[Bug 220069] [6.13.9] regression USB controller dies

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=220069

--- Comment #54 from Michał Pecio (michal.pecio@gmail.com) ---
avg_trb_len doesn't seem to be an issue for my ASMedia/Promontory HCs, they
still have their usual bugs after changing it to 8 as per spec.

By the way, the simplest way to break ASMedia chips is running
  while true ; do ifconfig ethN up ; ifconfig ethN down ; done
on a SuperSpeed NIC (AX88179 and RTL8153 known to work). This is what the limit
endpoint interval quirk appears to be helping with, though recently I found
some case (NIC behind a SuperSpeed hub) where it didn't help.

I am also not convinced that the "clearing xHC resources" patch is relevant,
because it's hard to explain why nothing bad happens when you run reset
manually. I found no problems resetting devices on my AMD B350, ASM3142 and
ASM1142, despite the dubious command sequence always being there. Your logs
show malfunction before we even get to this point. Indeed, the reason the hub
is being reset is because attempting to resume it normally suddenly failed.


So the outcome of testing is that the problem occurs if both the unlucky hub
and the unlucky webcam are present on the same X670 bus, and disappears if they
are on different X670 buses at the same time?

And it's happening on 6.13.9 or newer, but not on 6.13.7? I have no explanation
for that...

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

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