nbd minimal structured read

Pass handle parameter directly, as the whole request isn't needed.

Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>

---
 block/nbd-client.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Minimal implementation: drop most of additional error information.

Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>

---
 block/nbd-client.h  |   2 +
 include/block/nbd.h |  15 ++++-
 block/nbd-client.c  |  97 +++++++++++++++++++++++++-----
 nbd/client.c        | 169 +++++++++++++++++++++++++++++++++++++++++++++++-----
 4 files changed, 249 insertions(+), 34 deletions(-)

+++ b/include/block/nbd.h

include/block/nbd.h +2 -1

@@ -57,11 +57,17 @@ struct NBDRequest {
 };
 typedef struct NBDRequest NBDRequest;
 
-struct NBDReply {
+typedef struct NBDReply {
+    bool simple;
     uint64_t handle;

block/nbd-client.c +0 -1

     int ret;
     int i = HANDLE_TO_INDEX(s, handle);

@@ -191,29 +192,95 @@ static int nbd_co_receive_reply(NBDClientSession *s,
     qemu_coroutine_yield();
     s->requests[i].receiving = false;
     if (!s->ioc || s->quit) {
-        ret = -EIO;

block/nbd-client.c +10 -3

+    ret = -s->reply.error;
+    if (ret < 0) {
+        goto out;
+    }
+
+    if (s->reply.simple) {
+        if (qiov) {
             if (qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
-                                      NULL) < 0) {
-                ret = -EIO;
-                s->quit = true;
+                                      NULL) < 0)
+            {
+                goto fatal;
             }

Quoted 15 lines · Vladimir Sementsov-Ogievskiy

Minimal implementation: drop most of additional error information.

Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>

---
  block/nbd-client.h  |   2 +
  include/block/nbd.h |  15 ++++-
  block/nbd-client.c  |  97 +++++++++++++++++++++++++-----
  nbd/client.c        | 169 +++++++++++++++++++++++++++++++++++++++++++++++-----
  4 files changed, 249 insertions(+), 34 deletions(-)

+++ b/include/block/nbd.h

include/block/nbd.h +0 -0

@@ -57,11 +57,17 @@ struct NBDRequest {
  };
  typedef struct NBDRequest NBDRequest;
 

Quoted 4 lines

include/block/nbd.h +2 -1

-struct NBDReply {
+typedef struct NBDReply {
+    bool simple;
      uint64_t handle;

include/block/nbd.h struct NBDRequest { · line 70

Quoted 10 lines

include/block/nbd.h +7 -2

      uint32_t error;
-};
-typedef struct NBDReply NBDReply;
+
+    uint16_t flags;
+    uint16_t type;
+    uint32_t tail_length;
+    uint64_t offset;
+    uint32_t hole_size;
+} NBDReply;

This feels like it should be a discriminated union, rather than a struct containing fields that are only sometimes valid...

include/block/nbd.h enum { · line 189

Quoted 7 lines

include/block/nbd.h +2 -0

  #define NBD_SREP_TYPE_NONE          0
  #define NBD_SREP_TYPE_OFFSET_DATA   1
+#define NBD_SREP_TYPE_OFFSET_HOLE   2
  #define NBD_SREP_TYPE_ERROR         NBD_SREP_ERR(1)
+#define NBD_SREP_TYPE_ERROR_OFFSET  NBD_SREP_ERR(2)

...especially since there is more than one type of SREP packet layout.

I also wonder why you are defining constants in a piecemeal fashion, rather than all at once (even if your minimal server implementation doesn't send a particular constant, there's no harm in defining them all up front).

block/nbd-client.c err: · line 182

Quoted 10 lines

+++ b/block/nbd-client.c

block/nbd-client.c +1 -3

@@ -179,9 +179,10 @@ err:
      return rc;
  }
 
-static int nbd_co_receive_reply(NBDClientSession *s,
-                                uint64_t handle,
-                                QEMUIOVector *qiov)
+static int nbd_co_receive_1_reply_or_chunk(NBDClientSession *s,

Long name, and unusual to mix in "1" instead of "one". Would it be better to name it nbd_co_receive_chunk, where we declare that a simple reply is (roughly) the same amount of effort as a chunk in a structured reply?

block/nbd-client.c err: · line 186

Quoted 4 lines

block/nbd-client.c +3 -0

+                                           uint64_t handle,
+                                           bool *cont,
+                                           QEMUIOVector *qiov)
  {

No documentation of the function?

Quoted 14 lines

block/nbd-client.c +3 -5

      int ret;
      int i = HANDLE_TO_INDEX(s, handle);

@@ -191,29 +192,95 @@ static int nbd_co_receive_reply(NBDClientSession *s,
      qemu_coroutine_yield();
      s->requests[i].receiving = false;
      if (!s->ioc || s->quit) {
-        ret = -EIO;
-    } else {
-        assert(s->reply.handle == handle);
-        ret = -s->reply.error;
-        if (qiov && s->reply.error == 0) {
+        *cont = false;
+        return -EIO;
+    }

block/nbd-client.c static int nbd_co_receive_reply(NBDClientSession *s, · line 200

Quoted 3 lines

block/nbd-client.c +3 -0

+
+    assert(s->reply.handle == handle);
+    *cont = !(s->reply.simple || (s->reply.flags & NBD_SREP_FLAG_DONE));

We need to validate that the server is not sending us SREP chunks unless we negotiated them. I'm thinking it might be better to do it here (maybe you did it somewhere else, but I haven't seen it yet; I'm reviewing the patch in textual order rather than the order in which things are called).

Quoted 16 lines

block/nbd-client.c +10 -3

+    ret = -s->reply.error;
+    if (ret < 0) {
+        goto out;
+    }
+
+    if (s->reply.simple) {
+        if (qiov) {
              if (qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
-                                      NULL) < 0) {
-                ret = -EIO;
-                s->quit = true;
+                                      NULL) < 0)
+            {
+                goto fatal;
              }

block/nbd-client.c static int nbd_co_receive_reply(NBDClientSession *s, · line 220

Quoted 10 lines

block/nbd-client.c +6 -2

          }
+        goto out;
+    }
 
-        /* Tell the read handler to read another header.  */
-        s->reply.handle = 0;
+    /* here we deal with successful structured reply */
+    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command

callback function then parses the chunks.

Especially since the
definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile). This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

My review has to pause here for now...

Quoted 24 lines · Eric Blake

Quoted 23 lines · Vladimir Sementsov-Ogievskiy

Minimal implementation: drop most of additional error information.

Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>

---
  block/nbd-client.h  |   2 +
  include/block/nbd.h |  15 ++++- block/nbd-client.c  |  97 +++++++++++++++++++++++++----- nbd/client.c        | 169
+++++++++++++++++++++++++++++++++++++++++++++++-----

  4 files changed, 249 insertions(+), 34 deletions(-)

+++ b/include/block/nbd.h

include/block/nbd.h +3 -2

@@ -57,11 +57,17 @@ struct NBDRequest {
   };
   typedef struct NBDRequest NBDRequest;
   -struct NBDReply {
+typedef struct NBDReply {
+    bool simple;
       uint64_t handle;
       uint32_t error;
-};
-typedef struct NBDReply NBDReply;
+

Quoted 8 lines

include/block/nbd.h struct NBDRequest { · line 70

Quoted 6 lines

include/block/nbd.h +6 -0

+    uint16_t flags;
+    uint16_t type;
+    uint32_t tail_length;
+    uint64_t offset;
+    uint32_t hole_size;
+} NBDReply;

This feels like it should be a discriminated union, rather than a struct containing fields that are only sometimes valid...

