Skip to contentSkip to conversation

block: logical block provisioning enhancements

PATCH Superseded Peter Lieven 4 participants 47 messages Sep 24, 2013 – Oct 8, 2013 Block layer

v3 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v3 cover block: logical block provisioning enhancements

this patch adds the ability for targets to stay sparse during block migration (if the zero_blocks capability is set) and qemu-img convert even if the target does not have has_zero_init = 1.

the series was especially developed for iSCSI, but it should also work with other drivers with little or no adjustments. these adjustments should be limited to providing block provisioning information through get_block_info and/or honouring BDRV_REQ_MAY_UNMAP on writing zeroes.

there are also 4 patches which are fixes/enhancements for the get_block_status API discovered during development of this series.

v2->v3:
- fix merge conflict in block/qcow2_cluster.c
- changed return type of bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes to bool.
- moved alignment and limits info to a BlockLimits struct (Paolo).
- added magic constanst for default maximum in bdrv_co_do_write_zeroes and bdrv_co_discard (Eric).
- bdrv_co_do_write_zeroes: allocating the bounce buffer only once (Eric), fixed bounce iov_len in the fall back path.
- bdrv_zeroize: added inline docu (Eric) and do not mask flags passed to bdrv_write_zeroes (Eric).
- qemu-img: changed the default hint for -S (min_sparse) in the usage help to 4k. not changing the default as it is unclear why this default was set. size suffixes are already supported (Eric).

v1->v2:
- moved block max_discard and max_write_zeroes to BlockDriverState
- added discard_alignment and write_zeroes_alignment to BlockDriverState
- added bdrv_has_discard_zeroes() and bdrv_has_discard_write_zeroes()
- added logic to bdrv_co_discard and bdrv_co_do_write_zeroes to honour limit and alignment info.
- added support for -S 0 in qemu-img convert.

Peter Lieven (20):
block: make BdrvRequestFlags public
block: add flags to bdrv_*_write_zeroes
block: introduce BDRV_REQ_MAY_UNMAP request flag
block: introduce bdrv_has_discard_zeroes and
bdrv_has_discard_write_zeroes
block/raw: add bdrv_has_discard_zeroes and
bdrv_has_discard_write_zeroes
block: add BlockLimits structure to BlockDriverState
block: honour BlockLimits in bdrv_co_do_write_zeroes
block: honour BlockLimits in bdrv_co_discard
iscsi: simplify iscsi_co_discard
iscsi: set limits in BlockDriverState
iscsi: add bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes
iscsi: add bdrv_co_write_zeroes
block: introduce bdrv_zeroize
block/get_block_status: set *pnum = 0 on error block/get_block_status: avoid segfault if there is no backing_hd block/get_block_status: avoid redundant callouts on raw devices block/get_block_status: fix BDRV_BLOCK_ZERO for unallocated blocks
qemu-img: add support for fully allocated images
qemu-img: conditionally zero out target on convert block/raw: copy BlockLimits on raw_open

 block-migration.c         |    3 +-
 block.c                   |  205 +++++++++++++++++++++++++++++++++++++--------
 block/backup.c            |    3 +-
 block/iscsi.c             |  152 ++++++++++++++++++++++++---------
 block/qcow2-cluster.c     |    2 +-
 block/qcow2.c             |    2 +-
 block/qed.c               |    3 +-
 block/raw_bsd.c           |   62 ++++++++------
 block/vmdk.c              |    3 +-
 include/block/block.h     |   19 ++++-
 include/block/block_int.h |   32 ++++++-
 qemu-img.c                |   18 +++-
 qemu-io-cmds.c            |    2 +-
 13 files changed, 393 insertions(+), 113 deletions(-)
Signature · 1 line
1.7.9.5
#1 lore raw

PATCH v3 1/20 block: make BdrvRequestFlags public

---
 block.c               |    5 -----
 include/block/block.h |    5 +++++
 2 files changed, 5 insertions(+), 5 deletions(-)
block.c +0 -5
index ea4956d..8eefb97 100644
@@ -51,11 +51,6 @@
 
 #define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */
 
-typedef enum {
-    BDRV_REQ_COPY_ON_READ = 0x1,
-    BDRV_REQ_ZERO_WRITE   = 0x2,
-} BdrvRequestFlags;
-
 static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load);
 static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
         int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
include/block/block.h +5 -0
index f808550..9fb0c3d 100644
@@ -62,6 +62,11 @@ typedef struct BlockDevOps {
     void (*resize_cb)(void *opaque);
 } BlockDevOps;
 
+typedef enum {
+    BDRV_REQ_COPY_ON_READ = 0x1,
+    BDRV_REQ_ZERO_WRITE   = 0x2,
+} BdrvRequestFlags;
+
 #define BDRV_O_RDWR        0x0002
 #define BDRV_O_SNAPSHOT    0x0008 /* open the file read only and save writes in a snapshot */
 #define BDRV_O_NOCACHE     0x0020 /* do not use the host page cache */
Signature · 1 line
1.7.9.5
#2 lore raw

PATCH v3 2/20 block: add flags to bdrv_*_write_zeroes

In reply to Peter Lieven
---
 block-migration.c         |    2 +-
 block.c                   |   20 +++++++++++---------
 block/backup.c            |    3 ++-
 block/qcow2-cluster.c     |    2 +-
 block/qcow2.c             |    2 +-
 block/qed.c               |    3 ++-
 block/raw_bsd.c           |    5 +++--
 block/vmdk.c              |    3 ++-
 include/block/block.h     |    4 ++--
 include/block/block_int.h |    2 +-
 qemu-io-cmds.c            |    2 +-
 11 files changed, 27 insertions(+), 21 deletions(-)
block-migration.c +1 -1
index daf9ec1..713a8e3 100644
@@ -780,7 +780,7 @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
             }
 
             if (flags & BLK_MIG_FLAG_ZERO_BLOCK) {
-                ret = bdrv_write_zeroes(bs, addr, nr_sectors);
+                ret = bdrv_write_zeroes(bs, addr, nr_sectors, 0);
             } else {
                 buf = g_malloc(BLOCK_SIZE);
                 qemu_get_buffer(f, buf, BLOCK_SIZE);
block.c +11 -9
index 8eefb97..c453ec1 100644
@@ -79,7 +79,7 @@ static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                                bool is_write);
 static void coroutine_fn bdrv_co_do_rw(void *opaque);
 static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
-    int64_t sector_num, int nb_sectors);
+    int64_t sector_num, int nb_sectors, BdrvRequestFlags flags);
 
 static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
     QTAILQ_HEAD_INITIALIZER(bdrv_states);
@@ -2391,10 +2391,11 @@ int bdrv_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov)
     return bdrv_rwv_co(bs, sector_num, qiov, true, 0);
 }
 
-int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num, int nb_sectors)
+int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
+                      int nb_sectors, BdrvRequestFlags flags)
 {
     return bdrv_rw_co(bs, sector_num, NULL, nb_sectors, true,
-                      BDRV_REQ_ZERO_WRITE);
+                      BDRV_REQ_ZERO_WRITE | flags);
 }
 
 int bdrv_pread(BlockDriverState *bs, int64_t offset,
@@ -2576,7 +2577,7 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
     if (drv->bdrv_co_write_zeroes &&
         buffer_is_zero(bounce_buffer, iov.iov_len)) {
         ret = bdrv_co_do_write_zeroes(bs, cluster_sector_num,
-                                      cluster_nb_sectors);
+                                      cluster_nb_sectors, 0);
     } else {
         /* This does not change the data on the disk, it is not necessary
          * to flush even in cache=writethrough mode.
@@ -2710,7 +2711,7 @@ int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
 }
 
 static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
-    int64_t sector_num, int nb_sectors)
+    int64_t sector_num, int nb_sectors, BdrvRequestFlags flags)
 {
     BlockDriver *drv = bs->drv;
     QEMUIOVector qiov;
@@ -2722,7 +2723,7 @@ static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
 
     /* First try the efficient write zeroes operation */
     if (drv->bdrv_co_write_zeroes) {
-        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
+        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors, flags);
         if (ret != -ENOTSUP) {
             return ret;
         }
@@ -2777,7 +2778,7 @@ static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
     if (ret < 0) {
         /* Do nothing, write notifier decided to fail this request */
     } else if (flags & BDRV_REQ_ZERO_WRITE) {
-        ret = bdrv_co_do_write_zeroes(bs, sector_num, nb_sectors);
+        ret = bdrv_co_do_write_zeroes(bs, sector_num, nb_sectors, flags);
     } else {
         ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov);
     }
@@ -2811,12 +2812,13 @@ int coroutine_fn bdrv_co_writev(BlockDriverState *bs, int64_t sector_num,
 }
 
 int coroutine_fn bdrv_co_write_zeroes(BlockDriverState *bs,
-                                      int64_t sector_num, int nb_sectors)
+                                      int64_t sector_num, int nb_sectors,
+                                      BdrvRequestFlags flags)
 {
     trace_bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
 
     return bdrv_co_do_writev(bs, sector_num, nb_sectors, NULL,
-                             BDRV_REQ_ZERO_WRITE);
+                             BDRV_REQ_ZERO_WRITE | flags);
 }
 
