ide cd/dvd with 2.4.19-pre8

ide cd/dvd with 2.4.19-pre8

[Måns Rullgård]

I just noticed that reading from both my cdrom and dvd is a lot slower
with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared to
1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even impossible
to play DVDs. Any ideas?

-- 
Måns Rullgård
mru@users.sf.net

Re: ide cd/dvd with 2.4.19-pre8

[Alan Cox]

> I just noticed that reading from both my cdrom and dvd is a lot slower
> with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared=
>  to
> 1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even imp=
> ossible
> to play DVDs. Any ideas?

Did you include support for your controller. Are you in fact now in PIO
mode because you didnt ?

Re: ide cd/dvd with 2.4.19-pre8

[Måns Rullgård]

Alan Cox <alan@lxorguk.ukuu.org.uk> writes:

> > I just noticed that reading from both my cdrom and dvd is a lot slower
> > with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared=
> >  to
> > 1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even imp=
> > ossible
> > to play DVDs. Any ideas?
> 
> Did you include support for your controller. Are you in fact now in PIO
> mode because you didnt ?

I included support for PDC202xx and have a PDC20268 (Promise Ultra100
TX2).  Isn't that enough? hdparm claims the cd and dvd are using
dma. The configuration is exactly the same with both kernels.

However, I had to modify pdc202xx.c to make it use udma3 and higher
for the disk. It failed to set hwif->udma_four so other parts of the
ide driver refused to even try the fast modes. I pointed this out a couple
of weeks ago but didn't get any response.

There is also some trouble with udma4&5. Both the disk and the controller
claim they support it and it looks like it is properly detected. However
I get these messages when booting:
PDC20268: IDE controller on PCI bus 00 dev 60
PDC20268: chipset revision 2
PDC20268: not 100% native mode: will probe irqs later
PDC20268: ROM enabled at 0x09074000
    ide2: BM-DMA at 0x8090-0x8097, BIOS settings: hde:pio, hdf:pio
    ide3: BM-DMA at 0x8098-0x809f, BIOS settings: hdg:pio, hdh:pio
hde: ST340810A, ATA DISK drive
hdg: SAMSUNG COMBO SM-304B, ATAPI CD/DVD-ROM drive
ide2 at 0x80a0-0x80a7,0x80b2 on irq 32
ide3 at 0x80a8-0x80af,0x80b6 on irq 32
hde: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(100)
Partition check:
 /dev/ide/host2/bus0/target0/lun0:hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
PDC202XX: Primary channel reset.
ide2: reset: success
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
PDC202XX: Primary channel reset.
ide2: reset: success

After this the disk is in udma3 mode and works fine, though it should be able
to go faster. This problem has been present with all kernels from 2.4.10.
I haven't tried earlier ones. I have an Alpha in case it matters.

-- 
Måns Rullgård
mru@users.sf.net

Re: ide cd/dvd with 2.4.19-pre8

[Andre Hedrick]

The driver pulled your system back into safe data io ranges.
Either your cable routing is poor, your power supply is marginal, possible
but not likely (hardware combination does not like the odd or even cable
grounding setup), regardless it did the correct thing.

The problem is a feature to prevent the driver from dropping out of DMA
to PIO when it is better to down grade the transfer rate.

The next issue is whether your ATAPI is in DMA, and it should not be.
The driver core does not use split dma engines yet.

Cheers,

Andre Hedrick
LAD Storage Consulting Group

On 18 May 2002, Måns Rullgård wrote:

> Alan Cox <alan@lxorguk.ukuu.org.uk> writes:
> 
> > > I just noticed that reading from both my cdrom and dvd is a lot slower
> > > with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared=
> > >  to
> > > 1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even imp=
> > > ossible
> > > to play DVDs. Any ideas?
> > 
> > Did you include support for your controller. Are you in fact now in PIO
> > mode because you didnt ?
> 
> I included support for PDC202xx and have a PDC20268 (Promise Ultra100
> TX2).  Isn't that enough? hdparm claims the cd and dvd are using
> dma. The configuration is exactly the same with both kernels.
> 
> However, I had to modify pdc202xx.c to make it use udma3 and higher
> for the disk. It failed to set hwif->udma_four so other parts of the
> ide driver refused to even try the fast modes. I pointed this out a couple
> of weeks ago but didn't get any response.
> 
> There is also some trouble with udma4&5. Both the disk and the controller
> claim they support it and it looks like it is properly detected. However
> I get these messages when booting:
> PDC20268: IDE controller on PCI bus 00 dev 60
> PDC20268: chipset revision 2
> PDC20268: not 100% native mode: will probe irqs later
> PDC20268: ROM enabled at 0x09074000
>     ide2: BM-DMA at 0x8090-0x8097, BIOS settings: hde:pio, hdf:pio
>     ide3: BM-DMA at 0x8098-0x809f, BIOS settings: hdg:pio, hdh:pio
> hde: ST340810A, ATA DISK drive
> hdg: SAMSUNG COMBO SM-304B, ATAPI CD/DVD-ROM drive
> ide2 at 0x80a0-0x80a7,0x80b2 on irq 32
> ide3 at 0x80a8-0x80af,0x80b6 on irq 32
> hde: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(100)
> Partition check:
>  /dev/ide/host2/bus0/target0/lun0:hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> PDC202XX: Primary channel reset.
> ide2: reset: success
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x84 { DriveStatusError BadCRC }
> PDC202XX: Primary channel reset.
> ide2: reset: success
> 
> After this the disk is in udma3 mode and works fine, though it should be able
> to go faster. This problem has been present with all kernels from 2.4.10.
> I haven't tried earlier ones. I have an Alpha in case it matters.
> 
> -- 
> Måns Rullgård
> mru@users.sf.net
> -
> 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/
> 

Re: ide cd/dvd with 2.4.19-pre8

[Måns Rullgård]

Andre Hedrick <andre@linux-ide.org> writes:

> The driver pulled your system back into safe data io ranges.
> Either your cable routing is poor,
I use the cable that came with the controller. It runs as strait as the
box allows. How sensitive are these things really?

