Re: ext3-0.0.2a patch released

[Stephen C. Tweedie]

Hi,

On Tue, 19 Oct 1999 09:50:59 -0400, Daniel Veillard
<[EMAIL PROTECTED]> said:

>   The oops of the day :

> Oct 19 05:42:50 fr kernel: Assertion failure in journal_get_write_access() at 
>transaction.c line 436: "handle->h_buffer_credits > 0" 
...
> Oct 19 05:42:50 fr kernel: Call Trace: [cprt+18806/35456] [cprt+19583/35456] 
>[ext3_write_inode+197/424] [sync_all_inodes+131/188] [try_to_free_inodes+39/56] 
>[grow_inodes+30/364] [get_empty_inode+145/156]  

The problem is that an existing transaction is running before we call
ext3_new_inode(), and that get_empty_inode() has decided that there
are too many dirty inodes and that some need to be flushed to disk.
We're in the middle of an existing transaction, so those flushes get
accounted against the running transaction handle.  Oops.

This should be taken care of in 0.0.2b.  I've put a prerelease of that
up for ftp, and I'll make it official once it has had a bit more
testing.

--Stephen

Re: ext3-0.0.2a patch released

[Daniel Veillard]

  The oops of the day :

Oct 19 05:42:50 fr kernel: Assertion failure in journal_get_write_access() at 
transaction.c line 436: "handle->h_buffer_credits > 0" 
Oct 19 05:42:50 fr kernel: Unable to handle kernel NULL pointer dereference at virtual 
address  
Oct 19 05:42:50 fr kernel: current->tss.cr3 = 00da8000, %cr3 = 00da8000 
Oct 19 05:42:50 fr kernel: *pde =  
Oct 19 05:42:50 fr kernel: Oops: 0002 
Oct 19 05:42:50 fr kernel: CPU:0 
Oct 19 05:42:50 fr kernel: EIP:0010:[journal_get_write_access+707/852] 
Oct 19 05:42:50 fr kernel: EFLAGS: 00010296 
Oct 19 05:42:50 fr kernel: eax: 006d   ebx: c76180c0   ecx: 0001   edx: 
003a 
Oct 19 05:42:50 fr kernel: esi: c300bd60   edi: c586e5a0   ebp: c1dbb0e0   esp: 
c073bdfc 
Oct 19 05:42:50 fr kernel: ds: 0018   es: 0018   ss: 0018 
Oct 19 05:42:50 fr kernel: Process rsync (pid: 28461, process nr: 54, 
stackpage=c073b000) 
Oct 19 05:42:50 fr kernel: Stack: c01f62b6 01b4 c01f65bf  c300bd60 
c586e5a0 c1dbb0e0 c76180d8  
Oct 19 05:42:50 fr kernel: c4e3e5b0 c0148c59 c300bd60 c1dbb0e0 
c586e5a0 c6d8463c c6d84600  
Oct 19 05:42:50 fr kernel:c5948a20 c073be44 c1dbb0e0 c096fc00 008d 
c1dbb0e0 c096fc00 008d  
Oct 19 05:42:50 fr kernel: Call Trace: [cprt+18806/35456] [cprt+19583/35456] 
[ext3_write_inode+197/424] [sync_all_inodes+131/188] [try_to_free_inodes+39/56] 
[grow_inodes+30/364] [get_empty_inode+145/156]  
Oct 19 05:42:50 fr kernel:[ext3_new_inode+47/1784] [wake_up_process+58/68] 
[__wake_up+59/68] [start_this_handle+742/752] [journal_start+109/124] 
[ext3_create+61/576] [open_namei+352/844] [filp_open+66/240]  
Oct 19 05:42:50 fr kernel:[sys_open+51/152] [system_call+52/56]  
Oct 19 05:42:50 fr kernel: Code: c6 05 00 00 00 00 00 83 c4 14 8b 4c 24 20 8b 41 04 48 
8b 4c  

Later I got: 

Oct 19 08:22:21 fr kernel: VFS: file-max limit 4096 reached 


And then as a side effect:

Oct 19 08:29:00 fr kernel: Unable to load interpreter 
Oct 19 08:30:00 fr kernel: Unable to load interpreter 
Oct 19 08:32:01 fr kernel: Unable to load interpreter 
Oct 19 08:32:49 fr kernel: Unable to load interpreter 

Daniel

-- 
[EMAIL PROTECTED] | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux, WWW, rpmfind,
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | rpm2html, XML,
http://www.w3.org/People/W3Cpeople.html#Veillard | badminton, and Kaffe.

ext3-0.0.2a patch released

[Stephen C. Tweedie]

Hi,

Available at

ftp://ftp.uk.linux.org/pub/linux/sct/fs/jfs/ext3-0.0.2a.diff.gz

is a patch against ext3-0.0.2.tar.gz.  This patch fixes a couple of
problems in truncate(), especially deletion of large files where the
delete has to be split over multiple transactions.  Truncate of busy
files was also affected in some cases.  In the worst case, this bug may
cause inconsistencies in the on-disk filesystem bitmaps.

People who had problems running scp or netscape should hopefully find
that this release cures things.

This release also adds the exporting of a couple of kernel symbols to
allow a modular kernel build to work (modular ext3 is not yet supported,
though).

There will be another release within the next day or two to fix a
possible assert-failure in sync_inodes(), but that bug can never cause
any metadata inconsistency: you should apply the 0.0.2a update
immediately to fix the truncate bug rather than wait for 0.0.2b.

Finally, thanks for everybody who is sending me feedback on the 0.0.2
release. 

--Stephen