Re: 2.6.0-test9 VFAT problem

Re: 2.6.0-test9 VFAT problem

[Vid Strpic]

[-- Attachment #1: Type: text/plain, Size: 730 bytes --]

On Fri, Nov 14, 2003 at 09:39:23AM -0000, Patrick Beard wrote:
> I'm having trouble with my smartmedia. Whenever (occasionally it will
> mount - but very rare) I try to mount it I get the following;
> mount: wrong fs type, bad option, bad superblock on /dev/sda,
> or too many mounted file systems
> dmesg shows;
> FAT: Bogus number of reserved sectors
> VFS: Can't find a valid FAT filesystem on dev sda

Oh, no, not again... :(

We had the same problem few weeks ago... a patch is in the archives, it
worked for me.

-- 
           vms@bofhlet.net, IRC:*@Martin, /bin/zsh. C|N>K
Linux moria 2.6.0-test9 #1 Sat Oct 25 23:00:37 CEST 2003 i686
 12:31:15 up 8 days, 21:48,  1 user,  load average: 0.29, 0.29, 0.27

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.6.0-test9 VFAT problem

[Patrick Beard]

>> FAT: Bogus number of reserved sectors
>> VFS: Can't find a valid FAT filesystem on dev sda

>Oh, no, not again... :(

>We had the same problem few weeks ago... a patch is in the archives, it
>worked for me.

I did apply Andries's 'Relax FAT checking' patch last night but the only
difference I noticed was my belkin reader changed it's error to 'no media
found' the camera still gave 'wrong fs...'

After advice from Philippe and Andries and the following I found on a
website;
"The problem is that Linux only looks at the disk geometry the first time
the camera is plugged in. So when you unplug the camera and change the
memory card Linux does not check to see if the geometry has changed."

I really hope I've not chased my tail by hitting the following combination
'up-patched inode.c' + 'trying to use sda after a deja search' and finally
'Linux only checking the geometry once'. I'd like to put this thread on hold
until I take stock of the steps I've taken just to make sure I've not been a
right 'plum'.

--
Patrick

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 08:46, Patrick Beard wrote:
>>> FAT: Bogus number of reserved sectors
>>> VFS: Can't find a valid FAT filesystem on dev sda

I'm not getting these messages, but I did just now run 
into a weird vfat over usb_storage problem.

Camera is an Olympus C-3020, kernel is 2.6.0-test9-mm3, 
the usual configs to allow access to this camera.

The card was about full, so I turned it on and mounted it 
as usual.  Acessing it with filerunner, I started at the 
top of the list, deleting pix I had already downloaded to 
/usr/pix.  I deleted about 15 in the first batch, then 
around 10 in the next batch, and about 20 in the third batch.
Hi-lighting about another 20 and hitting delete got me a 
read-only file system error on the last one, but it did 
delete the rest of them.  I had about 60 left to delete, 
but had to umount the camera and turn it off, turn it back 
on about 10 secs later, and remount the camera (same device 
BTW, thanks guys) about 5 or 6 times before I was able to 
get that 64meg card all cleaned out.

And once again, I have NDI if its buggy software in the 
camera, or a usb_storage problem.

Here is one such section of the messages log:

Nov 14 09:19:51 coyote kernel: hub 3-2:1.0: new USB device on port 3, assigned address 9
Nov 14 09:19:51 coyote kernel: scsi4 : SCSI emulation for USB Mass Storage devices
Nov 14 09:19:52 coyote kernel:   Vendor: OLYMPUS   Model: C-3020ZOOM(U)     Rev: 1.00
Nov 14 09:19:52 coyote kernel:   Type:   Direct-Access                      ANSI SCSI revision: 02
Nov 14 09:19:52 coyote kernel: SCSI device sda: 128000 512-byte hdwr sectors (66 MB)
Nov 14 09:19:52 coyote kernel: sda: Write Protect is off
Nov 14 09:19:52 coyote kernel: sda: assuming drive cache: write through
Nov 14 09:19:52 coyote kernel:  sda: sda1
Nov 14 09:19:52 coyote kernel: Attached scsi removable disk sda at scsi4, channel 0, id 0, lun 0
Nov 14 09:19:52 coyote kernel: Attached scsi generic sg2 at scsi4, channel 0, id 0, lun 0,  type 0
Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF (i_pos 0)
Nov 14 09:20:34 coyote kernel:     File system has been set read-only

The "assigned address" was being incremented per disconnect.
And, /dev/camera is a link to /dev/sda1.  Is this wrong?
However attempting to mount it useing /dev/sg2 fails, not 
a valid block device.

