Re: ide-cd problem
From: Alan Chandler (alan_at_chandlerfamily.org.uk)
Date: 11/23/04
- Previous message: Karel Kulhavy: "description of struct sockaddr"
- In reply to: Jens Axboe: "Re: ide-cd problem"
- Next in thread: Alan Chandler: "Re: ide-cd problem"
- Reply: Alan Chandler: "Re: ide-cd problem"
- Reply: Alan Chandler: "Re: ide-cd problem"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
To: Jens Axboe <axboe@suse.de> Date: Tue, 23 Nov 2004 21:49:31 +0000
On Tuesday 23 November 2004 14:51, Jens Axboe wrote:
> On Tue, Nov 23 2004, Alan Chandler wrote:
> > On Monday 22 November 2004 23:48, Alan Chandler wrote:
> > ...
> >
> > > If I make the delay 600ns it works - I guess my hardware is a little
> > > off spec.
> >
> > I did a binary chop on the value to find the cut off point between what
> > works and what doesn't. Its approx 535ns (534 failed, 537 worked).
> >
> > All this was with 2.6.9,
> >
> > 2.6.10-rc2 is still failing during the cd initialisation on boot. Here I
> > tried with bot 600ns and 700ns delays in drive_is_ready, but both values
> > fail with what looks like missed interrupts. I'll try instrumenting this
> > a bit more to find out what is happening.
>
> It's getting more and more interesting! Look forward to hearing what
> your instrumentation brings.
>
> There are other reports of acpi causing interrupt problems with cdroms
> in 2.6.10-rc2, so it would be best if you stuck to 2.6.9 for testing
> this particular problem.
There is good and bad news related to 2.6.10-rc2
The good news is that the acpi problem was the cause of the startup issues.
adding pci=noacpi to the boot command line fixed that.
The bad news is that with the delay at 800ns in drive_is_ready() I am getting
the exact same symptoms I got with 2.6.9 before upping the delay to over
540ns.
Before, I thought my hardware was a little out of spec - now I think there is
something else at play here.
Firstly, the symptoms are the same between 2.6.9 and 2.6.10-rc2. The halt
seem to always be in exactly the same place. If it was a timing problem I
would have thought it would have varied.
Secondly, the command before seems to have an expectation of a 2048 transfer
rather than the 0 in the command, before the problem and then you get the
strange DRQ=1 but 0 in the len register.
Nov 23 20:37:33 kanger kernel: ide-cd:ide_do_rq_cdrom - cmd = 0x0
Nov 23 20:37:33 kanger kernel: ide-cd:cdrom_newpc_intr - cmd=0x0 stat=0x50
ireason=3 len=2048 rq len=0
Nov 23 20:37:33 kanger kernel: ide-cd:ide_do_rq_cdrom - cmd = 0x1b
Nov 23 20:37:33 kanger kernel: ide-cd:cdrom_newpc_intr - cmd=0x1b stat=0x58
ireason=2 len=0 rq len=0
I have got myself a copy of the ATA/ATAPI spec (document T13/1410D revision
3), I think I need to read more of it to understand what is the code is
trying to do.
-- Alan Chandler alan@chandlerfamily.org.uk First they ignore you, then they laugh at you, then they fight you, then you win. --Gandhi - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
- Previous message: Karel Kulhavy: "description of struct sockaddr"
- In reply to: Jens Axboe: "Re: ide-cd problem"
- Next in thread: Alan Chandler: "Re: ide-cd problem"
- Reply: Alan Chandler: "Re: ide-cd problem"
- Reply: Alan Chandler: "Re: ide-cd problem"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
