From: Eric Blake <eblake@redhat.com>
To: qemu-devel@nongnu.org
Cc: qemu-block@nongnu.org, libguestfs@redhat.com,
nbd@other.debian.org,
Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Subject: [PATCH v2 07/15] nbd/server: Initial support for extended headers
Date: Mon, 14 Nov 2022 16:48:40 -0600 [thread overview]
Message-ID: <20221114224848.2186298-8-eblake@redhat.com> (raw)
In-Reply-To: <20221114224848.2186298-1-eblake@redhat.com>
Even though the NBD spec has been altered to allow us to accept
NBD_CMD_READ larger than the max payload size (provided our response
is a hole or broken up over more than one data chunk), we are not
planning to take advantage of that, and continue to cap NBD_CMD_READ
to 32M regardless of header size.
For NBD_CMD_WRITE_ZEROES and NBD_CMD_TRIM, the block layer already
supports 64-bit operations without any effort on our part. For
NBD_CMD_BLOCK_STATUS, the client's length is a hint; the easiest
approach for now is to truncate our answer back to 32 bits, which lets
us delay the effort of implementing NBD_REPLY_TYPE_BLOCK_STATUS_EXT to
a separate patch.
Signed-off-by: Eric Blake <eblake@redhat.com>
---
nbd/nbd-internal.h | 7 ++-
nbd/server.c | 132 +++++++++++++++++++++++++++++++++++----------
2 files changed, 108 insertions(+), 31 deletions(-)
diff --git a/nbd/nbd-internal.h b/nbd/nbd-internal.h
index 0016793ff4..f9fe0b6ce3 100644
--- a/nbd/nbd-internal.h
+++ b/nbd/nbd-internal.h
@@ -1,7 +1,7 @@
/*
* NBD Internal Declarations
*
- * Copyright (C) 2016-2021 Red Hat, Inc.
+ * Copyright (C) 2016-2022 Red Hat, Inc.
*
* This work is licensed under the terms of the GNU GPL, version 2 or later.
* See the COPYING file in the top-level directory.
@@ -35,8 +35,11 @@
* https://github.com/yoe/nbd/blob/master/doc/proto.md
*/
-/* Size of all NBD_OPT_*, without payload */
+/* Size of all compact NBD_CMD_*, without payload */
#define NBD_REQUEST_SIZE (4 + 2 + 2 + 8 + 8 + 4)
+/* Size of all extended NBD_CMD_*, without payload */
+#define NBD_EXTENDED_REQUEST_SIZE (4 + 2 + 2 + 8 + 8 + 8)
+
/* Size of all NBD_REP_* sent in answer to most NBD_OPT_*, without payload */
#define NBD_REPLY_SIZE (4 + 4 + 8)
/* Size of reply to NBD_OPT_EXPORT_NAME */
diff --git a/nbd/server.c b/nbd/server.c
index 4d1400430b..b46655b4d8 100644
--- a/nbd/server.c
+++ b/nbd/server.c
@@ -141,7 +141,7 @@ struct NBDClient {
uint32_t check_align; /* If non-zero, check for aligned client requests */
- bool structured_reply;
+ bool structured_reply; /* also set true if extended_headers is set */
bool extended_headers;
NBDExportMetaContexts export_meta;
@@ -1260,6 +1260,10 @@ static int nbd_negotiate_options(NBDClient *client, Error **errp)
case NBD_OPT_STRUCTURED_REPLY:
if (length) {
ret = nbd_reject_length(client, false, errp);
+ } else if (client->extended_headers) {
+ ret = nbd_negotiate_send_rep_err(
+ client, NBD_REP_ERR_EXT_HEADER_REQD, errp,
+ "extended headers already negotiated");
} else if (client->structured_reply) {
ret = nbd_negotiate_send_rep_err(
client, NBD_REP_ERR_INVALID, errp,
@@ -1276,6 +1280,19 @@ static int nbd_negotiate_options(NBDClient *client, Error **errp)
errp);
break;
+ case NBD_OPT_EXTENDED_HEADERS:
+ if (length) {
+ ret = nbd_reject_length(client, false, errp);
+ } else if (client->extended_headers) {
+ ret = nbd_negotiate_send_rep_err(
+ client, NBD_REP_ERR_INVALID, errp,
+ "extended headers already negotiated");
+ } else {
+ ret = nbd_negotiate_send_rep(client, NBD_REP_ACK, errp);
+ client->structured_reply = client->extended_headers = true;
+ }
+ break;
+
default:
ret = nbd_opt_drop(client, NBD_REP_ERR_UNSUP, errp,
"Unsupported option %" PRIu32 " (%s)",
@@ -1411,11 +1428,13 @@ nbd_read_eof(NBDClient *client, void *buffer, size_t size, Error **errp)
static int nbd_receive_request(NBDClient *client, NBDRequest *request,
Error **errp)
{
- uint8_t buf[NBD_REQUEST_SIZE];
- uint32_t magic;
+ uint8_t buf[NBD_EXTENDED_REQUEST_SIZE];
+ uint32_t magic, expect;
int ret;
+ size_t size = client->extended_headers ? NBD_EXTENDED_REQUEST_SIZE
+ : NBD_REQUEST_SIZE;
- ret = nbd_read_eof(client, buf, sizeof(buf), errp);
+ ret = nbd_read_eof(client, buf, size, errp);
if (ret < 0) {
return ret;
}
@@ -1423,13 +1442,21 @@ static int nbd_receive_request(NBDClient *client, NBDRequest *request,
return -EIO;
}
- /* Request
- [ 0 .. 3] magic (NBD_REQUEST_MAGIC)
- [ 4 .. 5] flags (NBD_CMD_FLAG_FUA, ...)