Comments?  Screwed up kernel .config? Is mount "-t vfat" the 
correct filesystem?

Thanks.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[OGAWA Hirofumi]

Gene Heskett <gene.heskett@verizon.net> writes:

> On Friday 14 November 2003 08:46, Patrick Beard wrote:
> >>> FAT: Bogus number of reserved sectors
> >>> VFS: Can't find a valid FAT filesystem on dev sda

> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF (i_pos 0)
> Nov 14 09:20:34 coyote kernel:     File system has been set read-only

Is this reproduced easy?  Looks like reading the zeroed block.
In order to know details, could you try to read the data without mount?

For example, what happen by the following repetitions? 

	dd if=/dev/sda | md5sum
	unload driver(reset device) or unplug device
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 12:45, OGAWA Hirofumi wrote:
>Gene Heskett <gene.heskett@verizon.net> writes:
>> On Friday 14 November 2003 08:46, Patrick Beard wrote:
>> >>> FAT: Bogus number of reserved sectors
>> >>> VFS: Can't find a valid FAT filesystem on dev sda
>>
>> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
>> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF
>> (i_pos 0) Nov 14 09:20:34 coyote kernel:     File system has been
>> set read-only
>
>Is this reproduced easy?  Looks like reading the zeroed block.
>In order to know details, could you try to read the data without
> mount?
>
>For example, what happen by the following repetitions?
>
>	dd if=/dev/sda | md5sum
>	unload driver(reset device) or unplug device

dd if=/dev/sda1|md5sum <--note use of the same device as to read pix.
has been running for 3 or so minutes now, steadily reading the camera.  
I shoulda put a time in front of it!  Ok got it, heres the sum:
127945+0 records in
127945+0 records out
f6c568dd1f35bb37f3d667a2ab228e2f

its a 64 meg card.  I'll have to unplug the camera (and cycle its 
power too) as all this is compiled into the kernel.
Ok, repeat run, with time this time:
Mmm, didn't get the md5sum, just the time, which was 
real    2m39.807s
user    0m1.052s
sys     0m1.984s
So I'll do another run, without the reset and without the time:
[root@coyote root]# dd if=/dev/sda1 | md5sum
127945+0 records in
127945+0 records out
f6c568dd1f35bb37f3d667a2ab228e2f

so the sums are the same, one more repeat, this time from powerdown 
reset and unplugged.
Purrfect repeat of the above.  Now lets see if /dev/sda will read 
it...  Yes it does.  Am I to understand that the non-numbered version 
of the device is the raw read from LSN0 of the device?
Ahh, that of course gets a different sum, and number of records read:
This is without the reset:
[root@coyote root]# dd if=/dev/sda | md5sum
128000+0 records in
128000+0 records out
b074ba314ac23f234b6d27da78e1e093

reset and restart, replug etc and again:
[root@coyote root]# dd if=/dev/sda | md5sum
128000+0 records in
128000+0 records out
b074ba314ac23f234b6d27da78e1e093

Looks like another flawless repeat here, and I'm probably making a 
serious dent in the batteries by now, that thing is positively 
ravenous when it comes to eating batteries.

What does this tell us?

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.22% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[Maciej Zenczykowski]

> 
> What does this tell us?

Have you tried running a dosfsck on that?

Cheers,
MaZe.

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 14:13, Maciej Zenczykowski wrote:
>> What does this tell us?
>
>Have you tried running a dosfsck on that?
>
No, but here it is, on both sda and sda1:

[root@coyote root]# dosfsck /dev/sda
dosfsck 2.8, 28 Feb 2001, FAT32, LFN
Logical sector size is zero.