 /**
block/backup.c +2 -1
index 04c4b5c..5f6a642 100644
@@ -138,7 +138,8 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs,
 
         if (buffer_is_zero(iov.iov_base, iov.iov_len)) {
             ret = bdrv_co_write_zeroes(job->target,
-                                       start * BACKUP_SECTORS_PER_CLUSTER, n);
+                                       start * BACKUP_SECTORS_PER_CLUSTER,
+                                       n, 0);
         } else {
             ret = bdrv_co_writev(job->target,
                                  start * BACKUP_SECTORS_PER_CLUSTER, n,
block/qcow2-cluster.c +1 -1
index 738ff73..5215f5b 100644
@@ -1601,7 +1601,7 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
             }
 
             ret = bdrv_write_zeroes(bs->file, offset / BDRV_SECTOR_SIZE,
-                                    s->cluster_sectors);
+                                    s->cluster_sectors, 0);
             if (ret < 0) {
                 qcow2_free_clusters(bs, offset, s->cluster_size,
                         QCOW2_DISCARD_ALWAYS);
block/qcow2.c +1 -1
index 318d95d..6cb3e44 100644
@@ -1600,7 +1600,7 @@ static int qcow2_make_empty(BlockDriverState *bs)
 }
 
 static coroutine_fn int qcow2_co_write_zeroes(BlockDriverState *bs,
-    int64_t sector_num, int nb_sectors)
+    int64_t sector_num, int nb_sectors, BdrvRequestFlags flags)
 {
     int ret;
     BDRVQcowState *s = bs->opaque;
block/qed.c +2 -1
index 6c0cba0..adc2736 100644
@@ -1397,7 +1397,8 @@ static void coroutine_fn qed_co_write_zeroes_cb(void *opaque, int ret)
 
 static int coroutine_fn bdrv_qed_co_write_zeroes(BlockDriverState *bs,
                                                  int64_t sector_num,
-                                                 int nb_sectors)
+                                                 int nb_sectors,
+                                                 BdrvRequestFlags flags)
 {
     BlockDriverAIOCB *blockacb;
     BDRVQEDState *s = bs->opaque;
block/raw_bsd.c +3 -2
index d4ace60..d5ab295 100644
@@ -66,9 +66,10 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
 }
 
 static int coroutine_fn raw_co_write_zeroes(BlockDriverState *bs,
-                                            int64_t sector_num, int nb_sectors)
+                                            int64_t sector_num, int nb_sectors,
+                                            BdrvRequestFlags flags)
 {
-    return bdrv_co_write_zeroes(bs->file, sector_num, nb_sectors);
+    return bdrv_co_write_zeroes(bs->file, sector_num, nb_sectors, flags);
 }
 
 static int coroutine_fn raw_co_discard(BlockDriverState *bs,
block/vmdk.c +2 -1
index 96ef1b5..debf974 100644
@@ -1391,7 +1391,8 @@ static coroutine_fn int vmdk_co_write(BlockDriverState *bs, int64_t sector_num,
 
 static int coroutine_fn vmdk_co_write_zeroes(BlockDriverState *bs,
                                              int64_t sector_num,
-                                             int nb_sectors)
+                                             int nb_sectors,
+                                             BdrvRequestFlags flags)
 {
     int ret;
     BDRVVmdkState *s = bs->opaque;
include/block/block.h +2 -2
index 9fb0c3d..47c6475 100644
@@ -188,7 +188,7 @@ int bdrv_read_unthrottled(BlockDriverState *bs, int64_t sector_num,
 int bdrv_write(BlockDriverState *bs, int64_t sector_num,
                const uint8_t *buf, int nb_sectors);
 int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
-               int nb_sectors);
+               int nb_sectors, BdrvRequestFlags flags);
 int bdrv_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov);
 int bdrv_pread(BlockDriverState *bs, int64_t offset,
                void *buf, int count);
@@ -210,7 +210,7 @@ int coroutine_fn bdrv_co_writev(BlockDriverState *bs, int64_t sector_num,
  * because it may allocate memory for the entire region.
  */
 int coroutine_fn bdrv_co_write_zeroes(BlockDriverState *bs, int64_t sector_num,
-    int nb_sectors);
+    int nb_sectors, BdrvRequestFlags flags);
 BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
     const char *backing_file);
 int bdrv_get_backing_file_depth(BlockDriverState *bs);
include/block/block_int.h +1 -1
index 3eeb6fe..56e1994 100644
@@ -118,7 +118,7 @@ struct BlockDriver {
      * instead.
      */
     int coroutine_fn (*bdrv_co_write_zeroes)(BlockDriverState *bs,
-        int64_t sector_num, int nb_sectors);
+        int64_t sector_num, int nb_sectors, BdrvRequestFlags flags);
     int coroutine_fn (*bdrv_co_discard)(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors);
     int64_t coroutine_fn (*bdrv_co_get_block_status)(BlockDriverState *bs,
qemu-io-cmds.c +1 -1
index 8565d49..66c0b9e 100644
@@ -441,7 +441,7 @@ static void coroutine_fn co_write_zeroes_entry(void *opaque)
     CoWriteZeroes *data = opaque;
 
     data->ret = bdrv_co_write_zeroes(data->bs, data->offset / BDRV_SECTOR_SIZE,
-                                     data->count / BDRV_SECTOR_SIZE);
+                                     data->count / BDRV_SECTOR_SIZE, 0);
     data->done = true;
     if (data->ret < 0) {
         *data->total = data->ret;
Signature · 1 line
1.7.9.5
Reviewed-by on v3 2/20
#3 lore raw

Re: [Qemu-devel] [PATCHv3 02/20] block: add flags to bdrv_*_write_zeroes

Quoted 15 lines · Peter Lieven
---
 block-migration.c         |    2 +-
 block.c                   |   20 +++++++++++---------
 block/backup.c            |    3 ++-
 block/qcow2-cluster.c     |    2 +-
 block/qcow2.c             |    2 +-
 block/qed.c               |    3 ++-
 block/raw_bsd.c           |    5 +++--
 block/vmdk.c              |    3 ++-
 include/block/block.h     |    4 ++--
 include/block/block_int.h |    2 +-
 qemu-io-cmds.c            |    2 +-
 11 files changed, 27 insertions(+), 21 deletions(-)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#4 lore raw

PATCH v3 3/20 block: introduce BDRV_REQ_MAY_UNMAP request flag

In reply to Peter Lieven
---
 block-migration.c     |    3 ++-
 block.c               |    4 ++++
 block/backup.c        |    2 +-
 include/block/block.h |    7 +++++++
 4 files changed, 14 insertions(+), 2 deletions(-)
block-migration.c +2 -1
index 713a8e3..fc4ef93 100644
@@ -780,7 +780,8 @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
             }
 
             if (flags & BLK_MIG_FLAG_ZERO_BLOCK) {
-                ret = bdrv_write_zeroes(bs, addr, nr_sectors, 0);
+                ret = bdrv_write_zeroes(bs, addr, nr_sectors,
+                                        BDRV_REQ_MAY_UNMAP);
             } else {
                 buf = g_malloc(BLOCK_SIZE);
                 qemu_get_buffer(f, buf, BLOCK_SIZE);
block.c +4 -0
index c453ec1..e7bf6af 100644
@@ -2817,6 +2817,10 @@ int coroutine_fn bdrv_co_write_zeroes(BlockDriverState *bs,
 {
     trace_bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
 
+    if (!(bs->open_flags & BDRV_O_UNMAP)) {
+        flags &= ~BDRV_REQ_MAY_UNMAP;
+    }
+
     return bdrv_co_do_writev(bs, sector_num, nb_sectors, NULL,
                              BDRV_REQ_ZERO_WRITE | flags);
 }
block/backup.c +1 -1
index 5f6a642..1ab080d 100644
@@ -139,7 +139,7 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs,
         if (buffer_is_zero(iov.iov_base, iov.iov_len)) {
             ret = bdrv_co_write_zeroes(job->target,
                                        start * BACKUP_SECTORS_PER_CLUSTER,
-                                       n, 0);
+                                       n, BDRV_REQ_MAY_UNMAP);
         } else {
             ret = bdrv_co_writev(job->target,
                                  start * BACKUP_SECTORS_PER_CLUSTER, n,
include/block/block.h +7 -0
index 47c6475..b72c81a 100644
@@ -65,6 +65,13 @@ typedef struct BlockDevOps {
 typedef enum {
     BDRV_REQ_COPY_ON_READ = 0x1,
     BDRV_REQ_ZERO_WRITE   = 0x2,
+    /* The BDRV_REQ_MAY_UNMAP flag is used to indicate that the block driver
+     * is allowed to optimize a write zeroes request by unmapping (discarding)
+     * blocks if it is guaranteed that the result will read back as
+     * zeroes. The flag is only passed to the driver if the block device is
+     * opened with BDRV_O_UNMAP.
+     */
+    BDRV_REQ_MAY_UNMAP    = 0x4,
 } BdrvRequestFlags;
 
 #define BDRV_O_RDWR        0x0002
Signature · 1 line
1.7.9.5
#5 lore raw

PATCH v3 4/20 block: introduce bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

In reply to Peter Lieven
---
 block.c                   |   29 +++++++++++++++++++++++++++++
 include/block/block.h     |    2 ++
 include/block/block_int.h |   13 +++++++++++++
 3 files changed, 44 insertions(+)
block.c +29 -0
index e7bf6af..ac35cb5 100644
@@ -3101,6 +3101,35 @@ int bdrv_has_zero_init(BlockDriverState *bs)
     return 0;
 }
 