- [ 6 .. 7] type (NBD_CMD_READ, ...)
- [ 8 .. 15] handle
- [16 .. 23] from
- [24 .. 27] len
+ /*
+ * Compact request
+ * [ 0 .. 3] magic (NBD_REQUEST_MAGIC)
+ * [ 4 .. 5] flags (NBD_CMD_FLAG_FUA, ...)
+ * [ 6 .. 7] type (NBD_CMD_READ, ...)
+ * [ 8 .. 15] handle
+ * [16 .. 23] from
+ * [24 .. 27] len
+ * Extended request
+ * [ 0 .. 3] magic (NBD_EXTENDED_REQUEST_MAGIC)
+ * [ 4 .. 5] flags (NBD_CMD_FLAG_FUA, NBD_CMD_FLAG_PAYLOAD_LEN, ...)
+ * [ 6 .. 7] type (NBD_CMD_READ, ...)
+ * [ 8 .. 15] handle
+ * [16 .. 23] from
+ * [24 .. 31] len
*/
magic = ldl_be_p(buf);
@@ -1437,12 +1464,18 @@ static int nbd_receive_request(NBDClient *client, NBDRequest *request,
request->type = lduw_be_p(buf + 6);
request->handle = ldq_be_p(buf + 8);
request->from = ldq_be_p(buf + 16);
- request->len = ldl_be_p(buf + 24); /* widen 32 to 64 bits */
+ if (client->extended_headers) {
+ request->len = ldq_be_p(buf + 24);
+ expect = NBD_EXTENDED_REQUEST_MAGIC;
+ } else {
+ request->len = ldl_be_p(buf + 24); /* widen 32 to 64 bits */
+ expect = NBD_REQUEST_MAGIC;
+ }
trace_nbd_receive_request(magic, request->flags, request->type,
request->from, request->len);
- if (magic != NBD_REQUEST_MAGIC) {
+ if (magic != expect) {
error_setg(errp, "invalid magic (got 0x%" PRIx32 ")", magic);
return -EINVAL;
}
@@ -1890,14 +1923,37 @@ static int coroutine_fn nbd_co_send_iov(NBDClient *client, struct iovec *iov,
}
static inline void set_be_simple_reply(NBDClient *client, struct iovec *iov,
- uint64_t error, NBDRequest *request)
+ uint32_t error, NBDStructuredError *err,
+ NBDRequest *request)
{
- NBDSimpleReply *reply = iov->iov_base;
+ if (client->extended_headers) {
+ NBDExtendedReplyChunk *chunk = iov->iov_base;
- iov->iov_len = sizeof(*reply);
- stl_be_p(&reply->magic, NBD_SIMPLE_REPLY_MAGIC);
- stl_be_p(&reply->error, error);
- stq_be_p(&reply->handle, request->handle);
+ iov->iov_len = sizeof(*chunk);
+ stl_be_p(&chunk->magic, NBD_EXTENDED_REPLY_MAGIC);
+ stw_be_p(&chunk->flags, NBD_REPLY_FLAG_DONE);
+ stq_be_p(&chunk->handle, request->handle);
+ stq_be_p(&chunk->offset, request->from);
+ if (error) {
+ assert(!iov[1].iov_base);
+ iov[1].iov_base = err;
+ iov[1].iov_len = sizeof(*err);
+ stw_be_p(&chunk->type, NBD_REPLY_TYPE_ERROR);
+ stq_be_p(&chunk->length, sizeof(*err));
+ stl_be_p(&err->error, error);
+ stw_be_p(&err->message_length, 0);
+ } else {
+ stw_be_p(&chunk->type, NBD_REPLY_TYPE_NONE);
+ stq_be_p(&chunk->length, 0);
+ }
+ } else {
+ NBDSimpleReply *reply = iov->iov_base;
+
+ iov->iov_len = sizeof(*reply);
+ stl_be_p(&reply->magic, NBD_SIMPLE_REPLY_MAGIC);
+ stl_be_p(&reply->error, error);
+ stq_be_p(&reply->handle, request->handle);
+ }
}
static int nbd_co_send_simple_reply(NBDClient *client, NBDRequest *request,
@@ -1908,30 +1964,44 @@ static int nbd_co_send_simple_reply(NBDClient *client, NBDRequest *request,
{
NBDReply hdr;
int nbd_err = system_errno_to_nbd_errno(error);
+ NBDStructuredError err;
struct iovec iov[] = {
{.iov_base = &hdr},
{.iov_base = data, .iov_len = len}
};
+ assert(!len || !nbd_err);
trace_nbd_co_send_simple_reply(request->handle, nbd_err,
nbd_err_lookup(nbd_err), len);
- set_be_simple_reply(client, &iov[0], nbd_err, request);
+ set_be_simple_reply(client, &iov[0], nbd_err, &err, request);
- return nbd_co_send_iov(client, iov, len ? 2 : 1, errp);
+ return nbd_co_send_iov(client, iov, iov[1].iov_len ? 2 : 1, errp);
}
static inline void set_be_chunk(NBDClient *client, struct iovec *iov,
uint16_t flags, uint16_t type,
NBDRequest *request, uint32_t length)
{
- NBDStructuredReplyChunk *chunk = iov->iov_base;