[root@coyote root]# dosfsck /dev/sda1
dosfsck 2.8, 28 Feb 2001, FAT32, LFN
/dev/sda1: 2 files, 2/3997 clusters

That first one doesn't look kosher to me!
Q: Is FAT32 the same as VFAT?

>Cheers,
>MaZe.
>
>
>-
>To unsubscribe from this list: send the line "unsubscribe
> linux-kernel" in the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[Maciej Zenczykowski]

> No, but here it is, on both sda and sda1:
> 
> [root@coyote root]# dosfsck /dev/sda
> dosfsck 2.8, 28 Feb 2001, FAT32, LFN
> Logical sector size is zero.

We've already determined that /dev/sda is the partition table and should 
thus fail.

> [root@coyote root]# dosfsck /dev/sda1
> dosfsck 2.8, 28 Feb 2001, FAT32, LFN
> /dev/sda1: 2 files, 2/3997 clusters

Hmm so it passes.
Could you try passing -v (for verbose) to the fsck...
Also perhaps just do
  cat /dev/sda1 > /tmp/file
  dosfsck -v /tmp/file
  mount -o loop /tmp/file /mnt/somewhere
and see if that fails, if so the bug is pure vfat
then try bzipping the /tmp/file and posting it somewhere and pass the link 
and I'll take a look...

> That first one doesn't look kosher to me!
> Q: Is FAT32 the same as VFAT?

FAT32 is one possibility of VFAT - VFAT is any FAT(12,16,32) with long 
filenames.

Cheers,
MaZe.

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 15:30, Maciej Zenczykowski wrote:
>> No, but here it is, on both sda and sda1:
>>
>> [root@coyote root]# dosfsck /dev/sda
>> dosfsck 2.8, 28 Feb 2001, FAT32, LFN
>> Logical sector size is zero.
>
>We've already determined that /dev/sda is the partition table and
> should thus fail.
>
>> [root@coyote root]# dosfsck /dev/sda1
>> dosfsck 2.8, 28 Feb 2001, FAT32, LFN
>> /dev/sda1: 2 files, 2/3997 clusters
>
>Hmm so it passes.
>Could you try passing -v (for verbose) to the fsck...

I could, but a much larger problem is that cups managed to self 
destruct on a reboot, and I now have no printing facility at all, and 
thats more important since I like to print my bank activities for my 
records in case things get out of synch.  However, that shouldn't 
take too much time so here goes.

[root@coyote cups-1.1.20rc6]# dosfsck -v /dev/sda1
dosfsck 2.8 (28 Feb 2001)
dosfsck 2.8, 28 Feb 2001, FAT32, LFN
Boot sector contents:
System ID "        "
Media byte 0xf8 (hard disk)
       512 bytes per logical sector
     16384 bytes per cluster
         1 reserved sector
First FAT starts at byte 512 (sector 1)
         2 FATs, 12 bit entries
      6144 bytes per FAT (= 12 sectors)
Root directory starts at byte 12800 (sector 25)
       256 root directory entries
Data area starts at byte 20992 (sector 41)
      3997 data clusters (65486848 bytes)
32 sectors/track, 8 heads
        55 hidden sectors
    127945 sectors total
Checking for unused clusters.
/dev/sda1: 2 files, 2/3997 clusters
------------------------
The card is now empty, so that looks reasonable to me.  But what do I 
know about a messydos filesystem?  Zilch is what...


