Date: Wed, 4 Sep 2013 11:49:33 -0600 From: "Justin T. Gibbs" <gibbs@freebsd.org> To: =?iso-8859-1?Q?Roger_Pau_Monn=E9?= <roger.pau@citrix.com> Cc: "freebsd-xen@freebsd.org" <freebsd-xen@freebsd.org> Subject: Re: Another blkback issue related to flush Message-ID: <EAD57817-4430-41AA-B12A-DCDD07A61412@freebsd.org> In-Reply-To: <522768EA.6010809@citrix.com> References: <522768EA.6010809@citrix.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--Apple-Mail=_8FB0F9B6-B925-4672-AD4B-78D54204A510
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=iso-8859-1
On Sep 4, 2013, at 11:07 AM, Roger Pau Monn=E9 <roger.pau@citrix.com> =
wrote:
> Hello,
>=20
> I've found another issue with blkback handling of flush operations, it
> was incorrectly setting one of the bio parameters when using a block
> device as backend. The attached patch fixes it, and also includes a
> small fix to correctly set the operation when writing the response on
> the ring (all responses written by FreeBSD blkback were of type
> BLKIF_OP_READ, because nreq->operation was not set).
It seems that the req's pendcnt is unused and can be culled. Otherwise,
looks good to me. Here's the patch I have in my tree now.
--
Justin
--Apple-Mail=_8FB0F9B6-B925-4672-AD4B-78D54204A510
Content-Disposition: attachment;
filename=blkback.diff
Content-Type: application/octet-stream;
name="blkback.diff"
Content-Transfer-Encoding: 7bit
Index: blkback.c
===================================================================
--- blkback.c (revision 255179)
+++ blkback.c (working copy)
@@ -230,7 +230,7 @@ struct xbb_xen_reqlist {
int num_children;
/**
- * Number of I/O requests dispatched to the backend.
+ * Number of I/O requests still pending on the backend.
*/
int pendcnt;
@@ -327,13 +327,6 @@ struct xbb_xen_req {
int nr_512b_sectors;
/**
- * The number of struct bio requests still outstanding for this
- * request on the backend device. This field is only used for
- * device (rather than file) backed I/O.
- */
- int pendcnt;
-
- /**
* BLKIF_OP code for this request.
*/
int operation;
@@ -1240,6 +1233,7 @@ xbb_get_resources(struct xbb_softc *xbb, struct xb
nreq->reqlist = *reqlist;
nreq->req_ring_idx = ring_idx;
nreq->id = ring_req->id;
+ nreq->operation = ring_req->operation;
if (xbb->abi != BLKIF_PROTOCOL_NATIVE) {
bcopy(ring_req, &nreq->ring_req_storage, sizeof(*ring_req));
@@ -2062,7 +2056,6 @@ xbb_dispatch_dev(struct xbb_softc *xbb, struct xbb
{
struct xbb_dev_data *dev_data;
struct bio *bios[XBB_MAX_SEGMENTS_PER_REQLIST];
- struct xbb_xen_req *nreq;
off_t bio_offset;
struct bio *bio;
struct xbb_sg *xbb_sg;
@@ -2080,7 +2073,6 @@ xbb_dispatch_dev(struct xbb_softc *xbb, struct xbb
bio_idx = 0;
if (operation == BIO_FLUSH) {
- nreq = STAILQ_FIRST(&reqlist->contig_req_list);
bio = g_new_bio();
if (__predict_false(bio == NULL)) {
DPRINTF("Unable to allocate bio for BIO_FLUSH\n");
@@ -2094,10 +2086,10 @@ xbb_dispatch_dev(struct xbb_softc *xbb, struct xbb
bio->bio_offset = 0;
bio->bio_data = 0;
bio->bio_done = xbb_bio_done;
- bio->bio_caller1 = nreq;
+ bio->bio_caller1 = reqlist;
bio->bio_pblkno = 0;
- nreq->pendcnt = 1;
+ reqlist->pendcnt = 1;
SDT_PROBE1(xbb, kernel, xbb_dispatch_dev, flush,
device_get_unit(xbb->dev));
--Apple-Mail=_8FB0F9B6-B925-4672-AD4B-78D54204A510
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
charset=iso-8859-1
--Apple-Mail=_8FB0F9B6-B925-4672-AD4B-78D54204A510--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?EAD57817-4430-41AA-B12A-DCDD07A61412>