On 7/31/25 10:37 AM, Linus Torvalds wrote:
On Wed, 30 Jul 2025 at 21:58, Linus Torvalds
<torva...@linux-foundation.org> wrote:
d7b618bc41ee3d44c070212dff93949702ede997 is the first bad commit
drm/amd/display: Refactor DSC cap calculations
Let me go see how painful it is to just revert it from top-of-tree.
So with that reverted (didn't require a lot of fixing, only minor
unrelated context added nearby later), current top-of-tree works for
me again.
The revert I used for testing attached just so people can see exactly
what I did.
It's late here, I wasn't getting any more work done today anyway, so
I'll leave it like this for now.
I can continue to do the merge window with this revert purely local
for a while, so if somebody comes up with a proper fix, I can test
that out.
Linus
Pure guess from looking at the diff from
d7b618bc41ee3d44c070212dff93949702ede997, maybe the NULL ptr is a deref
on a pointer that doesn't have funcs declared being an older GPU. So
maybe check for funcs?
diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
index a454d16e6586..95d51de26c0b 100644
--- a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
+++ b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
@@ -673,7 +673,7 @@ static void get_dsc_enc_caps(
return;
/* check if reported cap global or only for a single DCN DSC enc */
- if (dsc->funcs->dsc_get_enc_caps) {
+ if (dsc->funcs && dsc->funcs->dsc_get_enc_caps) {
if (!dsc->ctx->dc->debug.disable_dsc)
dsc->funcs->dsc_get_enc_caps(dsc_enc_caps,
pixel_clock_100Hz);
} else {
