PR/52301 CVS commit: src/sys/ufs/lfs

["Taylor R Campbell" <riastradh%netbsd.org@localhost>]

The following reply was made to PR kern/52301; it has been noted by GNATS.

From: "Taylor R Campbell" <riastradh%netbsd.org@localhost>
To: gnats-bugs%gnats.NetBSD.org@localhost
Cc: 
Subject: PR/52301 CVS commit: src/sys/ufs/lfs
Date: Sun, 23 Feb 2020 08:40:37 +0000

 Module Name:	src
 Committed By:	riastradh
 Date:		Sun Feb 23 08:40:37 UTC 2020
 
 Modified Files:
 	src/sys/ufs/lfs: lfs_extern.h lfs_segment.c lfs_subr.c
 
 Log Message:
 Break deadlock in PR kern/52301.
 
 The lock order is lfs_writer -> lfs_seglock.  The problem in 52301 is
 that lfs_segwrite violates this lock order by sometimes doing
 lfs_seglock -> lfs_writer, either (a) when doing a checkpoint or (b),
 opportunistically, when there are no dirops pending.  Both cases can
 deadlock, because dirops sometimes take the seglock (lfs_truncate,
 lfs_valloc, lfs_vfree):
 
 (a) There may be dirops pending, and they may be waiting for the
 seglock, so we can't wait for them to complete while holding the
 seglock.
 
 (b) The test for fs->lfs_dirops == 0 happens unlocked, and the state
 may change by the time lfs_writer_enter acquires lfs_lock.
 
 To resolve this in each case:
 
 (a) Do lfs_writer_enter before lfs_seglock, since we will need it
 unconditionally anyway.  The worst performance impact of this should
 be that some dirops get delayed a little bit.
 
 (b) Create a new lfs_writer_tryenter to use at this point so that the
 test for fs->lfs_dirops == 0 and the acquisition of lfs_writer happen
 atomically under lfs_lock.
 
 
 To generate a diff of this commit:
 cvs rdiff -u -r1.115 -r1.116 src/sys/ufs/lfs/lfs_extern.h
 cvs rdiff -u -r1.284 -r1.285 src/sys/ufs/lfs/lfs_segment.c
 cvs rdiff -u -r1.98 -r1.99 src/sys/ufs/lfs/lfs_subr.c
 
 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.