diff options
| author | kib <kib@FreeBSD.org> | 2013-08-27 01:31:12 +0000 |
|---|---|---|
| committer | kib <kib@FreeBSD.org> | 2013-08-27 01:31:12 +0000 |
| commit | 9426e190e7d41951c9bafa9b1c8e6756607e475c (patch) | |
| tree | 4a0ff072d95f46af4aeae2178801b816e141a0e3 | |
| parent | 5f8ac855694c0f2266dab9f85530b4a331b47ed2 (diff) | |
| download | FreeBSD-src-9426e190e7d41951c9bafa9b1c8e6756607e475c.zip FreeBSD-src-9426e190e7d41951c9bafa9b1c8e6756607e475c.tar.gz | |
When allocating a pbuf for the cluster write, do not sleep waiting
for the available pbuf when passed vnode is backing md(4). Other i/o
directed to the same md device might already hold pbufs, and then we
could deadlock since only our progress can free a pbuf needed for
wakeup.
Obtained from: projects/vm6
Reminded and tested by: pho
MFC after: 1 week
| -rw-r--r-- | sys/kern/vfs_cluster.c | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/sys/kern/vfs_cluster.c b/sys/kern/vfs_cluster.c index c98db81..9601082 100644 --- a/sys/kern/vfs_cluster.c +++ b/sys/kern/vfs_cluster.c @@ -837,7 +837,9 @@ cluster_wbuild(struct vnode *vp, long size, daddr_t start_lbn, int len, (tbp->b_bcount != tbp->b_bufsize) || (tbp->b_bcount != size) || (len == 1) || - ((bp = getpbuf(&cluster_pbuf_freecnt)) == NULL)) { + ((bp = (vp->v_vflag & VV_MD) != 0 ? + trypbuf(&cluster_pbuf_freecnt) : + getpbuf(&cluster_pbuf_freecnt)) == NULL)) { totalwritten += tbp->b_bufsize; bawrite(tbp); ++start_lbn; |
