore the hardware.
- after wow wakeup clear away WoW events and query the
  WoW wakeup reason from the status register

*set_wakeup

- to call 'device_set_wakeup_enable' from cfg80211/mac80211
  when wow is configured and as per Rafael/Johannnes the
  right way to do so rather in the driver suspend/resume
  call back

Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: vadivel@qca.qualcomm.com
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      b11e640a
    • M
      ath: Add Wake-on-Wireless debug mask · b3ba6c52
      由 Mohammed Shafi Shajakhan 提交于 
      to help the developers and users to debug/know
whats happening with WoW

Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: vadivel@qca.qualcomm.com
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      b3ba6c52
    • M
      ath9k_hw: Add hardware code for WoW · 64875c63
      由 Mohammed Shafi Shajakhan 提交于 
      add a new file wow.c which takes care of the hardware code
for WoW.

*program the descriptors and data words to periodically
send Keep Alive Frames.
*program the user defined patterns/masks and pattern length
in the hardware registers.
*'ath9k_hw_wow_enable' is called during the drivers suspend
callback which takes care of the following
	- tracking wow event mask (to suppress spurious
	  wow events)
	- properly configure suspend/resume WAR registers
	- configure PCIE PM control register
	- configure MAC WoW registers and their timeouts
	- enabling wow configuration like magic packet,
	  user patterns based on users configuration
	- configuring timeouts for KAL, beacon miss,
	  aifs, slot time, backoff
	- create Keep Alive Pattern ('KAL')
*'ath9k_hw_wow_wakeup' is called during the drivers resume
callback which takes care of the following
	- primary task is to find the reason for wakeup
	  from the wow status register
	- configure/restore AR_PCIE_PM_CTRL register
	- clear all WoW events
	- configure/restore suspend/resume WAR registers

Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: vadivel@qca.qualcomm.com
Signed-off-by: NLuis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
      64875c63