+bool bdrv_has_discard_zeroes(BlockDriverState *bs)
+{
+    assert(bs->drv);
+
+    if (bs->backing_hd) {
+        return false;
+    }
+    if (bs->drv->bdrv_has_discard_zeroes) {
+        return bs->drv->bdrv_has_discard_zeroes(bs);
+    }
+
+    return false;
+}
+
+bool bdrv_has_discard_write_zeroes(BlockDriverState *bs)
+{
+    assert(bs->drv);
+
+    if (bs->backing_hd || !(bs->open_flags & BDRV_O_UNMAP)) {
+        return false;
+    }
+
+    if (bs->drv->bdrv_has_discard_write_zeroes) {
+        return bs->drv->bdrv_has_discard_write_zeroes(bs);
+    }
+
+    return false;
+}
+
 typedef struct BdrvCoGetBlockStatusData {
     BlockDriverState *bs;
     BlockDriverState *base;
include/block/block.h +2 -0
index b72c81a..2de226f 100644
@@ -310,6 +310,8 @@ int bdrv_discard(BlockDriverState *bs, int64_t sector_num, int nb_sectors);
 int bdrv_co_discard(BlockDriverState *bs, int64_t sector_num, int nb_sectors);
 int bdrv_has_zero_init_1(BlockDriverState *bs);
 int bdrv_has_zero_init(BlockDriverState *bs);
+bool bdrv_has_discard_zeroes(BlockDriverState *bs);
+bool bdrv_has_discard_write_zeroes(BlockDriverState *bs);
 int64_t bdrv_get_block_status(BlockDriverState *bs, int64_t sector_num,
                               int nb_sectors, int *pnum);
 int bdrv_is_allocated(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
include/block/block_int.h +13 -0
index 56e1994..cf78c3f 100644
@@ -211,6 +211,19 @@ struct BlockDriver {
      */
     int (*bdrv_has_zero_init)(BlockDriverState *bs);
 
+    /*
+     * Returns true if discarded blocks read back as zeroes.
+     */
+    bool (*bdrv_has_discard_zeroes)(BlockDriverState *bs);
+
+    /*
+     * Returns true if the driver can optimize writing zeroes by discarding
+     * sectors. It is additionally required that the block device is
+     * opened with BDRV_O_UNMAP and the that write zeroes request carries
+     * the BDRV_REQ_MAY_UNMAP flag for this to work.
+     */
+    bool (*bdrv_has_discard_write_zeroes)(BlockDriverState *bs);
+
     QLIST_ENTRY(BlockDriver) list;
 };
 
Signature · 1 line
1.7.9.5
Reviewed-by on v3 4/20
#6 lore raw

Re: [Qemu-devel] [PATCHv3 04/20] block: introduce bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

Quoted 7 lines · Peter Lieven
---
 block.c                   |   29 +++++++++++++++++++++++++++++
 include/block/block.h     |    2 ++
 include/block/block_int.h |   13 +++++++++++++
 3 files changed, 44 insertions(+)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#7 lore raw

PATCH v3 5/20 block/raw: add bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

In reply to Peter Lieven
---
 block/raw_bsd.c |   56 +++++++++++++++++++++++++++++++++----------------------
 1 file changed, 34 insertions(+), 22 deletions(-)
block/raw_bsd.c +34 -22
index d5ab295..8dc7bba 100644
@@ -131,6 +131,16 @@ static int raw_has_zero_init(BlockDriverState *bs)
     return bdrv_has_zero_init(bs->file);
 }
 
+static bool raw_has_discard_zeroes(BlockDriverState *bs)
+{
+    return bdrv_has_discard_zeroes(bs->file);
+}
+
+static bool raw_has_discard_write_zeroes(BlockDriverState *bs)
+{
+    return bdrv_has_discard_write_zeroes(bs->file);
+}
+
 static int raw_create(const char *filename, QEMUOptionParameter *options,
                       Error **errp)
 {
@@ -165,28 +175,30 @@ static int raw_probe(const uint8_t *buf, int buf_size, const char *filename)
 }
 
 static BlockDriver bdrv_raw = {
-    .format_name          = "raw",
-    .bdrv_probe           = &raw_probe,
-    .bdrv_reopen_prepare  = &raw_reopen_prepare,
-    .bdrv_open            = &raw_open,
-    .bdrv_close           = &raw_close,
-    .bdrv_create          = &raw_create,
-    .bdrv_co_readv        = &raw_co_readv,
-    .bdrv_co_writev       = &raw_co_writev,
-    .bdrv_co_write_zeroes = &raw_co_write_zeroes,
-    .bdrv_co_discard      = &raw_co_discard,
-    .bdrv_co_get_block_status = &raw_co_get_block_status,
-    .bdrv_truncate        = &raw_truncate,
-    .bdrv_getlength       = &raw_getlength,
-    .bdrv_get_info        = &raw_get_info,
-    .bdrv_is_inserted     = &raw_is_inserted,
-    .bdrv_media_changed   = &raw_media_changed,
-    .bdrv_eject           = &raw_eject,
-    .bdrv_lock_medium     = &raw_lock_medium,
-    .bdrv_ioctl           = &raw_ioctl,
-    .bdrv_aio_ioctl       = &raw_aio_ioctl,
-    .create_options       = &raw_create_options[0],
-    .bdrv_has_zero_init   = &raw_has_zero_init
+    .format_name                   = "raw",
+    .bdrv_probe                    = &raw_probe,
+    .bdrv_reopen_prepare           = &raw_reopen_prepare,
+    .bdrv_open                     = &raw_open,
+    .bdrv_close                    = &raw_close,
+    .bdrv_create                   = &raw_create,
+    .bdrv_co_readv                 = &raw_co_readv,
+    .bdrv_co_writev                = &raw_co_writev,
+    .bdrv_co_write_zeroes          = &raw_co_write_zeroes,
+    .bdrv_co_discard               = &raw_co_discard,
+    .bdrv_co_get_block_status      = &raw_co_get_block_status,
+    .bdrv_truncate                 = &raw_truncate,
+    .bdrv_getlength                = &raw_getlength,
+    .bdrv_get_info                 = &raw_get_info,
+    .bdrv_is_inserted              = &raw_is_inserted,
+    .bdrv_media_changed            = &raw_media_changed,
+    .bdrv_eject                    = &raw_eject,
+    .bdrv_lock_medium              = &raw_lock_medium,
+    .bdrv_ioctl                    = &raw_ioctl,
+    .bdrv_aio_ioctl                = &raw_aio_ioctl,
+    .create_options                = &raw_create_options[0],
+    .bdrv_has_zero_init            = &raw_has_zero_init,
+    .bdrv_has_discard_zeroes       = &raw_has_discard_zeroes,
+    .bdrv_has_discard_write_zeroes = &raw_has_discard_write_zeroes,
 };
 
 static void bdrv_raw_init(void)
Signature · 1 line
1.7.9.5
Reviewed-by on v3 5/20
#8 lore raw

Re: [Qemu-devel] [PATCHv3 05/20] block/raw: add bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

Quoted 5 lines · Peter Lieven
---
 block/raw_bsd.c |   56 +++++++++++++++++++++++++++++++++----------------------
 1 file changed, 34 insertions(+), 22 deletions(-)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#9 lore raw

PATCH v3 6/20 block: add BlockLimits structure to BlockDriverState

In reply to Peter Lieven

this patch adds BlockLimits which introduces discard and write_zeroes limits and alignment information to the BlockDriverState.

---
 include/block/block_int.h |   17 +++++++++++++++++
 1 file changed, 17 insertions(+)
include/block/block_int.h +17 -0
index cf78c3f..cda99f9 100644
@@ -227,6 +227,20 @@ struct BlockDriver {
     QLIST_ENTRY(BlockDriver) list;
 };
 
+struct BlockLimits {
+    /* maximum number of sectors that can be discarded at once */
+    int max_discard;
+
+    /* optimal alignment for discard requests in sectors */
+    int64_t discard_alignment;
+
+    /* maximum number of sectors that can zeroized at once */
+    int max_write_zeroes;
+
+    /* optimal alignment for write zeroes requests in sectors */
+    int64_t write_zeroes_alignment;
+};
+
 /*
  * Note: the function bdrv_append() copies and swaps contents of
  * BlockDriverStates, so if you add new fields to this struct, please
@@ -280,6 +294,9 @@ struct BlockDriverState {
     uint64_t total_time_ns[BDRV_MAX_IOTYPE];
     uint64_t wr_highest_sector;
 
+    /* I/O Limits */
+    struct BlockLimits bl;
+
     /* Whether the disk can expand beyond total_sectors */
     int growable;
 
Signature · 1 line
1.7.9.5
Reviewed-by on v3 6/20
#10 lore raw

Re: [Qemu-devel] [PATCHv3 06/20] block: add BlockLimits structure to BlockDriverState

Quoted 10 lines · Peter Lieven

this patch adds BlockLimits which introduces discard and write_zeroes limits and alignment information to the BlockDriverState.

---
 include/block/block_int.h |   17 +++++++++++++++++
 1 file changed, 17 insertions(+)

+struct BlockLimits {

Should this be a typedef? Either in include/qemu/typedefs.h (where BlockDriverState is listed), or locally (see BdrvTrackedRequest in the same file for an example)?

include/block/block_int.h struct BlockDriverState { · line 299

Quoted 7 lines
include/block/block_int.h +3 -0
@@ -280,6 +294,9 @@ struct BlockDriverState {
     uint64_t total_time_ns[BDRV_MAX_IOTYPE];
     uint64_t wr_highest_sector;
 
+    /* I/O Limits */
+    struct BlockLimits bl;
+

All other struct/pointer-to-struct members in BlockDriverState are listed by typedef name, rather than calling out 'struct foo'.

My question is one of style/consistency, not of C correctness; so unless a maintainer actually agrees that a typedef change is needed so that you comply with project coding standards, feel free to add:

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#11 lore raw

Re: [Qemu-devel] [PATCHv3 06/20] block: add BlockLimits structure to BlockDriverState

Quoted 27 lines · Eric Blake
Quoted 10 lines · Peter Lieven

this patch adds BlockLimits which introduces discard and write_zeroes limits and alignment information to the BlockDriverState.