No:

simple, handle and error are always valid
flags, type, tail_length are valid for all structured replies offset and hole_size are valid for structured hole reply

so, we have one case when all fields are valid.. how do you see it with union, what is the real difference? I think it would be just a complication.

Quoted 8 lines

Quoted 6 lines

include/block/nbd.h +1 -0

     #define NBD_SREP_TYPE_NONE 0
   #define NBD_SREP_TYPE_OFFSET_DATA   1
+#define NBD_SREP_TYPE_OFFSET_HOLE   2

  #define NBD_SREP_TYPE_ERROR         NBD_SREP_ERR(1) +#define NBD_SREP_TYPE_ERROR_OFFSET  NBD_SREP_ERR(2)

...especially since there is more than one type of SREP packet layout.

Quoted 4 lines

I also wonder why you are defining constants in a piecemeal fashion, rather than all at once (even if your minimal server implementation doesn't send a particular constant, there's no harm in defining them all up front).

hmm. just to not define unused constants. It doesn't matter, I can define them all if you prefer.

Quoted 36 lines

block/nbd-client.c err: · line 182

Quoted 9 lines

+++ b/block/nbd-client.c

block/nbd-client.c +1 -2

@@ -179,9 +179,10 @@ err:
       return rc;
   }
   -static int nbd_co_receive_reply(NBDClientSession *s,
-                                uint64_t handle,
-                                QEMUIOVector *qiov)
+static int nbd_co_receive_1_reply_or_chunk(NBDClientSession *s,

Long name, and unusual to mix in "1" instead of "one".  Would it be better to name it nbd_co_receive_chunk, where we declare that a simple reply is (roughly) the same amount of effort as a chunk in a structured reply?

block/nbd-client.c err: · line 186

Quoted 4 lines

block/nbd-client.c +3 -0

+ uint64_t handle,
+                                           bool *cont,
+                                           QEMUIOVector *qiov)
   {

No documentation of the function?

Quoted 16 lines

block/nbd-client.c +4 -5

       int ret;
       int i = HANDLE_TO_INDEX(s, handle);

@@ -191,29 +192,95 @@ static int 
 nbd_co_receive_reply(NBDClientSession *s,
       qemu_coroutine_yield();
       s->requests[i].receiving = false;
       if (!s->ioc || s->quit) {
-        ret = -EIO;
-    } else {
-        assert(s->reply.handle == handle);
-        ret = -s->reply.error;
-        if (qiov && s->reply.error == 0) {
+        *cont = false;
+        return -EIO;
+    }
+

Quoted 8 lines

Review static int · line 202

Quoted 3 lines

(quoted diff) +2 -0

+    assert(s->reply.handle == handle);
+    *cont = !(s->reply.simple || (s->reply.flags & 
 NBD_SREP_FLAG_DONE));

We need to validate that the server is not sending us SREP chunks unless we negotiated them.  I'm thinking it might be better to do it here (maybe you did it somewhere else, but I haven't seen it yet; I'm reviewing the patch in textual order rather than the order in which things are called).

No, I didn't. Will add (may be early, in reply_entry).

Quoted 25 lines

Quoted 19 lines

block/nbd-client.c +12 -3

+    ret = -s->reply.error;
+    if (ret < 0) {
+        goto out;
+    }
+
+    if (s->reply.simple) {
+        if (qiov) {
               if (qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
-                                      NULL) < 0) {
-                ret = -EIO;
-                s->quit = true;
+                                      NULL) < 0)
+            {
+                goto fatal;
               }
           }
+        goto out;
+    }

Quoted 6 lines

  -        /* Tell the read handler to read another header. */ -        s->reply.handle = 0;
+    /* here we deal with successful structured reply */ +    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

Quoted 4 lines

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command

Hmm. it was my idea to move all reading into one coroutine (in my refactoring series, but Paolo was against).

Or you mean to read a payload as raw? It would lead to double copying it to destination qiov, which I dislike..

Quoted 1 line

callback function then parses the chunks.

per-command? Then callback for CMD_READ would have all of these "smarts", so the whole code would not be simpler.. (However it will simplify adding block-status later).

Quoted 11 lines

   Especially since the
definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile).  This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

My review has to pause here for now...

Minimal implementation: drop most of additional error information.

Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>

---

Hi!

Here is a draft of how to refactor reply-payload receiving if you don't like my previous simple (but not flexible) try. Ofcourse, if we agree on this approach this patch should be splitted into several ones and many things (error handling) should be improved.

The idea is:

nbd_read_reply_entry reads only reply header through nbd/client.c code.

Then, the payload is read through block/nbd-client-cmds.c:
simple payload: generic per-command handler, however it should only exist
for CMD_READ
structured NONE: no payload, handle in nbd_co_receive_one_chunk structured error: read by nbd_handle_structured_error_payload
defined in block/nbd-client-cmds.c
structured success: read by per-[command X reply-type] handler
defined in block/nbd-client-cmds.c

For now nbd-client-cmds.c looks more like nbd-payload.c, but, may be, we should move command sending special-casing (CMD_WRITE) to it too..

Don't waste time on careful reviewing this patch, let's consider first the concept of nbd-client-cmds.c,

  block/nbd-client.h      |  10 +++
  include/block/nbd.h     |  82 ++++++++++++++++--
  nbd/nbd-internal.h      |  25 ------
  block/nbd-client-cmds.c | 220 ++++++++++++++++++++++++++++++++++++++++++++++++
  block/nbd-client.c      | 118 ++++++++++++++++++++------
  nbd/client.c            | 128 ++++++++++++++++------------
  block/Makefile.objs     |   2 +-
  7 files changed, 475 insertions(+), 110 deletions(-)

create mode 100644 block/nbd-client-cmds.c

block/nbd-client.h +10 -0

index b435754b82..abb88e4ea5 100644
@@ -35,6 +35,8 @@ typedef struct NBDClientSession {
      NBDClientRequest requests[MAX_NBD_REQUESTS];
      NBDReply reply;
      bool quit;
+
+    bool structured_reply;
  } NBDClientSession;
 
  NBDClientSession *nbd_get_client_session(BlockDriverState *bs);

@@ -60,4 +62,12 @@ void nbd_client_detach_aio_context(BlockDriverState *bs);
  void nbd_client_attach_aio_context(BlockDriverState *bs,
                                     AioContext *new_context);
 
