2001-04-03 00:08:51

by Stefan Linnemann

[permalink] [raw]
Subject: Sandisk flashcard reader on 2.4.2. It works. Sort of.

PLease Cc: all responses to [email protected].

Hiya,

since I'm not on the list, and my experience with and knowledge of kernel
internals are negligable, please keep in mind to Carbon copy all forcoming
relevant conversation to me. Thank you.

Ok, I bought a Nikon Coolpix 990, aand to download the ppix to my Linux PC at
home a SanDisk ImageMate SDDR-01. That's parallel IDE. At the time, I ran
2.2.17, which of course didn't touch the thing with a 10 parsecs pole. Som
searching on the web led me to two conclusions:

1. This thing is either new or scorned.
2, There is a patch at http://www.electricgod.net/~moomonk/epat/ which may or
may not help.

All the time during this story a 16 Mb SanDisk CompactFlash card is in the
reader, unles otherwise specified.

So I applied the patch to the 2.2.17 kernel, remade, and rebooted. No dice.
Several kernel reconfigurations and rebots later still nothing, it seemed.
So I downloaded and installed 2.4.2, which was mentioned to problably include
the necessary features. I copied .config from the 2.2.17, superficially
checked the config, and remade and rebooted.

This was where I noted, that the parport, paride, epat and pd modules didn't
get installed as modules at all. I havnet dug into the why of that, let
those familiar with the processes and Makefiles do that. So I reconfigured
to get those into the kernel, and remade and rebooted. No dice, so I
succesfully again applied the same patch, configured it into the kernel and
remade and rebooted. No SanDisk. For some reason or another I rebooted
again, and lo and behold, we have a SanDisk. I mount it ok, cd
/sandisk/dir/, mv * elsewhere, my system hangs. Reset. cp * hangs the
system as well, so I do a for i in *; do; cp $i /there; sync; done;

The system hangs whenever I try to copy a certain picture. (No, it doesn't
contain nudiepixels.) Luckily, I didn't care much about that one. Since the
start of this story and "now" I've aqcuired an 128 Mb card, and shot some
pictures using it. So I unmount my 16 Mb one, swicth cards, and mount the
device again. Or so I hoped. System hangs on the mount operation. And no
CTRL-ALT-DEL to go down more or less gracefully. It really hangs, like
whenever the patched system barfs on the SanDisk.

And today, several days later, the system even refuses to see the little
bugger, even the lasst time, without any card it it's waiting little mouth.

So the message is: Yes, it could work, but with the patch from
http://www.electricgod.net/~moomonk/epat/ it's slightly better working than
without it.

What am I to do?

With kind regards,
Stefan.
Please Cc: reactions to me.
--
Stefan Linnemann http://www.xs4all.nl/~mazur/
Systems programmer Unix ICQ: 25314387


2001-04-03 17:29:33

by Tim Waugh

[permalink] [raw]
Subject: Re: Sandisk flashcard reader on 2.4.2. It works. Sort of.

On Tue, Apr 03, 2001 at 02:08:13AM +0200, Stefan Linnemann wrote:

> the necessary features. I copied .config from the 2.2.17, superficially
> checked the config, and remade and rebooted.
>
> This was where I noted, that the parport, paride, epat and pd modules didn't
> get installed as modules at all. I havnet dug into the why of that, let
> those familiar with the processes and Makefiles do that.

It'll be because of the block device directory reorganisation I
expect, or something similar. Double-check your config.

> So I reconfigured to get those into the kernel, and remade and
> rebooted. No dice, so I succesfully again applied the same patch,
> configured it into the kernel and remade and rebooted. No
> SanDisk. For some reason or another I rebooted again, and lo and
> behold, we have a SanDisk.

So the kernel you run which can see the SanDisk is with, or without,
the C7/8 patch?

> I mount it ok, cd
> /sandisk/dir/, mv * elsewhere, my system hangs. Reset.

Enable magic-sysrq and see if Alt-SysRq-B reboots the machine or not.
Or, even better, jot down what Alt-SysRq-T says.

> So the message is: Yes, it could work, but with the patch from
> http://www.electricgod.net/~moomonk/epat/ it's slightly better working than
> without it.

This patch is in the queue, but behind the bug-fixes.

You might want to try fiddling with the BIOS options for the parallel
port and see if that makes any difference.

Tim.
*/


Attachments:
(No filename) (1.41 kB)
(No filename) (232.00 B)
Download all attachments

2001-04-04 00:48:43

by Stefan Linnemann

[permalink] [raw]
Subject: Re: Sandisk flashcard reader on 2.4.2. It works. Sort of.

On Tuesday 03 April 2001 19:16, Tim Waugh wrote:

> > On Tue, Apr 03, 2001 at 02:08:13AM +0200, Stefan Linnemann wrote:
> > the necessary features. I copied .config from the 2.2.17, superficially
> > checked the config, and remade and rebooted.