>Also perhaps just do
>  cat /dev/sda1 > /tmp/file
>  dosfsck -v /tmp/file
>  mount -o loop /tmp/file /mnt/somewhere
>and see if that fails, if so the bug is pure vfat
>then try bzipping the /tmp/file and posting it somewhere and pass
> the link and I'll take a look...
>
>> That first one doesn't look kosher to me!
>> Q: Is FAT32 the same as VFAT?
>
>FAT32 is one possibility of VFAT - VFAT is any FAT(12,16,32) with
> long filenames.
>
>Cheers,
>MaZe.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[OGAWA Hirofumi]

Gene Heskett <gene.heskett@verizon.net> writes:

> dd if=/dev/sda1|md5sum <--note use of the same device as to read pix.
> has been running for 3 or so minutes now, steadily reading the camera.  
> I shoulda put a time in front of it!  Ok got it, heres the sum:
> 127945+0 records in
> 127945+0 records out
> f6c568dd1f35bb37f3d667a2ab228e2f
> f6c568dd1f35bb37f3d667a2ab228e2f
[...]
> What does this tell us?

Thanks. I want to know it's not reading the randomly garbage.
Can this problem reproduce easy? If so, what operations?
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 15:02, OGAWA Hirofumi wrote:
>Gene Heskett <gene.heskett@verizon.net> writes:
>> dd if=/dev/sda1|md5sum <--note use of the same device as to read
>> pix. has been running for 3 or so minutes now, steadily reading
>> the camera. I shoulda put a time in front of it!  Ok got it, heres
>> the sum: 127945+0 records in
>> 127945+0 records out
>> f6c568dd1f35bb37f3d667a2ab228e2f
>> f6c568dd1f35bb37f3d667a2ab228e2f
>
>[...]
>
>> What does this tell us?
>
>Thanks. I want to know it's not reading the randomly garbage.
>Can this problem reproduce easy? If so, what operations?

NDI OGAWA.  And I rebooted to find that all my printers I had defined 
in cups were gone, and I cannot define new ones, getting 
sever-error-internal-error when I try.  And thats much more important 
to me ATM, sorry.

However, before I rebooted, those operations felt a solid as a rock, 
and always returned the same md5sum on a repeat.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[OGAWA Hirofumi]

Gene Heskett <gene.heskett@verizon.net> writes:

> On Friday 14 November 2003 15:02, OGAWA Hirofumi wrote:
> >Gene Heskett <gene.heskett@verizon.net> writes:
> >> dd if=/dev/sda1|md5sum <--note use of the same device as to read
> >> pix. has been running for 3 or so minutes now, steadily reading
> >> the camera. I shoulda put a time in front of it!  Ok got it, heres
> >> the sum: 127945+0 records in
> >> 127945+0 records out
> >> f6c568dd1f35bb37f3d667a2ab228e2f
> >> f6c568dd1f35bb37f3d667a2ab228e2f
[...]
> However, before I rebooted, those operations felt a solid as a rock, 
> and always returned the same md5sum on a repeat.

Umm... is you talking about my request? Sorry for confusing.
Let me start next step.


>> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
>> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF
>> (i_pos 0) Nov 14 09:20:34 coyote kernel:     File system has been
>> set read-only

I want to reproduce this on my machine. Can you reproduce this easy?
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Saturday 15 November 2003 03:41, OGAWA Hirofumi wrote:
>Gene Heskett <gene.heskett@verizon.net> writes:
>> On Friday 14 November 2003 15:02, OGAWA Hirofumi wrote:
>> >Gene Heskett <gene.heskett@verizon.net> writes:
>> >> dd if=/dev/sda1|md5sum <--note use of the same device as to
>> >> read pix. has been running for 3 or so minutes now, steadily
>> >> reading the camera. I shoulda put a time in front of it!  Ok
>> >> got it, heres the sum: 127945+0 records in
>> >> 127945+0 records out
>> >> f6c568dd1f35bb37f3d667a2ab228e2f
>> >> f6c568dd1f35bb37f3d667a2ab228e2f
>
>[...]
>
>> However, before I rebooted, those operations felt a solid as a
>> rock, and always returned the same md5sum on a repeat.
>
>Umm... is you talking about my request?

Yes.

Sorry for confusing.
>Let me start next step.

Please proceed.
>
>>> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
>>> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF
>>> (i_pos 0) Nov 14 09:20:34 coyote kernel:     File system has been
>>> set read-only
>
>I want to reproduce this on my machine. Can you reproduce this easy?

