| Message ID | 4A52776B.8000203@redhat.com |
|---|---|
| State | Superseded, archived |
| Headers | show |
On Mon, Jul 06, 2009 at 05:15:07PM -0500, Eric Sandeen wrote: > Commit 53422e8a5644e22ea3f6e0efba82a765b72e4308 moved > the new extent insertion in ext2fs_extent_set_bmap prior > to the modification of the original extent, but the > insert function left the handle pointing a the new > extent; this left us modifying the -new- extent not > the original one, and winding up with a corrupt extent > tree something like: > > BLOCKS: > (0-1):588791-588792, (0):588791 > > We need to move back to the previous extent prior > to modification, if we inserted a new one. Hmm, I just thought of something awful; what if the insert resulted in a node split? Instead of using ext2fs_extent_get(EXT2_EXTENT_PREV), we may need to use ext2fs_extent_goto() to seek to the correct logical block instead. - Ted -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Theodore Tso wrote: > On Mon, Jul 06, 2009 at 05:15:07PM -0500, Eric Sandeen wrote: >> Commit 53422e8a5644e22ea3f6e0efba82a765b72e4308 moved >> the new extent insertion in ext2fs_extent_set_bmap prior >> to the modification of the original extent, but the >> insert function left the handle pointing a the new >> extent; this left us modifying the -new- extent not >> the original one, and winding up with a corrupt extent >> tree something like: >> >> BLOCKS: >> (0-1):588791-588792, (0):588791 >> >> We need to move back to the previous extent prior >> to modification, if we inserted a new one. > > Hmm, I just thought of something awful; what if the insert resulted in > a node split? Instead of using ext2fs_extent_get(EXT2_EXTENT_PREV), > we may need to use ext2fs_extent_goto() to seek to the correct logical > block instead. > > - Ted Hm.... so PREV doesn't go to the node for the previous logical block, but to what, the node to the left at this level? Guess I need to read more carefully... -Eric -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/lib/ext2fs/extent.c b/lib/ext2fs/extent.c index 35b080e..9ea5c30 100644 --- a/lib/ext2fs/extent.c +++ b/lib/ext2fs/extent.c @@ -1257,6 +1257,11 @@ again: EXT2_EXTENT_INSERT_AFTER, &newextent); if (retval) goto done; + /* Now pointing at new extent; move back to prev */ + retval = ext2fs_extent_get(handle, EXT2_EXTENT_PREV, + &extent); + if (retval) + goto done; } extent.e_len--; retval = ext2fs_extent_replace(handle, 0, &extent);
Commit 53422e8a5644e22ea3f6e0efba82a765b72e4308 moved the new extent insertion in ext2fs_extent_set_bmap prior to the modification of the original extent, but the insert function left the handle pointing a the new extent; this left us modifying the -new- extent not the original one, and winding up with a corrupt extent tree something like: BLOCKS: (0-1):588791-588792, (0):588791 We need to move back to the previous extent prior to modification, if we inserted a new one. Signed-off-by: Eric Sandeen <sandeen@redhat.com> --- (aside: should there be functions to simply move the handle around, w/o necessarily populating an extent structure?) (aside2: maybe ext2fs_extent_insert should take a flag saying whether the handle should be moved after the insertion; moving it around to back where we want to be seems a bit tedious) -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html