Message ID | 20190506122322.90035-1-sgarzare@redhat.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | [v2] block/rbd: add preallocation support | expand |
On 06.05.19 14:23, Stefano Garzarella wrote: > This patch adds the support of preallocation (off/full) for the RBD > block driver. > If available, we use rbd_writesame() to quickly fill the image when > full preallocation is required. > > Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> > --- > v2: > - Use 4 KiB buffer for rbd_writesame() [Jason] > - Use "rbd_concurrent_management_ops" and "stripe_unit" to limit > concurrent ops to the backing cluster [Jason] > --- > block/rbd.c | 188 +++++++++++++++++++++++++++++++++++++++---- > qapi/block-core.json | 4 +- > 2 files changed, 175 insertions(+), 17 deletions(-) This patch conflicts a bit with the rbd file growth patch, so that would need to be resolved in a v3. But still, that doesn’t prevent me from reviewing it as it is: > diff --git a/block/rbd.c b/block/rbd.c > index 0c549c9935..bc54395e1c 100644 > --- a/block/rbd.c > +++ b/block/rbd.c [...] > @@ -331,6 +333,147 @@ static void qemu_rbd_memset(RADOSCB *rcb, int64_t offs) [...] > +static int qemu_rbd_do_truncate(rados_t cluster, rbd_image_t image, > + int64_t offset, PreallocMode prealloc, > + Error **errp) > +{ [...] > +#ifdef LIBRBD_SUPPORTS_WRITESAME > + uint64_t stripe_unit, writesame_max_size; > + int max_concurrent_ops; > + > + max_concurrent_ops = qemu_rbd_get_max_concurrent_ops(cluster); > + ret = rbd_get_stripe_unit(image, &stripe_unit); > + if (ret < 0) { > + error_setg_errno(errp, -ret, "Failed to get stripe unit"); > + goto out; > + } > + > + /* > + * We limit the rbd_writesame() size to avoid to spawn more then s/then/than/ > + * 'rbd_concurrent_management_ops' concurrent operations. > + */ > + writesame_max_size = MIN(stripe_unit * max_concurrent_ops, INT_MAX); > +#endif /* LIBRBD_SUPPORTS_WRITESAME */ > + > + buf = g_malloc(buf_size); > + /* > + * Some versions of rbd_writesame() discards writes of buffers with > + * all zeroes. In order to avoid this behaviour, we set the first byte > + * to one. > + */ > + buf[0] = 1; But I guess you’ll need to rewrite it as zero later, or the “bdrv_rbd.bdrv_has_zero_init = bdrv_has_zero_init_1” is no longer quite true. [...] > @@ -449,6 +603,16 @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, > BLOCK_OPT_CLUSTER_SIZE, 0); > rbd_opts->has_cluster_size = (rbd_opts->cluster_size != 0); > > + prealloc = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); > + rbd_opts->preallocation = qapi_enum_parse(&PreallocMode_lookup, prealloc, > + PREALLOC_MODE_OFF, &local_err); You also need to set rbd_opts->has_preallocation to true. > + g_free(prealloc); > + if (local_err) { > + ret = -EINVAL; > + error_propagate(errp, local_err); > + goto exit; > + } > + > options = qdict_new(); > qemu_rbd_parse_filename(filename, options, &local_err); > if (local_err) { [...] > diff --git a/qapi/block-core.json b/qapi/block-core.json > index 7ccbfff9d0..db25a4065b 100644 > --- a/qapi/block-core.json > +++ b/qapi/block-core.json > @@ -4277,13 +4277,15 @@ > # point to a snapshot. > # @size Size of the virtual disk in bytes > # @cluster-size RBD object size > +# @preallocation Preallocation mode (allowed values: off, full) There should be a "(Since: 4.1)" note here. Max > # > # Since: 2.12 > ## > { 'struct': 'BlockdevCreateOptionsRbd', > 'data': { 'location': 'BlockdevOptionsRbd', > 'size': 'size', > - '*cluster-size' : 'size' } } > + '*cluster-size' : 'size', > + '*preallocation': 'PreallocMode' } } > > ## > # @BlockdevVmdkSubformat: >