+int nbd_handle_structured_payload(QIOChannel *ioc, int cmd,
+                                  NBDStructuredReplyChunk *reply, void *opaque);
+int nbd_handle_simple_payload(QIOChannel *ioc, int cmd, void *opaque);
+
+int nbd_handle_structured_error_payload(QIOChannel *ioc,
+                                        NBDStructuredReplyChunk *reply,
+                                        int *request_ret);
+
  #endif /* NBD_CLIENT_H */

include/block/nbd.h +76 -6

index 314f2f9bbc..b9a4e1dfa9 100644
@@ -57,12 +57,6 @@ struct NBDRequest {
  };
  typedef struct NBDRequest NBDRequest;
 
-struct NBDReply {
-    uint64_t handle;
-    uint32_t error;
-};
-typedef struct NBDReply NBDReply;
-
  typedef struct NBDSimpleReply {
      uint32_t magic;  /* NBD_SIMPLE_REPLY_MAGIC */
      uint32_t error;

@@ -77,6 +71,24 @@ typedef struct NBDStructuredReplyChunk {
      uint32_t length; /* length of payload */
  } QEMU_PACKED NBDStructuredReplyChunk;
 
+typedef union NBDReply {
+    NBDSimpleReply simple;
+    NBDStructuredReplyChunk structured;
+    struct {
+        uint32_t magic;
+        uint32_t _skip;
+        uint64_t handle;
+    } QEMU_PACKED;
+} NBDReply;
+
+#define NBD_SIMPLE_REPLY_MAGIC      0x67446698
+#define NBD_STRUCTURED_REPLY_MAGIC  0x668e33ef
+
+static inline bool nbd_reply_is_simple(NBDReply *reply)
+{
+    return reply->magic == NBD_SIMPLE_REPLY_MAGIC;
+}
+
  typedef struct NBDStructuredRead {
      NBDStructuredReplyChunk h;
      uint64_t offset;

@@ -88,6 +100,11 @@ typedef struct NBDStructuredError {
      uint16_t message_length;
  } QEMU_PACKED NBDStructuredError;
 
+typedef struct NBDPayloadOffsetHole {
+    uint64_t offset;
+    uint32_t hole_size;
+} QEMU_PACKED NBDPayloadOffsetHole;
+
  /* Transmission (export) flags: sent from server to client during handshake,
     but describe what will happen during transmission */
  #define NBD_FLAG_HAS_FLAGS         (1 << 0) /* Flags are there */

@@ -178,12 +195,54 @@ enum {
 
  #define NBD_SREP_TYPE_NONE          0
  #define NBD_SREP_TYPE_OFFSET_DATA   1
+#define NBD_SREP_TYPE_OFFSET_HOLE   2
  #define NBD_SREP_TYPE_ERROR         NBD_SREP_ERR(1)
+#define NBD_SREP_TYPE_ERROR_OFFSET  NBD_SREP_ERR(2)
+
+static inline bool nbd_srep_type_is_error(int type)
+{
+    return type & (1 << 15);
+}
+
+/* NBD errors are based on errno numbers, so there is a 1:1 mapping,
+ * but only a limited set of errno values is specified in the protocol.
+ * Everything else is squashed to EINVAL.
+ */
+#define NBD_SUCCESS    0
+#define NBD_EPERM      1
+#define NBD_EIO        5
+#define NBD_ENOMEM     12
+#define NBD_EINVAL     22
+#define NBD_ENOSPC     28
+#define NBD_ESHUTDOWN  108
+
+static inline int nbd_errno_to_system_errno(int err)
+{
+    switch (err) {
+    case NBD_SUCCESS:
+        return 0;
+    case NBD_EPERM:
+        return EPERM;
+    case NBD_EIO:
+        return EIO;
+    case NBD_ENOMEM:
+        return ENOMEM;
+    case NBD_ENOSPC:
+        return ENOSPC;
+    case NBD_ESHUTDOWN:
+        return ESHUTDOWN;
+    case NBD_EINVAL:
+        return EINVAL;
+    }
+
+    return EINVAL;
+}
 
  /* Details collected by NBD_OPT_EXPORT_NAME and NBD_OPT_GO */
  struct NBDExportInfo {
      /* Set by client before nbd_receive_negotiate() */
      bool request_sizes;
+    bool structured_reply;
      /* Set by server results during nbd_receive_negotiate() */
      uint64_t size;
      uint16_t flags;

@@ -233,4 +292,15 @@ void nbd_client_put(NBDClient *client);
  void nbd_server_start(SocketAddress *addr, const char *tls_creds,
                        Error **errp);
 
+/* nbd_read
+ * Reads @size bytes from @ioc. Returns 0 on success.
+ */
+static inline int nbd_read(QIOChannel *ioc, void *buffer, size_t size,
+                           Error **errp)
+{
+    return qio_channel_read_all(ioc, buffer, size, errp) < 0 ? -EIO : 0;
+}
+
+int nbd_drop(QIOChannel *ioc, size_t size, Error **errp);
+
  #endif

nbd/nbd-internal.h +0 -25

index 6b0d1183ba..9f7b6b68e8 100644
@@ -47,8 +47,6 @@
  #define NBD_OLDSTYLE_NEGOTIATE_SIZE (8 + 8 + 8 + 4 + 124)
 
  #define NBD_REQUEST_MAGIC           0x25609513
-#define NBD_SIMPLE_REPLY_MAGIC      0x67446698
-#define NBD_STRUCTURED_REPLY_MAGIC  0x668e33ef
  #define NBD_OPTS_MAGIC              0x49484156454F5054LL
  #define NBD_CLIENT_MAGIC            0x0000420281861253LL
  #define NBD_REP_MAGIC               0x0003e889045565a9LL

@@ -65,18 +63,6 @@
  #define NBD_SET_TIMEOUT             _IO(0xab, 9)
  #define NBD_SET_FLAGS               _IO(0xab, 10)
 
-/* NBD errors are based on errno numbers, so there is a 1:1 mapping,
- * but only a limited set of errno values is specified in the protocol.
- * Everything else is squashed to EINVAL.
- */
-#define NBD_SUCCESS    0
-#define NBD_EPERM      1
-#define NBD_EIO        5
-#define NBD_ENOMEM     12
-#define NBD_EINVAL     22
-#define NBD_ENOSPC     28
-#define NBD_ESHUTDOWN  108
-
  /* nbd_read_eof
   * Tries to read @size bytes from @ioc.
   * Returns 1 on success

@@ -96,15 +82,6 @@ static inline int nbd_read_eof(QIOChannel *ioc, void *buffer, size_t size,
      return ret;
  }
 
-/* nbd_read
- * Reads @size bytes from @ioc. Returns 0 on success.
- */
-static inline int nbd_read(QIOChannel *ioc, void *buffer, size_t size,
-                           Error **errp)
-{
-    return qio_channel_read_all(ioc, buffer, size, errp) < 0 ? -EIO : 0;
-}
-
  /* nbd_write
   * Writes @size bytes to @ioc. Returns 0 on success.
   */

@@ -137,6 +114,4 @@ const char *nbd_rep_lookup(uint32_t rep);
  const char *nbd_info_lookup(uint16_t info);
  const char *nbd_cmd_lookup(uint16_t info);
 
