1
linux/fs/ceph
Henry C Chang d3d0720d4a ceph: do not use i_wrbuffer_ref as refcount for Fb cap
We increments i_wrbuffer_ref when taking the Fb cap. This breaks
the dirty page accounting and causes looping in
__ceph_do_pending_vmtruncate, and ceph client hangs.

This bug can be reproduced occasionally by running blogbench.

Add a new field i_wb_ref to inode and dedicate it to Fb reference
counting.

Signed-off-by: Henry C Chang <henry.cy.chang@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net>
2011-05-11 10:44:48 -07:00
..
addr.c ceph: handle ceph_osdc_new_request failure in ceph_writepages_start 2011-05-03 09:28:12 -07:00
caps.c ceph: do not use i_wrbuffer_ref as refcount for Fb cap 2011-05-11 10:44:48 -07:00
ceph_frag.c
debugfs.c ceph: remove debugfs debug cruft 2011-03-21 12:24:20 -07:00
dir.c ceph: rename dentry_release -> d_release, fix comment 2011-03-21 12:24:26 -07:00
export.c
file.c ceph: do not call __mark_dirty_inode under i_lock 2011-05-04 12:56:45 -07:00
inode.c ceph: do not use i_wrbuffer_ref as refcount for Fb cap 2011-05-11 10:44:48 -07:00
ioctl.c
ioctl.h
Kconfig
locks.c
Makefile
mds_client.c ceph: print debug message before put mds session 2011-05-11 10:44:34 -07:00
mds_client.h
mdsmap.c
snap.c ceph: fix list_add in ceph_put_snap_realm 2011-05-11 10:44:36 -07:00
strings.c
super.c ceph: Move secret key parsing earlier. 2011-03-29 12:11:16 -07:00
super.h ceph: do not use i_wrbuffer_ref as refcount for Fb cap 2011-05-11 10:44:48 -07:00
xattr.c ceph: do not call __mark_dirty_inode under i_lock 2011-05-04 12:56:45 -07:00