kpriv/linux-hardened
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

ocfs2: remove an overly aggressive BUG() in dlmfs

Browse source 
Don't BUG() user_dlm_unblock_lock() on the absence of the USER_LOCK_BLOCKED
flag - this turns out to be a valid case. Make some of the related BUG()
statements print more useful information.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
This commit is contained in:
Mark Fasheh 2006-03-29 10:33:35 -08:00
parent ab0920ce7e
commit 1f7bc828e3
1 changed files with 17 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
fs/ocfs2/dlm/userdlm.c
View file

@ -268,13 +268,26 @@ static void user_dlm_unblock_lock(void *opaque)
spin_lock(&lockres->l_lock); spin_lock(&lockres->l_lock);
BUG_ON(!(lockres->l_flags & USER_LOCK_BLOCKED)); mlog_bug_on_msg(!(lockres->l_flags & USER_LOCK_QUEUED),
BUG_ON(!(lockres->l_flags & USER_LOCK_QUEUED)); "Lockres %s, flags 0x%x\n",
lockres->l_name, lockres->l_flags);
/* notice that we don't clear USER_LOCK_BLOCKED here. That's /* notice that we don't clear USER_LOCK_BLOCKED here. If it's
* for user_ast to do. */ * set, we want user_ast clear it. */
lockres->l_flags &= ~USER_LOCK_QUEUED; lockres->l_flags &= ~USER_LOCK_QUEUED;
/* It's valid to get here and no longer be blocked - if we get
* several basts in a row, we might be queued by the first
* one, the unblock thread might run and clear the queued
* flag, and finally we might get another bast which re-queues
* us before our ast for the downconvert is called. */
if (!(lockres->l_flags & USER_LOCK_BLOCKED)) {
mlog(0, "Lockres %s, flags 0x%x: queued but not blocking\n",
lockres->l_name, lockres->l_flags);
spin_unlock(&lockres->l_lock);
goto drop_ref;
}
if (lockres->l_flags & USER_LOCK_IN_TEARDOWN) { if (lockres->l_flags & USER_LOCK_IN_TEARDOWN) {
mlog(0, "lock is in teardown so we do nothing\n"); mlog(0, "lock is in teardown so we do nothing\n");
spin_unlock(&lockres->l_lock); spin_unlock(&lockres->l_lock);