2020-07-27 11:27:34

by Viktor Jägersküpper

Subject: Please apply "ath9k: Fix general protection fault in ath9k_hif_usb_rx_cb" and the corresponding fix to all stable kernels

Hi Greg, Sasha and all the others,

Hans de Goede requested to revert "ath9k: Fix general protection fault in ath9k_hif_usb_rx_cb"
from all stable and longterm kernels because this commit broke certain devices
with Atheros 9271 and at that time it seemed that reverting the commit would be
done in the mainline kernel. The revert was done in kernel 5.7.9 etc., however
Mark O'Donovan found a fix for the original commit - which avoided the revert in
the mainline kernel - and this fix is now included in 5.8-rc7 with commit

92f53e2fda8bb9a559ad61d57bfb397ce67ed0ab ("ath9k: Fix regression with Atheros 9271").

To be consistent with the mainline kernel, please apply the original commit
again (or re-revert it, whatever is appropriate for stable kernels) and then
apply Mark's fix. I have tested this with the current 5.7.10 kernel to confirm
that it works because I was affected by the bug.

All relevant people are CC'ed if someone wants to object.

Thanks,
Viktor

2020-07-27 13:52:25

by Greg Kroah-Hartman

Subject: Re: Please apply "ath9k: Fix general protection fault in ath9k_hif_usb_rx_cb" and the corresponding fix to all stable kernels

On Mon, Jul 27, 2020 at 01:20:11PM +0200, Viktor J?gersk?pper wrote:
> Hi Greg, Sasha and all the others,
>
> Hans de Goede requested to revert "ath9k: Fix general protection fault in ath9k_hif_usb_rx_cb"
> from all stable and longterm kernels because this commit broke certain devices
> with Atheros 9271 and at that time it seemed that reverting the commit would be
> done in the mainline kernel. The revert was done in kernel 5.7.9 etc., however
> Mark O'Donovan found a fix for the original commit - which avoided the revert in
> the mainline kernel - and this fix is now included in 5.8-rc7 with commit
>
> 92f53e2fda8bb9a559ad61d57bfb397ce67ed0ab ("ath9k: Fix regression with Atheros 9271").
>
> To be consistent with the mainline kernel, please apply the original commit
> again (or re-revert it, whatever is appropriate for stable kernels) and then
> apply Mark's fix. I have tested this with the current 5.7.10 kernel to confirm
> that it works because I was affected by the bug.
>
> All relevant people are CC'ed if someone wants to object.

I've now queued both up, thanks!

greg k-h