-int nbd_drop(QIOChannel *ioc, size_t size, Error **errp);
-
  #endif

block/nbd-client-cmds.c new file +220 -0

new file mode 100644
index 0000000000..488a3dc267
@@ -0,0 +1,220 @@
+/*
+ * QEMU Block driver for NBD
+ *
+ * Copyright (c) 2017 Parallels International GmbH
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+#include "nbd-client.h"
+
+
+typedef int (*NBDCommandFn)(QIOChannel *ioc,
+                            NBDStructuredReplyChunk *chunk,
+                            void *opaque);
+typedef struct NBDCommand {
+    int (*read_simple_payload)(QIOChannel *ioc, void *opaque);
+    NBDCommandFn read_offset_data;
+    NBDCommandFn read_offset_hole;
+} NBDCommand;
+
+static int nbd_cmd_read__siple_payload(QIOChannel *ioc, void *opaque)
+{
+    QEMUIOVector *qiov = (QEMUIOVector *)opaque;
+    return qio_channel_readv_all(ioc, qiov->iov, qiov->niov, NULL);
+}
+
+static int nbd_cmd_read__offset_data(QIOChannel *ioc,
+                                     NBDStructuredReplyChunk *chunk,
+                                     void *opaque)
+{
+    QEMUIOVector *qiov = (QEMUIOVector *)opaque;
+    QEMUIOVector sub_qiov;
+    uint64_t offset;
+    size_t data_size;
+    int ret;
+
+    if (chunk->length < sizeof(offset)) {
+        return -1;
+    }
+
+    if (nbd_read(ioc, &offset, sizeof(offset), NULL) < 0) {
+        return -1;
+    }
+    be64_to_cpus(&offset);
+
+    data_size = chunk->length - sizeof(offset);
+    if (offset + data_size > qiov->size) {
+        return -1;
+    }
+
+    qemu_iovec_init(&sub_qiov, qiov->niov);
+    qemu_iovec_concat(&sub_qiov, qiov, offset, data_size);
+    ret = qio_channel_readv_all(ioc, sub_qiov.iov, sub_qiov.niov, NULL);
+    qemu_iovec_destroy(&sub_qiov);
+
+    return ret;
+}
+
+static int nbd_cmd_read__offset_hole(QIOChannel *ioc,
+                                     NBDStructuredReplyChunk *chunk,
+                                     void *opaque)
+{
+    QEMUIOVector *qiov = (QEMUIOVector *)opaque;
+    NBDPayloadOffsetHole pl;
+
+    if (chunk->length != sizeof(pl)) {
+        return -1;
+    }
+
+    if (nbd_read(ioc, &pl, sizeof(pl), NULL) < 0) {
+        return -1;
+    }
+
+    be64_to_cpus(&pl.offset);
+    be32_to_cpus(&pl.hole_size);
+
+    if (pl.offset + pl.hole_size > qiov->size) {
+        return -1;
+    }
+
+    qemu_iovec_memset(qiov, pl.offset, 0, pl.hole_size);
+
+    return 0;
+}
+
+NBDCommand nbd_cmd_read = {
+    .read_simple_payload = nbd_cmd_read__siple_payload,
+    .read_offset_data = nbd_cmd_read__offset_data,
+    .read_offset_hole = nbd_cmd_read__offset_hole,
+};
+
+static NBDCommand *nbd_get_cmd(int cmd)
+{
+    switch (cmd) {
+    case NBD_CMD_READ:
+        return &nbd_cmd_read;
+    }
+
+    return NULL;
+}
+
+static NBDCommandFn nbd_cmd_get_handler(int cmd, int reply_type)
+{
+    NBDCommand *command = nbd_get_cmd(cmd);
+
+    if (command == NULL) {
+        return NULL;
+    }
+
+    switch (reply_type) {
+    case NBD_SREP_TYPE_OFFSET_DATA:
+        return command->read_offset_data;
+    case NBD_SREP_TYPE_OFFSET_HOLE:
+        return command->read_offset_hole;
+    }
+
+    return NULL;
+}
+
+/* nbd_handle_structured_payload
+ * Find corresponding handler and call it.
+ * If not found return -1 (fail)
+ */
+int nbd_handle_structured_payload(QIOChannel *ioc, int cmd, NBDStructuredReplyChunk *chunk,
+                                  void *opaque)
+{
+    NBDCommandFn fn = nbd_cmd_get_handler(cmd, chunk->type);
+
+    if (fn == NULL) {
+        return -1;
+    }
+
+    return fn(ioc, chunk, opaque);
+}
+
+/* nbd_handle_simple_payload
+ * Find corresponding handler and call it.
+ * If not found return 0 (success)
+ */
+int nbd_handle_simple_payload(QIOChannel *ioc, int cmd, void *opaque)
+{
+    NBDCommand *command = nbd_get_cmd(cmd);
+
+    if (command == NULL || command->read_simple_payload == NULL) {
+        return 0;
+    }
+
+    return command->read_simple_payload(ioc, opaque);
+}
+
+int nbd_handle_structured_error_payload(QIOChannel *ioc, NBDStructuredReplyChunk *chunk,
+                                        int *request_ret)
+{
+    uint32_t error;
+    uint16_t message_size;
+    int ret;
+    assert(chunk->type & (1 << 15));
+
+    switch (chunk->type) {
+    case NBD_SREP_TYPE_ERROR:
+    case NBD_SREP_TYPE_ERROR_OFFSET:
+        ret = nbd_read(ioc, &error, sizeof(error), NULL);
+        if (ret < 0) {
+            return ret;
+        }
+        be32_to_cpus(&error);
+
+        ret = nbd_read(ioc, &message_size, sizeof(message_size), NULL);
+        if (ret < 0) {
+            return ret;
+        }
+        be16_to_cpus(&message_size);
+
+        if (message_size > 0) {
+            /* TODO: provide error message to user */
+            ret = nbd_drop(ioc, message_size, NULL);
+            if (ret < 0) {
+                return ret;
+            }
+        }
+
+        if (chunk->type == NBD_SREP_TYPE_ERROR_OFFSET) {
+            /* drop 64bit offset */
+            ret = nbd_drop(ioc, 8, NULL);
+            if (ret < 0) {
+                return ret;
+            }
+        }
+        break;
+    default:
+        /* unknown error */
+        ret = nbd_drop(ioc, chunk->length, NULL);
+        if (ret < 0) {
+            return ret;
+        }
+
+        error = NBD_EINVAL;
+    }
+
+    *request_ret = nbd_errno_to_system_errno(error);
+    return 0;
+}

block/nbd-client.c +94 -24

index e4f0c789f4..cf80564a83 100644
@@ -179,9 +179,9 @@ err:
      return rc;
  }
 
-static int nbd_co_receive_reply(NBDClientSession *s,
-                                uint64_t handle,
-                                QEMUIOVector *qiov)
+static int nbd_co_receive_one_chunk(NBDClientSession *s, uint64_t handle,
+                                    int request_cmd, void *request_opaque,
+                                    bool *cont)
  {
      int ret;
      int i = HANDLE_TO_INDEX(s, handle);

