[AMD Public Use]

-----Original Message-----
From: Andre Tomt <redacted>
Sent: Tuesday, January 5, 2021 11:32 AM
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-kernel@vger.kernel.org; stable@vger.kernel.org; Wentland, Harry
[off-list ref]; Deucher, Alexander
[off-list ref]; Kazlauskas, Nicholas
[off-list ref]; Wang, Chao-kai (Stylon)
[off-list ref]
Subject: Re: [PATCH 5.10 637/717] drm/amd/display: Fix memory leaks in S3
resume

On 05.01.2021 07:54, Greg Kroah-Hartman wrote:

quoted

On Mon, Jan 04, 2021 at 08:04:08PM +0100, Andre Tomt wrote:

quoted

On 28.12.2020 13:50, Greg Kroah-Hartman wrote:

quoted

From: Stylon Wang <redacted>

commit a135a1b4c4db1f3b8cbed9676a40ede39feb3362 upstream.

EDID parsing in S3 resume pushes new display modes to probed_modes
list but doesn't consolidate to actual mode list. This creates a
race condition when
amdgpu_dm_connector_ddc_get_modes() re-initializes the list head
without walking the list and results in  memory leak.

This commit is causing me problems on 5.10.4: when I turn off the
display (a LG TV in this case), and turn it back on again later there
is no video output and I get the following in the kernel log:

[ 8245.259628] [drm:dm_restore_drm_connector_state [amdgpu]]

*ERROR*

quoted

quoted

Restoring old state failed with -12

I've found another report on this commit as well:

https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbug

quoted

quoted

zilla.kernel.org%2Fshow_bug.cgi%3Fid%3D211033&data=04%7C01%7Cal
ex

quoted

quoted

ander.deucher%40amd.com%7Cad673e351bab4f6af94508d8b1977ed8%7C3d
d8961f

quoted

quoted

e4884e608e11a82d994e183d%7C0%7C0%7C637454612140560971%7CUnknow
n%7CTWF

quoted

quoted

pbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXV
CI6

quoted

quoted

Mn0%3D%7C3000&sdata=8Rsnbfh4P5GmFUlybb31mT7C0Ee4vDInxJ1gt
C3jrVI%3

quoted

quoted

D&reserved=0

And I suspect this is the same:

https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbug

quoted

quoted

s.archlinux.org%2Ftask%2F69202&data=04%7C01%7Calexander.deuche
r%4

quoted

quoted

0amd.com%7Cad673e351bab4f6af94508d8b1977ed8%7C3dd8961fe4884e608
e11a82

quoted

quoted

d994e183d%7C0%7C0%7C637454612140560971%7CUnknown%7CTWFpbGZsb
3d8eyJWIj

quoted

quoted

oiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3
000&a

quoted

quoted

mp;sdata=1149tCcm3rcaj1MkVPbWqWhWFIPgkeBoYxo0oVv%2FzNI%3D&a
mp;reserve

quoted

quoted

d=0

Reverting it from 5.10.4 makes things behave again.

Have not tested 5.4.86 or 5.11-rc.

I'm using a RX570 Polaris based card.

Can you test 5.11-rc to see if this issue is there as well?

Just did, and have the same issue on 5.11-rc2. Reverting it also solves the
problem on 5.11-rc2, as it does on 5.10.4

FWIW one easy way to reproduce seems to be unplugging and re-plugging
the HDMI.

We are looking into the root cause, but I'll send out the revert for now.

Thanks,

Alex