---
 include/block/block_int.h |   17 +++++++++++++++++
 1 file changed, 17 insertions(+)

+struct BlockLimits {

Should this be a typedef? Either in include/qemu/typedefs.h (where BlockDriverState is listed), or locally (see BdrvTrackedRequest in the same file for an example)?

include/block/block_int.h struct BlockDriverState { · line 299

Quoted 7 lines
include/block/block_int.h +3 -0
@@ -280,6 +294,9 @@ struct BlockDriverState {
     uint64_t total_time_ns[BDRV_MAX_IOTYPE];
     uint64_t wr_highest_sector;
 
+    /* I/O Limits */
+    struct BlockLimits bl;
+

All other struct/pointer-to-struct members in BlockDriverState are listed by typedef name, rather than calling out 'struct foo'.

Quoted 3 lines

My question is one of style/consistency, not of C correctness; so unless a maintainer actually agrees that a typedef change is needed so that you comply with project coding standards, feel free to add:

QEMU coding style requires use of typedef:

typedef struct {
    ...
} BlockLimits;

Stefan

#12 lore raw

PATCH v3 7/20 block: honour BlockLimits in bdrv_co_do_write_zeroes

In reply to Peter Lieven
---
 block.c |   65 +++++++++++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 49 insertions(+), 16 deletions(-)
block.c +49 -16
index ac35cb5..580570a 100644
@@ -2710,32 +2710,65 @@ int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
                             BDRV_REQ_COPY_ON_READ);
 }
 
+/* if no limit is specified in the BlockLimits use a default
+ * of 32768 512-byte sectors (16 MiB) per request.
+ */
+#define MAX_WRITE_ZEROES_DEFAULT 32768
+
 static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
     int64_t sector_num, int nb_sectors, BdrvRequestFlags flags)
 {
     BlockDriver *drv = bs->drv;
     QEMUIOVector qiov;
-    struct iovec iov;
-    int ret;
+    struct iovec iov = {0};
+    int ret = 0;
 
-    /* TODO Emulate only part of misaligned requests instead of letting block
-     * drivers return -ENOTSUP and emulate everything */
+    int max_write_zeroes = bs->bl.max_write_zeroes ?
+                           bs->bl.max_write_zeroes : MAX_WRITE_ZEROES_DEFAULT;
 
-    /* First try the efficient write zeroes operation */
-    if (drv->bdrv_co_write_zeroes) {
-        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors, flags);
-        if (ret != -ENOTSUP) {
-            return ret;
+    while (nb_sectors > 0 && !ret) {
+        int num = nb_sectors;
+
+        /* align request */
+        if (bs->bl.write_zeroes_alignment &&
+            num >= bs->bl.write_zeroes_alignment &&
+            sector_num % bs->bl.write_zeroes_alignment) {
+            if (num > bs->bl.write_zeroes_alignment) {
+                num = bs->bl.write_zeroes_alignment;
+            }
+            num -= sector_num % bs->bl.write_zeroes_alignment;
         }
-    }
 
-    /* Fall back to bounce buffer if write zeroes is unsupported */
-    iov.iov_len  = nb_sectors * BDRV_SECTOR_SIZE;
-    iov.iov_base = qemu_blockalign(bs, iov.iov_len);
-    memset(iov.iov_base, 0, iov.iov_len);
-    qemu_iovec_init_external(&qiov, &iov, 1);
+        /* limit request size */
+        if (num > max_write_zeroes) {
+            num = max_write_zeroes;
+        }
+
+        ret = -ENOTSUP;
+        /* First try the efficient write zeroes operation */
+        if (drv->bdrv_co_write_zeroes) {
+            ret = drv->bdrv_co_write_zeroes(bs, sector_num, num, flags);
+        }
+
+        if (ret == -ENOTSUP) {
+            /* Fall back to bounce buffer if write zeroes is unsupported */
+            iov.iov_len  = num * BDRV_SECTOR_SIZE;
+            if (iov.iov_base == NULL) {
+                /* allocate bounce buffer only once and ensure that it
+                 * is big enough for this and all future requests.
+                 */
+                size_t bufsize = num <= nb_sectors ? num : max_write_zeroes;
+                iov.iov_base = qemu_blockalign(bs, bufsize * BDRV_SECTOR_SIZE);
+                memset(iov.iov_base, 0, bufsize * BDRV_SECTOR_SIZE);
+            }
+            qemu_iovec_init_external(&qiov, &iov, 1);
 
-    ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, &qiov);
+            ret = drv->bdrv_co_writev(bs, sector_num, num, &qiov);
+        }
+
+        sector_num += num;
+        nb_sectors -= num;
+    }
 
     qemu_vfree(iov.iov_base);
     return ret;
Signature · 1 line
1.7.9.5
Reviewed-by on v3 7/20
#13 lore raw

Re: [Qemu-devel] [PATCHv3 07/20] block: honour BlockLimits in bdrv_co_do_write_zeroes

Quoted 10 lines · Peter Lieven
---
 block.c |   65 +++++++++++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 49 insertions(+), 16 deletions(-)
block.c +0 -0
index ac35cb5..580570a 100644

block.c int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs, · line 2718

Quoted 9 lines
block.c +4 -0
@@ -2710,32 +2710,65 @@ int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
                             BDRV_REQ_COPY_ON_READ);
 }
 
 
 
+
+        if (ret == -ENOTSUP) {
+            /* Fall back to bounce buffer if write zeroes is unsupported */
+            iov.iov_len  = num * BDRV_SECTOR_SIZE;

Why 2 spaces?

But that's trivial.

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#14 lore raw

PATCH v3 8/20 block: honour BlockLimits in bdrv_co_discard

In reply to Peter Lieven
---
 block.c |   37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)
block.c +36 -1
index 580570a..784e545 100644
@@ -4224,6 +4224,11 @@ static void coroutine_fn bdrv_discard_co_entry(void *opaque)
     rwco->ret = bdrv_co_discard(rwco->bs, rwco->sector_num, rwco->nb_sectors);
 }
 
+/* if no limit is specified in the BlockLimits use a default
+ * of 32768 512-byte sectors (16 MiB) per request.
+ */
+#define MAX_DISCARD_DEFAULT 32768
+
 int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
                                  int nb_sectors)
 {
@@ -4245,7 +4250,37 @@ int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
     }
 
     if (bs->drv->bdrv_co_discard) {
-        return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors);
+        int max_discard = bs->bl.max_discard ?
+                          bs->bl.max_discard : MAX_DISCARD_DEFAULT;
+
+        while (nb_sectors > 0) {
+            int ret;
+            int num = nb_sectors;
+
+            /* align request */
+            if (bs->bl.discard_alignment &&
+                num >= bs->bl.discard_alignment &&
+                sector_num % bs->bl.discard_alignment) {
+                if (num > bs->bl.discard_alignment) {
+                    num = bs->bl.discard_alignment;
+                }
+                num -= sector_num % bs->bl.discard_alignment;
+            }
+
+            /* limit request size */
+            if (num > max_discard) {
+                num = max_discard;
+            }
+
+            ret = bs->drv->bdrv_co_discard(bs, sector_num, num);
+            if (ret) {
+                return ret;
+            }
+
+            sector_num += num;
+            nb_sectors -= num;
+        }
+        return 0;
     } else if (bs->drv->bdrv_aio_discard) {
         BlockDriverAIOCB *acb;
         CoroutineIOCompletion co = {
Signature · 1 line
1.7.9.5
Reviewed-by on v3 8/20
#15 lore raw

Re: [Qemu-devel] [PATCHv3 08/20] block: honour BlockLimits in bdrv_co_discard

Quoted 5 lines · Peter Lieven
---
 block.c |   37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#16 lore raw

Re: [Qemu-devel] [PATCHv3 08/20] block: honour BlockLimits in bdrv_co_discard

In reply to Peter Lieven
Quoted 11 lines · Peter Lieven
block.c +6 -1
@@ -4245,7 +4250,37 @@ int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
     }
 
     if (bs->drv->bdrv_co_discard) {
-        return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors);
+        int max_discard = bs->bl.max_discard ?
+                          bs->bl.max_discard : MAX_DISCARD_DEFAULT;
+
+        while (nb_sectors > 0) {
+            int ret;
+            int num = nb_sectors;

block.c int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num, · line 4268

Quoted 10 lines
block.c +10 -0
+
+            /* align request */
+            if (bs->bl.discard_alignment &&
+                num >= bs->bl.discard_alignment &&
+                sector_num % bs->bl.discard_alignment) {
+                if (num > bs->bl.discard_alignment) {
+                    num = bs->bl.discard_alignment;
+                }
+                num -= sector_num % bs->bl.discard_alignment;
+            }

Is it always possible to discard at arbitrary sector offsets?

#17 lore raw

Re: [Qemu-devel] [PATCHv3 08/20] block: honour BlockLimits in bdrv_co_discard

Am 07.10.2013 um 10:29 schrieb Stefan Hajnoczi <[email protected]>:

Quoted 16 lines · Peter Lieven
Quoted 15 lines
block.c +10 -1
@@ -4245,7 +4250,37 @@ int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
    }
 
    if (bs->drv->bdrv_co_discard) {
-        return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors);
+        int max_discard = bs->bl.max_discard ?
+                          bs->bl.max_discard : MAX_DISCARD_DEFAULT;
+
+        while (nb_sectors > 0) {
+            int ret;
+            int num = nb_sectors;
+
+            /* align request */
+            if (bs->bl.discard_alignment &&
+                num >= bs->bl.discard_alignment &&
Quoted 8 lines

block.c int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num, · line 4268

Quoted 6 lines
block.c +6 -0
+                sector_num % bs->bl.discard_alignment) {
+                if (num > bs->bl.discard_alignment) {
+                    num = bs->bl.discard_alignment;
+                }
+                num -= sector_num % bs->bl.discard_alignment;
+            }

Is it always possible to discard at arbitrary sector offsets?