> your power supply is marginal, possible
It's 300 W. Isn't that enough?

> but not likely (hardware combination does not like the odd or even cable
> grounding setup), regardless it did the correct thing.
> 
> The problem is a feature to prevent the driver from dropping out of DMA
> to PIO when it is better to down grade the transfer rate.
> 
> The next issue is whether your ATAPI is in DMA, and it should not be.
> The driver core does not use split dma engines yet.

Was there some change from 2.4.18 to 19-pre8 in this area?

Is there any way to use the disk in udma5? Can there be compatibility
problems with the disk?

-- 
Måns Rullgård
mru@users.sf.net

Re: ide cd/dvd with 2.4.19-pre8

[G Sandine]

On Sat, May 18, 2002 at 02:42:00PM +0200, Måns Rullgård wrote:
> Is there any way to use the disk in udma5? Can there be compatibility
> problems with the disk?

I have recently been struggling with similar issues.  My troubles seem
for one thing to be related to VIA chipset problems, mentioned at
  
 http://www.tecchannel.de/hardware/817/index.html
  
With various kernels (and IDE patch versions), I see the errors you see,
and switching the kernel and/or IDE patch version makes the problem go away
(e.g. 2.4.19-pre5-ac3 is good for Promise controllers with hard disks).
Mainly, I only see the errors you report if I use both master and slave in
the primary channel on the Promise card; no troubles (except too slow)
if I use primary channel only on the Promise card, and I can do whatever
in the secondary channel (none, master, master + slave, slave).  This
holds for PDC20269 ATA 133 cards and Maxtor ATA 133 drives, or PDC2026x
ATA 100 cards and various ATA 100 drives.
  
Here are a few chipsets and what I have seen them do (all using
2.4.19-pre5-ac3 because it eliminates the errors during boot):
  
1. VIA VT8233CD south bridge (VIA KT266A chipset), ATA 100 devices
   and ATA 133 devices hanging off ATA 100 and ATA 133 Promise cards
   (respectively) are reported at udma2 by hdparm.  This is a PC2100
   AMD board.
  
2. VIA VT8233A south bridge (VIA KT333 chipset), ATA 100 devices and
   ATA 133 devices hanging off of the Promise cards are reported at
   udma5 and udma6 (respectively) by hdparm.  This is a PC2700 AMD
   board.  ATA 100 and ATA 133 devices run as they should (udma5 and
   udma6) using the onboard ATA 133 IDE (by the VT8233A chip).

3. AMD 766 south bridge (AMD 760MP chipset), ATA 133 devices are reported
   at udma6 by hdparm when hanging off of the ATA 133 Promise controller.
   This is a PC2100 AMD board.

Disk benchmarks by bonnie++ have verified that the disks are running in
modes reported by hdparm in each case.

Item 1. is particularly troubling.  I have found many references to the
tecChannel article, but I know of no solutions.

Re: ide cd/dvd with 2.4.19-pre8

[Måns Rullgård]

Andre Hedrick <andre@linux-ide.org> writes:

> The driver pulled your system back into safe data io ranges.
> Either your cable routing is poor, your power supply is marginal, possible
> but not likely (hardware combination does not like the odd or even cable
> grounding setup), regardless it did the correct thing.
> 
> The problem is a feature to prevent the driver from dropping out of DMA
> to PIO when it is better to down grade the transfer rate.
> 
> The next issue is whether your ATAPI is in DMA, and it should not be.
> The driver core does not use split dma engines yet.

I just tried connecting it to a cmd646 controller. It runs at full speed
but no dma because of cmd bugs. Is there a problem with the new pdc202xx
driver in 2.4.19 or is it somehow my fault?

-- 
Måns Rullgård
mru@users.sf.net

Re: ide cd/dvd with 2.4.19-pre8

[Denis Vlasenko]

On 17 May 2002 16:18, M?ns Rullg?rd wrote:
> I just noticed that reading from both my cdrom and dvd is a lot slower
> with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared to
> 1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even
> impossible to play DVDs. Any ideas?

Do you know at which preN this has happened?
--
vda

Re: ide cd/dvd with 2.4.19-pre8

[Måns Rullgård]

Denis Vlasenko <vda@port.imtp.ilyichevsk.odessa.ua> writes:

> On 17 May 2002 16:18, M?ns Rullg?rd wrote:
> > I just noticed that reading from both my cdrom and dvd is a lot slower
> > with 2.4.19-pre8 than 2.4.18. Now hdparm reports ~800 kbytes/s compared to
> > 1.7 MBytes/s for CD and >2 MBytes/s for DVD with 2.4.18. It is even
> > impossible to play DVDs. Any ideas?
> 
> Do you know at which preN this has happened?

No, and really don't have the time to try them all just to find out.
I do suspect it is related to the major changes in the ide drivers and
particularly pdc202xx.c.

-- 
Måns Rullgård
mru@users.sf.net