Why can't I flush /dev/ram0?

Why can't I flush /dev/ram0?

[Kelledin Tane]

At this point, I'm trying to get an initrd working properly.  So far, it
works, the system boots, etc. etc., but whenever I try to do a "blockdev
--flushbufs /dev/ram0", I get "device or resource busy"

When I mount the filesystem to check it out, nothing appears to have
anything open on the filesystem.  So why am I not able to flush it
clean?

This is kernel 2.4.5 stock, btw.

Kelledin

Re: Why can't I flush /dev/ram0?

[Richard B. Johnson]

On Mon, 18 Jun 2001, Kelledin Tane wrote:

> At this point, I'm trying to get an initrd working properly.  So far, it
> works, the system boots, etc. etc., but whenever I try to do a "blockdev
> --flushbufs /dev/ram0", I get "device or resource busy"
> 
> When I mount the filesystem to check it out, nothing appears to have
> anything open on the filesystem.  So why am I not able to flush it
> clean?
> 
> This is kernel 2.4.5 stock, btw.
> 
> Kelledin
> 

If you have a directory called /initrd, in your root file-system,
you may find that the old initrd is still mounted:

Script started on Mon Jun 18 17:22:20 2001
# ls /initrd
bin  dev  etc  lib  linuxrc  sbin
# umount /initrd
# ls /initrd
# exit
exit
Script done on Mon Jun 18 17:22:44 2001

Unmount it first.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.

Re: Why can't I flush /dev/ram0?

[Kelledin Tane]

> If you have a directory called /initrd, in your root file-system,
> you may find that the old initrd is still mounted:
>
> Script started on Mon Jun 18 17:22:20 2001
> # ls /initrd
> bin  dev  etc  lib  linuxrc  sbin
> # umount /initrd
> # ls /initrd
> # exit
> exit
> Script done on Mon Jun 18 17:22:44 2001
>
> Unmount it first.

I actually had something to that effect in my rc.sysinit file.  It unmounted
/initrd, removed it, and attempted to flush ram0.  I started trying it
manually when that didn't work =(

Funny thing though (not entirely unexpected)--/initrd's not in /etc/mtab.
I can see why.  Also, /proc/sys/kernel/real-root-dev is still 0x0100
(/dev/ram0).  Changing it to 0x0301 (/dev/hda1) doesn't help, but it might be
a clue for those who know this stuff better than I do.

Kelledin

Re: Why can't I flush /dev/ram0?

[Jan Rekorajski]

On Mon, 18 Jun 2001, Kelledin Tane wrote:

> At this point, I'm trying to get an initrd working properly.  So far, it
> works, the system boots, etc. etc., but whenever I try to do a "blockdev
> --flushbufs /dev/ram0", I get "device or resource busy"
> 
> When I mount the filesystem to check it out, nothing appears to have
> anything open on the filesystem.  So why am I not able to flush it
> clean?

Because of a bug present in Linus tree. Try this patch:

--- linux.orig/drivers/block/rd.c	Mon Nov 20 02:07:47 2000
+++ linux/drivers/block/rd.c	Mon Nov 20 04:03:42 2000
@@ -690,6 +690,7 @@
 done:
 	if (infile.f_op->release)
 		infile.f_op->release(inode, &infile);
+	blkdev_put(out_inode->i_bdev, BDEV_FILE);
 	set_fs(fs);
 	return;
 free_inodes: /* free inodes on error */ 

BTW, it's fixed in -ac patches.

Jan
-- 
Jan Rękorajski            |  ALL SUSPECTS ARE GUILTY. PERIOD!
baggins<at>mimuw.edu.pl   |  OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
BOFH, MANIAC              |                   -- TROOPS by Kevin Rubio

Re: Why can't I flush /dev/ram0?

[Kelledin Tane]

> > When I mount the filesystem to check it out, nothing appears to have
> > anything open on the filesystem.  So why am I not able to flush it
> > clean?
>
> Because of a bug present in Linus tree. Try this patch:

Thanks, that seems to have fixed it.  There's something else I'm curious about
though...

Before, when I was having this problem, I would check mem usage, and there would
be about 5MB of physical memory marked for "buffers" (I assume 4MB for ram0).

Now, using an identical kernel configuration, an identical module set loaded,
and a roughly identical process table, there's about 18MB of memory marked for
"buffers!"