It should be. This just makes sure that the unaligned part of a request gets executed first (eventually resulting in an -ENOTSUP) and consecutive calls are aligned.

For example. The alignment is 1MB (2048 sectors) and max_discard its 2048 sectors as well.. Lets say you want to discard 4096 sectors starting at offset 1024.

Then there will be 3 calls
1) discard at 1024, len 1024
2) discard at 2048, len 2048
3) discard at 4096, len 1024.

Peter

#18 lore raw

PATCH v3 9/20 iscsi: simplify iscsi_co_discard

In reply to Peter Lieven

now that bdrv_co_discard can handle limits we do not need the request split logic here anymore.

---
 block/iscsi.c |   67 +++++++++++++++++++++------------------------------------
 1 file changed, 25 insertions(+), 42 deletions(-)
block/iscsi.c +25 -42
index 4460382..ce8823d 100644
@@ -87,7 +87,6 @@ typedef struct IscsiAIOCB {
 #define NOP_INTERVAL 5000
 #define MAX_NOP_FAILURES 3
 #define ISCSI_CMD_RETRIES 5
-#define ISCSI_MAX_UNMAP 131072
 
 static void
 iscsi_bh_cb(void *p)
@@ -912,8 +911,6 @@ coroutine_fn iscsi_co_discard(BlockDriverState *bs, int64_t sector_num,
     IscsiLun *iscsilun = bs->opaque;
     struct IscsiTask iTask;
     struct unmap_list list;
-    uint32_t nb_blocks;
-    uint32_t max_unmap;
 
     if (!is_request_lun_aligned(sector_num, nb_sectors, iscsilun)) {
         return -EINVAL;
@@ -925,52 +922,38 @@ coroutine_fn iscsi_co_discard(BlockDriverState *bs, int64_t sector_num,
     }
 
     list.lba = sector_qemu2lun(sector_num, iscsilun);
-    nb_blocks = sector_qemu2lun(nb_sectors, iscsilun);
+    list.num = sector_qemu2lun(nb_sectors, iscsilun);
 
-    max_unmap = iscsilun->bl.max_unmap;
-    if (max_unmap == 0xffffffff) {
-        max_unmap = ISCSI_MAX_UNMAP;
-    }
-
-    while (nb_blocks > 0) {
-        iscsi_co_init_iscsitask(iscsilun, &iTask);
-        list.num = nb_blocks;
-        if (list.num > max_unmap) {
-            list.num = max_unmap;
-        }
+    iscsi_co_init_iscsitask(iscsilun, &iTask);
 retry:
-        if (iscsi_unmap_task(iscsilun->iscsi, iscsilun->lun, 0, 0, &list, 1,
-                         iscsi_co_generic_cb, &iTask) == NULL) {
-            return -EIO;
-        }
-
-        while (!iTask.complete) {
-            iscsi_set_events(iscsilun);
-            qemu_coroutine_yield();
-        }
+    if (iscsi_unmap_task(iscsilun->iscsi, iscsilun->lun, 0, 0, &list, 1,
+                     iscsi_co_generic_cb, &iTask) == NULL) {
+        return -EIO;
+    }
 
-        if (iTask.task != NULL) {
-            scsi_free_scsi_task(iTask.task);
-            iTask.task = NULL;
-        }
+    while (!iTask.complete) {
+        iscsi_set_events(iscsilun);
+        qemu_coroutine_yield();
+    }
 
-        if (iTask.do_retry) {
-            goto retry;
-        }
+    if (iTask.task != NULL) {
+        scsi_free_scsi_task(iTask.task);
+        iTask.task = NULL;
+    }
 
-        if (iTask.status == SCSI_STATUS_CHECK_CONDITION) {
-            /* the target might fail with a check condition if it
-               is not happy with the alignment of the UNMAP request
-               we silently fail in this case */
-            return 0;
-        }
+    if (iTask.do_retry) {
+        goto retry;
+    }
 
-        if (iTask.status != SCSI_STATUS_GOOD) {
-            return -EIO;
-        }
+    if (iTask.status == SCSI_STATUS_CHECK_CONDITION) {
+        /* the target might fail with a check condition if it
+           is not happy with the alignment of the UNMAP request
+           we silently fail in this case */
+        return 0;
+    }
 
-        list.lba += list.num;
-        nb_blocks -= list.num;
+    if (iTask.status != SCSI_STATUS_GOOD) {
+        return -EIO;
     }
 
     return 0;
Signature · 1 line
1.7.9.5
#19 lore raw

PATCH v3 10/20 iscsi: set limits in BlockDriverState

In reply to Peter Lieven
---
 block/iscsi.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)
block/iscsi.c +14 -0
index ce8823d..90dc7c2 100644
@@ -1367,6 +1367,20 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,
                sizeof(struct scsi_inquiry_block_limits));
         scsi_free_scsi_task(task);
         task = NULL;
+
+        if (iscsilun->bl.max_unmap < 0xffffffff) {
+            bs->bl.max_discard = sector_lun2qemu(iscsilun->bl.max_unmap,
+                                                 iscsilun);
+        }
+        bs->bl.discard_alignment = sector_lun2qemu(iscsilun->bl.opt_unmap_gran,
+                                                   iscsilun);
+
+        if (iscsilun->bl.max_ws_len < 0xffffffff) {
+            bs->bl.max_write_zeroes = sector_lun2qemu(iscsilun->bl.max_ws_len,
+                                                      iscsilun);
+        }
+        bs->bl.write_zeroes_alignment = sector_lun2qemu(iscsilun->bl.opt_unmap_gran,
+                                                        iscsilun);
     }
 
 #if defined(LIBISCSI_FEATURE_NOP_COUNTER)
Signature · 1 line
1.7.9.5
Reviewed-by on v3 10/20
#20 lore raw

Re: [Qemu-devel] [PATCHv3 10/20] iscsi: set limits in BlockDriverState

Quoted 4 lines · Peter Lieven
---
 block/iscsi.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#21 lore raw

PATCH v3 11/20 iscsi: add bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

In reply to Peter Lieven
---
 block/iscsi.c |   16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
block/iscsi.c +15 -1
index 90dc7c2..f6cd322 100644
@@ -1449,6 +1449,18 @@ static int iscsi_has_zero_init(BlockDriverState *bs)
     return 0;
 }
 
+static bool iscsi_has_discard_zeroes(BlockDriverState *bs)
+{
+    IscsiLun *iscsilun = bs->opaque;
+    return !!iscsilun->lbprz;
+}
+
+static bool iscsi_has_discard_write_zeroes(BlockDriverState *bs)
+{
+    IscsiLun *iscsilun = bs->opaque;
+    return iscsilun->lbprz && iscsilun->lbp.lbpws;
+}
+
 static int iscsi_create(const char *filename, QEMUOptionParameter *options,
                         Error **errp)
 {
@@ -1534,7 +1546,9 @@ static BlockDriver bdrv_iscsi = {
     .bdrv_aio_writev = iscsi_aio_writev,
     .bdrv_aio_flush  = iscsi_aio_flush,
 
-    .bdrv_has_zero_init = iscsi_has_zero_init,
+    .bdrv_has_zero_init            = iscsi_has_zero_init,
+    .bdrv_has_discard_zeroes       = iscsi_has_discard_zeroes,
+    .bdrv_has_discard_write_zeroes = iscsi_has_discard_write_zeroes,
 
 #ifdef __linux__
     .bdrv_ioctl       = iscsi_ioctl,
Signature · 1 line
1.7.9.5
Reviewed-by on v3 11/20
#22 lore raw

Re: [Qemu-devel] [PATCHv3 11/20] iscsi: add bdrv_has_discard_zeroes and bdrv_has_discard_write_zeroes

Quoted 5 lines · Peter Lieven
---
 block/iscsi.c |   16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#23 lore raw

PATCH v3 12/20 iscsi: add bdrv_co_write_zeroes

In reply to Peter Lieven
---
 block/iscsi.c |   59 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
block/iscsi.c +59 -0
index f6cd322..8f52ef8 100644
@@ -56,6 +56,7 @@ typedef struct IscsiLun {
     uint8_t lbprz;
     struct scsi_inquiry_logical_block_provisioning lbp;
     struct scsi_inquiry_block_limits bl;
+    unsigned char *zeroblock;
 } IscsiLun;
 
 typedef struct IscsiTask {
@@ -959,6 +960,62 @@ retry:
     return 0;
 }
 
+
+static int
+coroutine_fn iscsi_co_write_zeroes(BlockDriverState *bs, int64_t sector_num,
+                                   int nb_sectors, BdrvRequestFlags flags)
+{
+    IscsiLun *iscsilun = bs->opaque;
+    struct IscsiTask iTask;
+    uint64_t lba;
+    uint32_t nb_blocks;
+
+    if (!is_request_lun_aligned(sector_num, nb_sectors, iscsilun)) {
+        return -EINVAL;
+    }
+
+    if (!iscsilun->lbp.lbpws) {
+        /* WRITE SAME is not supported by the target */
+        return -ENOTSUP;
+    }
+
+    lba = sector_qemu2lun(sector_num, iscsilun);
+    nb_blocks = sector_qemu2lun(nb_sectors, iscsilun);
+
+    if (iscsilun->zeroblock == NULL) {
+        iscsilun->zeroblock = g_malloc0(iscsilun->block_size);
+    }
+
+    iscsi_co_init_iscsitask(iscsilun, &iTask);
+retry:
+    if (iscsi_writesame16_task(iscsilun->iscsi, iscsilun->lun, lba,
+                               iscsilun->zeroblock, iscsilun->block_size,
+                               nb_blocks, 0, !!(flags & BDRV_REQ_MAY_UNMAP),
+                               0, 0, iscsi_co_generic_cb, &iTask) == NULL) {
+        return -EIO;
+    }
+
+    while (!iTask.complete) {
+        iscsi_set_events(iscsilun);
+        qemu_coroutine_yield();
+    }
+
+    if (iTask.task != NULL) {
+        scsi_free_scsi_task(iTask.task);
+        iTask.task = NULL;
+    }
+
+    if (iTask.do_retry) {
+        goto retry;
+    }
+
+    if (iTask.status != SCSI_STATUS_GOOD) {
+        return -EIO;
+    }
+
+    return 0;
+}
+
 static int parse_chap(struct iscsi_context *iscsi, const char *target)
 {
     QemuOptsList *list;
@@ -1421,6 +1478,7 @@ static void iscsi_close(BlockDriverState *bs)
     }
     qemu_aio_set_fd_handler(iscsi_get_fd(iscsi), NULL, NULL, NULL);
     iscsi_destroy_context(iscsi);
+    g_free(iscsilun->zeroblock);
     memset(iscsilun, 0, sizeof(IscsiLun));
 }
 