On Tue, Jun 25, 2019 at 06:06:02PM +0200, Max Reitz wrote: > On 06.05.19 14:23, Stefano Garzarella wrote: > > This patch adds the support of preallocation (off/full) for the RBD > > block driver. > > If available, we use rbd_writesame() to quickly fill the image when > > full preallocation is required. > > > > Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> > > --- > > v2: > > - Use 4 KiB buffer for rbd_writesame() [Jason] > > - Use "rbd_concurrent_management_ops" and "stripe_unit" to limit > > concurrent ops to the backing cluster [Jason] > > --- > > block/rbd.c | 188 +++++++++++++++++++++++++++++++++++++++---- > > qapi/block-core.json | 4 +- > > 2 files changed, 175 insertions(+), 17 deletions(-) > > This patch conflicts a bit with the rbd file growth patch, so that would > need to be resolved in a v3. Sure, I'll rebase this patch in a v3! > > But still, that doesn’t prevent me from reviewing it as it is: > > > diff --git a/block/rbd.c b/block/rbd.c > > index 0c549c9935..bc54395e1c 100644 > > --- a/block/rbd.c > > +++ b/block/rbd.c > > [...] > > > @@ -331,6 +333,147 @@ static void qemu_rbd_memset(RADOSCB *rcb, int64_t offs) > > [...] > > > +static int qemu_rbd_do_truncate(rados_t cluster, rbd_image_t image, > > + int64_t offset, PreallocMode prealloc, > > + Error **errp) > > +{ > > [...] > > > +#ifdef LIBRBD_SUPPORTS_WRITESAME > > + uint64_t stripe_unit, writesame_max_size; > > + int max_concurrent_ops; > > + > > + max_concurrent_ops = qemu_rbd_get_max_concurrent_ops(cluster); > > + ret = rbd_get_stripe_unit(image, &stripe_unit); > > + if (ret < 0) { > > + error_setg_errno(errp, -ret, "Failed to get stripe unit"); > > + goto out; > > + } > > + > > + /* > > + * We limit the rbd_writesame() size to avoid to spawn more then > > s/then/than/ Ooo, I'll fix it! > > > + * 'rbd_concurrent_management_ops' concurrent operations. > > + */ > > + writesame_max_size = MIN(stripe_unit * max_concurrent_ops, INT_MAX); > > +#endif /* LIBRBD_SUPPORTS_WRITESAME */ > > + > > + buf = g_malloc(buf_size); > > + /* > > + * Some versions of rbd_writesame() discards writes of buffers with > > + * all zeroes. In order to avoid this behaviour, we set the first byte > > + * to one. > > + */ > > + buf[0] = 1; > > But I guess you’ll need to rewrite it as zero later, or the > “bdrv_rbd.bdrv_has_zero_init = bdrv_has_zero_init_1” is no longer quite > true. Yes, and I should use g_malloc0. I'll check if the next rewrites is too slow, in this case I'll use rbd_writesame() only for ceph version where zeroed buffer is supported. > > [...] > > > @@ -449,6 +603,16 @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, > > BLOCK_OPT_CLUSTER_SIZE, 0); > > rbd_opts->has_cluster_size = (rbd_opts->cluster_size != 0); > > > > + prealloc = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); > > + rbd_opts->preallocation = qapi_enum_parse(&PreallocMode_lookup, prealloc, > > + PREALLOC_MODE_OFF, &local_err); > > You also need to set rbd_opts->has_preallocation to true. I'll fix. > > > + g_free(prealloc); > > + if (local_err) { > > + ret = -EINVAL; > > + error_propagate(errp, local_err); > > + goto exit; > > + } > > + > > options = qdict_new(); > > qemu_rbd_parse_filename(filename, options, &local_err); > > if (local_err) { > > [...] > > > diff --git a/qapi/block-core.json b/qapi/block-core.json > > index 7ccbfff9d0..db25a4065b 100644 > > --- a/qapi/block-core.json > > +++ b/qapi/block-core.json > > @@ -4277,13 +4277,15 @@ > > # point to a snapshot. > > # @size Size of the virtual disk in bytes > > # @cluster-size RBD object size > > +# @preallocation Preallocation mode (allowed values: off, full) > > There should be a "(Since: 4.1)" note here. I'll add the note! Thanks for the review, Stefano