> > This was where I noted, that the parport, paride, epat and pd modules
> > didn't get installed as modules at all. I haven't dug into the why of
> > that, let those familiar with the processes and Makefiles do that.

> It'll be because of the block device directory reorganisation I
> expect, or something similar. Double-check your config.

Config is fine, it's just make modules_install that's ignoring them.

> > So I reconfigured to get those into the kernel, and remade and
> > rebooted. No dice, so I succesfully again applied the same patch,
> > configured it into the kernel and remade and rebooted. No
> > SanDisk. For some reason or another I rebooted again, and lo and
> > behold, we have a SanDisk.

> So the kernel you run which can see the SanDisk is with, or without,
> the C7/8 patch?

With both 2.2.17 and 2.4.2, only with the patch, and it reports a c7 chip.
The only times it did get recognized the 16 Mb SanDisk CompactFlash card
(EC-16CF) was in the reader. Though even that now doesn't seem to help
anymore. One clue only remains to be told: ever since installing the patch I
get one error message lots of time: "invalid character 46 in exportstr for
pd.drive0". It's even printed at bootup from almost every init script.

> > I mount it ok, cd
> > /sandisk/dir/, mv * elsewhere, my system hangs. Reset.

> Enable magic-sysrq and see if Alt-SysRq-B reboots the machine or not.
> Or, even better, jot down what Alt-SysRq-T says.

It is in, and was in, I only had completely forgotten about that, never
having had a need for it yet.

> > So the message is: Yes, it could work, but with the patch from
> > http://www.electricgod.net/~moomonk/epat/ it's slightly better working
> > than without it.

> This patch is in the queue, but behind the bug-fixes.

That, I figured. Which is why I bothered the mailing list in the first
place, so you know there are some issues with the patch as it is.

> You might want to try fiddling with the BIOS options for the parallel
> port and see if that makes any difference.

The only options I get in BIOS for my parallel port are Output-Only,
Bi-Directional, EPP and ECP. ECP was the setting, and changing that to EPP
and Bi_Directional only removed some of the protocols reported by the OS, so
I'm back to ECP now.

I'll include a dmesg diff between one time he did recognize the thing and the
current one:

*** dmesg Wed Apr 4 02:03:56 2001
--- dmesg.sandisk Fri Mar 30 16:45:40 2001
***************
*** 9,17 ****
zone(0): 4096 pages.
zone(1): 36864 pages.
zone(2): 0 pages.
! Kernel command line: BOOT_IMAGE=linux ro root=301 hisax=3,2,10,0x180,HiSax
opl3sa2=0x370,5,0,3,0x530,0x330 pd.drive0=0x378
Initializing CPU#0
! Detected 233.290 MHz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 465.30 BogoMIPS
Memory: 158892k/163840k available (1118k kernel code, 4560k reserved, 374k
data, 84k init, 0k highmem)
--- 9,17 ----
zone(0): 4096 pages.
zone(1): 36864 pages.
zone(2): 0 pages.
! Kernel command line: auto BOOT_IMAGE=linux ro root=301
hisax=3,2,10,0x180,HiSax opl3sa2=0x370,5,0,3,0x530,0x330 pd.drive0=0x378
Initializing CPU#0
! Detected 233.294 MHz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 465.30 BogoMIPS
Memory: 158892k/163840k available (1118k kernel code, 4560k reserved, 374k
data, 84k init, 0k highmem)
***************
*** 72,79 ****
hdc: hdc1 hdc2
paride: epat registered as protocol 0
pd: pd version 1.05, major 45, cluster 64, nice 0
! epat_init_protopda: Autoprobe failed
! pd: no valid drive found
Floppy drive(s): fd0 is 1.44M
FDC 0 is a National Semiconductor PC87306
Serial driver version 5.02 (2000-08-09) with MANY_PORTS MULTIPORT SHARE_IRQ
ISAPNP enabled
--- 72,81 ----
hdc: hdc1 hdc2
paride: epat registered as protocol 0
pd: pd version 1.05, major 45, cluster 64, nice 0
! epat_init_protopda: Sharing parport0 at 0x378
! pda: epat 1.02, Shuttle EPAT chip c7 at 0x378, mode 2 (8-bit), delay 1
! pda: SanDisk SDCFB-, master, 31360 blocks [15M], (490/2/32), removable media
! pda: pda1
Floppy drive(s): fd0 is 1.44M
FDC 0 is a National Semiconductor PC87306
Serial driver version 5.02 (2000-08-09) with MANY_PORTS MULTIPORT SHARE_IRQ
ISAPNP enabled

Thanks for the reply, anyway,
Stefan.
--
Stefan Linnemann http://www.xs4all.nl/~mazur/
Systeem programmeur Unix ICQ: 25314387