Tegra PCIe system hangs

Tegra PCIe system hangs

[Lucas Stach]

Hi all,

I've been experimenting with the Nouveau driver on Tegra for a few days
now and made some promising progress. The one thing that keeps throwing
me back is that I get seemingly random system lockups, which are not
recoverable.

I think this could be the same issue that was reported by Jay Agarwal
earlier. (Message-id:
<C79B248886DD134989C8FF6B096A91AB91B603D82B@BGMAIL01.nvidia.com>)
I also found a WAR for a system hang issue in the NVidia downstream
kernel [1], but when tried on upstream it had no effect on the issue.

I'm reasonably sure that this is not driver software related, as I've
already hunted down most of the caching issues and acceleration seems to
work at this point. So either this is a random HW hang, or there is some
software condition I'm not aware of.

I'm running a NV92 connected to the NVidia Beaver board. Kernel is
next-20130809 with the PCIe series + the LP2 workaround applied.

Could you please investigate if there are any other known errata for the
Tegra PCIe controller, which could hang the entire system?

Thanks,
Lucas

[1]
http://nv-tegra.nvidia.com/gitweb/?p=linux-2.6.git;a=blobdiff;f=arch/arm/mach-tegra/pcie.c;h=9fab5de016cfd3cd2ebd8176da5bdbedfac73cd8;hp=44749685e2bf822665c9345aec45151b8d2447c4;hb=c5d1bcfb7249b291453a80122148155a00e5456a;hpb=704125ef90e12bc41de057001cc96d81d7cc0376

Re: Tegra PCIe system hangs

[Lucas Stach]

Ok, turn out this seems to be an issue with legacy PCI interrupts.
Nouveau wasn't really using MSI interrupts, switching them on seems to
fix the stability issue.

Still I would be interested in the root cause and would like to see this
documented somewhere.

Thanks,
Lucas

Am Sonntag, den 25.08.2013, 19:14 +0200 schrieb Lucas Stach:
> Hi all,
> 
> I've been experimenting with the Nouveau driver on Tegra for a few days
> now and made some promising progress. The one thing that keeps throwing
> me back is that I get seemingly random system lockups, which are not
> recoverable.
> 
> I think this could be the same issue that was reported by Jay Agarwal
> earlier. (Message-id:
> <C79B248886DD134989C8FF6B096A91AB91B603D82B@BGMAIL01.nvidia.com>)
> I also found a WAR for a system hang issue in the NVidia downstream
> kernel [1], but when tried on upstream it had no effect on the issue.
> 
> I'm reasonably sure that this is not driver software related, as I've
> already hunted down most of the caching issues and acceleration seems to
> work at this point. So either this is a random HW hang, or there is some
> software condition I'm not aware of.
> 
> I'm running a NV92 connected to the NVidia Beaver board. Kernel is
> next-20130809 with the PCIe series + the LP2 workaround applied.
> 
> Could you please investigate if there are any other known errata for the
> Tegra PCIe controller, which could hang the entire system?
> 
> Thanks,
> Lucas
> 
> [1]
> http://nv-tegra.nvidia.com/gitweb/?p=linux-2.6.git;a=blobdiff;f=arch/arm/mach-tegra/pcie.c;h=9fab5de016cfd3cd2ebd8176da5bdbedfac73cd8;hp=44749685e2bf822665c9345aec45151b8d2447c4;hb=c5d1bcfb7249b291453a80122148155a00e5456a;hpb=704125ef90e12bc41de057001cc96d81d7cc0376
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Tegra PCIe system hangs

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 2404 bytes --]

On Mon, Aug 26, 2013 at 01:50:34AM +0200, Lucas Stach wrote:
> Ok, turn out this seems to be an issue with legacy PCI interrupts.
> Nouveau wasn't really using MSI interrupts, switching them on seems to
> fix the stability issue.
> 
> Still I would be interested in the root cause and would like to see this
> documented somewhere.

Hi Lucas,

Thanks for reporting. It's quite possible that there is a general issue
with legacy interrupts since I haven't tested that much. I'll see if I
can reproduce the lockups when using legacy interrupts on any of my
setups.

Thanks,
Thierry

> 
> Thanks,
> Lucas
> 
> Am Sonntag, den 25.08.2013, 19:14 +0200 schrieb Lucas Stach:
> > Hi all,
> > 
> > I've been experimenting with the Nouveau driver on Tegra for a few days
> > now and made some promising progress. The one thing that keeps throwing
> > me back is that I get seemingly random system lockups, which are not
> > recoverable.
> > 
> > I think this could be the same issue that was reported by Jay Agarwal
> > earlier. (Message-id:
> > <C79B248886DD134989C8FF6B096A91AB91B603D82B@BGMAIL01.nvidia.com>)
> > I also found a WAR for a system hang issue in the NVidia downstream
> > kernel [1], but when tried on upstream it had no effect on the issue.
> > 
> > I'm reasonably sure that this is not driver software related, as I've
> > already hunted down most of the caching issues and acceleration seems to
> > work at this point. So either this is a random HW hang, or there is some
> > software condition I'm not aware of.
> > 
> > I'm running a NV92 connected to the NVidia Beaver board. Kernel is
> > next-20130809 with the PCIe series + the LP2 workaround applied.
> > 
> > Could you please investigate if there are any other known errata for the
> > Tegra PCIe controller, which could hang the entire system?
> > 
> > Thanks,
> > Lucas
> > 
> > [1]
> > http://nv-tegra.nvidia.com/gitweb/?p=linux-2.6.git;a=blobdiff;f=arch/arm/mach-tegra/pcie.c;h=9fab5de016cfd3cd2ebd8176da5bdbedfac73cd8;hp=44749685e2bf822665c9345aec45151b8d2447c4;hb=c5d1bcfb7249b291453a80122148155a00e5456a;hpb=704125ef90e12bc41de057001cc96d81d7cc0376
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]