Re: SATA Conflict with PATA DMA

Re: SATA Conflict with PATA DMA

[Jeff Garzik]

Esben Stien wrote:
> I'm having problems enabling DMA for my PATA HD.
> 
> hdparm -d1 /dev/hdb reports: 
> HDIO_SET_DMA failed: Operation not permitted
> 
> Of course, I'm super user. Nothing is printed in dmesg. 
> 
> I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
> ICH5. I also have a SATA hd in the computer and this only happens when
> the SATA hd is there. If I remove the SATA HD, then I can enable DMA
> for the PATA hd.

Disabled combined mode in BIOS.

	Jeff

SATA Conflict with PATA DMA

[Esben Stien]

I'm having problems enabling DMA for my PATA HD.

hdparm -d1 /dev/hdb reports: 
HDIO_SET_DMA failed: Operation not permitted

Of course, I'm super user. Nothing is printed in dmesg. 

I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
ICH5. I also have a SATA hd in the computer and this only happens when
the SATA hd is there. If I remove the SATA HD, then I can enable DMA
for the PATA hd.

Both the SATA and the PATA are very new. 

# zcat /proc/config.gz |grep -i dma
CONFIG_GENERIC_ISA_DMA=y
CONFIG_ISA_DMA_API=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
CONFIG_IDEDMA_PCI_AUTO=y
# CONFIG_IDEDMA_ONLYDISK is not set
CONFIG_BLK_DEV_IDEDMA=y
# CONFIG_IDEDMA_IVB is not set
CONFIG_IDEDMA_AUTO=y
# CONFIG_SCSI_PDC_ADMA is not set

Any pointers as to what I can try?

-- 
Esben Stien is b0ef@e     s      a             
         http://www. s     t    n m
          irc://irc.  b  -  i  .   e/%23contact
          [sip|iax]:   e     e 
           jid:b0ef@    n     n

Re: SATA Conflict with PATA DMA

[Bill Waddington]

On Sat, 15 Apr 2006 16:42:22 UTC, in fa.linux.kernel you wrote:

>Esben Stien wrote:
>> I'm having problems enabling DMA for my PATA HD.
>> 
>> hdparm -d1 /dev/hdb reports: 
>> HDIO_SET_DMA failed: Operation not permitted
>> 
>> Of course, I'm super user. Nothing is printed in dmesg. 
>> 
>> I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
>> ICH5. I also have a SATA hd in the computer and this only happens when
>> the SATA hd is there. If I remove the SATA HD, then I can enable DMA
>> for the PATA hd.
>
>Disabled combined mode in BIOS.

If only that was possible on my fscking T43.  *sigh*

Bill
-- 
William D Waddington
william.waddington@beezmo.com
"Even bugs...are unexpected signposts on
the long road of creativity..." - Ken Burtch

Re: SATA Conflict with PATA DMA

[Robert Hancock]

Bill Waddington wrote:
> On Sat, 15 Apr 2006 16:42:22 UTC, in fa.linux.kernel you wrote:
> 
>> Esben Stien wrote:
>>> I'm having problems enabling DMA for my PATA HD.
>>>
>>> hdparm -d1 /dev/hdb reports: 
>>> HDIO_SET_DMA failed: Operation not permitted
>>>
>>> Of course, I'm super user. Nothing is printed in dmesg. 
>>>
>>> I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
>>> ICH5. I also have a SATA hd in the computer and this only happens when
>>> the SATA hd is there. If I remove the SATA HD, then I can enable DMA
>>> for the PATA hd.
>> Disabled combined mode in BIOS.
> 
> If only that was possible on my fscking T43.  *sigh*

The other way around is to make libata handle the PATA devices. I'm not 
sure if the support for that on combined mode is in mainline or if you 
need Alan's libata PATA patch for that?

-- 
Robert Hancock      Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@nospamshaw.ca
Home Page: http://www.roberthancock.com/

Re: SATA Conflict with PATA DMA

[Alan Cox]

On Sad, 2006-04-15 at 17:33 -0600, Robert Hancock wrote:
> The other way around is to make libata handle the PATA devices. I'm not 
> sure if the support for that on combined mode is in mainline or if you 
> need Alan's libata PATA patch for that?

There should be enough in 2.6.17-rc1 to just add the needed PCI
identifiers for recent chips to work SATA/PATA. There are some ugly bugs
left in the base code that the -ac patches fix in ata_piix but the core
stuff needed is in the base tree

Re: SATA Conflict with PATA DMA

[Alistair John Strachan]

On Saturday 15 April 2006 19:38, Bill Waddington wrote:
> On Sat, 15 Apr 2006 16:42:22 UTC, in fa.linux.kernel you wrote:
> >Esben Stien wrote:
> >> I'm having problems enabling DMA for my PATA HD.
> >>
> >> hdparm -d1 /dev/hdb reports:
> >> HDIO_SET_DMA failed: Operation not permitted
> >>
> >> Of course, I'm super user. Nothing is printed in dmesg.
> >>
> >> I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
> >> ICH5. I also have a SATA hd in the computer and this only happens when
> >> the SATA hd is there. If I remove the SATA HD, then I can enable DMA
> >> for the PATA hd.
> >
> >Disabled combined mode in BIOS.
>
> If only that was possible on my fscking T43.  *sigh*

Not sure if this is universal, but if Linux doesn't claim the PATA interface, 
the SATA seems to drive optical drives on (presumably) the other channel 
(ICH7 here, on a Dell laptop with a similar BIOS limitation).

