RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-01-10 05:00:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

librbd: handle parent change while async I/Os are in flight

Browse Source 
During a test_librbd_fsx run including flatten, ImageCtx->parent
was being dereferenced while null.  Between the time the parent
overlap is calculated and the time the guard+write completes
with ENOENT and submits the copyup+write, the parent image
could have changed (by resize) or been made irrelevant (by
child flatten) such that the parent overlap is now incorrect.

Handle "no parent" by just sending the copyup+write; the copyup
part will be a no-op.  Move to WRITE_FLAT state in this case
because there's no more child to deal with.

Handle "overlap changed" by recalculating overlap before
reading parent data; if none is left, don't read, but rather
just clear m_object_image_extents, in which case the copyup
will again be a no-op because it will be of zero length.
However we still have a parent, so stay in WRITE_COPYUP state
and come back through as usual.

Signed-off-by: Dan Mick <dan.mick@inktank.com>
Fixes: #3524
This commit is contained in:
Dan Mick 2012-11-30 18:11:09 -08:00
parent 917a6f2963
commit 41e16a3b40
1 changed files with 37 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
src/librbd/AioRequest.cc
View File

@ -157,15 +157,47 @@ namespace librbd {
ldout(m_ictx->cct, 20) << "WRITE_CHECK_GUARD" << dendl;
if (r == -ENOENT) {
Mutex::Locker l(m_ictx->snap_lock);
Mutex::Locker l2(m_ictx->parent_lock);
// copyup the entire object up to the overlap point
ldout(m_ictx->cct, 20) << "reading from parent " << m_object_image_extents << dendl;
assert(m_object_image_extents.size());
/*
* Parent may have disappeared; if so, recover by using
* send_copyup() to send the original write req (the copyup
* operation itself will be a no-op, since someone must have
* populated the child object while we weren't looking).
* Move to WRITE_FLAT state as we'll be done with the
* operation once the null copyup completes.
*/
m_state = LIBRBD_AIO_WRITE_COPYUP;
read_from_parent(m_object_image_extents);
if (m_ictx->parent == NULL) {
ldout(m_ictx->cct, 20) << "parent is gone; do null copyup " << dendl;
m_state = LIBRBD_AIO_WRITE_FLAT;
send_copyup();
finished = false;
break;
}
// If parent still exists, overlap might also have changed.
uint64_t newlen = m_ictx->prune_parent_extents(
m_object_image_extents, m_ictx->parent_md.overlap);
// copyup the entire object up to the overlap point, if any
if (newlen != 0) {
ldout(m_ictx->cct, 20) << "should_complete(" << this << ") overlap "
<< m_ictx->parent_md.overlap << " newlen "
<< newlen << " image_extents"
<< m_object_image_extents << dendl;
m_state = LIBRBD_AIO_WRITE_COPYUP;
read_from_parent(m_object_image_extents);
} else {
ldout(m_ictx->cct, 20) << "should_complete(" << this
<< "): parent overlap now 0" << dendl;
m_object_image_extents.clear();
m_state = LIBRBD_AIO_WRITE_FLAT;
send_copyup();
}
finished = false;
break;
}