+ if (client->extended_headers) {
+ NBDExtendedReplyChunk *chunk = iov->iov_base;
- iov->iov_len = sizeof(*chunk);
- stl_be_p(&chunk->magic, NBD_STRUCTURED_REPLY_MAGIC);
- stw_be_p(&chunk->flags, flags);
- stw_be_p(&chunk->type, type);
- stq_be_p(&chunk->handle, request->handle);
- stl_be_p(&chunk->length, length);
+ iov->iov_len = sizeof(*chunk);
+ stl_be_p(&chunk->magic, NBD_EXTENDED_REPLY_MAGIC);
+ stw_be_p(&chunk->flags, flags);
+ stw_be_p(&chunk->type, type);
+ stq_be_p(&chunk->handle, request->handle);
+ stq_be_p(&chunk->offset, request->from);
+ stq_be_p(&chunk->length, length);
+ } else {
+ NBDStructuredReplyChunk *chunk = iov->iov_base;
+
+ iov->iov_len = sizeof(*chunk);
+ stl_be_p(&chunk->magic, NBD_STRUCTURED_REPLY_MAGIC);
+ stw_be_p(&chunk->flags, flags);
+ stw_be_p(&chunk->type, type);
+ stq_be_p(&chunk->handle, request->handle);
+ stl_be_p(&chunk->length, length);
+ }
}
static int coroutine_fn nbd_co_send_structured_done(NBDClient *client,
@@ -2595,7 +2665,11 @@ static coroutine_fn int nbd_handle_request(NBDClient *client,
return nbd_send_generic_reply(client, request, -EINVAL,
"need non-zero length", errp);
}
- assert(request->len <= UINT32_MAX);
+ if (request->len > UINT32_MAX) {
+ /* For now, truncate our response to a 32 bit window */
+ request->len = QEMU_ALIGN_DOWN(BDRV_REQUEST_MAX_BYTES,
+ client->check_align ?: 1);
+ }
if (client->export_meta.count) {
bool dont_fragment = request->flags & NBD_CMD_FLAG_REQ_ONE;
int contexts_remaining = client->export_meta.count;
--
2.38.1
next prev parent reply other threads:[~2022-11-14 23:27 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-14 22:41 [cross-project PATCH v2] NBD 64-bit extensions Eric Blake
2022-11-14 22:46 ` [PATCH v2 0/6] NBD spec changes for " Eric Blake
2022-11-14 22:46 ` [PATCH v2 1/6] spec: Recommend cap on NBD_REPLY_TYPE_BLOCK_STATUS length Eric Blake
2022-12-16 19:32 ` Vladimir Sementsov-Ogievskiy
2023-03-03 22:17 ` Eric Blake
2023-03-05 8:41 ` Wouter Verhelst
2023-03-06 8:48 ` [Libguestfs] " Laszlo Ersek
2023-03-06 13:48 ` Nir Soffer
2022-11-14 22:46 ` [PATCH v2 2/6] spec: Tweak description of maximum block size Eric Blake
2022-12-16 20:22 ` Vladimir Sementsov-Ogievskiy
2023-03-03 22:20 ` Eric Blake
2023-02-21 15:21 ` Wouter Verhelst
2023-03-03 22:26 ` Eric Blake
2023-03-05 8:45 ` Wouter Verhelst
2022-11-14 22:46 ` [PATCH v2 3/6] spec: Add NBD_OPT_EXTENDED_HEADERS Eric Blake
2022-12-19 18:26 ` Vladimir Sementsov-Ogievskiy
2023-02-22 9:49 ` Wouter Verhelst
2023-03-03 22:36 ` Eric Blake
2023-03-05 8:49 ` Wouter Verhelst
2022-11-14 22:46 ` [PATCH v2 4/6] spec: Allow 64-bit block status results Eric Blake
2022-11-14 22:46 ` [PATCH v2 5/6] spec: Introduce NBD_FLAG_BLOCK_STATUS_PAYLOAD Eric Blake
2023-02-22 10:05 ` Wouter Verhelst
2023-03-03 22:40 ` Eric Blake
2023-03-05 8:50 ` Wouter Verhelst
2022-11-14 22:46 ` [PATCH v2 6/6] RFC: spec: Introduce NBD_REPLY_TYPE_OFFSET_HOLE_EXT Eric Blake
2022-11-14 22:48 ` [PATCH v2 00/15] qemu patches for 64-bit NBD extensions Eric Blake
2022-11-14 22:48 ` [PATCH v2 01/15] nbd/client: Add safety check on chunk payload length Eric Blake
2022-11-14 22:48 ` [PATCH v2 02/15] nbd/server: Prepare for alternate-size headers Eric Blake
2022-11-14 22:48 ` [PATCH v2 03/15] nbd: Prepare for 64-bit request effect lengths Eric Blake
2022-11-14 22:48 ` [PATCH v2 04/15] nbd: Add types for extended headers Eric Blake
2022-11-14 22:48 ` [PATCH v2 05/15] nbd/server: Refactor handling of request payload Eric Blake
2022-11-14 22:48 ` [PATCH v2 06/15] nbd/server: Refactor to pass full request around Eric Blake