Try CONFIG_IDE=n and boot with libata.atapi_enable=1 and see what happens...

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: SATA Conflict with PATA DMA

[Randy.Dunlap]

On Sun, 16 Apr 2006 04:48:47 +0100 Alistair John Strachan wrote:

> On Saturday 15 April 2006 19:38, Bill Waddington wrote:
> > On Sat, 15 Apr 2006 16:42:22 UTC, in fa.linux.kernel you wrote:
> > >Esben Stien wrote:
> > >> I'm having problems enabling DMA for my PATA HD.
> > >>
> > >> hdparm -d1 /dev/hdb reports:
> > >> HDIO_SET_DMA failed: Operation not permitted
> > >>
> > >> Of course, I'm super user. Nothing is printed in dmesg.
> > >>
> > >> I'm on linux-2.6.16 and motherboard is Fujitsu Siemens D1561 with an
> > >> ICH5. I also have a SATA hd in the computer and this only happens when
> > >> the SATA hd is there. If I remove the SATA HD, then I can enable DMA
> > >> for the PATA hd.
> > >
> > >Disabled combined mode in BIOS.
> >
> > If only that was possible on my fscking T43.  *sigh*
> 
> Not sure if this is universal, but if Linux doesn't claim the PATA interface, 
> the SATA seems to drive optical drives on (presumably) the other channel 
> (ICH7 here, on a Dell laptop with a similar BIOS limitation).
> 
> Try CONFIG_IDE=n and boot with libata.atapi_enable=1 and see what happens...

make that libata.atapi_enabled=1

---
~Randy

Re: SATA Conflict with PATA DMA

[Esben Stien]

Jeff Garzik <jeff@garzik.org> writes:

> Disabled combined mode in BIOS.

There is no such option in my BIOS. Does it maybe go by another name?.

Phoenix BIOS 4.06 Rev. 1.09.1561

-- 
Esben Stien is b0ef@e     s      a             
         http://www. s     t    n m
          irc://irc.  b  -  i  .   e/%23contact
          [sip|iax]:   e     e 
           jid:b0ef@    n     n

Re: SATA Conflict with PATA DMA

[Esben Stien]

Esben Stien <b0ef@esben-stien.name> writes:

> There is no such option in my BIOS. Does it maybe go by another
> name?.

Sorry, I see some replies where sent to this thread from MUAs' with
broken headers, so I didn't catch them. 

It seems it will be fixed in 2.6.17

-- 
Esben Stien is b0ef@e     s      a             
         http://www. s     t    n m
          irc://irc.  b  -  i  .   e/%23contact
          [sip|iax]:   e     e 
           jid:b0ef@    n     n

Re: SATA Conflict with PATA DMA

[Esben Stien]

Esben Stien <b0ef@esben-stien.name> writes:

> It seems it will be fixed in 2.6.17

Hmm, I'm still seeing this with 2.6.17-rc6-rt3. 

-- 
Esben Stien is b0ef@e     s      a             
         http://www. s     t    n m
          irc://irc.  b  -  i  .   e/%23contact
           sip:b0ef@   e     e 
           jid:b0ef@    n     n

Re: SATA Conflict with PATA DMA

[Bill Waddington]

On Sat, 15 Apr 2006 18:39:20 UTC, in fa.linux.kernel you wrote:

>On Sat, 15 Apr 2006 16:42:22 UTC, in fa.linux.kernel you wrote:
>
>>Esben Stien wrote:
>>> I'm having problems enabling DMA for my PATA HD.
>>> 
>>> hdparm -d1 /dev/hdb reports: 
>>> HDIO_SET_DMA failed: Operation not permitted

>>Disabled combined mode in BIOS.
>
>If only that was possible on my fscking T43.  *sigh*

(for others struggling with this) Per the fix reported at

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=163418

Updating my T43/FC5 to kernel 2.6.16-1.2122_FC5 and booting
with "combined_mode=libata" gets the DVD drive running at speed
as /dev/scd0 (although there doesn't seem to be any explicit
"dma enabled" indication).

Thanks Jeff et al.

Bill
-- 
William D Waddington
william.waddington@beezmo.com
"Even bugs...are unexpected signposts on
the long road of creativity..." - Ken Burtch

Re: SATA Conflict with PATA DMA

[Esben Stien]

Esben Stien <b0ef@esben-stien.name> writes:

> It seems it will be fixed in 2.6.17

I ran 2.6.17-rc6-rt3 with combined_mode=libata and this worked to make
the hd work at full speed, but something weird is happening:

It lists both hd's as the same device: 

#df -h
/dev/sda1             276G  276G  727M 100% /
/dev/sda1              73G  1,1G   72G   2% /mnt/fs1

#ls /
bck-sys-20051017  boot  etc-2006-02-16  lib    nfs   root   src  tmp   var
bin               dev   etc-2006-02-17  media  opt   sbin   srv  udev
blfs              etc   home            mnt    proc  share  sys  usr

#ls /mnt/fs1
b0ef  bacula

This must clearly be a bug.

-- 
Esben Stien is b0ef@e     s      a             
         http://www. s     t    n m
          irc://irc.  b  -  i  .   e/%23contact
           sip:b0ef@   e     e 
           jid:b0ef@    n     n