[Qemu-devel] [PATCH v2 0/2] sheepdog: add user-defined redundancy option

[Qemu-devel] [PATCH v2 0/2] sheepdog: add user-defined redundancy option

[Liu Yuan]

v2:
 - fix a typo in comment and commit log

This patch set add one sheepdog specific option for qemu-img to control
redundancy.

This patch set is on top of Kevin's block tree.

Liu Yuan (2):
  sheepdog: pass copy_policy in the request
  sheepdog: refactor do_sd_create()

 block/sheepdog.c |   45 ++++++++++++++++++++++-----------------------
 1 file changed, 22 insertions(+), 23 deletions(-)

-- 
1.7.9.5

[Qemu-devel] [PATCH v2 1/2] sheepdog: pass copy_policy in the request

[Liu Yuan]

Currently copy_policy isn't used. Recent sheepdog supports erasure coding, which
make use of copy_policy internally, but require client explicitly passing
copy_policy from base inode to newly creately inode for snapshot related
operations.

If connected sheep daemon doesn't utilize copy_policy, passing it to sheep
daemon is just one extra null effect operation. So no compatibility problem.

With this patch, sheepdog can provide erasure coded volume for QEMU VM.

Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Liu Yuan <namei.unix@gmail.com>
---
 block/sheepdog.c |   20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

diff --git a/block/sheepdog.c b/block/sheepdog.c
index b8a2985..9f0757b 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -126,7 +126,8 @@ typedef struct SheepdogObjReq {
     uint64_t oid;
     uint64_t cow_oid;
     uint8_t copies;
-    uint8_t reserved[7];
+    uint8_t copy_policy;
+    uint8_t reserved[6];
     uint64_t offset;
 } SheepdogObjReq;
 
@@ -139,7 +140,8 @@ typedef struct SheepdogObjRsp {
     uint32_t data_length;
     uint32_t result;
     uint8_t copies;
-    uint8_t reserved[3];
+    uint8_t copy_policy;
+    uint8_t reserved[2];
     uint32_t pad[6];
 } SheepdogObjRsp;
 
@@ -153,7 +155,8 @@ typedef struct SheepdogVdiReq {
     uint64_t vdi_size;
     uint32_t vdi_id;
     uint8_t copies;
-    uint8_t reserved[3];
+    uint8_t copy_policy;
+    uint8_t reserved[2];
     uint32_t snapid;
     uint32_t pad[3];
 } SheepdogVdiReq;
@@ -1346,7 +1349,8 @@ out:
 }
 
 static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
-                        uint32_t base_vid, uint32_t *vdi_id, int snapshot)
+                        uint32_t base_vid, uint32_t *vdi_id, int snapshot,
+                        uint8_t copy_policy)
 {
     SheepdogVdiReq hdr;
     SheepdogVdiRsp *rsp = (SheepdogVdiRsp *)&hdr;
@@ -1376,6 +1380,7 @@ static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
 
     hdr.data_length = wlen;
     hdr.vdi_size = vdi_size;
+    hdr.copy_policy = copy_policy;
 
     ret = do_req(fd, (SheepdogReq *)&hdr, buf, &wlen, &rlen);
 
@@ -1528,7 +1533,8 @@ static int sd_create(const char *filename, QEMUOptionParameter *options,
         bdrv_unref(bs);
     }
 
-    ret = do_sd_create(s, vdi, vdi_size, base_vid, &vid, 0);
+    /* TODO: allow users to specify copy number */
+    ret = do_sd_create(s, vdi, vdi_size, base_vid, &vid, 0, 0);
     if (!prealloc || ret) {
         goto out;
     }
@@ -1718,7 +1724,7 @@ static int sd_create_branch(BDRVSheepdogState *s)
      */
     deleted = sd_delete(s);
     ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &vid,
-                       !deleted);
+                       !deleted, s->inode.copy_policy);
     if (ret) {
         goto out;
     }
@@ -2008,7 +2014,7 @@ static int sd_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info)
     }
 
     ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &new_vid,
-                       1);
+                       1, s->inode.copy_policy);
     if (ret < 0) {
         error_report("failed to create inode for snapshot. %s",
                      strerror(errno));
-- 
1.7.9.5

[Qemu-devel] [PATCH v2 2/2] sheepdog: refactor do_sd_create()

[Liu Yuan]

We can actually use BDRVSheepdogState *s to pass most of the parameters.

Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Liu Yuan <namei.unix@gmail.com>
---
 block/sheepdog.c |   37 +++++++++++++++----------------------
 1 file changed, 15 insertions(+), 22 deletions(-)

diff --git a/block/sheepdog.c b/block/sheepdog.c
index 9f0757b..e66d2f8 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -1348,9 +1348,7 @@ out:
     return ret;
 }
 
-static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
-                        uint32_t base_vid, uint32_t *vdi_id, int snapshot,
-                        uint8_t copy_policy)
+static int do_sd_create(BDRVSheepdogState *s, uint32_t *vdi_id, int snapshot)
 {
     SheepdogVdiReq hdr;
     SheepdogVdiRsp *rsp = (SheepdogVdiRsp *)&hdr;
@@ -1367,11 +1365,11 @@ static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
      * does not fit in buf?  For now, just truncate and avoid buffer overrun.
      */
     memset(buf, 0, sizeof(buf));
-    pstrcpy(buf, sizeof(buf), filename);
+    pstrcpy(buf, sizeof(buf), s->name);
 
     memset(&hdr, 0, sizeof(hdr));
     hdr.opcode = SD_OP_NEW_VDI;
-    hdr.vdi_id = base_vid;
+    hdr.vdi_id = s->inode.vdi_id;
 
     wlen = SD_MAX_VDI_LEN;
 
@@ -1379,8 +1377,8 @@ static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
     hdr.snapid = snapshot;
 
     hdr.data_length = wlen;
-    hdr.vdi_size = vdi_size;
-    hdr.copy_policy = copy_policy;
+    hdr.vdi_size = s->inode.vdi_size;
+    hdr.copy_policy = s->inode.copy_policy;
 
     ret = do_req(fd, (SheepdogReq *)&hdr, buf, &wlen, &rlen);
 
@@ -1391,7 +1389,7 @@ static int do_sd_create(BDRVSheepdogState *s, char *filename, int64_t vdi_size,
     }
 
     if (rsp->result != SD_RES_SUCCESS) {
-        error_report("%s, %s", sd_strerror(rsp->result), filename);
+        error_report("%s, %s", sd_strerror(rsp->result), s->inode.name);
         return -EIO;
     }
 
@@ -1452,23 +1450,21 @@ static int sd_create(const char *filename, QEMUOptionParameter *options,
                      Error **errp)
 {
     int ret = 0;
-    uint32_t vid = 0, base_vid = 0;
-    int64_t vdi_size = 0;
+    uint32_t vid = 0;
     char *backing_file = NULL;
     BDRVSheepdogState *s;
-    char vdi[SD_MAX_VDI_LEN], tag[SD_MAX_VDI_TAG_LEN];
+    char tag[SD_MAX_VDI_TAG_LEN];
     uint32_t snapid;
     bool prealloc = false;
     Error *local_err = NULL;
 
     s = g_malloc0(sizeof(BDRVSheepdogState));
 
-    memset(vdi, 0, sizeof(vdi));
     memset(tag, 0, sizeof(tag));
     if (strstr(filename, "://")) {
-        ret = sd_parse_uri(s, filename, vdi, &snapid, tag);
+        ret = sd_parse_uri(s, filename, s->name, &snapid, tag);
     } else {
-        ret = parse_vdiname(s, filename, vdi, &snapid, tag);
+        ret = parse_vdiname(s, filename, s->name, &snapid, tag);
     }
     if (ret < 0) {
         goto out;
@@ -1476,7 +1472,7 @@ static int sd_create(const char *filename, QEMUOptionParameter *options,
 
     while (options && options->name) {
         if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
-            vdi_size = options->value.n;
+            s->inode.vdi_size = options->value.n;
         } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FILE)) {
             backing_file = options->value.s;
         } else if (!strcmp(options->name, BLOCK_OPT_PREALLOC)) {
@@ -1494,7 +1490,7 @@ static int sd_create(const char *filename, QEMUOptionParameter *options,
         options++;
     }
 
-    if (vdi_size > SD_MAX_VDI_SIZE) {
+    if (s->inode.vdi_size > SD_MAX_VDI_SIZE) {
         error_report("too big image size");
         ret = -EINVAL;
         goto out;
@@ -1529,12 +1525,11 @@ static int sd_create(const char *filename, QEMUOptionParameter *options,
             goto out;
         }
 
-        base_vid = s->inode.vdi_id;
         bdrv_unref(bs);
     }
 
     /* TODO: allow users to specify copy number */
-    ret = do_sd_create(s, vdi, vdi_size, base_vid, &vid, 0, 0);
+    ret = do_sd_create(s, &vid, 0);
     if (!prealloc || ret) {
         goto out;
     }
@@ -1723,8 +1718,7 @@ static int sd_create_branch(BDRVSheepdogState *s)
      * false bail out.
      */
     deleted = sd_delete(s);
-    ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &vid,
-                       !deleted, s->inode.copy_policy);
+    ret = do_sd_create(s, &vid, !deleted);
     if (ret) {
         goto out;
     }
@@ -2013,8 +2007,7 @@ static int sd_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info)
         goto cleanup;
     }
 
-    ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &new_vid,
-                       1, s->inode.copy_policy);
+    ret = do_sd_create(s, &new_vid, 1);
     if (ret < 0) {
         error_report("failed to create inode for snapshot. %s",
                      strerror(errno));
-- 
1.7.9.5

Re: [Qemu-devel] [PATCH v2 0/2] sheepdog: add user-defined redundancy option

[Liu Yuan]

On Tue, Oct 29, 2013 at 04:23:33PM +0800, Liu Yuan wrote:
> v2:
>  - fix a typo in comment and commit log
> 
> This patch set add one sheepdog specific option for qemu-img to control
> redundancy.
> 
> This patch set is on top of Kevin's block tree.
 
Oops, please ignore this patch set because I formated the wrong patch.

Thanks
Yuan