[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 #37 from Michał Pecio (michal.pecio@gmail.com) ---
Can you try this magic trick which solves all recent problems with AMD xHCI?

echo on > /sys/bus/pci/devices/0000:c4:00.4/power/control

You log shows the PCI device being runtime autosuspended and resumed
immediately before suspending again for s2idle. The above disables autosuspend.

Other than that, nothing really stands out. Suspend clears the RUN bit and INTE
bit (my addition), on resume registers are at their defaults and get restored
to pre-suspend state, INTE and RUN are enabled again. A timeout they are still
the same, except for sts/erdp flags indicating that interrupt is pending.

(In reply to Alexander F from comment #35)
> So I returned to vanilla-sources-6.19.6 manually built kernel and 20260110
> firmware, patched the kernel with the register dump patch. Interestingly it
> was a little harder to trigger than before, but I didn't do enough runs to
> say definitively. The files are klog-pecio-patch and klog-pecio-patch2.
Maybe it's a race condition and logging slows things down. Bottom line, it's
not a matter of xHCI interrupt control bits being cleared.

> >There are some error flags set on DevSta 
> 
> These flags only appear after the "HC died" occurs. (That event also adds
> (warning) taint.)
The warning is unimportant, it's just USB resume taking unexpectedly long. We
know it already :)

I think all we got is just more evidence that it's a PCI or x86 architecture
problem, not USB. I would mail linux-pci, or at least reassign the bug to
drivers/PCI (but not sure if this subsystem pays attention to bugzilla).

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

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