Re: XenVif div by zero on Tx path after resume.

["Durrant, Paul" <xadimgnik@xxxxxxxxx>]

On 22/04/2022 15:09, Martin Harvey wrote:
> -----Original Message-----
> From: win-pv-devel <win-pv-devel-bounces@xxxxxxxxxxxxxxxxxxxx> On Behalf Of 
> Durrant, Paul
> Sent: 22 April 2022 13:32
> To: paul@xxxxxxx; win-pv-devel@xxxxxxxxxxxxxxxxxxxx
> Subject: Re: XenVif div by zero on Tx path after resume.
>
>
> > Oh ffs... the call to KeLowerIrql(DISPATCH_LEVEL) is being done at the end of 
> > SyncEnableInterrupts() rather than after the late callbacks at the beginning of 
> > SyncRelease(). That will be the problem... and now I need to dig for why it is 
> > like that.
> Yes, I thought something was a bit wrong.
>
> On the running machine, I notice all the CPU's are at IRQL zero (small sample 
> here of a few CPU's of 32), and a dump of the running threads on the system, 
> one of which points fairly conclusively to SyncRelease:
... and unfortunately it seems to have been broken for so long my git 
history won't tell me when :-(
  Paul