@@ -191,29 +191,93 @@ static int nbd_co_receive_reply(NBDClientSession *s,
      qemu_coroutine_yield();
      s->requests[i].receiving = false;
      if (!s->ioc || s->quit) {
-        ret = -EIO;
-    } else {
-        assert(s->reply.handle == handle);
-        ret = -s->reply.error;
-        if (qiov && s->reply.error == 0) {
-            if (qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
-                                      NULL) < 0) {
-                ret = -EIO;
-                s->quit = true;
+        *cont = false;
+        return -EIO;
+    }
+
+    assert(s->reply.handle == handle);
+
+    if (nbd_reply_is_simple(&s->reply)) {
+        *cont = false;
+        ret = -s->reply.simple.error;
+        if (s->structured_reply && request_cmd == NBD_CMD_READ) {
+            goto fatal;
+        }
+        if (ret == 0) {
+            if (nbd_handle_simple_payload(s->ioc, request_cmd,
+                                          request_opaque) < 0)
+            {
+                goto fatal;
              }
          }
+        goto out;
+    }
+
+    /* handle structured reply chunk */
+
+    *cont = !(s->reply.structured.flags & NBD_SREP_FLAG_DONE);
 
-        /* Tell the read handler to read another header.  */
-        s->reply.handle = 0;
+    if (s->reply.structured.type == NBD_SREP_TYPE_NONE) {
+        goto out;
      }
 
-    s->requests[i].coroutine = NULL;
+    if (nbd_srep_type_is_error(s->reply.structured.type)) {
+        if (nbd_handle_structured_error_payload(s->ioc, &s->reply.structured,
+                                                &ret) < 0)
+        {
+            goto fatal;
+        }
+        goto out;
+    }
+
+    /* here we deal with successful not-NONE structured reply */
+    if (nbd_handle_structured_payload(s->ioc, request_cmd,
+                                      &s->reply.structured,
+                                      request_opaque) < 0)
+    {
+        goto fatal;
+    }
+
+out:
+    /* For assert at loop start in nbd_read_reply_entry */
+    s->reply.handle = 0;
 
-    /* Kick the read_reply_co to get the next reply.  */
      if (s->read_reply_co) {
          aio_co_wake(s->read_reply_co);
      }
 
+    return ret;
+
+fatal:
+    /* protocol or ioc failure */
+    *cont = false;
+    s->quit = true;
+    if (s->read_reply_co) {
+        aio_co_wake(s->read_reply_co);
+    }
+
+    return -EIO;
+}
+
+static int nbd_co_receive_reply(NBDClientSession *s,
+                                uint64_t handle,
+                                int request_cmd,
+                                void *request_opaque)
+{
+    int ret = 0;
+    int i = HANDLE_TO_INDEX(s, handle);
+    bool cont = true;
+
+    while (cont) {
+        int rc = nbd_co_receive_one_chunk(s, handle, request_cmd,
+                                          request_opaque, &cont);
+        if (rc < 0 && ret == 0) {
+            ret = rc;
+        }
+    }
+
+    s->requests[i].coroutine = NULL;
+
      qemu_co_mutex_lock(&s->send_mutex);
      s->in_flight--;
      qemu_co_queue_next(&s->free_sema);

@@ -224,22 +288,28 @@ static int nbd_co_receive_reply(NBDClientSession *s,
 
  static int nbd_co_request(BlockDriverState *bs,
                            NBDRequest *request,
-                          QEMUIOVector *qiov)
+                          void *request_opaque)
  {
      NBDClientSession *client = nbd_get_client_session(bs);
      int ret;
+    QEMUIOVector *send_qiov = NULL;
+
+    if (request->type == NBD_CMD_WRITE) {
+        /* TODO: move request sending special casing to nbd-client-cmds.c like
+         * receiving part. */
+        send_qiov = request_opaque;
+        request_opaque = NULL;
+        assert(send_qiov &&
+               request->len == iov_size(send_qiov->iov, send_qiov->niov));
+    }
 
-    assert(!qiov || request->type == NBD_CMD_WRITE ||
-           request->type == NBD_CMD_READ);
-    assert(!qiov || request->len == iov_size(qiov->iov, qiov->niov));
-    ret = nbd_co_send_request(bs, request,
-                              request->type == NBD_CMD_WRITE ? qiov : NULL);
+    ret = nbd_co_send_request(bs, request, send_qiov);
      if (ret < 0) {
          return ret;
      }
 
-    return nbd_co_receive_reply(client, request->handle,
-                                request->type == NBD_CMD_READ ? qiov : NULL);
+    return nbd_co_receive_reply(client, request->handle, request->type,
+                                request_opaque);
  }
 
  int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,

nbd/client.c +74 -54

index 51ae492e92..d0e9b8f138 100644
@@ -22,38 +22,6 @@
  #include "trace.h"
  #include "nbd-internal.h"
 
-static int nbd_errno_to_system_errno(int err)
-{
-    int ret;
-    switch (err) {
-    case NBD_SUCCESS:
-        ret = 0;
-        break;
-    case NBD_EPERM:
-        ret = EPERM;
-        break;
-    case NBD_EIO:
-        ret = EIO;
-        break;
-    case NBD_ENOMEM:
-        ret = ENOMEM;
-        break;
-    case NBD_ENOSPC:
-        ret = ENOSPC;
-        break;
-    case NBD_ESHUTDOWN:
-        ret = ESHUTDOWN;
-        break;
-    default:
-        trace_nbd_unknown_error(err);
-        /* fallthrough */
-    case NBD_EINVAL:
-        ret = EINVAL;
-        break;
-    }
-    return ret;
-}
-
  /* Definitions for opaque data types */
 
  static QTAILQ_HEAD(, NBDExport) exports = QTAILQ_HEAD_INITIALIZER(exports);

@@ -719,6 +687,13 @@ int nbd_receive_negotiate(QIOChannel *ioc, const char *name,
          if (fixedNewStyle) {
              int result;
 
+            result = nbd_request_simple_option(ioc, NBD_OPT_STRUCTURED_REPLY,
+                                               errp);
+            if (result < 0) {
+                goto fail;
+            }
+            info->structured_reply = result > 0;
+
              /* Try NBD_OPT_GO first - if it works, we are done (it
               * also gives us a good message if the server requires
               * TLS).  If it is not available, fall back to

@@ -759,6 +734,12 @@ int nbd_receive_negotiate(QIOChannel *ioc, const char *name,
              goto fail;
          }
          be16_to_cpus(&info->flags);
+
+        if (info->structured_reply && !(info->flags & NBD_CMD_FLAG_DF)) {
+            error_setg(errp, "Structured reply is negotiated, "
+                             "but DF flag is not.");
+            goto fail;
+        }
      } else if (magic == NBD_CLIENT_MAGIC) {
          uint32_t oldflags;
 

@@ -942,6 +923,46 @@ int nbd_send_request(QIOChannel *ioc, NBDRequest *request)
      return nbd_write(ioc, buf, sizeof(buf), NULL);
  }
 
