Possibility of RX570 responsible for spontaneous reboots (MCE) with Ryzen 3700x?

[Bridgman, John]

[AMD Official Use Only - Internal Distribution Only]

>I know RX570 (polaris) should stay at PCI3 as far as I know.

Yep... thought I remembered you mentioning having a 5700XT though... is that in a different system ?

________________________________
From: Clemens Eisserer <linuxhippy at gmail.com>
Sent: March 9, 2020 2:30 AM
To: Bridgman, John <John.Bridgman at amd.com>; amd-gfx at lists.freedesktop.org <amd-gfx at lists.freedesktop.org>
Subject: Re: Possibility of RX570 responsible for spontaneous reboots (MCE) with Ryzen 3700x?

Hi John,

Thanks a lot for taking the time to look at this, even if it doesn't
seem to be GPU related at first.

> OK, that's a bit strange... I found mce log and MCE-Ryzen-Decoder as options for decoding.
Sorry for omitting that information - indeed I was using
MCE-Ryzen-Decoder, thanks for pointing to mcelog.
The mce log output definitivly makes more sense, I'll try to
experiment a bit with RAM.

Thanks also for the link to the forum, seems of all the affected users,
no one reported success in that thread.

> For something as simple as the GPU bus interface not responding to an access
> by the CPU I think you would get a different error (bus error) but not 100% sure about that.
>
> My first thought would be to see if your mobo BIOS has an option to force PCIE
> gen3 instead of 4 and see if that makes a difference. There are some amdgpu module parms
> related to PCIE as well but I'm not sure which ones to recommend.

I'll give it a try and have a look at the pcie options - but as far as
I know RX570 (polaris) should stay at PCI3 as far as I know.
Disabling IOMMU didn't help as far as I recall.

Thanks & best regards, Clemens
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20200309/075ea9f3/attachment.htm>