Compact flash on mpc8349eITX

I'm trying to get the compact flash up and running on an mpc8349eITX,
and am continually running into issues where the IRQ is not processed,
and the mmio commands are not getting response. At this point, I am
starting to suspect a HW issue, but I was hoping to verify that I am
using a known good kernel configuration.

Has anyone successfully communicated with a compact flash device in true
IDE mode on the MPC8349eITX? If so, do you happen to know the kernel
configuration and/or revision?

I appreciate any help, as I am really getting stuck here.
Regards,
Sam

On Fri, Aug 01, 2008 at 10:58:58AM -0500, Sparks, Sam wrote:

I'm trying to get the compact flash up and running on an mpc8349eITX,
and am continually running into issues where the IRQ is not processed,
and the mmio commands are not getting response. At this point, I am
starting to suspect a HW issue, but I was hoping to verify that I am
using a known good kernel configuration.

Has anyone successfully communicated with a compact flash device in true
IDE mode on the MPC8349eITX? If so, do you happen to know the kernel
configuration and/or revision?

Works like a charm here. I use quite outdated u-boot though. Which one
you use? I could try it (if it is community u-boot).

As for the backtrace you posted recently, I remeber I've noticed the
similar issue long ago on some mITX board. I believe the board had
hw issues indeed (or CF card I tested with). The problem could be in
the IRQ line itself (for some reason it always low), or it could be
IDE cmd/data lines at fault, so that IDE can't receive status read
command and does not deassert the interrupt line... quite hard to
tell.

The logs below.

(Notice that u-boot says that UPMA is configured to work with CF IDE).

U-Boot 1.2.0-g5b1313fb (Jun  9 2007 - 17:29:42) MPC83XX

CPU:   e300c1, MPC8349E, Rev: 11 at 533.333 MHz, CSB:  266 MHz
Board: Freescale MPC8349E-mITX
UPMA:  Configured for compact flash
I2C:   ready
DRAM:
   DDR DIMM: data bus width is 64 bit without ECC
   DDRC ECC mode: OFF
   DDR RAM: 256 MB
FLASH: 16 MB
PCI:   Bus Dev VenId DevId Class Int
        00  10  1095  3114  0180  00
In:    serial
Out:   serial
Err:   serial
Board revision: 1.0 (PCF8475A)
Net:   TSEC0
IDE:   Bus 0: OK
  Device 0: Model: KINGSTON Firm: 20060713 Ser#: KINGSTON CF 00003E2C
            Type: Hard Disk
            Capacity: 495.1 MB = 0.4 GB (1014048 x 512)
MPC8349E-mITX> boot
[...]

30 more lines

From: Anton Vorontsov
Sent: Friday, August 01, 2008 11:51 AM

Works like a charm here. I use quite outdated u-boot though. Which one
you use? I could try it (if it is community u-boot).

I'm using the head, which I grabbed last week: U-Boot
1.3.4-rc1-00012-g1953d12
TIA for checking your itx with this version.

As for the backtrace you posted recently, I remeber I've noticed the
similar issue long ago on some mITX board. I believe the board had
hw issues indeed (or CF card I tested with). The problem could be in
the IRQ line itself (for some reason it always low), or it could be
IDE cmd/data lines at fault, so that IDE can't receive status read
command and does not deassert the interrupt line... quite hard to
tell.

Thanks for confirming.. I was losing my mind :-)
This looks like it is indeed a HW problem with the ITX. Are you aware of
any threads/documents describing the issue? We based our custom HW off
the ITX schematic with regards to the CF, and I hope we didn't copy over
_that_ problem :-/

Cheers,
Sam

On Fri, Aug 01, 2008 at 12:36:39PM -0500, Sparks, Sam wrote:

quoted

From: Anton Vorontsov
Sent: Friday, August 01, 2008 11:51 AM

Works like a charm here. I use quite outdated u-boot though. Which one
you use? I could try it (if it is community u-boot).

I'm using the head, which I grabbed last week: U-Boot
1.3.4-rc1-00012-g1953d12
TIA for checking your itx with this version.

U-Boot 1.3.4-rc1-00012-g1953d12 (Aug  1 2008 - 21:45:42) MPC83XX

The CF still works.

quoted

As for the backtrace you posted recently, I remeber I've noticed the
similar issue long ago on some mITX board. I believe the board had
hw issues indeed (or CF card I tested with). The problem could be in
the IRQ line itself (for some reason it always low), or it could be
IDE cmd/data lines at fault, so that IDE can't receive status read
command and does not deassert the interrupt line... quite hard to
tell.

Thanks for confirming.. I was losing my mind :-)
This looks like it is indeed a HW problem with the ITX. Are you aware of
any threads/documents describing the issue? We based our custom HW off
the ITX schematic with regards to the CF, and I hope we didn't copy over
_that_ problem :-/

I don't think it was a hw design problem, more likely somebody
fried the CF slot on that board (either by hot-plugging the CF, or via
inserting CF 5V only flash.. either of this is forbidden for the
8349mITX boards).

-- 
Anton Vorontsov
email: cbouatmailru@gmail.com
irc://irc.freenode.net/bd2

On Friday 01 August 2008, Sparks, Sam wrote:

Has anyone successfully communicated with a compact flash device in true
IDE mode on the MPC8349eITX? If so, do you happen to know the kernel
configuration and/or revision?

There are some CF cards and some IDE-CF adapters that are missing the lines
for DMA mode, even if the controller on the card supports DMA.
If a specific card doesn't work, try using a different one, or use PIO mode
only.

	Arnd <><