Re: [parisc-linux] problems with PCI IDE controller

Re: [parisc-linux] problems with PCI IDE controller

[John Marvin]

> On Sat, 2002-12-14 at 13:10, Matthew Wilcox wrote:
> > > hp-c240 login: Uniform Multi-Platform E-IDE driver Revision: 6.31
> > > ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
> > > PDC20268: IDE controller on PCI bus 00 dev 08
> > > PDC20268: chipset revision 2
> > > PDC20268: not 100% native mode: will probe irqs later
>
> Is this a plug in card. For some reason it has been left in non native
> mode so won't work as a plug in board.

I don't think this is a problem. I have a similar card in my home PC, and
it prints the same thing (re: not 100% native mode). Here is an excerpt
from my PC's boot messages running 2.4.19:

    Uniform Multi-Platform E-IDE driver Revision: 6.31
    ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
    PDC20267: IDE controller on PCI bus 02 dev 08
    PCI: Found IRQ 10 for device 02:01.0
    PDC20267: chipset revision 2
    PDC20267: not 100% native mode: will probe irqs later
    PDC20267: (U)DMA Burst Bit ENABLED Primary PCI Mode Secondary PCI Mode.
	ide0: BM-DMA at 0xa400-0xa407, BIOS settings: hda:pio, hdb:pio
	ide1: BM-DMA at 0xa408-0xa40f, BIOS settings: hdc:pio, hdd:pio

All of my hard drives are on this card, so I know it works just fine (I
use ide=reverse so that my hard drives start with hda. I keep my CD and
DVD drives on the motherboard IDE buses).

John

Re: [parisc-linux] problems with PCI IDE controller

[John David Anglin]

> > On Sat, 2002-12-14 at 13:10, Matthew Wilcox wrote:
> > > > hp-c240 login: Uniform Multi-Platform E-IDE driver Revision: 6.31
> > > > ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
> > > > PDC20268: IDE controller on PCI bus 00 dev 08
> > > > PDC20268: chipset revision 2
> > > > PDC20268: not 100% native mode: will probe irqs later

See the same on the pc that I have.

Dave
-- 
J. David Anglin                                  dave.anglin@nrc.ca
National Research Council of Canada              (613) 990-0752 (FAX: 952-6605)

Re: [parisc-linux] problems with PCI IDE controller

[Alan Cox]

On Mon, 2002-12-16 at 06:09, John Marvin wrote:
> > On Sat, 2002-12-14 at 13:10, Matthew Wilcox wrote:
> > > > hp-c240 login: Uniform Multi-Platform E-IDE driver Revision: 6.31
> > > > ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
> > > > PDC20268: IDE controller on PCI bus 00 dev 08
> > > > PDC20268: chipset revision 2
> > > > PDC20268: not 100% native mode: will probe irqs later
> >
> > Is this a plug in card. For some reason it has been left in non native
> > mode so won't work as a plug in board.
> 
> I don't think this is a problem. I have a similar card in my home PC, and
> it prints the same thing (re: not 100% native mode). Here is an excerpt
> from my PC's boot messages running 2.4.19:

A PC is a very different thing

In legacy mode the PCI BAR registers are ignored, the I/O decode is at
0x170 etc and interrupts are flagged by the ISA legacy not by the PCI
bus.

That isnt going to work very well on your average PA-RISC box. Most
probably you want to force the chip into native mode on pa-risc
hardware.

Re: [parisc-linux] problems with PCI IDE controller

[Joerg Steindlberger]

Hi Alan! Great things happened!

I could sweare, that a few weeks ago nothing worked, neither with the nat=
ive=20
IDE drivers nor with PDC20268 options. (That kernel was 2.4.19-pa14). Now=
 it=20
only doesn't work with the special IDE options for this card. Is there a=20
roadmap to fix that prolem? ;-)

Thanks to all who helped me!
Joerg

hp-c240:~# modprobe -k ide-mod
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=3D=
xx
PDC20268: IDE controller on PCI bus 00 dev 08
PDC20268: chipset revision 2
PDC20268: not 100% native mode: will probe irqs later
    ide0: BM-DMA at 0xfd00-0xfd07, BIOS settings: hda:pio, hdb:pio
    ide1: BM-DMA at 0xfd08-0xfd0f, BIOS settings: hdc:pio, hdd:pio
hp-c240:~# modprobe -k ide-probe-mod
hdd: IBM-DTTA-351680, ATA DISK drive
ide1 at 0xfb00-0xfb07,0xfc02 on irq 97
hp-c240:~# modprobe -k ide-disk
hdd: 33022080 sectors (16907 MB) w/462KiB Cache, CHS=3D32760/16/63
 /dev/ide/host0/bus1/target1/lun0: [PTBL] [2055/255/63] p1
hp-c240:~# mount /dev/ide/host0/bus1/target1/lun0/part1 /mnt/tmp
(works with and without devfs)

On Monday 16 December 2002 16:03, you wrote:
> In legacy mode the PCI BAR registers are ignored, the I/O decode is at
> 0x170 etc and interrupts are flagged by the ISA legacy not by the PCI
> bus.
> That isnt going to work very well on your average PA-RISC box. Most
> probably you want to force the chip into native mode on pa-risc
> hardware.

Re: [parisc-linux] still problems with PCI IDE controller

[Joerg Steindlberger]

Hi all!