@@ -1541,6 +1599,7 @@ static BlockDriver bdrv_iscsi = {
     .bdrv_co_get_block_status = iscsi_co_get_block_status,
 #endif
     .bdrv_co_discard      = iscsi_co_discard,
+    .bdrv_co_write_zeroes = iscsi_co_write_zeroes,
 
     .bdrv_aio_readv  = iscsi_aio_readv,
     .bdrv_aio_writev = iscsi_aio_writev,
Signature · 1 line
1.7.9.5
#24 lore raw

PATCH v3 13/20 block: introduce bdrv_zeroize

In reply to Peter Lieven

this patch adds a call to completely zero out a block device. the operation is sped up by checking the block status and only writing zeroes to the device if they currently do not return zeroes. optionally the zero writing can be sped up by setting the flag BDRV_REQ_MAY_UNMAP to emulate the zero write by unmapping if the driver supports it.

---
 block.c               |   37 +++++++++++++++++++++++++++++++++++++
 include/block/block.h |    1 +
 2 files changed, 38 insertions(+)
block.c +37 -0
index 784e545..66b9eae 100644
@@ -2398,6 +2398,43 @@ int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
                       BDRV_REQ_ZERO_WRITE | flags);
 }
 
+/*
+ * Completely zero out a block device with the help of bdrv_write_zeroes.
+ * The operation is sped up by checking the block status and only writing
+ * zeroes to the device if they currently do not return zeroes. Optional
+ * flags are passed through to bdrv_write_zeroes (e.g. BDRV_REQ_MAY_UNMAP).
+ *
+ * Returns < 0 on error, 0 on success. For error codes see bdrv_write().
+ */
+int bdrv_zeroize(BlockDriverState *bs, BdrvRequestFlags flags)
+{
+    int64_t target_size = bdrv_getlength(bs) / BDRV_SECTOR_SIZE;
+    int64_t ret, nb_sectors, sector_num = 0;
+    int n;
+
+    for (;;) {
+        nb_sectors = target_size - sector_num;
+        if (nb_sectors <= 0) {
+            return 0;
+        }
+        if (nb_sectors > INT_MAX) {
+            nb_sectors = INT_MAX;
+        }
+        ret = bdrv_get_block_status(bs, sector_num, nb_sectors, &n);
+        if (ret & BDRV_BLOCK_ZERO) {
+            sector_num += n;
+            continue;
+        }
+        ret = bdrv_write_zeroes(bs, sector_num, n, flags);
+        if (ret < 0) {
+            error_report("error writing zeroes at sector %" PRId64 ": %s",
+                         sector_num, strerror(-ret));
+            return ret;
+        }
+        sector_num += n;
+    }
+}
+
 int bdrv_pread(BlockDriverState *bs, int64_t offset,
                void *buf, int count1)
 {
include/block/block.h +1 -0
index 2de226f..e059188 100644
@@ -196,6 +196,7 @@ int bdrv_write(BlockDriverState *bs, int64_t sector_num,
                const uint8_t *buf, int nb_sectors);
 int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
                int nb_sectors, BdrvRequestFlags flags);
+int bdrv_zeroize(BlockDriverState *bs, BdrvRequestFlags flags);
 int bdrv_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov);
 int bdrv_pread(BlockDriverState *bs, int64_t offset,
                void *buf, int count);
Signature · 1 line
1.7.9.5
Reviewed-by on v3 13/20
#25 lore raw

Re: [Qemu-devel] [PATCHv3 13/20] block: introduce bdrv_zeroize

Quoted 7 lines · Peter Lieven

this patch adds a call to completely zero out a block device. the operation is sped up by checking the block status and only writing zeroes to the device if they currently do not return zeroes. optionally the zero writing can be sped up by setting the flag BDRV_REQ_MAY_UNMAP to emulate the zero write by unmapping if the driver supports it.

Quoted 6 lines
---
 block.c               |   37 +++++++++++++++++++++++++++++++++++++
 include/block/block.h |    1 +
 2 files changed, 38 insertions(+)
Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#26 lore raw

Re: [Qemu-devel] [PATCHv3 13/20] block: introduce bdrv_zeroize

In reply to Peter Lieven
Quoted 20 lines · Peter Lieven

this patch adds a call to completely zero out a block device. the operation is sped up by checking the block status and only writing zeroes to the device if they currently do not return zeroes. optionally the zero writing can be sped up by setting the flag BDRV_REQ_MAY_UNMAP to emulate the zero write by unmapping if the driver supports it.

---
 block.c               |   37 +++++++++++++++++++++++++++++++++++++
 include/block/block.h |    1 +
 2 files changed, 38 insertions(+)
block.c +0 -0
index 784e545..66b9eae 100644
@@ -2398,6 +2398,43 @@ int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
                       BDRV_REQ_ZERO_WRITE | flags);
 }

block.c int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num, · line 2409

Quoted 10 lines
block.c +9 -0
 
+/*
+ * Completely zero out a block device with the help of bdrv_write_zeroes.
+ * The operation is sped up by checking the block status and only writing
+ * zeroes to the device if they currently do not return zeroes. Optional
+ * flags are passed through to bdrv_write_zeroes (e.g. BDRV_REQ_MAY_UNMAP).
+ *
+ * Returns < 0 on error, 0 on success. For error codes see bdrv_write().
+ */
+int bdrv_zeroize(BlockDriverState *bs, BdrvRequestFlags flags)

BlockDriver->bdrv_make_empty() implements zeroing the entire disk for image formats. Please extend that function prototype instead of adding a new interface.

Stefan

#27 lore raw

Re: [Qemu-devel] [PATCHv3 13/20] block: introduce bdrv_zeroize

Am 07.10.2013 um 10:34 schrieb Stefan Hajnoczi <[email protected]>:

Quoted 22 lines · Peter Lieven
Quoted 21 lines

this patch adds a call to completely zero out a block device. the operation is sped up by checking the block status and only writing zeroes to the device if they currently do not return zeroes. optionally the zero writing can be sped up by setting the flag BDRV_REQ_MAY_UNMAP to emulate the zero write by unmapping if the driver supports it.

---
block.c | 37 +++++++++++++++++++++++++++++++++++++
include/block/block.h | 1 +

2 files changed, 38 insertions(+)
block.c +0 -0
index 784e545..66b9eae 100644
@@ -2398,6 +2398,43 @@ int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num,
                      BDRV_REQ_ZERO_WRITE | flags);
 }
 
Quoted 13 lines

block.c int bdrv_write_zeroes(BlockDriverState *bs, int64_t sector_num, · line 2409

Quoted 9 lines
block.c +9 -0
+/*
+ * Completely zero out a block device with the help of bdrv_write_zeroes.
+ * The operation is sped up by checking the block status and only writing
+ * zeroes to the device if they currently do not return zeroes. Optional
+ * flags are passed through to bdrv_write_zeroes (e.g. BDRV_REQ_MAY_UNMAP).
+ *
+ * Returns < 0 on error, 0 on success. For error codes see bdrv_write().
+ */
+int bdrv_zeroize(BlockDriverState *bs, BdrvRequestFlags flags)

BlockDriver->bdrv_make_empty() implements zeroing the entire disk for image formats. Please extend that function prototype instead of adding a new interface.

You mean, adding bdrv_make_empty as a function to block.c calling drv->bdrv_make_empty if it is defined?

Peter

Quoted 2 lines

Stefan

#28 lore raw

PATCH v3 14/20 block/get_block_status: set *pnum = 0 on error

In reply to Peter Lieven

if the call is invoked through bdrv_is_allocated the caller might expect *pnum = 0 on error. however, a new implementation of bdrv_get_block_status might only return a negative exit value on error while keeping *pnum untouched.

