If anyone is suffering from the dreaded "dmaproc error 14: unsupported"
error and want to test a code that could get you out of that deadlock
please speak up.
Basically this is an Intel 440BX PIIX4 issues, but the solution is global
and should work for all cases.
Regards,
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
This isnt the same as "ide_dmaproc: chipset supported ide_dma_timeout
func only: 14" is it? This was on a PIIX4 w/ test11-pre5. I've since
upgraded to test11-pre7 and the problem hasn't surfaced again.
Andre Hedrick wrote:
>
> If anyone is suffering from the dreaded "dmaproc error 14: unsupported"
> error and want to test a code that could get you out of that deadlock
> please speak up.
>
> Basically this is an Intel 440BX PIIX4 issues, but the solution is global
> and should work for all cases.
>
> Regards,
>
> Andre Hedrick
> CTO Timpanogas Research Group
> EVP Linux Development, TRG
> Linux ATA Development
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> Please read the FAQ at http://www.tux.org/lkml/
--
=====================================================================
Mohammad A. Haque http://www.haque.net/
[email protected]
"Alcohol and calculus don't mix. Project Lead
Don't drink and derive." --Unknown http://wm.themes.org/
[email protected]
=====================================================================
On Sat, 18 Nov 2000 12:23:05 -0800 (PST), Andre Hedrick <[email protected]> wrote:
>
> If anyone is suffering from the dreaded "dmaproc error 14: unsupported"
> error and want to test a code that could get you out of that deadlock
> please speak up.
>
> Basically this is an Intel 440BX PIIX4 issues, but the solution is global
> and should work for all cases.
Interestingly enough, I get it on a VIA MVP3 with 2.2.18pre + bkz's patch.
Since it's eaten two filesystems by now, I'm not overly eager to play
with it again... On a second thought though, I'd rather play with it now
in a controlled environment, so feel free to send the patch my way.
If you want details, just ask.
Thanks,
Ion
--
It is better to keep your mouth shut and be thought a fool,
than to open it and remove all doubt.
It is on kernel.org and the README tells you what to do to enable the stub
in ide-dma.c If it works let me know!
Cheers,
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
> It is on kernel.org and the README tells you what to do to enable the stub
> in ide-dma.c If it works let me know!
Andre,
Where on kernel.org are you hiding the README and patch files you
mention?
Regards,
Sean
/pub/linux/kernel/people/hedrick/ide-2.2.17/README
/pub/linux/kernel/people/hedrick/ide-2.2.17/ide.2.2.17.all.20001118.patch.bz2
There you go Sean, hope that helps.
Cheers,
On Sun, 19 Nov 2000, Sean B. Estabrooks wrote:
>
> > It is on kernel.org and the README tells you what to do to enable the stub
> > in ide-dma.c If it works let me know!
>
> Andre,
>
> Where on kernel.org are you hiding the README and patch files you
> mention?
>
> Regards,
> Sean
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> Please read the FAQ at http://www.tux.org/lkml/
>
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
On Sun, Nov 19, 2000 at 03:37:56PM -0800, Andre Hedrick wrote:
> /pub/linux/kernel/people/hedrick/ide-2.2.17/README
> /pub/linux/kernel/people/hedrick/ide-2.2.17/ide.2.2.17.all.20001118.patch.bz2
> There you go Sean, hope that helps.
It didn't help me :)
Axialys have a squid/socks/NFS server which therefore has a fair bit of
disk activity.
It died randomly a while ago, so I put a serial console on it and this
morning a coworker had to reboot it.
The last message was apparently the very same you are trying to prevent:
"hdg: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14"
(usual caveat about hand copied)
This was 2.2.17+ide (20000904)+RAID(2.2.17-A0).
This morning I updated to this patch and I tried to kill it just now.
It died :(
(That is with the 20001118 ide patch instead of the 20000904 one)
This time only with a "hdg: timeout waiting for DMA" with nothing
following it on the serial console.
It seems that
hdparm -t /dev/hde & hdparm -t /dev/hde & hdparm -t /dev/hde & hdparm -t
/dev/hde & hdparm -t /dev/hdg & hdparm -t /dev/hdg & hdparm -t /dev/hdg
& hdparm -t /dev/hdg
kills it in this way. (That might be excessive but I'll get odd looks
is the proxy is down again because I'm testing exactly how many hdparms
it takes).
