Re: dma ripping

[Philip Dodd]

Jens Axboe wrote:
> On Tue, May 18 2004, Philip Dodd wrote:
> > Hugo Mills wrote:
> > >  Put me down for this latter one, too. I'm using a vanilla 2.6.[56]
> > > on amd64. Controller is VIA.
> > >  It seems to be related to hard-to-read CDs (dirty/scratched/badly-
> > > made) -- I've got a couple here that I'm pretty sure I can use as test
> > > cases to trigger the problem instantly.
> > Hi,
> > OK - I don't know if any of this helps, but I guess a little more 
> > precision won't do anyone any harm.
> > Intel i820 Chipset on P3C-D mobo.
> > ide0: BM-DMA at 0xa800-0xa807, BIOS settings: hda:DMA, hdb:pio
> > hda: ASUS DVD-ROM E616, ATAPI CD/DVD-ROM drive
> > hda: ATAPI 48X DVD-ROM drive, 512kB Cache, UDMA(33)
> > ide1: BM-DMA at 0xa808-0xa80f, BIOS settings: hdc:DMA, hdd:pio
> > hdc: RICOH CD-R/RW MP7060A, ATAPI CD/DVD-ROM drive
> > hdc: ATAPI 24X CD-ROM CD-R/RW drive, 2048kB Cache, DMA
> > Now hda is the one that bogs out, ripping silence after the "cdrom:
> > dropping to single frame dma" error. hdc can rip for hours and hardly 
> > ever get cdparanoia errors - even on "problematic" CDs that would appear 
> > to be a declenching factor for the single frame dma switch for hda.

> Any chance you can see if this makes any difference (on 2.6.6-BK)?

Hi,

Sorry it took a while for me to get you feedback on this patch.  I have 
just applied this patch against 2.6.7-rc2 (some fuzz and some offset - I 
can get you details if you would like).  It exhibits exactly the same 
symptoms as before - ie. certain CDs will cause the "kernel: cdrom: 
dropping to single frame dma", and all ripping form that point on until 
reboot will just rip to silence (I have one test case that does it 75% 
of the way through track 4, as regular as clockwork, but several other 
do to, and some even appear not to do it all the time - physically the 
CD is in good shape and rips fine using Win32 ripping tools on my laptop).

My next guess is that this is a hardware problem, though I'd appreciate 
your feedback on whether that patch should have fixed or not.

Thanks again for all your help on this,

Phil
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/