[PATCH 3.0+] ath9k: Fix softlockup in AR9485(backported)

[PATCH 3.0+] ath9k: Fix softlockup in AR9485(backported)

[Mohammed Shafi Shajakhan]

From: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>

Please note this is the backported version for the linux
stable tree, while the patch for wireless-testing tree
http://permalink.gmane.org/gmane.linux.kernel.wireless.general/92608

steps to recreate:
load latest ath9k driver with AR9485
stop the network-manager and wpa_supplicant
bring the interface up

	Call Trace:
	[<ffffffffa0517490>] ? ath_hw_check+0xe0/0xe0 [ath9k]
	[<ffffffff812cd1e8>] __const_udelay+0x28/0x30
	[<ffffffffa03bae7a>] ar9003_get_pll_sqsum_dvc+0x4a/0x80 [ath9k_hw]
	[<ffffffffa05174eb>] ath_hw_pll_work+0x5b/0xe0 [ath9k]
	[<ffffffff810744fe>] process_one_work+0x11e/0x470
	[<ffffffff8107530f>] worker_thread+0x15f/0x360
	[<ffffffff810751b0>] ? manage_workers+0x230/0x230
	[<ffffffff81079af3>] kthread+0x93/0xa0
	[<ffffffff815fd3a4>] kernel_thread_helper+0x4/0x10
	[<ffffffff81079a60>] ? kthread_freezable_should_stop+0x70/0x70
	[<ffffffff815fd3a0>] ? gs_change+0x13/0x13

ensure that the PLL-WAR for AR9485/AR9340 is executed only if the STA is
associated (or) IBSS/AP mode had started beaconing. Ideally this WAR
is needed to recover from some rare beacon stuck during stress testing.
Before the STA is associated/IBSS had started beaconing, PLL4(0x1618c)
always seem to have zero even though we had configured PLL3(0x16188) to
query about PLL's locking status. When we keep on polling infinitely PLL4's
8th bit(ie check for PLL locking measurements is done), machine hangs
due to softlockup.

fixes https://bugzilla.redhat.com/show_bug.cgi?id=811142

Reported-by: Rolf Offermanns <rolf.offermanns@gmx.net>
Cc: stable@vger.kernel.org [3.0+]
Tested-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/main.c |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/main.c b/drivers/net/wireless/ath/ath9k/main.c
index 4de4473..41096f6 100644
--- a/drivers/net/wireless/ath/ath9k/main.c
+++ b/drivers/net/wireless/ath/ath9k/main.c
@@ -971,6 +971,15 @@ void ath_hw_pll_work(struct work_struct *work)
 					    hw_pll_work.work);
 	u32 pll_sqsum;
 
+	/*
+	 * ensure that the PLL WAR is executed only
+	 * after the STA is associated (or) if the
+	 * beaconing had started in interfaces that
+	 * uses beacons.
+	 */
+	if (!(sc->sc_flags & SC_OP_BEACONS))
+		return;
+
 	if (AR_SREV_9485(sc->sc_ah)) {
 
 		ath9k_ps_wakeup(sc);
-- 
1.7.4.1

[PATCH 3.0+] ath9k: Fix softlockup in AR9485(backported)

[Sujith Manoharan]

Mohammed Shafi Shajakhan wrote:
> From: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
> 
> Please note this is the backported version for the linux
> stable tree, while the patch for wireless-testing tree
> http://permalink.gmane.org/gmane.linux.kernel.wireless.general/92608
> 
> steps to recreate:
> load latest ath9k driver with AR9485
> stop the network-manager and wpa_supplicant
> bring the interface up
> 
> 	Call Trace:
> 	[<ffffffffa0517490>] ? ath_hw_check+0xe0/0xe0 [ath9k]
> 	[<ffffffff812cd1e8>] __const_udelay+0x28/0x30
> 	[<ffffffffa03bae7a>] ar9003_get_pll_sqsum_dvc+0x4a/0x80 [ath9k_hw]
> 	[<ffffffffa05174eb>] ath_hw_pll_work+0x5b/0xe0 [ath9k]
> 	[<ffffffff810744fe>] process_one_work+0x11e/0x470
> 	[<ffffffff8107530f>] worker_thread+0x15f/0x360
> 	[<ffffffff810751b0>] ? manage_workers+0x230/0x230
> 	[<ffffffff81079af3>] kthread+0x93/0xa0
> 	[<ffffffff815fd3a4>] kernel_thread_helper+0x4/0x10
> 	[<ffffffff81079a60>] ? kthread_freezable_should_stop+0x70/0x70
> 	[<ffffffff815fd3a0>] ? gs_change+0x13/0x13
> 
> ensure that the PLL-WAR for AR9485/AR9340 is executed only if the STA is
> associated (or) IBSS/AP mode had started beaconing. Ideally this WAR
> is needed to recover from some rare beacon stuck during stress testing.
> Before the STA is associated/IBSS had started beaconing, PLL4(0x1618c)
> always seem to have zero even though we had configured PLL3(0x16188) to
> query about PLL's locking status. When we keep on polling infinitely PLL4's
> 8th bit(ie check for PLL locking measurements is done), machine hangs
> due to softlockup.