---
 block.c |    1 +
 1 file changed, 1 insertion(+)
block.c +1 -0
index 66b9eae..3499c90 100644
@@ -3259,6 +3259,7 @@ static int64_t coroutine_fn bdrv_co_get_block_status(BlockDriverState *bs,
 
     ret = bs->drv->bdrv_co_get_block_status(bs, sector_num, nb_sectors, pnum);
     if (ret < 0) {
+        *pnum = 0;
         return ret;
     }
 
Signature · 1 line
1.7.9.5
#29 lore raw

PATCH v3 15/20 block/get_block_status: avoid segfault if there is no backing_hd

In reply to Peter Lieven
---
 block.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
block.c +1 -1
index 3499c90..3bf1163 100644
@@ -3266,7 +3266,7 @@ static int64_t coroutine_fn bdrv_co_get_block_status(BlockDriverState *bs,
     if (!(ret & BDRV_BLOCK_DATA)) {
         if (bdrv_has_zero_init(bs)) {
             ret |= BDRV_BLOCK_ZERO;
-        } else {
+        } else if (bs->backing_hd) {
             BlockDriverState *bs2 = bs->backing_hd;
             int64_t length2 = bdrv_getlength(bs2);
             if (length2 >= 0 && sector_num >= (length2 >> BDRV_SECTOR_BITS)) {
Signature · 1 line
1.7.9.5
#30 lore raw

PATCH v3 16/20 block/get_block_status: avoid redundant callouts on raw devices

In reply to Peter Lieven

if a raw device like an iscsi target or host device is used the current implementation makes a second call out to get the block status of bs->file. however, the raw driver already has called bdrv_get_block_status on bs->file.

---
 block.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
block.c +2 -1
index 3bf1163..f56660d 100644
@@ -3277,7 +3277,8 @@ static int64_t coroutine_fn bdrv_co_get_block_status(BlockDriverState *bs,
 
     if (bs->file &&
         (ret & BDRV_BLOCK_DATA) && !(ret & BDRV_BLOCK_ZERO) &&
-        (ret & BDRV_BLOCK_OFFSET_VALID)) {
+        (ret & BDRV_BLOCK_OFFSET_VALID) &&
+        strcmp(bs->drv->format_name, "raw")) {
         ret2 = bdrv_co_get_block_status(bs->file, ret >> BDRV_SECTOR_BITS,
                                         *pnum, pnum);
         if (ret2 >= 0) {
Signature · 1 line
1.7.9.5
#31 lore raw

PATCH v3 17/20 block/get_block_status: fix BDRV_BLOCK_ZERO for unallocated blocks

In reply to Peter Lieven

this patch does 2 things:
a) only do additional call outs if BDRV_BLOCK_ZERO is not already set.
b) use the newly introduced bdrv_has_discard_zeroes() to return the zero state of an unallocated block. the used callout to bdrv_has_zero_init() is only valid right after bdrv_create.

---
 block.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
block.c +2 -2
index f56660d..032809a 100644
@@ -3263,8 +3263,8 @@ static int64_t coroutine_fn bdrv_co_get_block_status(BlockDriverState *bs,
         return ret;
     }
 
-    if (!(ret & BDRV_BLOCK_DATA)) {
-        if (bdrv_has_zero_init(bs)) {
+    if (!(ret & BDRV_BLOCK_DATA) && !(ret & BDRV_BLOCK_ZERO)) {
+        if (bdrv_has_discard_zeroes(bs)) {
             ret |= BDRV_BLOCK_ZERO;
         } else if (bs->backing_hd) {
             BlockDriverState *bs2 = bs->backing_hd;
Signature · 1 line
1.7.9.5
#32 lore raw

PATCH v3 18/20 qemu-img: add support for fully allocated images

In reply to Peter Lieven
---
 qemu-img.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
qemu-img.c +5 -3
index 926f0a0..c6eff15 100644
@@ -100,8 +100,10 @@ static void help(void)
            "  '-h' with or without a command shows this help and lists the supported formats\n"
            "  '-p' show progress of command (only certain commands)\n"
            "  '-q' use Quiet mode - do not print any output (except errors)\n"
-           "  '-S' indicates the consecutive number of bytes that must contain only zeros\n"
-           "       for qemu-img to create a sparse image during conversion\n"
+           "  '-S' indicates the consecutive number of bytes (defaults to 4k) that must\n"
+           "       contain only zeros for qemu-img to create a sparse image during\n"
+           "       conversion. if the number of bytes is 0 sparse files are disabled and\n"
+           "       images will always be fully allocated\n"
            "  '--output' takes the format in which the output must be done (human or json)\n"
            "  '-n' skips the target volume creation (useful if the volume is created\n"
            "       prior to running qemu-img)\n"
@@ -1465,7 +1467,7 @@ static int img_convert(int argc, char **argv)
         /* signal EOF to align */
         bdrv_write_compressed(out_bs, 0, NULL, 0);
     } else {
-        int has_zero_init = bdrv_has_zero_init(out_bs);
+        int has_zero_init = min_sparse ? bdrv_has_zero_init(out_bs) : 0;
 
         sector_num = 0; // total number of sectors converted so far
         nb_sectors = total_sectors - sector_num;
Signature · 1 line
1.7.9.5
#33 lore raw

Re: [Qemu-devel] [PATCHv3 18/20] qemu-img: add support for fully allocated images

Quoted 4 lines · Peter Lieven
---
 qemu-img.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

Missing a counterpart change to qemu-img.texi.

Quoted 6 lines
qemu-img.c +0 -0
index 926f0a0..c6eff15 100644

qemu-img.c static void help(void) · line 106

Quoted 9 lines
qemu-img.c +4 -2
@@ -100,8 +100,10 @@ static void help(void)
            "  '-h' with or without a command shows this help and lists the supported formats\n"
            "  '-p' show progress of command (only certain commands)\n"
            "  '-q' use Quiet mode - do not print any output (except errors)\n"
-           "  '-S' indicates the consecutive number of bytes that must contain only zeros\n"
-           "       for qemu-img to create a sparse image during conversion\n"
+           "  '-S' indicates the consecutive number of bytes (defaults to 4k) that must\n"
+           "       contain only zeros for qemu-img to create a sparse image during\n"
+           "       conversion. if the number of bytes is 0 sparse files are disabled and\n"
+           "       images will always be fully allocated\n"

the texi page should have similar wording.

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#34 lore raw

PATCH v3 19/20 qemu-img: conditionally zero out target on convert

In reply to Peter Lieven

If the target has_zero_init = 0, but supports efficiently writing zeroes by unmapping we call bdrv_zeroize to avoid fully allocating the target. Ihis currently is designed especially for iscsi.

---
 qemu-img.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
qemu-img.c +9 -1
index c6eff15..2a41942 100644
@@ -1353,7 +1353,7 @@ static int img_convert(int argc, char **argv)
         }
     }
 
-    flags = BDRV_O_RDWR;
+    flags = min_sparse ? (BDRV_O_RDWR | BDRV_O_UNMAP) : BDRV_O_RDWR;
     ret = bdrv_parse_cache_flags(cache, &flags);
     if (ret < 0) {
         error_report("Invalid cache option: %s", cache);
@@ -1469,6 +1469,14 @@ static int img_convert(int argc, char **argv)
     } else {
         int has_zero_init = min_sparse ? bdrv_has_zero_init(out_bs) : 0;
 
+        if (!has_zero_init && bdrv_has_discard_write_zeroes(out_bs)) {
+            ret = bdrv_zeroize(out_bs, BDRV_REQ_MAY_UNMAP);
+            if (ret < 0) {
+                goto out;
+            }
+            has_zero_init = 1;
+        }
+
         sector_num = 0; // total number of sectors converted so far
         nb_sectors = total_sectors - sector_num;
         if (nb_sectors != 0) {
Signature · 1 line
1.7.9.5
#35 lore raw

PATCH v3 20/20 block/raw: copy BlockLimits on raw_open

In reply to Peter Lieven
---
 block/raw_bsd.c |    1 +
 1 file changed, 1 insertion(+)
block/raw_bsd.c +1 -0
index 8dc7bba..7af26ad 100644
@@ -159,6 +159,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
                     Error **errp)
 {
     bs->sg = bs->file->sg;
+    memcpy(&bs->bl, &bs->file->bl, sizeof(struct BlockLimits));
     return 0;
 }
 
Signature · 1 line
1.7.9.5
#36 lore raw

Re: [Qemu-devel] [PATCHv3 20/20] block/raw: copy BlockLimits on raw_open

Quoted 10 lines · Peter Lieven
---
 block/raw_bsd.c |    1 +
 1 file changed, 1 insertion(+)
block/raw_bsd.c +0 -0
index 8dc7bba..7af26ad 100644

block/raw_bsd.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 162

Quoted 4 lines
block/raw_bsd.c +1 -0
@@ -159,6 +159,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
                     Error **errp)
 {
     bs->sg = bs->file->sg;
+    memcpy(&bs->bl, &bs->file->bl, sizeof(struct BlockLimits));

Personally, I think that sizeof(var) is more robust, because if the declaration of var is ever changed, you don't have to remember to also touch up the memcpy. As in:

memcpy(&bs->bl, &bs->file->bl, sizeof(bs->bl));

But there's plenty of examples of sizeof(type) in the codebase even when a var is handy, so you are not alone, and that's not a reason for me to request a respin.

On the other hand, why use memcpy() at all?

bs->bl = bs->file->bl;

should do the same trick, with less typing.

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#37 lore raw

Re: [Qemu-devel] [PATCHv3 20/20] block/raw: copy BlockLimits on raw_open

Quoted 26 lines · Eric Blake
Quoted 10 lines · Peter Lieven
---
 block/raw_bsd.c |    1 +
 1 file changed, 1 insertion(+)
block/raw_bsd.c +0 -0
index 8dc7bba..7af26ad 100644

block/raw_bsd.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 162

Quoted 4 lines
block/raw_bsd.c +1 -0
@@ -159,6 +159,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
                     Error **errp)
 {
     bs->sg = bs->file->sg;
+    memcpy(&bs->bl, &bs->file->bl, sizeof(struct BlockLimits));

Personally, I think that sizeof(var) is more robust, because if the declaration of var is ever changed, you don't have to remember to also touch up the memcpy. As in:

memcpy(&bs->bl, &bs->file->bl, sizeof(bs->bl));

But there's plenty of examples of sizeof(type) in the codebase even when a var is handy, so you are not alone, and that's not a reason for me to request a respin.