+/* nbd_receive_simple_reply
+ * Read simple reply except magic field (which should be already read)
+ */
+static int nbd_receive_simple_reply(QIOChannel *ioc, NBDSimpleReply *reply,
+                                   Error **errp)
+{
+    int ret = nbd_read(ioc, (uint8_t *)reply + sizeof(reply->magic),
+                       sizeof(*reply) - sizeof(reply->magic), errp);
+    if (ret < 0) {
+        return ret;
+    }
+
+    be32_to_cpus(&reply->error);
+    be64_to_cpus(&reply->handle);
+
+    return 0;
+}
+
+/* nbd_receive_structured_reply_chunk
+ * Read structured reply chunk except magic field (which should be already read)
+ * Payload is not read.
+ */
+static int nbd_receive_structured_reply_chunk(QIOChannel *ioc,
+                                              NBDStructuredReplyChunk *chunk,
+                                              Error **errp)
+{
+    int ret = nbd_read(ioc, (uint8_t *)chunk + sizeof(chunk->magic),
+                       sizeof(*chunk) - sizeof(chunk->magic), errp);
+    if (ret < 0) {
+        return ret;
+    }
+
+    be16_to_cpus(&chunk->flags);
+    be16_to_cpus(&chunk->type);
+    be64_to_cpus(&chunk->handle);
+    be32_to_cpus(&chunk->length);
+
+    return 0;
+}
+
  /* nbd_receive_reply
   * Returns 1 on success
   *         0 on eof, when no data was read (errp is not set)

@@ -949,39 +970,38 @@ int nbd_send_request(QIOChannel *ioc, NBDRequest *request)
   */
  int nbd_receive_reply(QIOChannel *ioc, NBDReply *reply, Error **errp)
  {
-    uint8_t buf[NBD_REPLY_SIZE];
-    uint32_t magic;
      int ret;
 
-    ret = nbd_read_eof(ioc, buf, sizeof(buf), errp);
+    ret = nbd_read_eof(ioc, &reply->magic, sizeof(reply->magic), errp);
      if (ret <= 0) {
          return ret;
      }
 
-    /* Reply
-       [ 0 ..  3]    magic   (NBD_SIMPLE_REPLY_MAGIC)
-       [ 4 ..  7]    error   (0 == no error)
-       [ 7 .. 15]    handle
-     */
+    be32_to_cpus(&reply->magic);
 
-    magic = ldl_be_p(buf);
-    reply->error  = ldl_be_p(buf + 4);
-    reply->handle = ldq_be_p(buf + 8);
+    switch (reply->magic) {
+    case NBD_SIMPLE_REPLY_MAGIC:
+        ret = nbd_receive_simple_reply(ioc, &reply->simple, errp);
+        reply->simple.error = nbd_errno_to_system_errno(reply->simple.error);
 
-    reply->error = nbd_errno_to_system_errno(reply->error);
-
-    if (reply->error == ESHUTDOWN) {
-        /* This works even on mingw which lacks a native ESHUTDOWN */
-        error_setg(errp, "server shutting down");
+        if (reply->simple.error == ESHUTDOWN) {
+            /* This works even on mingw which lacks a native ESHUTDOWN */
+            error_setg(errp, "server shutting down");
+            return -EINVAL;
+        }
+        trace_nbd_receive_reply(reply->magic, reply->simple.error,
+                                reply->simple.handle);
+        break;
+    case NBD_STRUCTURED_REPLY_MAGIC:
+        ret = nbd_receive_structured_reply_chunk(ioc, &reply->structured, errp);
+        break;
+    default:
+        error_setg(errp, "invalid magic (got 0x%" PRIx32 ")", reply->magic);
          return -EINVAL;
      }
-    trace_nbd_receive_reply(magic, reply->error, reply->handle);
-
-    if (magic != NBD_SIMPLE_REPLY_MAGIC) {
-        error_setg(errp, "invalid magic (got 0x%" PRIx32 ")", magic);
-        return -EINVAL;
+    if (ret < 0) {
+        return ret;
      }
 
      return 1;
  }
-

block/Makefile.objs +1 -1

index 6eaf78a046..c99420f42b 100644
@@ -12,7 +12,7 @@ block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o
  block-obj-y += null.o mirror.o commit.o io.o
  block-obj-y += throttle-groups.o
 
-block-obj-y += nbd.o nbd-client.o sheepdog.o
+block-obj-y += nbd.o nbd-client.o nbd-client-cmds.o sheepdog.o
  block-obj-$(CONFIG_LIBISCSI) += iscsi.o
  block-obj-$(if $(CONFIG_LIBISCSI),y,n) += iscsi-opts.o
  block-obj-$(CONFIG_LIBNFS) += nfs.o

block/nbd-client.c static int nbd_co_receive_reply(NBDClientSession *s, · line 220

Quoted 4 lines

block/nbd-client.c +4 -0

+    /* here we deal with successful structured reply */
+    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command callback function then parses the chunks. Especially since the definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile). This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

Quoted 4 lines · Eric Blake

block/nbd-client.c static int nbd_co_receive_reply(NBDClientSession *s, · line 220

Quoted 4 lines

block/nbd-client.c +4 -0

+    /* here we deal with successful structured reply */
+    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

Quoted 11 lines

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command callback function then parses the chunks. Especially since the definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile). This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

I think there should be two kinds of replies: 1) read directly into a QEMUIOVector, using structured replies only as an encapsulation of the

payload; 2) read a chunk at a time into malloc-ed memory, yielding back to the calling coroutine after receiving one complete chunk.

In the end this probably means that you have a read_chunk_header function and a read_chunk function. READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

Maybe qio_channel_readv/writev_full could have "offset" and "bytes" arguments. Most code in iov_send_recv could be cut-and-pasted. (When sheepdog is converted to QIOChannel, iov_send_recv can go away).

Paolo

Quoted 16 lines · Paolo Bonzini

Quoted 15 lines · Eric Blake

Quoted 4 lines

+    /* here we deal with successful structured reply */ +    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command callback function then parses the chunks.  Especially since the definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile).  This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

Quoted 2 lines

I think there should be two kinds of replies: 1) read directly into a QEMUIOVector, using structured replies only as an encapsulation of the

who should read to qiov? reply_entry, or calling coroutine?.. reply_entry should anyway
parse reply, to understand should it read it all or read it to qiov (or yield back, and then
calling coroutine will read to qiov)..

Quoted 13 lines

payload; 2) read a chunk at a time into malloc-ed memory, yielding back to the calling coroutine after receiving one complete chunk.

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

Maybe qio_channel_readv/writev_full could have "offset" and "bytes" arguments.  Most code in iov_send_recv could be cut-and-pasted.  (When sheepdog is converted to QIOChannel, iov_send_recv can go away).

Paolo

Quoted 4 lines · Eric Blake

block/nbd-client.c static int nbd_co_receive_reply(NBDClientSession *s, · line 220

Quoted 4 lines

block/nbd-client.c +4 -0

+    /* here we deal with successful structured reply */
+    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

Quoted 11 lines

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command callback function then parses the chunks. Especially since the definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile). This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

I think there should be two kinds of replies: 1) read directly into a QEMUIOVector, using structured replies only as an encapsulation of the payload; 2) read a chunk at a time into malloc-ed memory, yielding back to the calling coroutine after receiving one complete chunk.

