Re: Thinkpad T23 60 second stall on resuming 8.0-RELEASE/i386

Ian Smith wrote:
> On Tue, 15 Dec 2009, Nate Lawson wrote:
>  > Ian Smith wrote:
>  > > -wakeup from sleeping state (slept 00:00:07)
>  > > +t_delta 15.f9ad99f01204edd8 too short                        <<<<<<<
>  > > +t_delta 16.07bb5b66ef900000 too long                         <<<<<<<
>  > > +t_delta 15.f9ad90918acc0000 too short                        <<<<<<<
>  > > +ct_to_ts([2009-12-13 17:10:39]) = 1260724239.000000000
>  > > +wakeup from sleeping state (slept 00:01:07)
>  > >  ata0: reiniting channel ..
>  > >  ata0: reset tp1 mask=03 ostat0=50 ostat1=00
>  > >  ata0: stat0=0x50 err=0x01 lsb=0x00 msb=0x00
>  > >  ata0: stat1=0x00 err=0x01 lsb=0x00 msb=0x00
>  > > -ata0: reset tp2 stat0=50 stat1=00 devices=0x1<ATA_MASTER>
>  > > +ata0: reset tp2 stat0=50 stat1=00 devices=0x1
>  > >  ad0: setting PIO4 on ICH3 chip
>  > >  ad0: setting UDMA100 on ICH3 chip
>  > >  ata0: reinit done ..
>  > > @@ -36,7 +37,7 @@
>  > >  ata1: reset tp1 mask=03 ostat0=00 ostat1=00
>  > >  ata1: stat0=0x00 err=0x01 lsb=0x14 msb=0xeb
>  > >  ata1: stat1=0x00 err=0x00 lsb=0x00 msb=0x00
>  > > -ata1: reset tp2 stat0=00 stat1=00 devices=0x4<ATAPI_MASTER>
>  > > +ata1: reset tp2 stat0=00 stat1=00 devices=0x10000
>  > 
>  > I think it's ATA timing out for some reason.
> 
> Thanks Nate.  I nearly clipped that part of the diff, figuring only the 
> <*_MASTER> messages had changed, but I now see ata1 devices differ too.
> 
> Have similar regressions re ATA been turning up elsewere, do you know?
> 
> I've been digging, still on the trail of those t_delta messages, but now 
> figure these might be spurious, some timecounter missing/gaining a tick 
> or something, if ATA is maybe what's hanging meanwhile?
> 
> There's no HD light activity at all from resume button/switch through 
> that 60 second wait, when 15? seconds after the third t_delta line is 
> written to console, the HD light flashes while apparently simultaneously 
> writing the rest of the resume messages, and it comes alive.
> 
> Guess I should next upgrade the 7.0-R slice to 7.2-STABLE to find out if 
> this problem has been mfc'd :)

I'd try to revert or move ATA forward some to see if that fixes things.
Usually that can be done without too many other changes to the kernel.

-- 
Nate