Apparently so, but I'd have to go out and waste a set of batteries 
loading the card up again, its presently empty.  However, I suspect 
you may have something else on your mind. :)

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[OGAWA Hirofumi]

Gene Heskett <gene.heskett@verizon.net> writes:

> >>> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
> >>> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF
> >>> (i_pos 0) Nov 14 09:20:34 coyote kernel:     File system has been
> >>> set read-only
> >
> >I want to reproduce this on my machine. Can you reproduce this easy?
> 
> Apparently so, but I'd have to go out and waste a set of batteries 
> loading the card up again, its presently empty.  However, I suspect 
> you may have something else on your mind. :)

What meaning is it? What do I misunderstand?
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: 2.6.0-test9 VFAT problem

[Andries Brouwer]

On Fri, Nov 14, 2003 at 09:45:36AM -0500, Gene Heskett wrote:

> Nov 14 09:19:51 coyote kernel: hub 3-2:1.0: new USB device on port 3, assigned address 9
> Nov 14 09:19:51 coyote kernel: scsi4 : SCSI emulation for USB Mass Storage devices
> Nov 14 09:19:52 coyote kernel:   Vendor: OLYMPUS   Model: C-3020ZOOM(U)     Rev: 1.00
> Nov 14 09:19:52 coyote kernel:   Type:   Direct-Access                      ANSI SCSI revision: 02
> Nov 14 09:19:52 coyote kernel: SCSI device sda: 128000 512-byte hdwr sectors (66 MB)
> Nov 14 09:19:52 coyote kernel: sda: assuming drive cache: write through
> Nov 14 09:19:52 coyote kernel:  sda: sda1
> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF (i_pos 0)
> Nov 14 09:20:34 coyote kernel:     File system has been set read-only
> 
> Comments?  Screwed up kernel .config? Is mount "-t vfat" the 
> correct filesystem?

It would have been interesting to see the filesystem after this error message.
Can you reproduce the error?

(vfat? I don't know - most cameras just use msdos, but vfat doesnt harm,
I suppose)

The error message means that the fatfs followed a chain of clusters in order
to delete them all and found a free cluster before finding an end-of-file mark.

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 15:23, Andries Brouwer wrote:
>On Fri, Nov 14, 2003 at 09:45:36AM -0500, Gene Heskett wrote:
>> Nov 14 09:19:51 coyote kernel: hub 3-2:1.0: new USB device on port
>> 3, assigned address 9 Nov 14 09:19:51 coyote kernel: scsi4 : SCSI
>> emulation for USB Mass Storage devices Nov 14 09:19:52 coyote
>> kernel:   Vendor: OLYMPUS   Model: C-3020ZOOM(U)     Rev: 1.00 Nov
>> 14 09:19:52 coyote kernel:   Type:   Direct-Access                
>>      ANSI SCSI revision: 02 Nov 14 09:19:52 coyote kernel: SCSI
>> device sda: 128000 512-byte hdwr sectors (66 MB) Nov 14 09:19:52
>> coyote kernel: sda: assuming drive cache: write through Nov 14
>> 09:19:52 coyote kernel:  sda: sda1
>> Nov 14 09:20:34 coyote kernel: FAT: Filesystem panic (dev sda1)
>> Nov 14 09:20:34 coyote kernel:     fat_free: deleting beyond EOF
>> (i_pos 0) Nov 14 09:20:34 coyote kernel:     File system has been
>> set read-only
>>
>> Comments?  Screwed up kernel .config? Is mount "-t vfat" the
>> correct filesystem?
>
>It would have been interesting to see the filesystem after this
> error message. Can you reproduce the error?
>
>(vfat? I don't know - most cameras just use msdos, but vfat doesnt
> harm, I suppose)
>
>The error message means that the fatfs followed a chain of clusters
> in order to delete them all and found a free cluster before finding
> an end-of-file mark.

And I could possibly have created that empty sector by not treating 
the card as a LIFO I suppose...  Makes a certain amount of sense, but 
why did a powerdown and a reconnect fix it?  Some compacting routine 
in the cameras soft maybe?

