[Bug 221073] xHCI host controller dies on resume from s2idle on AMD Strix Halo [1022:1587]

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=221073

--- Comment #53 from Alexander F (superveridical@gmail.com) ---
Created attachment 310441
  --> https://bugzilla.kernel.org/attachment.cgi?id=310441&action=edit
Power dependency patch attempt klog

For reasons that are unknown to me, the flags variable here

2572    if (flags & AMD_IS_APU)
2573            amdgpu_setup_apu_usb_power_dep(adev);

doesn't have the appropriate bits set, so the code doesn't get executed at all.

To discover this I instrumented the code with the two pirintks, one just before
the if with the dump of the flags, one inside the function before the while
loop.

2379         printk(KERN_INFO "OLOLOLOLOLOLOL 2\n");

2561         unsigned long testt = AMD_IS_APU;
2562
2563         printk(KERN_INFO "LALALALALALALA 1 flags:%*pbl AMD_IS_APU: %*pbl
\n", 64, &flags, 64, &testt);

The flags are indeed different:

Jul 13 08:29:36 rescue-flow kernel: LALALALALALALA 1 flags:2,5 AMD_IS_APU: 17 

Then I commented the offending if out, the second printk appeared in the logs,
but nothing from the loop got printed.

Just in case I supplied klog with debug turned on.

I reverted the previous patch btw. I'm pretty sure AMD is my APU, and 3d
acceleration works nominally(checked with firefox/shadertoy and amdgpu_top). If
it's some misconfiguration on my part(for example my kernel config), please
provide further instructions.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.