In the end this probably means that you have a read_chunk_header function and a read_chunk function. READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

Maybe qio_channel_readv/writev_full could have "offset" and "bytes" arguments. Most code in iov_send_recv could be cut-and-pasted. (When sheepdog is converted to QIOChannel, iov_send_recv can go away).

Paolo

Quoted 21 lines · Paolo Bonzini

Quoted 4 lines · Eric Blake

Quoted 4 lines

+    /* here we deal with successful structured reply */ +    switch (s->reply.type) {
+        QEMUIOVector sub_qiov;
+    case NBD_SREP_TYPE_OFFSET_DATA:

Quoted 11 lines

This is putting a LOT of smarts directly into the receive routine. Here's where I was previously wondering (and I think Paolo as well) whether it might be better to split the efforts: the generic function reads off the chunk information and any payload, but a per-command callback function then parses the chunks.  Especially since the definition of the chunks differs on a per-command basis (yes, the NBD spec will try to not reuse an SREP chunk type across multiple commands unless the semantics are similar, but that's a bit more fragile).  This particularly matters given my statement above that you want a discriminated union, rather than a struct that contains unused fields, for handling different SREP chunk types.

I think there should be two kinds of replies: 1) read directly into a QEMUIOVector, using structured replies only as an encapsulation of the payload; 2) read a chunk at a time into malloc-ed memory, yielding back to the calling coroutine after receiving one complete chunk.

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request, so loop, receiving them should be common for all requests I think

Quoted 7 lines

Maybe qio_channel_readv/writev_full could have "offset" and "bytes" arguments.  Most code in iov_send_recv could be cut-and-pasted. (When sheepdog is converted to QIOChannel, iov_send_recv can go away).

Paolo

Quoted 9 lines

Quoted 5 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Quoted 8 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

Quoted 4 lines · Vladimir Sementsov-Ogievskiy

Quoted 4 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

Quoted 4 lines

Quoted 3 lines

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

Quoted 8 lines · Paolo Bonzini

Quoted 7 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

Quoted 4 lines

Quoted 1 line

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

So, you want a loop in READ, and separate loop for other commands? Then we will have separate loop for BLOCK_STATUS and for all future commands with specific replies?

Quoted 9 lines · Vladimir Sementsov-Ogievskiy

Quoted 8 lines · Paolo Bonzini

Quoted 7 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

Quoted 8 lines

Quoted 4 lines

Quoted 1 line

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

So, you want a loop in READ, and separate loop for other commands? Then we will have separate loop for BLOCK_STATUS and for all future commands with specific replies?

There should be a separate loop for each command.

The only difference between READ and other commands is that READ receives directly in QEMUIOVector, while other commands can use a common function to to receive each structured reply chunk into malloc-ed memory.

Quoted 14 lines · Paolo Bonzini

Quoted 13 lines · Vladimir Sementsov-Ogievskiy

Quoted 12 lines · Paolo Bonzini

Quoted 8 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

Quoted 8 lines

Quoted 3 lines

So, you want a loop in READ, and separate loop for other commands? Then we will have separate loop for BLOCK_STATUS and for all future commands with specific replies?

There should be a separate loop for each command.

The only difference between READ and other commands is that READ receives directly in QEMUIOVector, while other commands can use a common function to to receive each structured reply chunk into malloc-ed memory.

To make sure we're on the same page, here's how I see it. At a high level, we have:

Each command calls nbd_co_send_request once, then calls nbd_co_receive_reply in a loop until there is an indication of the last packet. nbd_co_receive_reply waits for data to come from the server, and parses the header.

If the packet is unrecognized, report failure and request a quit (negative return value)

If it is old-style:
- if the command is read, and we did not negotiate structured read, then we also read the payload into qiov
- if the command is read, but we negotiated structured read, the server is buggy, so report the bug and request a quit
- for all other commands, there is no payload, return success or failure to the caller based on the header payload
- at any rate, the reply to the caller is that this is the final packet, and there is no payload returned (so we return negative or 1, but never 0)

Otherwise, it is new-style:
- if we did not negotiate structured reply, the server is buggy, so report the bug and request a quit (negative return)
- if the chunk is an error, we process the entire packet and report the error; if we have any commands that care about the error details, we could return the error in a malloc'd discriminated union, but we can probably get by without that. If callers don't care about details, but the error chunk is not final, it may be easier to just store the fact that an error occurred and return 0 to tell the caller to keep looping, and return the negative value later when the final chunk is finally received
- if the chunk is NBD_REPLY_TYPE_NONE, there is no payload, and this should be the final chunk, so the return to the caller can be the same as for old-style (return 1 if we had no earlier error packets, or the saved negative value corresponding to the first error received)
- if the command is read, we can read the payload into qiov (saves malloc'ing space for the reply only to copy it into the qiov), so we don't have to return any data
- for any other command, we malloc space for the non-error payload, and then it is up to the command's loop to process the payload

so the signature can be something like:

int nbd_co_receive_reply(NBDClientSession *s, QEMUIOVector *qiov,
void **payload)

where it returns -errno on failure, 0 if the command is not complete, and 1 if the command is done. READ passes qiov, which is fully populated when the function returns 1; all other commands pass NULL. Commands pass NULL for payload if they don't expect a payload return (this includes READ); but a command that expects a payload (BLOCK_STATUS) passes a pointer in payload and gets malloc'd space stored there if return is 0 or 1.

Does that sound like we're on the right design track?

Quoted 77 lines · Eric Blake

Quoted 14 lines · Paolo Bonzini

Quoted 13 lines · Vladimir Sementsov-Ogievskiy

Quoted 12 lines · Paolo Bonzini

Quoted 8 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

Quoted 8 lines

Quoted 3 lines

So, you want a loop in READ, and separate loop for other commands? Then we will have separate loop for BLOCK_STATUS and for all future commands with specific replies?

There should be a separate loop for each command.

The only difference between READ and other commands is that READ receives directly in QEMUIOVector, while other commands can use a common function to to receive each structured reply chunk into malloc-ed memory.

To make sure we're on the same page, here's how I see it.  At a high level, we have:

Each command calls nbd_co_send_request once, then calls nbd_co_receive_reply in a loop until there is an indication of the last packet.  nbd_co_receive_reply waits for data to come from the server, and parses the header.