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[Parick Beard]

> I'd like to put this thread on hold
> until I take stock of the steps I've taken just to make sure I've not been a
> right 'plum'.

Hi,
Right, I'm not sure if I've been a 'plum' or not. In fact I'm not sure
about anything now.
Friday when I got home I tried to mount my belkin reader with sda in the
fstab. I got 'no media found'. I then tried to mount my Olympus camera and
got 'wrong fs...'. I changed my fstab to sda1 and tried again. This gave
the same two errors. I rebooted the pc and this time I could mount the
camera, but the belkin usb reader still gave 'no media found'. I then put
the original inode.c back into my source and recompiled. And to my
complete surprise the camera mounted. The belkin however gave 'no media
found'.
All the above would suggest I've been a 'plum' - however I've done this
before with 2.4.  I do tend to be able to get things working by reading
and not bothering good people like yourselves.

Now I have done some further tests and I have found something strange. I
can mount my camera with a 64mb and 16mb smart media. The belkin reader,
if I put the 64mb card in, I get 'no media found'. If I then put the 16mb
card in I also get 'no media found'. If I disconnect and reconnect the
belkin I can mount the 16mb card. If I then put the 64mb in I get 'no
media found'. Putting the 16mb back in gives 'no media found'.
disconnecting and reconnecting the reader allows me to mount the 16mb card.

What I'm trying to say is the reader won't mount the 64mb card. I could
mount it with no problems under 2.4 with this reader. 
The other strange thing is that prior to me patching inode.c both the
camera and reader gave the same 'wrong fs...' errors with the 64mb card.
Yet as I've said I put the original inode.c back and recompiled. yet the
reader with the 64mb card will only give 'no media found'.

with the 64mb card mounted in the camera fdisk -l /dev/sda give me this;

/dev/sda1	*	1	500	63972+	1	FAT12

Any advice would be appreciated.

__
Patrick

Re: 2.6.0-test9 VFAT problem

[Vid Strpic]

[-- Attachment #1: Type: text/plain, Size: 1698 bytes --]

On Fri, Nov 14, 2003 at 01:46:51PM -0000, Patrick Beard wrote:
> >> FAT: Bogus number of reserved sectors
> >> VFS: Can't find a valid FAT filesystem on dev sda
> >Oh, no, not again... :(
> >We had the same problem few weeks ago... a patch is in the archives, it
> >worked for me.
> I did apply Andries's 'Relax FAT checking' patch last night but the only
> difference I noticed was my belkin reader changed it's error to 'no media
> found' the camera still gave 'wrong fs...'

Hmmm.

> After advice from Philippe and Andries and the following I found on a
> website;
> "The problem is that Linux only looks at the disk geometry the first time
> the camera is plugged in. So when you unplug the camera and change the
> memory card Linux does not check to see if the geometry has changed."

I didn't know that... it shouldn't happen?  I don't connect my camera by
usb-storage, I have a card reader, and it really behaves as removable
media device, as it should - checks size everytime.  Maybe we should
change something here?

> I really hope I've not chased my tail by hitting the following combination
> 'up-patched inode.c' + 'trying to use sda after a deja search' and finally
> 'Linux only checking the geometry once'. I'd like to put this thread on hold
> until I take stock of the steps I've taken just to make sure I've not been a
> right 'plum'.

Well, those things sometimes take a lot of time.

-- 
           vms@bofhlet.net, IRC:*@Martin, /bin/zsh. C|N>K
Linux moria 2.6.0-test9 #1 Sat Oct 25 23:00:37 CEST 2003 i686
 10:05:31 up 9 days, 19:22,  1 user,  load average: 0.30, 0.48, 0.29
Eagles may soar, but weasels don't get sucked into jet engines.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

RE: 2.6.0-test9 VFAT problem

[Patrick Beard]

> > My fstab entry is;
> > /dev/sda    /mnt/smedia    vfat    rw,user,noauto    0,0
> 
> I would guess that you have to mount /dev/sda1 or perhaps /dev/sda4.
> Isn't that what you do under 2.4?
> 
> Andries

Hi Andries,

Yes, with 2.4 I used sda1. When I first compiled 2.6 I used sda1 but I got
the 'wrong fs..' error. This was a clean install of debian so I didn't have
my original fstab. I checked the web and noticed people using sda. so I
tried that - same error. In trying to get this to work I've used sda and
sda1 at different times both gave the same errors.

