I've been banging on my test setup for the past week. I haven't
tracked down the cause yet, but something in my prior build system
results in this problem being masked with some builds. Doing pure
vanilla builds from git I'm able to reproduce with 3.6 and the 3.7-rc
series reliably. Support for
Given 3.7 has shipped, should I rebase on release and put a new RFC
post up, rebase on linux-next, wait for a 3.8-rc or just post it up as
v2 for submission this time as is?
Josh C
On Mon, Nov 26, 2012 at 4:38 PM, Josh Coombs josh.coo...@gmail.com wrote:
I've been banging on my test setup
Starting with 3.6.5 on a Marvell Kirkwood based GoFlex Net I began
observing scheduler bugs when using a USB based RTL8712 WiFi NIC.
These would eventually overwhelm systemd's logger under moderate
network activity and crash the box.
[ 64.312377] BUG: scheduling while atomic:
On Sun, Nov 18, 2012 at 12:47 PM, Russell King - ARM Linux
li...@arm.linux.org.uk wrote:
On Sun, Nov 18, 2012 at 12:06:40PM -0500, Josh Coombs wrote:
Starting with 3.6.5 on a Marvell Kirkwood based GoFlex Net I began
observing scheduler bugs when using a USB based RTL8712 WiFi NIC.
These would
Testing now, will re-submit in the correct format once I confirm it
works as intended.
On Sun, Nov 18, 2012 at 1:20 PM, Andrew Lunn and...@lunn.ch wrote:
diff -ruN a/drivers/staging/rtl8712/rtl871x_sta_mgt.c
b/drivers/staging/rtl8712/rtl871x_sta_mgt.c
---
With the _malloc() adjusted as suggested, I'm still not seeing
scheduler bugs tracing back to the rtl8172 driver. I am however
seeing allocation errors occasionally now in cryptodev on my test
system, so all is still not right. This behavior was not present in
3.6.4, I haven't seen it in the 3.7
I only have access to Marvell's SATA controller for testing, so that
is why I only targeted it. The Kconfig depends change makes perfect
sense.
Changing the name of the trigger might cause hardship for those
already using it, as they will have to update scripts/etc to account
for the name
To: Josh Coombs
Cc: coolo...@gmail.com; linux-kernel@vger.kernel.org; linux-
i...@vger.kernel.org; rpur...@rpsys.net; linux ARM; jgar...@pobox.com;
linux-l...@vger.kernel.org
Subject: Re: [PATCH] Allow Marvell SATA driver to work with
LEDS_TRIGGER_IDE_DISK
On Tue, Jan 08, 2013 at 02:18:04PM -0500
Digging back through the archives, this came up once in 2007, Alan Cox
did not like the idea of globally adding this call into the critical
paths of libata at the time. (I've seen it done via external patches
in ata_qc_issue and ata_qc_complete for example.) If that concern is
no longer in place,
I only have access to Marvell's SATA controller for testing, so that
is why I only targeted it. The Kconfig depends change makes perfect
sense.
Changing the name of the trigger might cause hardship for those
already using it, as they will have to update scripts/etc to account
for the name
DS_TRIGGER_IDE_DISK
>>
>> On Tue, Jan 08, 2013 at 02:18:04PM -0500, Josh Coombs wrote:
>> > Would it make more sense to add a second trigger
>> > for SATA instead?
>>
>> I'll leave that up to the led guys, I just wanted to raise the point
>> that the
Digging back through the archives, this came up once in 2007, Alan Cox
did not like the idea of globally adding this call into the critical
paths of libata at the time. (I've seen it done via external patches
in ata_qc_issue and ata_qc_complete for example.) If that concern is
no longer in place,
I've been banging on my test setup for the past week. I haven't
tracked down the cause yet, but something in my prior build system
results in this problem being masked with some builds. Doing pure
vanilla builds from git I'm able to reproduce with 3.6 and the 3.7-rc
series reliably. Support for
Given 3.7 has shipped, should I rebase on release and put a new RFC
post up, rebase on linux-next, wait for a 3.8-rc or just post it up as
v2 for submission this time as is?
Josh C
On Mon, Nov 26, 2012 at 4:38 PM, Josh Coombs wrote:
> I've been banging on my test setup for the past week.
Starting with 3.6.5 on a Marvell Kirkwood based GoFlex Net I began
observing scheduler bugs when using a USB based RTL8712 WiFi NIC.
These would eventually overwhelm systemd's logger under moderate
network activity and crash the box.
[ 64.312377] BUG: scheduling while atomic:
On Sun, Nov 18, 2012 at 12:47 PM, Russell King - ARM Linux
wrote:
> On Sun, Nov 18, 2012 at 12:06:40PM -0500, Josh Coombs wrote:
>> Starting with 3.6.5 on a Marvell Kirkwood based GoFlex Net I began
>> observing scheduler bugs when using a USB based RTL8712 WiFi NIC.
>> T
Testing now, will re-submit in the correct format once I confirm it
works as intended.
On Sun, Nov 18, 2012 at 1:20 PM, Andrew Lunn wrote:
>> >> diff -ruN a/drivers/staging/rtl8712/rtl871x_sta_mgt.c
>> >> b/drivers/staging/rtl8712/rtl871x_sta_mgt.c
>> >> ---
With the _malloc() adjusted as suggested, I'm still not seeing
scheduler bugs tracing back to the rtl8172 driver. I am however
seeing allocation errors occasionally now in cryptodev on my test
system, so all is still not right. This behavior was not present in
3.6.4, I haven't seen it in the 3.7
18 matches
Mail list logo