I still do have problems with using my IDE controller with parisc-linux. =
I=20
get the following message onto my console when generation large traffic t=
o /=20
from my IDE disks:
hp-c240:~# dd if=3D/dev/ide/host0/bus0/target0/lun0/disc of=3D/dev/null
Dino f2000000: IRQ base 96, stuck IRQ lines? 0x2
Dino f2000000: IRQ base 96, stuck IRQ lines? 0x2
Dino f2000000: IRQ base 96, stuck IRQ lines? 0x2
[...]
The device using the IRQ 96 is the internal ethernet controller (not the =
IDE=20
controller!):
hp-c240:~# cat /proc/interrupts | grep 96
 96:      17424            Dino  eth0

Seems that there is still nobody else who tried to use cheap IDE disks wi=
th a=20
parisc machine :-(

For those who are interrested anyway: The reason that causes the kernel c=
rash=20
is the multimode option. You can either select it for native IDE support=20
(CONFIG_IDEDISK_MULTI_MODE=3Dy) or it is automatically selected with the=20
special drivers for i.e. Promise controllers.

Regards
Joerg

> hp-c240:~# modprobe -k ide-mod
> Uniform Multi-Platform E-IDE driver Revision: 6.31
> ide: Assuming 33MHz system bus speed for PIO modes; override with idebu=
s=3Dxx
> PDC20268: IDE controller on PCI bus 00 dev 08
> PDC20268: chipset revision 2
> PDC20268: not 100% native mode: will probe irqs later
>     ide0: BM-DMA at 0xfd00-0xfd07, BIOS settings: hda:pio, hdb:pio
>     ide1: BM-DMA at 0xfd08-0xfd0f, BIOS settings: hdc:pio, hdd:pio
> hp-c240:~# modprobe -k ide-probe-mod
> hdd: IBM-DTTA-351680, ATA DISK drive
> ide1 at 0xfb00-0xfb07,0xfc02 on irq 97
> hp-c240:~# modprobe -k ide-disk
> hdd: 33022080 sectors (16907 MB) w/462KiB Cache, CHS=3D32760/16/63
>  /dev/ide/host0/bus1/target1/lun0: [PTBL] [2055/255/63] p1
> hp-c240:~# mount /dev/ide/host0/bus1/target1/lun0/part1 /mnt/tmp
> (works with and without devfs)

Re: [parisc-linux] still problems with PCI IDE controller

[Grant Grundler]

On Wed, Dec 18, 2002 at 10:08:34PM +0100, Joerg Steindlberger wrote:
> Dino f2000000: IRQ base 96, stuck IRQ lines? 0x2

The loop that reads and deMuxes the IRQ lines for Dino exits after
a few times in order to avoid monopolizing the CPU.
What that msg tells me is we are in the interrupt handler long
enough that the next eth0 interrupt is outstanding again.
I would guess this is because IDE is handling IO on the interrupt
stack (reading block data via PIO).

IDE in PIO Mode (ie not using DMA) does not play well with the
rest of the system.

You can usually ignore this msg.


> Seems that there is still nobody else who tried to use cheap IDE disks with a 
> parisc machine :-(

Nope. IDE is EEEeevil ;^)

grant

Re: [parisc-linux] still problems with PCI IDE controller

[Joerg Steindlberger]

Hi Grant

On Wednesday 18 December 2002 22:24, Grant Grundler wrote:
> IDE in PIO Mode (ie not using DMA) does not play well with the
> rest of the system.
Okay. I now set hdparm -c 1 /dev/... and hdparm -d 1 /dev/... -- before=20
starting the softwareRAID. The RAID1 is syncing this moment with about 2=20
Mbyte/s. I think I'll stop it now and mount the shorter 80pin cable again=
 and=20
play with some speed settings. My earlier problem with DMA was, that the=20
kernal crashed when setting the CONFIG_IDEDMA_AUTO=3Dy.
> Nope. IDE is EEEeevil ;^)
Yes, I know. But what would You pay for two 80 GB disks?

Regards
Joerg

Re: [parisc-linux] still problems with PCI IDE controller

[Grant Grundler]

On Wed, Dec 18, 2002 at 11:18:14PM +0100, Joerg Steindlberger wrote:
...
> Okay. I now set hdparm -c 1 /dev/... and hdparm -d 1 /dev/... -- before 
> starting the softwareRAID. The RAID1 is syncing this moment with about 2 
> Mbyte/s.

Excellent!
TBH, I wasn't expecting that to work.

> I think I'll stop it now and mount the shorter 80pin cable again and 
> play with some speed settings. My earlier problem with DMA was, that the 
> kernal crashed when setting the CONFIG_IDEDMA_AUTO=y.

My guess is the IDE code is poking at registers that don't exist.
Did you post console output from the crash?
I'm just curious if it was an HPMC or kernel trap (data page fault).

> > Nope. IDE is EEEeevil ;^)
> Yes, I know. But what would You pay for two 80 GB disks?

The pro/cons of both are well documented.
I'm not going to re-hash them here.  :^)

grant

ps. To answer your question: Nothing. I collect 20 or so 9GB 10K rpm SCSI
drives from the scrap area and setup raid0/1 as needed. ;^)

Re: [parisc-linux] still problems with PCI IDE controller

[Alan Cox]

On Wed, 2002-12-18 at 22:18, Joerg Steindlberger wrote:
> Okay. I now set hdparm -c 1 /dev/... and hdparm -d 1 /dev/... -- before 
> starting the softwareRAID. The RAID1 is syncing this moment with about 2 
> Mbyte/s. I think I'll stop it now and mount the shorter 80pin cable again and 
> play with some speed settings. My earlier problem with DMA was, that the 
> kernal crashed when setting the CONFIG_IDEDMA_AUTO=y.
> > Nope. IDE is EEEeevil ;^)
> Yes, I know. But what would You pay for two 80 GB disks?

To be fair PIO IDE is evil but UDMA is pretty nice. The big thing
lacking which is slowly coming is TCQ. Given TCQ and SATA hotswap (which
we dont handle at all yet) SCSI is I think rather doomed on the low end.

Alan