Thanks for your reply,

--
Patrick 


Whilst we have taken every effort to ensure that all attachments 
are virus free, use is at the recipients own risk.

_________________ANTI_VIRUS PROTECTION___________________ 

This message has been checked for all known viruses by Norman Anti-Virus. 
For further information visit http://www.norman.com 

Re: 2.6.0-test9 VFAT problem

[Andries Brouwer]

On Fri, Nov 14, 2003 at 11:51:57AM -0000, Patrick Beard wrote:
> > > My fstab entry is;
> > > /dev/sda    /mnt/smedia    vfat    rw,user,noauto    0,0
> > 
> > I would guess that you have to mount /dev/sda1 or perhaps /dev/sda4.
> > Isn't that what you do under 2.4?
> 
> Yes, with 2.4 I used sda1. When I first compiled 2.6 I used sda1 but I got
> the 'wrong fs..' error. This was a clean install of debian so I didn't have
> my original fstab. I checked the web and noticed people using sda. so I
> tried that - same error. In trying to get this to work I've used sda and
> sda1 at different times both gave the same errors.

Good.

Maybe you know already, but just to be sure:
You must use sda if the thing has no partition table.
You must use sda1 if the thing has a partition table.

So, if sda1 works under 2.4, then sda is certainly wrong under 2.6 -
your device would just look like garbage and the error messages are
meaningless.
Only try sda1, and report whatever goes wrong in that case.

Andries

Re: 2.6.0-test9 VFAT problem

[Gene Heskett]

On Friday 14 November 2003 14:03, Andries Brouwer wrote:
>On Fri, Nov 14, 2003 at 11:51:57AM -0000, Patrick Beard wrote:
>> > > My fstab entry is;
>> > > /dev/sda    /mnt/smedia    vfat    rw,user,noauto    0,0
>> >
>> > I would guess that you have to mount /dev/sda1 or perhaps
>> > /dev/sda4. Isn't that what you do under 2.4?
>>
>> Yes, with 2.4 I used sda1. When I first compiled 2.6 I used sda1
>> but I got the 'wrong fs..' error. This was a clean install of
>> debian so I didn't have my original fstab. I checked the web and
>> noticed people using sda. so I tried that - same error. In trying
>> to get this to work I've used sda and sda1 at different times both
>> gave the same errors.
>
>Good.
>
>Maybe you know already, but just to be sure:
>You must use sda if the thing has no partition table.
>You must use sda1 if the thing has a partition table.
>
>So, if sda1 works under 2.4, then sda is certainly wrong under 2.6 -
>your device would just look like garbage and the error messages are
>meaningless.
>Only try sda1, and report whatever goes wrong in that case.
>
>Andries

I just played 210 monkeys, and /dev/sda doesn't work for mount, but 
sda1 does.  Both work for a read by dd FWTW.  Links suitably 
restored.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.

Re: 2.6.0-test9 VFAT problem

[OGAWA Hirofumi]

Andries Brouwer <aebr@win.tue.nl> writes:

> > Yes, with 2.4 I used sda1. When I first compiled 2.6 I used sda1 but I got
> > the 'wrong fs..' error. This was a clean install of debian so I didn't have
> > my original fstab. I checked the web and noticed people using sda. so I
> > tried that - same error. In trying to get this to work I've used sda and
> > sda1 at different times both gave the same errors.
[...]
> Only try sda1, and report whatever goes wrong in that case.

Ah, I see. His error message are...
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: 2.6.0-test9 VFAT problem

[Patrick Beard]

> So, if sda1 works under 2.4, then sda is certainly wrong under 2.6 -
> your device would just look like garbage and the error messages are
> meaningless.
> Only try sda1, and report whatever goes wrong in that case.