If the packet is unrecognized, report failure and request a quit (negative return value)

If it is old-style:
- if the command is read, and we did not negotiate structured read, then we also read the payload into qiov
- if the command is read, but we negotiated structured read, the server is buggy, so report the bug and request a quit
- for all other commands, there is no payload, return success or failure to the caller based on the header payload
- at any rate, the reply to the caller is that this is the final packet, and there is no payload returned (so we return negative or 1, but never 0)

Otherwise, it is new-style:
- if we did not negotiate structured reply, the server is buggy, so report the bug and request a quit (negative return)
- if the chunk is an error, we process the entire packet and report the error; if we have any commands that care about the error details, we could return the error in a malloc'd discriminated union, but we can probably get by without that. If callers don't care about details, but the error chunk is not final, it may be easier to just store the fact that an error occurred and return 0 to tell the caller to keep looping, and return the negative value later when the final chunk is finally received
- if the chunk is NBD_REPLY_TYPE_NONE, there is no payload, and this should be the final chunk, so the return to the caller can be the same as for old-style (return 1 if we had no earlier error packets, or the saved negative value corresponding to the first error received)
- if the command is read, we can read the payload into qiov (saves malloc'ing space for the reply only to copy it into the qiov), so we don't have to return any data
- for any other command, we malloc space for the non-error payload, and then it is up to the command's loop to process the payload

so the signature can be something like:

int nbd_co_receive_reply(NBDClientSession *s, QEMUIOVector *qiov, void **payload)

where it returns -errno on failure, 0 if the command is not complete, and 1 if the command is done.  READ passes qiov, which is fully populated when the function returns 1; all other commands pass NULL. Commands pass NULL for payload if they don't expect a payload return (this includes READ); but a command that expects a payload (BLOCK_STATUS) passes a pointer in payload and gets malloc'd space stored there if return is 0 or 1.

Quoted 2 lines

Does that sound like we're on the right design track?

Hmm. and what is the difference with my patch? Except payload? The only command with payload now is READ (except errors), but you (and me in my patch) suggest to fill this qiov in nbd_co_receive_reply (nbd_co_receive_1_reply_or_chunk in my patch), so payload will be unused for now, we can add it later if it will be needed for BLOCK_STATUS.

Quoted 14 lines · Paolo Bonzini

Quoted 13 lines · Vladimir Sementsov-Ogievskiy

Quoted 12 lines · Paolo Bonzini

Quoted 8 lines · Vladimir Sementsov-Ogievskiy

Quoted 7 lines

Quoted 4 lines

In the end this probably means that you have a read_chunk_header function and a read_chunk function.  READ has a loop that calls read_chunk_header followed by direct reading into the QEMUIOVector, while everyone else calls read_chunk.

accordingly to spec, we can receive several error reply chunks to any request,
so loop, receiving them should be common for all requests I think

as well as handling error chunks should be common..

Yes, reading error chunks should be part of read_chunk_header.

Paolo

Quoted 8 lines

Quoted 3 lines

So, you want a loop in READ, and separate loop for other commands? Then we will have separate loop for BLOCK_STATUS and for all future commands with specific replies?

There should be a separate loop for each command.

The only difference between READ and other commands is that READ receives directly in QEMUIOVector, while other commands can use a common function to to receive each structured reply chunk into malloc-ed memory.

To make sure we're on the same page, here's how I see it. At a high level, we have:

Each command calls nbd_co_send_request once, then calls nbd_co_receive_reply in a loop until there is an indication of the last packet. nbd_co_receive_reply waits for data to come from the server, and parses the header.

If the packet is unrecognized, report failure and request a quit (negative return value)

If it is old-style:
- if the command is read, and we did not negotiate structured read, then we also read the payload into qiov
- if the command is read, but we negotiated structured read, the server is buggy, so report the bug and request a quit
- for all other commands, there is no payload, return success or failure to the caller based on the header payload
- at any rate, the reply to the caller is that this is the final packet, and there is no payload returned (so we return negative or 1, but never 0)

Otherwise, it is new-style:
- if we did not negotiate structured reply, the server is buggy, so report the bug and request a quit (negative return)
- if the chunk is an error, we process the entire packet and report the error; if we have any commands that care about the error details, we could return the error in a malloc'd discriminated union, but we can probably get by without that. If callers don't care about details, but the error chunk is not final, it may be easier to just store the fact that an error occurred and return 0 to tell the caller to keep looping, and return the negative value later when the final chunk is finally received
- if the chunk is NBD_REPLY_TYPE_NONE, there is no payload, and this should be the final chunk, so the return to the caller can be the same as for old-style (return 1 if we had no earlier error packets, or the saved negative value corresponding to the first error received)
- if the command is read, we can read the payload into qiov (saves malloc'ing space for the reply only to copy it into the qiov), so we

don't have to return any data
- for any other command, we malloc space for the non-error payload, and then it is up to the command's loop to process the payload

so the signature can be something like:

int nbd_co_receive_reply(NBDClientSession *s, QEMUIOVector *qiov,
void **payload)

where it returns -errno on failure, 0 if the command is not complete, and 1 if the command is done. READ passes qiov, which is fully populated when the function returns 1; all other commands pass NULL. Commands pass NULL for payload if they don't expect a payload return (this includes READ); but a command that expects a payload (BLOCK_STATUS) passes a pointer in payload and gets malloc'd space stored there if return is 0 or 1.

Does that sound like we're on the right design track?

Quoted 6 lines

- if the chunk is NBD_REPLY_TYPE_NONE, there is no payload, and this should be the final chunk, so the return to the caller can be the same as for old-style (return 1 if we had no earlier error packets, or the saved negative value corresponding to the first error received)
- if the command is read, we can read the payload into qiov (saves malloc'ing space for the reply only to copy it into the qiov), so we

but here you said "This is putting a LOT of smarts directly into the receive routine"

Hi all. Here is minimal structured-read extension implementation for nbd. It is needed mostly to implement nbd block-status extension over it, so it is minimal.

It's based on my three patches, currently on pull-request stage:
block/nbd-client: refactor nbd_co_receive_reply block/nbd-client: simplify check in nbd_co_receive_reply block/nbd-client: nbd_co_send_request: fix return code

clone: tag up-nbd-minimal-structured-read-v1 from https://src.openvz.org/scm/~vsementsov/qemu.git
online: https://src.openvz.org/users/vsementsov/repos/qemu/browse?at=up-nbd-minimal-structured-read-v1

Note: actually it's a continuation of a part of my series "nbd: BLOCK_STATUS", but it was sent about half a year ago, so let's consider it as new series.

Vladimir Sementsov-Ogievskiy (8):
block/nbd-client: assert qiov len once in nbd_co_request block/nbd-client: refactor nbd_co_receive_reply
nbd: rename NBD_REPLY_MAGIC to NBD_SIMPLE_REPLY_MAGIC
nbd-server: refactor simple reply sending
nbd: header constants indenting
nbd: Minimal structured read for server nbd/client: refactor nbd_receive_starttls
nbd: Minimal structured read for client

  block/nbd-client.h                       |   2 +
  include/block/nbd.h                      |  67 +++++++--
  nbd/nbd-internal.h                       |  44 +++---
  block/nbd-client.c                       | 104 +++++++++++---
  nbd/client.c                             | 233 ++++++++++++++++++++++++++-----
  nbd/server.c                             | 189 +++++++++++++++++--------
  nbd/trace-events                         |   7 +-
  tests/qemu-iotests/nbd-fault-injector.py |   4 +-
  8 files changed, 507 insertions(+), 143 deletions(-)