diff --git a/block/rbd.c b/block/rbd.c index 0c549c9935..bc54395e1c 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -13,6 +13,7 @@ #include "qemu/osdep.h" +#include "qemu/units.h" #include <rbd/librbd.h> #include "qapi/error.h" #include "qemu/error-report.h" @@ -63,6 +64,7 @@ #define OBJ_MAX_SIZE (1UL << OBJ_DEFAULT_OBJ_ORDER) #define RBD_MAX_SNAPS 100 +#define RBD_DEFAULT_CONCURRENT_OPS 10 /* The LIBRBD_SUPPORTS_IOVEC is defined in librbd.h */ #ifdef LIBRBD_SUPPORTS_IOVEC @@ -331,6 +333,147 @@ static void qemu_rbd_memset(RADOSCB *rcb, int64_t offs) } } +static int qemu_rbd_get_max_concurrent_ops(rados_t cluster) +{ + char buf[16]; + int ret, max_concurrent_ops; + + ret = rados_conf_get(cluster, "rbd_concurrent_management_ops", buf, + sizeof(buf)); + if (ret < 0) { + return RBD_DEFAULT_CONCURRENT_OPS; + } + + ret = qemu_strtoi(buf, NULL, 10, &max_concurrent_ops); + if (ret < 0) { + return RBD_DEFAULT_CONCURRENT_OPS; + } + + return max_concurrent_ops; +} + +static int qemu_rbd_do_truncate(rados_t cluster, rbd_image_t image, + int64_t offset, PreallocMode prealloc, + Error **errp) +{ + uint64_t current_length; + char *buf = NULL; + int ret; + + ret = rbd_get_size(image, ¤t_length); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to get file length"); + goto out; + } + + if (current_length > offset && prealloc != PREALLOC_MODE_OFF) { + error_setg(errp, "Cannot use preallocation for shrinking files"); + ret = -ENOTSUP; + goto out; + } + + switch (prealloc) { + case PREALLOC_MODE_FULL: { + uint64_t current_offset = current_length; + int buf_size = 4 * KiB; + ssize_t bytes; + +#ifdef LIBRBD_SUPPORTS_WRITESAME + uint64_t stripe_unit, writesame_max_size; + int max_concurrent_ops; + + max_concurrent_ops = qemu_rbd_get_max_concurrent_ops(cluster); + ret = rbd_get_stripe_unit(image, &stripe_unit); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to get stripe unit"); + goto out; + } + + /* + * We limit the rbd_writesame() size to avoid to spawn more then + * 'rbd_concurrent_management_ops' concurrent operations. + */ + writesame_max_size = MIN(stripe_unit * max_concurrent_ops, INT_MAX); +#endif /* LIBRBD_SUPPORTS_WRITESAME */ + + buf = g_malloc(buf_size); + /* + * Some versions of rbd_writesame() discards writes of buffers with + * all zeroes. In order to avoid this behaviour, we set the first byte + * to one. + */ + buf[0] = 1; + + ret = rbd_resize(image, offset); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to resize file"); + goto out; + } + +#ifdef LIBRBD_SUPPORTS_WRITESAME + while (offset - current_offset > buf_size) { + bytes = MIN(offset - current_offset, writesame_max_size); + /* + * rbd_writesame() supports only request where the size of the + * operation is multiple of buffer size. + */ + bytes -= bytes % buf_size; + + bytes = rbd_writesame(image, current_offset, bytes, buf, buf_size, + 0); + if (bytes < 0) { + ret = bytes; + error_setg_errno(errp, -ret, + "Failed to write for preallocation"); + goto out; + } + + current_offset += bytes; + } +#endif /* LIBRBD_SUPPORTS_WRITESAME */ + + while (current_offset < offset) { + bytes = rbd_write(image, current_offset, + MIN(offset - current_offset, buf_size), buf); + if (bytes < 0) { + ret = bytes; + error_setg_errno(errp, -ret, + "Failed to write for preallocation"); + goto out; + } + + current_offset += bytes; + } + + ret = rbd_flush(image); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to flush the file"); + goto out; + } + + break; + } + case PREALLOC_MODE_OFF: + ret = rbd_resize(image, offset); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to resize file"); + goto out; + } + break; + default: + error_setg(errp, "Unsupported preallocation mode: %s", + PreallocMode_str(prealloc)); + ret = -ENOTSUP; + goto out; + } + + ret = 0; + +out: + g_free(buf); + return ret; +} + static