Andries,
Thanks for your answer.
After what you and Philippe told me I wanted to do a sanity check on what
I'd done.

I don't leave my system on, and so when I have photos to get off the camera
I start the system insert my 64mb Olympus SM into my Belkin SM reader (USB)
mount it, and get the photos off of it. This explains why I've not run into
'Linux only checking the geometry on first connection' thing. As it turns
out sda1 does work on my system, even without your patch. However I can only
mount my 16mb Olympus SM card in my Belkin reader (something I rarely do).
If I try to mount my 64mb card I get 'no media found'. As I normally don't
use the 16mb card I think this problem together with the 'geometry checking'
led to my confusion of the problem. I say this as once I've tried to mount
the 64mb card in my reader and get the 'no media...', then Linux will report
this error for every card until the reader is disconnected and reconnected.

With that confusion now clarified in my own head here is what I know for
sure and can reproduce every time;

1. I am now using sda1 (sorry to the group for using sda and reporting this
as a problem).
2. I can mount both my 16mb and 64mb Olympus SM cards when using my Olympus
300-zoom camera as the reader.
3. I can mount the 16mb card when using my Belkin USB SM reader.
4. I cannot mount my 64mb card using the Belkin reader. Linux reports 'No
Media Found'.
5. I cannot mount my 16mb card in the Belkin reader after I've tried to
mount the 64mb card. However disconnecting and reconnecting the reader
allows me once again to mount the 16mb card.

Both smartmedia cards I have (16 & 64mb), are 'Olympus' branded and were
both formatted in the Olympus 300-zoom camera. I have reformatted both of
them but this didn't help.
Both cards worked in the Belkin reader when I was using the Debian 2.4.19
kernel source.

Philippe asked me to do a fdisk -l on /dev/sda He asked this when I reported
my first (non problem), and so this might not be relevant but here is what I
get back when I mount the 64mb using the camera (with the reader I obviously
don't get anything back);

/dev/sda1 * 1 500 63972+ 1 FAT12

Any advice would be appreciated.

__
Patrick

2.6.0-test9 VFAT problem

[Patrick Beard]

Hi,

First, I'm still very much a newbie to Linux - but I try my best.

I'm having trouble with my smartmedia. Whenever (occasionally it will
mount - but very rare) I try to mount it I get the following;

mount: wrong fs type, bad option, bad superblock on /dev/sda,
or too many mounted file systems

dmesg shows;

FAT: Bogus number of reserved sectors
VFS: Can't find a valid FAT filesystem on dev sda

These are Olympus sm cards, my camera is a Olympus 300-zoom. I use both the
camera (usb connection) and a belkin usb smartmedia reader.
These cards camera and reader all worked well using Kernel 2.4.18.

My distrib is Debian (testing) and the 2.6 source is the 'Debianized' source
within testing. I noticed that Andries Brouwer has created a patch to relax
the FAT checking within /fs/fat/inode.c. I have applied that diff and
recompiled but if anything it now appears worse. That is to say I still get
the original error when trying to mount the camera, but this time the belkin
reader returns 'no media found'. I will try to confirm this by recompiling
with the original inode.c.

My fstab entry is;
/dev/sda    /mnt/smedia    vfat    rw,user,noauto    0,0

As I said at the start the media does mount occasionally so I would presume
I haven't missed something.

Can anyone help?

--
Patrick

Re: 2.6.0-test9 VFAT problem

[Andries Brouwer]

On Fri, Nov 14, 2003 at 09:39:23AM -0000, Patrick Beard wrote:

> I'm having trouble with my smartmedia.
> 
> FAT: Bogus number of reserved sectors
> VFS: Can't find a valid FAT filesystem on dev sda
> 
> These are Olympus sm cards, my camera is a Olympus 300-zoom.
> These cards camera and reader all worked well using Kernel 2.4.18.
> 
> My fstab entry is;
> /dev/sda    /mnt/smedia    vfat    rw,user,noauto    0,0

I would guess that you have to mount /dev/sda1 or perhaps /dev/sda4.
Isn't that what you do under 2.4?

Andries