Let me know if there is anything I can test (in the evening/this weekend).
Here's some information:
Nov 24 11:50:15 orange kernel: Uniform Multi-Platform E-IDE driver Revision: 6.30
Nov 24 11:50:15 orange kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
Nov 24 11:50:15 orange kernel: PIIX4: IDE controller on PCI bus 00 dev 39
Nov 24 11:50:15 orange kernel: PIIX4: chipset revision 1
Nov 24 11:50:15 orange kernel: PIIX4: not 100%% native mode: will probe irqs later
Nov 24 11:50:15 orange kernel: ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:pio, hdb:DMA
Nov 24 11:50:15 orange kernel: ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:pio, hdd:pio
Nov 24 11:50:15 orange kernel: HPT366: onboard version of chipset, pin1=1 pin2=2
Nov 24 11:50:15 orange kernel: PCI: HPT366: Fixing interrupt 11 pin 2 to ZERO
Nov 24 11:50:15 orange kernel: HPT366: IDE controller on PCI bus 00 dev 98
Nov 24 11:50:15 orange kernel: HPT366: chipset revision 1
Nov 24 11:50:15 orange kernel: HPT366: not 100%% native mode: will probe irqs later
Nov 24 11:50:15 orange kernel: ide2: BM-DMA at 0xe000-0xe007, BIOS settings: hde:DMA, hdf:pio
Nov 24 11:50:15 orange kernel: HPT366: IDE controller on PCI bus 00 dev 99
Nov 24 11:50:15 orange kernel: HPT366: chipset revision 1
Nov 24 11:50:15 orange kernel: HPT366: not 100%% native mode: will probe irqs later
Nov 24 11:50:15 orange kernel: ide3: BM-DMA at 0xec00-0xec07, BIOS settings: hdg:DMA, hdh:pio
Nov 24 11:50:15 orange kernel: hdb: FX4824T, ATAPI CDROM drive
Nov 24 11:50:15 orange kernel: hde: Maxtor 54098H8, ATA DISK drive
Nov 24 11:50:15 orange kernel: hdg: Maxtor 54098H8, ATA DISK drive
Nov 24 11:50:15 orange kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Nov 24 11:50:15 orange kernel: ide2 at 0xd800-0xd807,0xdc02 on irq 11
Nov 24 11:50:15 orange kernel: ide3 at 0xe400-0xe407,0xe802 on irq 11
Nov 24 11:50:15 orange kernel: hde: Maxtor 54098H8, 39082MB w/2048kB Cache, CHS=79406/16/63, UDMA(66)
Nov 24 11:50:15 orange kernel: hdg: Maxtor 54098H8, 39082MB w/2048kB Cache, CHS=79406/16/63, UDMA(66)
Nov 24 11:50:15 orange kernel: hdb: ATAPI 48X CD-ROM drive, 128kB Cache,
UDMA(33)
Nov 24 11:50:15 orange kernel: Uniform CD-ROM driver Revision: 3.11
orange:/home/huggie# hdparm -i /dev/hde
/dev/hde:
Model=Maxtor 54098H8, FwRev=DAC10SC0, SerialNo=K80B8WCC
Config={ Fixed }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
BuffType=3(DualPortCache), BuffSize=2048kB, MaxMultSect=16, MultSect=16
DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=0
CurCHS=65535/1/63, CurSects=4128705, LBA=yes, LBAsects=80041248
tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
IORDY=on/off, tPIO={min:120,w/IORDY:120}, PIO modes: mode3 mode4
UDMA modes: mode0 mode1 mode2 mode3 *mode4 mode5
orange:/home/huggie# hdparm -i /dev/hdg
/dev/hdg:
Model=Maxtor 54098H8, FwRev=DAC10SC0, SerialNo=K80AHT0C
Config={ Fixed }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
BuffType=3(DualPortCache), BuffSize=2048kB, MaxMultSect=16, MultSect=16
DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=0
CurCHS=65535/1/63, CurSects=4128705, LBA=yes, LBAsects=80041248
tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
IORDY=on/off, tPIO={min:120,w/IORDY:120}, PIO modes: mode3 mode4
UDMA modes: mode0 mode1 mode2 mode3 *mode4 mode5
Simon.
