Xen devel - Feb 2010 - [PATCH 0 of 8] Linux/pvops blktap + blkfront updates.

First 4 stop blktap barfing once and blkfront twice during shutdown,
plus a memory leak.

16 A fix-double-free.diff: blktap: Fix tap kobj double unref in sysfs_destroy.
17 A fix-blk-cleanup.diff: blkfront: Fix backtrace in blk_cleanup_queue.
18 A fix-gendisk-leak.diff: blkfront: Stop leaking the gendisk struct.
19 A fix-bd-oops.diff: blkfront: Fix an unlikely xenbus crasher.

The rest is decoration.

20 A blkfront-close.diff: blkfront: Clean up closing transition.
21 A xs-write-unlocked.diff: blkfront: Move xenstore accesses off the device
release path.
22 A xlvbd-release-gendisk.diff: blkfront: Clean up blkfront_closing.
23 A debug-common.diff: blkfront: Clean up debug statements.

Cheers,
Daniel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

It''s that final put_device in device_unregister which is supposed to
clear the caller''s reference. Subtle indeed, because refcount went
2->3 when after sysfs_destroy:

 +1 for our own ring->dev
 +1 for the open sysfs ./remove attribute
 +1 for the callback scheduled

Drops to 1 in device_unregister and 0 only after leaving callbacks.

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>
CC: stable@kernel.org




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>
CC: stable@kernel.org




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Getting partition 0 should not fail, but if it does we won''t continue.

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>
CC: stable@kernel.org




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Just gained enough locals to justify a proper subroutine.

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>
Cc: stable@kernel.org




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Not holding the bdev mutex should be slightly more robust. Also helps
getting blkfront_closing to complement xlvbd_alloc_gendisk.

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

* Call blkfront_closing the xlvbd_release_gendisk it is.
 * Fix comments regarding device disconnects.

Signed-off-by: Daniel Stodden <daniel.stodden@citrix.com>
Cc: stable@kernel.org




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

* Use dev_dbg everywhere.
 * Fix excess line withs.
 * Wrap dev_dbg to take a blkfront_info handle.
 * Separate datapath debug statements (DEV_QDBG), conditionally.
 * Verbose disk remove/release dbg_info statements.




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel