From: Eric Blake <eblake@redhat.com>
To: qemu-devel@nongnu.org
Cc: qemu-block@nongnu.org, pbonzini@redhat.com, mreitz@redhat.com,
Kevin Wolf <kwolf@redhat.com>
Subject: [Qemu-devel] [PATCH v6 15/15] nbd: Implement NBD_CMD_WRITE_ZEROES on client
Date: Thu, 13 Oct 2016 15:58:55 -0500 [thread overview]
Message-ID: <1476392335-9256-16-git-send-email-eblake@redhat.com> (raw)
In-Reply-To: <1476392335-9256-1-git-send-email-eblake@redhat.com>
Upstream NBD protocol recently added the ability to efficiently
write zeroes without having to send the zeroes over the wire,
along with a flag to control whether the client wants a hole.
The generic block code takes care of falling back to the obvious
write of lots of zeroes if we return -ENOTSUP because the server
does not have WRITE_ZEROES.
Ideally, since NBD_CMD_WRITE_ZEROES does not involve any data
over the wire, we want to support transactions that are much
larger than the normal 32M limit imposed on NBD_CMD_WRITE. But
the server may still have a limit smaller than UINT_MAX, so
until experimental NBD protocol additions for advertising various
command sizes is finalized (see [1], [2]), for now we just stick to
the same limits as normal writes.
[1] https://github.com/yoe/nbd/blob/extension-info/doc/proto.md
[2] https://sourceforge.net/p/nbd/mailman/message/35081223/
Signed-off-by: Eric Blake <eblake@redhat.com>
---
v6: rebase
v5: enhance commit message
v4: rebase to byte-based limits
v3: rebase, tell block layer about our support
---
block/nbd-client.h | 2 ++
block/nbd-client.c | 35 +++++++++++++++++++++++++++++++++++
block/nbd.c | 4 ++++
3 files changed, 41 insertions(+)
diff --git a/block/nbd-client.h b/block/nbd-client.h
index 78e8e57..e51df22 100644
--- a/block/nbd-client.h
+++ b/block/nbd-client.h
@@ -48,6 +48,8 @@ int nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, int count);
int nbd_client_co_flush(BlockDriverState *bs);
int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
uint64_t bytes, QEMUIOVector *qiov, int flags);
+int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
+ int count, BdrvRequestFlags flags);
int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,
uint64_t bytes, QEMUIOVector *qiov, int flags);
diff --git a/block/nbd-client.c b/block/nbd-client.c
index 8e89add..31db557 100644
--- a/block/nbd-client.c
+++ b/block/nbd-client.c
@@ -275,6 +275,41 @@ int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
return -reply.error;
}
+int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
+ int count, BdrvRequestFlags flags)
+{
+ ssize_t ret;
+ NBDClientSession *client = nbd_get_client_session(bs);
+ NBDRequest request = {
+ .type = NBD_CMD_WRITE_ZEROES,
+ .from = offset,
+ .len = count,
+ };
+ NBDReply reply;
+
+ if (!(client->nbdflags & NBD_FLAG_SEND_WRITE_ZEROES)) {
+ return -ENOTSUP;
+ }
+
+ if (flags & BDRV_REQ_FUA) {
+ assert(client->nbdflags & NBD_FLAG_SEND_FUA);
+ request.flags |= NBD_CMD_FLAG_FUA;
+ }
+ if (!(flags & BDRV_REQ_MAY_UNMAP)) {
+ request.flags |= NBD_CMD_FLAG_NO_HOLE;
+ }
+
+ nbd_coroutine_start(client, &request);
+ ret = nbd_co_send_request(bs, &request, NULL);
+ if (ret < 0) {
+ reply.error = -ret;
+ } else {
+ nbd_co_receive_reply(client, &request, &reply, NULL);
+ }
+ nbd_coroutine_end(client, &request);
+ return -reply.error;
+}
+
int nbd_client_co_flush(BlockDriverState *bs)
{
NBDClientSession *client = nbd_get_client_session(bs);
diff --git a/block/nbd.c b/block/nbd.c
index e227490..6c7bbc8 100644
--- a/block/nbd.c
+++ b/block/nbd.c
@@ -403,6 +403,7 @@ static int nbd_co_flush(BlockDriverState *bs)
static void nbd_refresh_limits(BlockDriverState *bs, Error **errp)
{
bs->bl.max_pdiscard = NBD_MAX_BUFFER_SIZE;
+ bs->bl.max_pwrite_zeroes = NBD_MAX_BUFFER_SIZE;
bs->bl.max_transfer = NBD_MAX_BUFFER_SIZE;
}
@@ -491,6 +492,7 @@ static BlockDriver bdrv_nbd = {
.bdrv_file_open = nbd_open,
.bdrv_co_preadv = nbd_client_co_preadv,
.bdrv_co_pwritev = nbd_client_co_pwritev,
+ .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_pdiscard = nbd_client_co_pdiscard,
@@ -509,6 +511,7 @@ static BlockDriver bdrv_nbd_tcp = {
.bdrv_file_open = nbd_open,
.bdrv_co_preadv = nbd_client_co_preadv,
.bdrv_co_pwritev = nbd_client_co_pwritev,
+ .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_pdiscard = nbd_client_co_pdiscard,
@@ -527,6 +530,7 @@ static BlockDriver bdrv_nbd_unix = {
.bdrv_file_open = nbd_open,
.bdrv_co_preadv = nbd_client_co_preadv,
.bdrv_co_pwritev = nbd_client_co_pwritev,
+ .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes,
.bdrv_close = nbd_close,
.bdrv_co_flush_to_os = nbd_co_flush,
.bdrv_co_pdiscard = nbd_client_co_pdiscard,
--
2.7.4
next prev parent reply other threads:[~2016-10-13 20:59 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-13 20:58 [Qemu-devel] [PATCH v6 00/15] nbd: efficient write zeroes Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 01/15] nbd: Add qemu-nbd -D for human-readable description Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 02/15] nbd: Treat flags vs. command type as separate fields Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 03/15] nbd: Rename NBDRequest to NBDRequestData Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 04/15] nbd: Rename NbdClientSession to NBDClientSession Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 05/15] nbd: Rename struct nbd_request and nbd_reply Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 06/15] nbd: Share common reply-sending code in server Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 07/15] nbd: Send message along with server NBD_REP_ERR errors Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 08/15] nbd: Share common option-sending code in client Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 09/15] nbd: Let server know when client gives up negotiation Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 10/15] nbd: Let client skip portions of server reply Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 11/15] nbd: Less allocation during NBD_OPT_LIST Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 12/15] nbd: Support shorter handshake Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 13/15] nbd: Improve server handling of shutdown requests Eric Blake
2016-10-13 20:58 ` [Qemu-devel] [PATCH v6 14/15] nbd: Implement NBD_CMD_WRITE_ZEROES on server Eric Blake
2016-10-13 20:58 ` Eric Blake [this message]
2016-10-14 0:00 ` [Qemu-devel] [PATCH v6 00/15] nbd: efficient write zeroes no-reply
2016-10-14 2:06 ` Eric Blake
2016-10-14 4:59 ` no-reply
2016-10-14 14:19 ` Eric Blake
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1476392335-9256-16-git-send-email-eblake@redhat.com \
--to=eblake@redhat.com \
--cc=kwolf@redhat.com \
--cc=mreitz@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).