Quoted 5 lines

On the other hand, why use memcpy() at all?

bs->bl = bs->file->bl;

should do the same trick, with less typing.

Yes, please use struct assignment.

Stefan

#38 lore raw

Re: [Qemu-devel] [PATCHv3 20/20] block/raw: copy BlockLimits on raw_open

Am 07.10.2013 um 10:38 schrieb Stefan Hajnoczi <[email protected]>:

Quoted 27 lines · Eric Blake
Quoted 26 lines · Peter Lieven
Quoted 10 lines

---
block/raw_bsd.c | 1 +

1 file changed, 1 insertion(+)
block/raw_bsd.c +0 -0
index 8dc7bba..7af26ad 100644

block/raw_bsd.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 162

Quoted 4 lines
block/raw_bsd.c +1 -0
@@ -159,6 +159,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags,
                    Error **errp)
 {
    bs->sg = bs->file->sg;
+    memcpy(&bs->bl, &bs->file->bl, sizeof(struct BlockLimits));

Personally, I think that sizeof(var) is more robust, because if the declaration of var is ever changed, you don't have to remember to also touch up the memcpy. As in:

memcpy(&bs->bl, &bs->file->bl, sizeof(bs->bl));

But there's plenty of examples of sizeof(type) in the codebase even when a var is handy, so you are not alone, and that's not a reason for me to request a respin.

Quoted 7 lines
Quoted 5 lines

On the other hand, why use memcpy() at all?

bs->bl = bs->file->bl;

should do the same trick, with less typing.

Yes, please use struct assignment.

Okay, I was unsure because when looking at bdrv_move_feature_fields I found that there memcpy was used.

Peter

#39 lore raw
In reply to Peter Lieven
Quoted 8 lines · Peter Lieven

this patch adds the ability for targets to stay sparse during block migration (if the zero_blocks capability is set) and qemu-img convert even if the target does not have has_zero_init = 1.

the series was especially developed for iSCSI, but it should also work with other drivers with little or no adjustments. these adjustments should be limited to providing block provisioning information through get_block_info and/or honouring BDRV_REQ_MAY_UNMAP on writing zeroes.

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

Stefan

#40 lore raw

Am 07.10.2013 um 10:42 schrieb Stefan Hajnoczi <[email protected]>:

Quoted 15 lines · Peter Lieven
Quoted 8 lines

this patch adds the ability for targets to stay sparse during block migration (if the zero_blocks capability is set) and qemu-img convert even if the target does not have has_zero_init = 1.

the series was especially developed for iSCSI, but it should also work with other drivers with little or no adjustments. these adjustments should be limited to providing block provisioning information through get_block_info and/or honouring BDRV_REQ_MAY_UNMAP on writing zeroes.

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Quoted 7 lines

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

There was a big discussing going on on this flags some time ago with Paolo and Kevin and they wanted to have it. (thread was "qemu-img: conditionally discard target on convert"). For my use case I don't need the flag.

Peter

#41 lore raw
In reply to Stefan Hajnoczi

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 11 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero, otherwise it's emulated with bdrv_aio_writev)

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Paolo

#42 lore raw
Quoted 20 lines · Paolo Bonzini

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 11 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero, otherwise it's emulated with bdrv_aio_writev)

Quoted 8 lines

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Okay, then let's keep the patches to expose the flag.

Stefan

#43 lore raw
Quoted 24 lines · Stefan Hajnoczi
Quoted 23 lines · Paolo Bonzini

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 11 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero, otherwise it's emulated with bdrv_aio_writev)

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Quoted 5 lines
Quoted 4 lines

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Okay, then let's keep the patches to expose the flag.

Okay, then I can keep those.

Can you give a short hint if my approach with brdv_make_empty is what you want? I would like to not change the parameters, so use BDRV_REQ_MAY_UNMAP unconditionally.

int bdrv_make_empty(BlockDriverState *bs)

{
    int64_t target_size = bdrv_getlength(bs) / BDRV_SECTOR_SIZE;
    int64_t ret, nb_sectors, sector_num = 0;
    int n;

    if (bs->drv->bdrv_make_empty) {
        return bs->drv->bdrv_make_empty(bs);
    }

    for (;;) {
        nb_sectors = target_size - sector_num;
        if (nb_sectors <= 0) {
            return 0;
        }
        if (nb_sectors > INT_MAX) {
            nb_sectors = INT_MAX;
        }
        ret = bdrv_get_block_status(bs, sector_num, nb_sectors, &n);
        if (ret & BDRV_BLOCK_ZERO) {
            sector_num += n;
            continue;
        }
        ret = bdrv_write_zeroes(bs, sector_num, n, BDRV_REQ_MAY_UNMAP);
        if (ret < 0) {
            error_report("error writing zeroes at sector %" PRId64 ": %s",
                         sector_num, strerror(-ret));
            return ret;
        }
        sector_num += n;
    }
}
#44 lore raw
Quoted 37 lines · Peter Lieven
Quoted 29 lines · Stefan Hajnoczi
Quoted 26 lines · Paolo Bonzini

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 6 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Quoted 6 lines

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero, otherwise it's emulated with bdrv_aio_writev)

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Quoted 6 lines
Quoted 4 lines

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Okay, then let's keep the patches to expose the flag.

Quoted 8 lines

Okay, then I can keep those.

Can you give a short hint if my approach with brdv_make_empty is what you want? I would like to not change the parameters, so use BDRV_REQ_MAY_UNMAP
unconditionally.

int bdrv_make_empty(BlockDriverState *bs)

The semantics of bdrv_make_empty() today are: deallocate all data in the top layer of the image file. If there is a backing file, reads will fall back to the backing file.

The semantics that you want are zeroing the entire disk image (efficiently, when possible).

A flags argument is needed to support both of sets of semantics. If you don't like that, then I suggest creating a new function called bdrv_make_zero().

Stefan

#45 lore raw
Quoted 44 lines · Stefan Hajnoczi
Quoted 31 lines · Peter Lieven
Quoted 25 lines · Stefan Hajnoczi
Quoted 23 lines · Paolo Bonzini

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 11 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero, otherwise it's emulated with bdrv_aio_writev)

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Quoted 5 lines
Quoted 4 lines

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Okay, then let's keep the patches to expose the flag.

Quoted 8 lines

Okay, then I can keep those.

Can you give a short hint if my approach with brdv_make_empty is what you want? I would like to not change the parameters, so use BDRV_REQ_MAY_UNMAP
unconditionally.

int bdrv_make_empty(BlockDriverState *bs)

The semantics of bdrv_make_empty() today are: deallocate all data in the top layer of the image file. If there is a backing file, reads will fall back to the backing file.

Quoted 6 lines

The semantics that you want are zeroing the entire disk image (efficiently, when possible).

A flags argument is needed to support both of sets of semantics. If you don't like that, then I suggest creating a new function called bdrv_make_zero().

Ok, that is what I would like to do. In this case I only have to rename bdrv_zeroize to bdrv_make_zero. Ok ?

Peter

#46 lore raw
Quoted 59 lines · Peter Lieven
Quoted 58 lines · Stefan Hajnoczi
Quoted 40 lines · Peter Lieven
Quoted 31 lines · Stefan Hajnoczi
Quoted 28 lines · Paolo Bonzini

Il 07/10/2013 10:42, Stefan Hajnoczi ha scritto:

Quoted 7 lines

Could you make bdrv_co_write_zeroes() always use UNMAP, if possible, and
avoid adding the new BDRV_REQ_MAY_UNMAP flag? While reading the first few patches in this series I wondered why there is a need to expose flags at all...

Quoted 6 lines

Sometimes it is useful to distinguish between zeroing at the image format level from discarding at the device level, but I don't think we make use of that yet. I'd prefer to keep the interface simple for now and add flags later, if necessary.

Or maybe I just missed something ;)

The flag is needed to implement the right semantics for the SCSI WRITE SAME command, which are:

- if the UNMAP bit is off, always write the sectors (that's bdrv_aio_write_zeroes without BDRV_REQ_MAY_UNMAP if the payload is zero,
otherwise it's emulated with bdrv_aio_writev)

- if the target can "discard and write the specified payload", you can discard, else you must write the sectors with the correct payload (that's bdrv_aio_write_zeroes with BDRV_REQ_MAY_UNMAP).

Quoted 6 lines
Quoted 4 lines

Contrast this with the UNMAP command, which does not make any guarantee on the content of the sectors after the command is completed (a few months ago we agreed that, even if you have discard_zeroes=true in the target, it is fine for UNMAP to do nothing).

Okay, then let's keep the patches to expose the flag.

Quoted 8 lines

Okay, then I can keep those.

Can you give a short hint if my approach with brdv_make_empty is what you want? I would like to not change the parameters, so use BDRV_REQ_MAY_UNMAP
unconditionally.

int bdrv_make_empty(BlockDriverState *bs)

The semantics of bdrv_make_empty() today are: deallocate all data in the top layer of the image file. If there is a backing file, reads will fall back to the backing file.

The semantics that you want are zeroing the entire disk image (efficiently, when possible).

Quoted 6 lines
Quoted 3 lines

A flags argument is needed to support both of sets of semantics. If you don't like that, then I suggest creating a new function called bdrv_make_zero().

Ok, that is what I would like to do. In this case I only have to rename bdrv_zeroize to bdrv_make_zero. Ok ?

Yes, please.

Stefan