This patchset untangles some of the locking in reiserfs. It has seen more testing as part of the maintenance cycle in SLE 11 SP2. - We push the write lock out of the xattr code. It doesn't need it and we can simplify locking by releasing and reacquiring the locks around the xattr calls. - Handle nested locks properly. There's confusion on when a lock is nested, when it's not, and how to drop it across schedules like the BKL it is modeled after. We make the distinction between taking/releasing the lock and when to drop it for schedules and simplify the logic. This fixes a number of deadlocks that happen because the intention was to drop the write lock but it really only decremented the use count. - Fix the deadlocks with the quota code. This involves dropping the write lock before quota calls and reacquring it afterwards. Without this patch reiserfs quotas are essentially unusable. Previous postings of this patchset were mismerged. This revision integrates Jan's advice about search_by_key_reada returning the depth instead of a bool. Jan - This series is in the for-3.12 branch of git.kernel.org:/pub/scm/linux/kernel/git/jeffm/linux-reiserfs.git Thanks. -Jeff -- To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
