Re: [PATCH v4 1/2] ath10k: add refcount for ath10k_core_restart

[Justin Capella <justincapella@gmail.com>]

Sorry, that response was intended for the start retries.

On Mon, Jan 20, 2020 at 5:34 AM Wen Gong <wgong@codeaurora.org> wrote:
>
> On 2020-01-20 17:38, Justin Capella wrote:
> > Ok thanks. I am still trying to familiarize myself with ath10k. A
> > couple things come to mind:
> >
> > firstly recently there was a patch that set stations back to
> > associated during recovery, do you know if the branch you're working
> > against includes that? I will try to figure that out myself but it's
> > at least worth considering.
> >
> can you give the patch link or patch?
>
> > If I remember correctly, the sdio rx involves peeking and checking to
> > see if the frame/packet continues/has more. In the case of firmware
> > recovery/reconfig I think current code enables all sorts of debug
> > pktlog stuff, historically it was maybe used as a hack to check for hw
> > changes... But I mention it because I think what might be happening in
> > some cases is fw crash follows a frame that would continue/extend,
> > which is messing with the content and act_len (maybe?)
> does it have relation with this patch?
> >
> > I also noticed that with aggregation I wind up getting a
> > deauth/dissasoc followed by mlme delete sta/peer/key.
> >
> does it have relation with this patch?
> > I think a retry count could be useful but what do you think of maybe
> > using an event / timeout?
> >
> does it have relation with this patch?
>
> >
> >