Sorry to be a bother, but I can't help wondering if this patch might have gotten
rid of one problem and replaced it with another?  Or maybe I'm missing some
other detail here...

Kelledin

bash-2.05 $ kill -9 1
init: Just what do you think you're doing, Dave?

Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Shawn Starr]

When diffing 2.4.6-pre2 & pre3 I noticed some reiserfs code was changed.
This seems to cause VFS to panic via reiserfs.

Anyone else notice this?

Shawn.

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Olivier Galibert]

On Mon, Jun 18, 2001 at 10:58:57PM -0400, Shawn Starr wrote:
> When diffing 2.4.6-pre2 & pre3 I noticed some reiserfs code was changed.
> This seems to cause VFS to panic via reiserfs.
> 
> Anyone else notice this?

I don't, and I boot on reiserfs.

  OG.

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Shawn Starr]

Two things:

1) It broke apparently with gcc 2.95.3 when patching from 2.4.6-pre2 ->
2.4.6pre3

2) I tried building it with gcc 3.00 and had same result.

3) I now have gcc 3.00 and going to rebuild 2.4.6-pre2 and see if reiserfs
panics if it doesn't there's an issue with the new pre3 modifications.

I hope ReiserFS *MAINTAINS* compatability from slightly older revisions,
or even migrates systems over to handle new issues.

Shawn.

On Mon, 18 Jun 2001, Olivier Galibert wrote:

> On Mon, Jun 18, 2001 at 10:58:57PM -0400, Shawn Starr wrote:
> > When diffing 2.4.6-pre2 & pre3 I noticed some reiserfs code was changed.
> > This seems to cause VFS to panic via reiserfs.
> >
> > Anyone else notice this?
>
> I don't, and I boot on reiserfs.
>
>   OG.
> -
> 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/
>
>

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Chris Mason]

On Monday, June 18, 2001 10:58:57 PM -0400 Shawn Starr
<spstarr@sh0n.net> wrote:

> 
> When diffing 2.4.6-pre2 & pre3 I noticed some reiserfs code was
> changed. This seems to cause VFS to panic via reiserfs.
> 
> Anyone else notice this?

What is the panic message?

-chris

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Shawn Starr]

read_super_block: can't find a reiserfs filesystem on dev 03:42
read_old_super_block: try to find super block in old location
read_old_super_block: can't find a reiserfs filesystem on dev 03:42
Kernel Panic: VFS: Unable to mount root fs on 03:42

my super block broke somewhere?

Shawn.

On Mon, 18 Jun 2001, Shawn Starr wrote:

>
> Two things:
>
> 1) It broke apparently with gcc 2.95.3 when patching from 2.4.6-pre2 ->
> 2.4.6pre3
>
> 2) I tried building it with gcc 3.00 and had same result.
>
> 3) I now have gcc 3.00 and going to rebuild 2.4.6-pre2 and see if reiserfs
> panics if it doesn't there's an issue with the new pre3 modifications.
>
> I hope ReiserFS *MAINTAINS* compatability from slightly older revisions,
> or even migrates systems over to handle new issues.
>
> Shawn.
>
> On Mon, 18 Jun 2001, Olivier Galibert wrote:
>
> > On Mon, Jun 18, 2001 at 10:58:57PM -0400, Shawn Starr wrote:
> > > When diffing 2.4.6-pre2 & pre3 I noticed some reiserfs code was changed.
> > > This seems to cause VFS to panic via reiserfs.
> > >
> > > Anyone else notice this?
> >
> > I don't, and I boot on reiserfs.
> >
> >   OG.
> > -
> > 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/
> >
> >
>
>

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Chris Mason]

On Monday, June 18, 2001 11:57:16 PM -0400 Shawn Starr <spstarr@sh0n.net>
wrote:

> 
> read_super_block: can't find a reiserfs filesystem on dev 03:42
> read_old_super_block: try to find super block in old location
> read_old_super_block: can't find a reiserfs filesystem on dev 03:42
> Kernel Panic: VFS: Unable to mount root fs on 03:42
> 
> my super block broke somewhere?

Well, this one is usually due to massive corruption or a configuration 
error. Are the correct drives and controllers found during boot?

