[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 #30 from Michał Pecio (michal.pecio@gmail.com) ---
(In reply to Alexander F from comment #19)
> The xhci_hcd.quirks=0x40 quirk resolves the 10 second resume issue, the
> devices on this bus also don't disappear and work nominally. Not sure if
> there are side effects to this quirk.
This should be harmless in principle, it just switches from MSI to (emulated)
PCI INTx interrupt.

> - The device could accrue "brokenness" that is preserved between "soft"
> resets. This is also rare. For example one time it was systematically doing
> very long boots due to getting stuck on a btusb device, and it was cleared
> only by hard reset.
Unrelated. Onboard bluetooth chips are known to get flakey and fail to
enumerate until power cycled, which is tricky if they run off a standby rail.

(In reply to David Hubbard from comment #23)
> I'm fascinated with diagnosing the xHC USBSTS with the EINT bit set, using
> the instructions in comment #5. I'm zeroing in on that. Maybe the issues
> with resume taking 10 seconds in comment #19 can be considered separately in
> a different bug?
The delay is a consequence of missing interrupts *and* having some USB device
connected. Your case is similar to the other one.

Sorry about the wrong instructions. Indeed, tar doesn't work with debugfs.

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

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