While I do agree that this patch fixes the regression for AR9485 chipsets,
the code in ar9003_get_pll_sqsum_dvc needs to be fixed, i.e., to not loop
till the end of time and have a timeout.

Sujith

Re: [PATCH 3.0+] ath9k: Fix softlockup in AR9485(backported)

[Mohammed Shafi Shajakhan]

Hi Sujith,

On Wednesday 13 June 2012 09:36 PM, Sujith Manoharan wrote:
> Mohammed Shafi Shajakhan wrote:
>> From: Mohammed Shafi Shajakhan<mohammed@qca.qualcomm.com>
>>
>> Please note this is the backported version for the linux
>> stable tree, while the patch for wireless-testing tree
>> http://permalink.gmane.org/gmane.linux.kernel.wireless.general/92608
>>
>> steps to recreate:
>> load latest ath9k driver with AR9485
>> stop the network-manager and wpa_supplicant
>> bring the interface up
>>
>> 	Call Trace:
>> 	[<ffffffffa0517490>] ? ath_hw_check+0xe0/0xe0 [ath9k]
>> 	[<ffffffff812cd1e8>] __const_udelay+0x28/0x30
>> 	[<ffffffffa03bae7a>] ar9003_get_pll_sqsum_dvc+0x4a/0x80 [ath9k_hw]
>> 	[<ffffffffa05174eb>] ath_hw_pll_work+0x5b/0xe0 [ath9k]
>> 	[<ffffffff810744fe>] process_one_work+0x11e/0x470
>> 	[<ffffffff8107530f>] worker_thread+0x15f/0x360
>> 	[<ffffffff810751b0>] ? manage_workers+0x230/0x230
>> 	[<ffffffff81079af3>] kthread+0x93/0xa0
>> 	[<ffffffff815fd3a4>] kernel_thread_helper+0x4/0x10
>> 	[<ffffffff81079a60>] ? kthread_freezable_should_stop+0x70/0x70
>> 	[<ffffffff815fd3a0>] ? gs_change+0x13/0x13
>>
>> ensure that the PLL-WAR for AR9485/AR9340 is executed only if the STA is
>> associated (or) IBSS/AP mode had started beaconing. Ideally this WAR
>> is needed to recover from some rare beacon stuck during stress testing.
>> Before the STA is associated/IBSS had started beaconing, PLL4(0x1618c)
>> always seem to have zero even though we had configured PLL3(0x16188) to
>> query about PLL's locking status. When we keep on polling infinitely PLL4's
>> 8th bit(ie check for PLL locking measurements is done), machine hangs
>> due to softlockup.
>
> While I do agree that this patch fixes the regression for AR9485 chipsets,
> the code in ar9003_get_pll_sqsum_dvc needs to be fixed, i.e., to not loop
> till the end of time and have a timeout.
>

agreed, i already have a patch for that and did some tests too :)

diff --git a/drivers/net/wireless/ath/ath9k/hw.c 
b/drivers/net/wireless/ath/ath9k/hw.c
index 6fa8128..932199c 100644
--- a/drivers/net/wireless/ath/ath9k/hw.c
+++ b/drivers/net/wireless/ath/ath9k/hw.c
@@ -719,13 +719,26 @@ static void ath9k_hw_init_qos(struct ath_hw *ah)

  u32 ar9003_get_pll_sqsum_dvc(struct ath_hw *ah)
  {
+       int i = 0;
+       struct ath_common *common = ath9k_hw_common(ah);
+
         REG_CLR_BIT(ah, PLL3, PLL3_DO_MEAS_MASK);
         udelay(100);
         REG_SET_BIT(ah, PLL3, PLL3_DO_MEAS_MASK);

-       while ((REG_READ(ah, PLL4) & PLL4_MEAS_DONE) == 0)
+       while ((REG_READ(ah, PLL4) & PLL4_MEAS_DONE) == 0) {
+
+               i++;
                 udelay(100);

+               if (i > 100) {
+                       ath_err(common, "PLL4 meaurement not done");
+                       WARN_ON(1);
+                       break;
+               }
+
+       }
+
         return (REG_READ(ah, PLL3) & SQSUM_DVC_MASK) >> 3;
  }
  EXPORT_SYMBOL(ar9003_get_pll_sqsum_dvc);


during my testing i found that, when things are fine we would usually 
break out when i = 2. will send this patch too. thanks for your thoughts!


-- 
thanks,
shafi