-chris

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Jeff Mahoney]

On Mon, Jun 18, 2001 at 11:57:16PM -0400, Shawn Starr wrote:
> 
> read_super_block: can't find a reiserfs filesystem on dev 03:42
> read_old_super_block: try to find super block in old location
> read_old_super_block: can't find a reiserfs filesystem on dev 03:42
> Kernel Panic: VFS: Unable to mount root fs on 03:42
> 
> my super block broke somewhere?

    Out of curiousity, what device are you trying to boot from? 03:42, at least
    according to linux/Documentation/devices.txt, corresponds to /dev/hda42.

    Is that really the disk you're trying to mount? I'm not familiar with how
    some IDE RAID controllers present disks, but it was the first thing I
    noticed.

    -Jeff

-- 
Jeff Mahoney
jeffm@suse.com
jeffm@csh.rit.edu

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Neil Brown]

On Tuesday June 19, jeffm@suse.com wrote:
> On Mon, Jun 18, 2001 at 11:57:16PM -0400, Shawn Starr wrote:
> > 
> > read_super_block: can't find a reiserfs filesystem on dev 03:42
> > read_old_super_block: try to find super block in old location
> > read_old_super_block: can't find a reiserfs filesystem on dev 03:42
> > Kernel Panic: VFS: Unable to mount root fs on 03:42
> > 
> > my super block broke somewhere?
> 
>     Out of curiousity, what device are you trying to boot from? 03:42, at least
>     according to linux/Documentation/devices.txt, corresponds to /dev/hda42.

or, noting that kdevname used hexadecimal, 
  /dev/hdb2

NeilBrown

> 
>     Is that really the disk you're trying to mount? I'm not familiar with how
>     some IDE RAID controllers present disks, but it was the first thing I
>     noticed.
> 
>     -Jeff
> 
> -- 
> Jeff Mahoney
> jeffm@suse.com
> jeffm@csh.rit.edu
> -
> 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/

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Jeff Mahoney]

On Tue, Jun 19, 2001 at 03:31:06PM +1000, Neil Brown wrote:
> On Tuesday June 19, jeffm@suse.com wrote:
> > On Mon, Jun 18, 2001 at 11:57:16PM -0400, Shawn Starr wrote:
> > > 
> > > read_super_block: can't find a reiserfs filesystem on dev 03:42
> > > read_old_super_block: try to find super block in old location
> > > read_old_super_block: can't find a reiserfs filesystem on dev 03:42
> > > Kernel Panic: VFS: Unable to mount root fs on 03:42
> > > 
> > > my super block broke somewhere?
> > 
> >     Out of curiousity, what device are you trying to boot from? 03:42, at least
> >     according to linux/Documentation/devices.txt, corresponds to /dev/hda42.
> 
> or, noting that kdevname used hexadecimal, 
>   /dev/hdb2

    Ugh. Ignore me, I should've known that.

    -Jeff

-- 
Jeff Mahoney
jeffm@suse.com
jeffm@csh.rit.edu

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Justin Guyett]

On Mon, 18 Jun 2001, Shawn Starr wrote:

> 1) It broke apparently with gcc 2.95.3 when patching from 2.4.6-pre2 ->
> 2.4.6pre3
>
> 2) I tried building it with gcc 3.00 and had same result.

hmm, how ironic, i just had 2.4.5ac15 lock up when untarring gcc 3.0 on a
reiserfs partition (/)

This after only using ac15 for a few hours... I've never seen anything
like that with ac13, which I've used for days.


Justin

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Alan Cox]

> 
> This after only using ac15 for a few hours... I've never seen anything
> like that with ac13, which I've used for days.

Is ac14 stable for you ?

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Chris Mason]

On Tuesday, June 19, 2001 10:33:49 AM +0100 Alan Cox
<alan@lxorguk.ukuu.org.uk> wrote:

>> 
>> This after only using ac15 for a few hours... I've never seen anything
>> like that with ac13, which I've used for days.
> 
> Is ac14 stable for you ?
> 

Hi Justin,

ac14 was the first with a big reiserfs cleanup patch from me, so it would
be great if you could try it to help narrow things down a bit.

I can also just send you the patch to reverse, it might be faster, just let
me know which way you'd prefer.