2022-11-14 22:48 ` Eric Blake [this message]
2022-11-14 22:48 ` [PATCH v2 08/15] nbd/server: Support 64-bit block status Eric Blake
2022-11-14 22:48 ` [PATCH v2 09/15] nbd/client: Initial support for extended headers Eric Blake
2022-11-14 22:48 ` [PATCH v2 10/15] nbd/client: Accept 64-bit block status chunks Eric Blake
2022-11-14 22:48 ` [PATCH v2 11/15] nbd/client: Request extended headers during negotiation Eric Blake
2022-11-14 22:48 ` [PATCH v2 12/15] nbd/server: Prepare for per-request filtering of BLOCK_STATUS Eric Blake
2022-11-14 22:48 ` [PATCH v2 13/15] nbd/server: Add FLAG_PAYLOAD support to CMD_BLOCK_STATUS Eric Blake
2022-11-14 22:48 ` [PATCH v2 14/15] RFC: nbd/client: Accept 64-bit hole chunks Eric Blake
2022-11-14 22:48 ` [PATCH v2 15/15] RFC: nbd/server: Send 64-bit hole chunk Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 00/23] libnbd 64-bit NBD extensions Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 01/23] block_status: Refactor array storage Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 02/23] internal: Refactor layout of replies in sbuf Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 03/23] protocol: Add definitions for extended headers Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 04/23] states: Prepare to send 64-bit requests Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 05/23] states: Prepare to receive 64-bit replies Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 06/23] states: Break deadlock if server goofs on extended replies Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 07/23] generator: Add struct nbd_extent in prep for 64-bit extents Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 08/23] block_status: Track 64-bit extents internally Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 09/23] block_status: Accept 64-bit extents during block status Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 10/23] api: Add [aio_]nbd_block_status_64 Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 11/23] api: Add several functions for controlling extended headers Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 12/23] copy: Update nbdcopy to use 64-bit block status Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 13/23] dump: Update nbddump " Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 14/23] info: Expose extended-headers support through nbdinfo Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 15/23] info: Update nbdinfo --map to use 64-bit block status Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 16/23] examples: Update copy-libev " Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 17/23] ocaml: Add example for 64-bit extents Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 18/23] generator: Actually request extended headers Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 19/23] api: Add nbd_[aio_]opt_extended_headers() Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 20/23] interop: Add test of 64-bit block status Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 21/23] api: Add nbd_can_block_status_payload() Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 22/23] api: Add nbd_[aio_]block_status_filter() Eric Blake
2022-11-14 22:51 ` [libnbd PATCH v2 23/23] RFC: pread: Accept 64-bit holes 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=20221114224848.2186298-8-eblake@redhat.com \
--to=eblake@redhat.com \
--cc=libguestfs@redhat.com \
--cc=nbd@other.debian.org \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=vsementsov@yandex-team.ru \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.