PROBLEM: 2.6.9-rc3 Bug in NTFS code
initially I applied only the second patch (since I didn't realize there were two different patches), and that fixed the umount problem. afterwards I applied both, checked again, and there were no problems as well.
thanks.
Quoted 100 lines · Anton Altaparmakov
Quoted 2 lines · Dino Klein
Quoted 1 line
below is what I had in the logs when attempting to umount a readonly
NTFS.
Quoted 3 lines
[snip]
I suspect that this is caused by a bug I introduced in 2.1.18 release
and
Quoted 3 lines
that I already fixed in my development tree. I unfortunately completely forgot that the code containing the bug was already in the mainstream kernels so I didn't think of submitting the fix straight away. )))-:
Ouch. Having slept over it I awoke this morning with the realization that there is in fact another bug, which would definitely cause exactly the BUG check to trigger that Dino reported. )-: Sorry about that. Just goes to show that when I was porting my code from libntfs to the kernel and thinking "wow, this is really too easy", I was actually missing the subtleties involved so I broke it. )-:
Linus, please do a
bk pull bk://linux-ntfs.bkbits.net/ntfs-2.6to apply the fix which is also shown below in diff style patch. Thanks!
Dino, could you apply the below patch and just check that the bug goes away in case I have missed yet another problem... I am pretty confident that this is it this time but you never know. (You can just add the lines "if (ni !=
base_ni)unmap_extent_mft_record(ni);" tofs/ntfs/attrib.cas shown in the patch.) Thanks a lot in advance!Note you also want the fix from my previous email as the two bugs/fixes are related...
Best regards,
Anton
--
Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @) Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK Linux NTFS maintainer / IRC: #ntfs on irc.freenode.net
WWW: http://linux-ntfs.sf.net/ & http://www-stu.christs.cam.ac.uk/~aia21/This will update the following files:
Documentation/filesystems/ntfs.txt | 2 +- fs/ntfs/ChangeLog | 4 ++++ fs/ntfs/attrib.c | 4 +++- 3 files changed, 8 insertions(+), 2 deletions(-)through these ChangeSets:
<[email protected]> (04/10/03 1.2034.1.1)
NTFS: Fix another stupid bug in
fs/ntfs/attrib.c::ntfs_external_attr_find()
where we forgot to unmap the extent mft record when we had
finished
enumerating an attribute which caused a bug check to trigger when
theVFS calls ->clear_inode.Signed-off-by Anton Altaparmakov <[email protected]>===================================================================
diff -Nru
a/Documentation/filesystems/ntfs.txtb/Documentation/filesystems/ntfs.txtDocumentation/filesystems/ntfs.txt +0
-0@@ -278,7 +278,7 @@ Note, a technical ChangeLog aimed at kernel hackers is in
fs/ntfs/ChangeLog.2.1.20:
Documentation/filesystems/ntfs.txt +1
-1- - Fix a stupid bug introduced in 2.1.18 release. + - Fix two stupid bugs introduced in 2.1.18 release. 2.1.19: - Minor bugfix in handling of the default upcase table. - Many internal cleanups and improvements. Many thanks to Linusdiff -Nru a/fs/ntfs/ChangeLog b/fs/ntfs/ChangeLog
fs/ntfs/ChangeLog +4
-0@@ -27,6 +27,10 @@ where we did not clear ctx->al_entry but it was still set due to changes in ntfs_attr_lookup() and ntfs_external_attr_find() in particular. + - Fix another stupid bug in fs/ntfs/attrib.c::ntfs_external_attr_find() + where we forgot to unmap the extent mft record when we had finished + enumerating an attribute which caused a bug check to trigger when the + VFS calls ->clear_inode. 2.1.19 - Many cleanups, improvements, and a minor bug fix.diff -Nru
a/fs/ntfs/attrib.cb/fs/ntfs/attrib.cfs/ntfs/attrib.c +1
-0@@ -1738,11 +1738,13 @@* correctly yet as we do not know what @ctx->attr will be set to by * the call to ntfs_attr_find() below. */ +if (ni != base_ni)
fs/ntfs/attrib.c line 1750
Quoted 10 lines
fs/ntfs/attrib.c +2
-1+unmap_extent_mft_record (ni); ctx->mrec = ctx->base_mrec; ctx->attr = (ATTR_RECORD *)((u8 *)ctx->mrec +le16_to_cpu (ctx->mrec->attrs_offset)); ctx->is_first = TRUE; - ctx->ntfs_ino = ctx->base_ntfs_ino; + ctx->ntfs_ino = base_ni; ctx->base_ntfs_ino =NULL ; ctx->base_mrec =NULL ; ctx->base_attr =NULL ;
_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to get there! http://lifeevents.msn.com/category.aspx?cid=Retirement
Quoted 5 lines · Dino Klein
initially I applied only the second patch (since I didn't realize there were two different patches), and that fixed the umount problem. afterwards I applied both, checked again, and there were no problems as well.
thanks.
Great, thanks for letting me know.
Best regards,
Anton
Signature · 4 lines
Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @) Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK Linux NTFS maintainer / IRC: #ntfs on irc.freenode.net WWW: http://linux-ntfs.sf.net/ & http://www-stu.christs.cam.ac.uk/~aia21/