--
[ "Peace and understanding through brute force." -- David Parsons ]
Black Cat Networks. http://www.blackcatnetworks.co.uk/
Hi!
I would like to ask if the tagged command queueing capability in the
decent ATA standards is utilized in the linux IDE driver (2.2 2.2ide
patches, or 2.4 maybe...)?
Another question, a little bit offtopic is if anybody on this list is able
to point me to some pci UltraATA controller card, which has more than 2
channels (more than 4 drives w/o cascade magic) (preferrably 6 or 8
channels)? If the exists one, I do not need no HW RAID or anything like
that... I only need a lot of channels, in only one PCI slot.
There have been a lot of rant around here, and why I am addressing Andre
personally too is that in that debate his stated, that because such cards
exist, IDE can fall into line with SCSI in drives possible / PCI slot.
Your help is very much appreciated,
--
SaPE - Peter, Sasi - mailto:[email protected] - http://sape.iq.rulez.org/
On Mon, Nov 27 2000, Sasi Peter wrote:
> Hi!
>
> I would like to ask if the tagged command queueing capability in the
> decent ATA standards is utilized in the linux IDE driver (2.2 2.2ide
^^^^^^
> patches, or 2.4 maybe...)?
I hope that is supposed to be 'recent', because with the current TCQ
implementation listed in the specs Linux might as well not support it :)
It's simply not worth it.
--
* Jens Axboe <[email protected]>
* SuSE Labs
On Mon, 27 Nov 2000, Jens Axboe wrote:
> > I would like to ask if the tagged command queueing capability in the
> > decent ATA standards is utilized in the linux IDE driver (2.2 2.2ide
> ^^^^^^
> > patches, or 2.4 maybe...)?
> I hope that is supposed to be 'recent', because with the current TCQ
If you think I should have...
> implementation listed in the specs Linux might as well not support it :)
> It's simply not worth it.
But seriously, how come?
I thought they just somewhat like copied the SCSI implementation...
PS: Anybody knowing about a multi (>2) channel UATA host adapter?
--
SaPE - Peter, Sasi - mailto:[email protected] - http://sape.iq.rulez.org/
On Mon, Nov 27 2000, Sasi Peter wrote:
> > implementation listed in the specs Linux might as well not support it :)
> > It's simply not worth it.
>
> But seriously, how come?
>
> I thought they just somewhat like copied the SCSI implementation...
I wish they would have, and based it on atapi. But they didn't...
Basically it requires you to poll for completion of tags with
a service command.
--
* Jens Axboe <[email protected]>
* SuSE Labs
On Mon, 27 Nov 2000, Jens Axboe wrote:
> On Mon, Nov 27 2000, Sasi Peter wrote:
> > Hi!
> >
> > I would like to ask if the tagged command queueing capability in the
> > decent ATA standards is utilized in the linux IDE driver (2.2 2.2ide
> ^^^^^^
> > patches, or 2.4 maybe...)?
>
> I hope that is supposed to be 'recent', because with the current TCQ
> implementation listed in the specs Linux might as well not support it :)
> It's simply not worth it.
Exactly, Jens has seen the ugly beast because I have worked on coding it.
I am working to get IBM to change the method of doing this to make it
sane, but its not now.
Cheers,
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
On Mon, 27 Nov 2000, Jens Axboe wrote:
> On Mon, Nov 27 2000, Sasi Peter wrote:
> > > implementation listed in the specs Linux might as well not support it :)
> > > It's simply not worth it.
> > But seriously, how come?
> > I thought they just somewhat like copied the SCSI implementation...
> I wish they would have, and based it on atapi. But they didn't...
> Basically it requires you to poll for completion of tags with a
> service command.
On Sun, 26 Nov 2000, Andre Hedrick wrote:
> On Mon, 27 Nov 2000, Jens Axboe wrote:
> > On Mon, Nov 27 2000, Sasi Peter wrote:
> > > I would like to ask if the tagged command queueing capability in the
> > > decent ATA standards is utilized in the linux IDE driver (2.2 2.2ide
> > ^^^^^^
> > > patches, or 2.4 maybe...)?
Ok, so anybody knows a multichannel adapter? At least 4 channels?
Andre? Where do I get one? Where does the ATA revolution start?
--
SaPE - Peter, Sasi - mailto:[email protected] - http://sape.iq.rulez.org/