RE: Problem with PCMCIA and RPXCLASSIC

Hi Mathew Locke,

The pcmcia-3.1.22 has compilation problems as the mpc8xx_bdinfo >is not
defined.  It is used only for the frequency, hence we hard >coded it to
40MHz(which is our boards operating freq).

yeah, i forgot to export that in my patch.

Then we gave _IO_BASE = 0x80000000
This still did not work and the kernel panics.  We had to >comment out
a line in ll_rw_blk.c( where and outb_p is called)..
After this the kernel comes up.

So, IO_BASE = 0x80000000 does work, ll_rw_blk.c is an unrelated bug.

Wpe suspect some goof up in the file: >arch/ppc/kernel/m8xx_setup.c
Here ATA_FLASH is defined to 1.(and the comment is:
 /* Define this to make a PCMCIA ATA Flash card work.
 */
)
But with this define, the function rpx_pcmcia_config() doesnt >get
called at all.

It may be a little confusing, but since it works for me I doubt there is a
goof up. The confusing part is that there is code in there to bypass card
services for ATA cards.  Instead of deleting it, I left it so that people
who don't want the overhead of card services (my guess is that it doesn't
really affect perfomance that much) don't have to use it.

Does anyone have a patch for this file?

you have the patch that works.

Also, what should be the values in the config.opts for this?

config.opts is not used for the 8xx.

you can send me a microdrive and I will figure out what the problem is.

I HAVE READ IN THE MAILING LIST from Dan dated TUE 02, may 2000
that he has successfully used IBM Micro drive with MPC860 PCMCIA.
has anybody got the detail information ?

I don't think Dan uses card services.  He has code in the kernel that
connects the ide driver directly to the 8xx pcmcia controller.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Matthew Locke wrote:

quoted

Then we gave _IO_BASE = 0x80000000
This still did not work and the kernel panics.  We had to >comment out
a line in ll_rw_blk.c( where and outb_p is called)..
After this the kernel comes up.

So, IO_BASE = 0x80000000 does work, ll_rw_blk.c is an unrelated bug.

What are you guys using for a kernel?  This outb_p "bug" was fixed
in ancient times (or you didn't properly configure the kernel).

It may be a little confusing, but since it works for me I doubt there is a
goof up. The confusing part is that there is code in there to bypass card
services for ATA cards.

Yes, this works.  It sounds like you are mixing the internal ATA hack
with card services.  You can't use both........

quoted

I HAVE READ IN THE MAILING LIST from Dan dated TUE 02, may 2000
that he has successfully used IBM Micro drive with MPC860 PCMCIA.
has anybody got the detail information ?

I don't think Dan uses card services.  He has code in the kernel that
connects the ide driver directly to the 8xx pcmcia controller.

That's right, no card services when I wrote that message.  I just
used the PCMCIA interface as a dedicated ATA interface.  The purpose
was simply to verify the hardware would work correctly, and to provide
a method for embedded systems to use this interface directly without
card services.


	-- Dan

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/