-chris

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[David Ford]

I have a server on ac12 and it crashes nearly every day, BUG in slab.c:1244.

invalid operand: 0000
CPU:    0
EIP:    0010:[<c01261f3>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010082
eax: 0000001b   ebx: c13bf768   ecx: c033e160   edx: 0000262a
esi: c9a5e000   edi: 00001000   ebp: 00000246   esp: ca1ddf2c
ds: 0018   es: 0018   ss: 0018
Process proftpd (pid: 27352, stackpage=ca1dd000)
Stack: c02beba5 000004dc c596faa0 00000000 c596fbac c596fb78 00001000 
c9a5e000
       c027c042 00000810 00000007 c596faa0 c35f2b54 ffffffea bffff39c 
00000004
       c0293d2e c596faa0 c35f2b54 00000001 080628b8 c025919b c35f2b54 
00000001
Call Trace: [<c027c042>] [<c0293d2e>] [<c025919b>] [<c0259b8e>] 
[<c0259bf3>]
   [<c0106aa7>]
Code: 0f 0b 83 c4 08 f6 43 11 04 74 55 b8 a5 c2 0f 17 87 46 00 3d

 >>EIP; c01261f3 <kmalloc+123/1bc>   <=====
Trace; c027c042 <tcp_listen_start+5e/144>
Trace; c0293d2e <inet_listen+5a/bc>
Trace; c025919b <sys_listen+37/50>
Trace; c0259b8e <sys_socketcall+32/1bc>
Trace; c0259bf3 <sys_socketcall+97/1bc>
Trace; c0106aa7 <system_call+37/40>
Code;  c01261f3 <kmalloc+123/1bc>
00000000 <_EIP>:
Code;  c01261f3 <kmalloc+123/1bc>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  c01261f5 <kmalloc+125/1bc>
   2:   83 c4 08                  add    $0x8,%esp
Code;  c01261f8 <kmalloc+128/1bc>
   5:   f6 43 11 04               testb  $0x4,0x11(%ebx)
Code;  c01261fc <kmalloc+12c/1bc>
   9:   74 55                     je     60 <_EIP+0x60> c0126253 
<kmalloc+183/1bc>
Code;  c01261fe <kmalloc+12e/1bc>
   b:   b8 a5 c2 0f 17            mov    $0x170fc2a5,%eax
Code;  c0126203 <kmalloc+133/1bc>
  10:   87 46 00                  xchg   %eax,0x0(%esi)
Code;  c0126206 <kmalloc+136/1bc>
  13:   3d 00 00 00 00            cmp    $0x0,%eax


It's the same bug as the previous kernel I tried running, ac8 was it..

-d


Alan Cox wrote:

>>This after only using ac15 for a few hours... I've never seen anything
>>like that with ac13, which I've used for days.
>>
>
>Is ac14 stable for you ?
>
>
>-
>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/
>

Re: Linux 2.4.6-pre3 breaks ReiserFS mount on boot

[Shawn Starr]

Not /dev/hda42, thats odd. From 2.4.5 -> 2.4.6 ReiserFS would refuse to
mount the drive on startup.

I noticed in pre5 there was a reiserfs fix to something but im not sure if
its related or not.

My domain is also back so I'm going to resubscribe.

Shawn.

On Tue, 19 Jun 2001, Neil Brown wrote:

> On Tuesday June 19, jeffm@suse.com wrote:
> > On Mon, Jun 18, 2001 at 11:57:16PM -0400, Shawn Starr wrote:
> > >
> > > read_super_block: can't find a reiserfs filesystem on dev 03:42
> > > read_old_super_block: try to find super block in old location
> > > read_old_super_block: can't find a reiserfs filesystem on dev 03:42
> > > Kernel Panic: VFS: Unable to mount root fs on 03:42
> > >
> > > my super block broke somewhere?
> >
> >     Out of curiousity, what device are you trying to boot from? 03:42, at least
> >     according to linux/Documentation/devices.txt, corresponds to /dev/hda42.
>
> or, noting that kdevname used hexadecimal,
>   /dev/hdb2
>
> NeilBrown
>
> >
> >     Is that really the disk you're trying to mount? I'm not familiar with how
> >     some IDE RAID controllers present disks, but it was the first thing I
> >     noticed.
> >
> >     -Jeff
> >
> > --
> > Jeff Mahoney
> > jeffm@suse.com
> > jeffm@csh.rit.edu
> > -
> > 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/
>
>