QemuOptsList runtime_opts = { .name = "rbd", .head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head), @@ -376,6 +519,7 @@ static int qemu_rbd_do_create(BlockdevCreateOptions *options, BlockdevCreateOptionsRbd *opts = &options->u.rbd; rados_t cluster; rados_ioctx_t io_ctx; + rbd_image_t image; int obj_order = 0; int ret; @@ -404,13 +548,22 @@ static int qemu_rbd_do_create(BlockdevCreateOptions *options, return ret; } - ret = rbd_create(io_ctx, opts->location->image, opts->size, &obj_order); + ret = rbd_create(io_ctx, opts->location->image, 0, &obj_order); if (ret < 0) { error_setg_errno(errp, -ret, "error rbd create"); goto out; } - ret = 0; + ret = rbd_open(io_ctx, opts->location->image, &image, NULL); + if (ret < 0) { + error_setg_errno(errp, -ret, "error rbd open"); + goto out; + } + + ret = qemu_rbd_do_truncate(cluster, image, opts->size, opts->preallocation, + errp); + + rbd_close(image); out: rados_ioctx_destroy(io_ctx); rados_shutdown(cluster); @@ -431,6 +584,7 @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, BlockdevOptionsRbd *loc; Error *local_err = NULL; const char *keypairs, *password_secret; + char *prealloc; QDict *options = NULL; int ret = 0; @@ -449,6 +603,16 @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, BLOCK_OPT_CLUSTER_SIZE, 0); rbd_opts->has_cluster_size = (rbd_opts->cluster_size != 0); + prealloc = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); + rbd_opts->preallocation = qapi_enum_parse(&PreallocMode_lookup, prealloc, + PREALLOC_MODE_OFF, &local_err); + g_free(prealloc); + if (local_err) { + ret = -EINVAL; + error_propagate(errp, local_err); + goto exit; + } + options = qdict_new(); qemu_rbd_parse_filename(filename, options, &local_err); if (local_err) { @@ -1052,21 +1216,8 @@ static int coroutine_fn qemu_rbd_co_truncate(BlockDriverState *bs, Error **errp) { BDRVRBDState *s = bs->opaque; - int r; - if (prealloc != PREALLOC_MODE_OFF) { - error_setg(errp, "Unsupported preallocation mode '%s'", - PreallocMode_str(prealloc)); - return -ENOTSUP; - } - - r = rbd_resize(s->image, offset); - if (r < 0) { - error_setg_errno(errp, -r, "Failed to resize file"); - return r; - } - - return 0; + return qemu_rbd_do_truncate(s->cluster, s->image, offset, prealloc, errp); } static int qemu_rbd_snap_create(BlockDriverState *bs, @@ -1219,6 +1370,11 @@ static QemuOptsList qemu_rbd_create_opts = { .type = QEMU_OPT_SIZE, .help = "RBD object size" }, + { + .name = BLOCK_OPT_PREALLOC, + .type = QEMU_OPT_STRING, + .help = "Preallocation mode (allowed values: off, full)" + }, { .name = "password-secret", .type = QEMU_OPT_STRING, diff --git a/qapi/block-core.json b/qapi/block-core.json index 7ccbfff9d0..db25a4065b 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -4277,13 +4277,15 @@ # point to a snapshot. # @size Size of the virtual disk in bytes # @cluster-size RBD object size +# @preallocation Preallocation mode (allowed values: off, full) # # Since: 2.12 ## { 'struct': 'BlockdevCreateOptionsRbd', 'data': { 'location': 'BlockdevOptionsRbd', 'size': 'size', - '*cluster-size' : 'size' } } + '*cluster-size' : 'size', + '*preallocation': 'PreallocMode' } } ## # @BlockdevVmdkSubformat:
This patch adds the support of preallocation (off/full) for the RBD block driver. If available, we use rbd_writesame() to quickly fill the image when full preallocation is required. Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> --- v2: - Use 4 KiB buffer for rbd_writesame() [Jason] - Use "rbd_concurrent_management_ops" and "stripe_unit" to limit concurrent ops to the backing cluster [Jason] --- block/rbd.c | 188 +++++++++++++++++++++++++++++++++++++++---- qapi/block-core.json | 4 +- 2 files changed, 175 insertions(+), 17 deletions(-)