Very slow disk access via PCMCIA adapter

[Peter Stuge]

On Sat, Aug 11, 2007 at 08:35:56AM +0100, Simon Atkinson wrote:
> The stages described above take similar lengths of time when I
> issue the equivalent commands via a terminal window at the command
> line.

This is the only metric that should be used to diagnose the problem.


> Any ideas what the problem is?

I think there are several.


> cardmgr

This is the first. cardmgr is deprecated and should not be used.


> Linux version 2.6.19

There are many newer kernels out there..


> Yenta: ISA IRQ mask 0x06b8, PCI irq 11

Everything in your laptop seems to use IRQ 11.


> pcmcia: Detected deprecated PCMCIA ioctl usage from process: cardmgr.
> pcmcia: This interface will soon be removed from the kernel; please
> expect breakage unless you upgrade to new tools.
> pcmcia: see http://www.kernel.org/pub/linux/utils/kernel/pcmcia/pcmcia.html
> for details.

So, please check that webpage out, and get rid of cardmgr.


> pccard: PCMCIA card inserted into slot 1
> cs: memory probe 0xa0000000-0xa0ffffff: excluding 0xa0000000-0xa0ffffff
> cs: memory probe 0x60000000-0x60ffffff: clean.
> pcmcia: registering new device pcmcia1.0
> Probing IDE interface ide2...
> hde: IBM-DSCM-10512, CFA DISK drive
> ide2 at 0x100-0x107,0x10e on irq 3
> hde: max request size: 128KiB
> hde: 1052352 sectors (538 MB) w/60KiB Cache, CHS=1044/16/63

For some reason the card gets irq 3 while the controller has irq 11.


>  hde:<4>hde: lost interrupt
> hde: lost interrupt
> hde: lost interrupt
> hde: lost interrupt
>  hde1
> ide-cs: hde: Vpp = 0.0
> hde: lost interrupt
> hde: lost interrupt

And here is some fun. No interrupts are coming through, which
probably leads to long timeouts here and there in the IDE code.


Try 2.6.22.1 with pata_pcmcia, get udev and pcmciautils and get rid
of cardmgr. That may just work out of the box.


//Peter