From nobody Mon Apr 29 06:20:58 2024 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=fail(p=none dis=none) header.from=virtuozzo.com Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1541006128014550.4067813353433; Wed, 31 Oct 2018 10:15:28 -0700 (PDT) Received: from localhost ([::1]:60909 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gHu5S-0001Ba-9c for importer@patchew.org; Wed, 31 Oct 2018 13:15:22 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53954) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gHter-0006j7-8P for qemu-devel@nongnu.org; Wed, 31 Oct 2018 12:47:54 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gHten-0004da-1a for qemu-devel@nongnu.org; Wed, 31 Oct 2018 12:47:53 -0400 Received: from relay.sw.ru ([185.231.240.75]:55774) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1gHtem-0004cp-NI; Wed, 31 Oct 2018 12:47:48 -0400 Received: from [172.16.25.136] (helo=localhost.sw.ru) by relay.sw.ru with esmtp (Exim 4.90_1) (envelope-from ) id 1gHteQ-0005jc-K4; Wed, 31 Oct 2018 19:47:26 +0300 From: Andrey Shinkevich To: qemu-devel@nongnu.org, qemu-block@nongnu.org Date: Wed, 31 Oct 2018 19:47:19 +0300 Message-Id: <1541004440-182262-2-git-send-email-andrey.shinkevich@virtuozzo.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1541004440-182262-1-git-send-email-andrey.shinkevich@virtuozzo.com> References: <1541004440-182262-1-git-send-email-andrey.shinkevich@virtuozzo.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x X-Received-From: 185.231.240.75 X-Mailman-Approved-At: Wed, 31 Oct 2018 13:12:12 -0400 Subject: [Qemu-devel] [PATCH 1/2] The discard flag for block stream operation X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kwolf@redhat.com, vsementsov@virtuozzo.com, jcody@redhat.com, armbru@redhat.com, dgilbert@redhat.com, andrey.shinkevich@virtuozzo.com, den@openvz.org, mreitz@redhat.com Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Adding a parameter to QMP block-stream command to allow discarding blocks in the backing chain while blocks are being copied to the active layer. Signed-off-by: Andrey Shinkevich --- block/stream.c | 2 +- blockdev.c | 8 +++++++- hmp-commands.hx | 4 ++-- hmp.c | 4 +++- include/block/block_int.h | 2 +- qapi/block-core.json | 5 ++++- 6 files changed, 18 insertions(+), 7 deletions(-) diff --git a/block/stream.c b/block/stream.c index 81a7ec8..db81df4 100644 --- a/block/stream.c +++ b/block/stream.c @@ -221,7 +221,7 @@ static const BlockJobDriver stream_job_driver =3D { =20 void stream_start(const char *job_id, BlockDriverState *bs, BlockDriverState *base, const char *backing_file_str, - int creation_flags, int64_t speed, + int creation_flags, int64_t speed, bool discard, BlockdevOnError on_error, Error **errp) { StreamBlockJob *s; diff --git a/blockdev.c b/blockdev.c index 574adbc..04aecf5 100644 --- a/blockdev.c +++ b/blockdev.c @@ -3122,6 +3122,7 @@ void qmp_block_stream(bool has_job_id, const char *jo= b_id, const char *device, bool has_base_node, const char *base_node, bool has_backing_file, const char *backing_file, bool has_speed, int64_t speed, + bool has_discard, bool discard, bool has_on_error, BlockdevOnError on_error, bool has_auto_finalize, bool auto_finalize, bool has_auto_dismiss, bool auto_dismiss, @@ -3138,6 +3139,10 @@ void qmp_block_stream(bool has_job_id, const char *j= ob_id, const char *device, on_error =3D BLOCKDEV_ON_ERROR_REPORT; } =20 + if (!has_discard) { + discard =3D false; + } + bs =3D bdrv_lookup_bs(device, device, errp); if (!bs) { return; @@ -3202,7 +3207,8 @@ void qmp_block_stream(bool has_job_id, const char *jo= b_id, const char *device, } =20 stream_start(has_job_id ? job_id : NULL, bs, base_bs, base_name, - job_flags, has_speed ? speed : 0, on_error, &local_err); + job_flags, has_speed ? speed : 0, + discard, on_error, &local_err); if (local_err) { error_propagate(errp, local_err); goto out; diff --git a/hmp-commands.hx b/hmp-commands.hx index db0c681..b455e0d 100644 --- a/hmp-commands.hx +++ b/hmp-commands.hx @@ -95,8 +95,8 @@ ETEXI =20 { .name =3D "block_stream", - .args_type =3D "device:B,speed:o?,base:s?", - .params =3D "device [speed [base]]", + .args_type =3D "device:B,speed:o?,base:s?,discard:o?", + .params =3D "device [speed [base]] [discard]", .help =3D "copy data from a backing file into a block device= ", .cmd =3D hmp_block_stream, }, diff --git a/hmp.c b/hmp.c index 7828f93..c63e806 100644 --- a/hmp.c +++ b/hmp.c @@ -1920,9 +1920,11 @@ void hmp_block_stream(Monitor *mon, const QDict *qdi= ct) const char *device =3D qdict_get_str(qdict, "device"); const char *base =3D qdict_get_try_str(qdict, "base"); int64_t speed =3D qdict_get_try_int(qdict, "speed", 0); + bool discard =3D qdict_get_try_bool(qdict, "discard", false); =20 qmp_block_stream(true, device, device, base !=3D NULL, base, false, NU= LL, - false, NULL, qdict_haskey(qdict, "speed"), speed, tru= e, + false, NULL, qdict_haskey(qdict, "speed"), speed, + qdict_haskey(qdict, "discard"), discard, true, BLOCKDEV_ON_ERROR_REPORT, false, false, false, false, &error); =20 diff --git a/include/block/block_int.h b/include/block/block_int.h index 92ecbd8..e531d03 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -970,7 +970,7 @@ int is_windows_drive(const char *filename); */ void stream_start(const char *job_id, BlockDriverState *bs, BlockDriverState *base, const char *backing_file_str, - int creation_flags, int64_t speed, + int creation_flags, int64_t speed, bool discard, BlockdevOnError on_error, Error **errp); =20 /** diff --git a/qapi/block-core.json b/qapi/block-core.json index cfb37f8..3f50b88 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -2329,6 +2329,9 @@ # # @speed: the maximum speed, in bytes per second # +# @discard: true to delete blocks duplicated in old backing files. +# (default: false). Since 3.1. +# # @on-error: the action to take on an error (default report). # 'stop' and 'enospc' can only be used if the block device # supports io-status (see BlockInfo). Since 1.3. @@ -2361,7 +2364,7 @@ { 'command': 'block-stream', 'data': { '*job-id': 'str', 'device': 'str', '*base': 'str', '*base-node': 'str', '*backing-file': 'str', '*speed': 'int', - '*on-error': 'BlockdevOnError', + '*discard': 'bool', '*on-error': 'BlockdevOnError', '*auto-finalize': 'bool', '*auto-dismiss': 'bool' } } =20 ## --=20 1.8.3.1 From nobody Mon Apr 29 06:20:58 2024 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=fail(p=none dis=none) header.from=virtuozzo.com Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1541006227482302.2067122623928; Wed, 31 Oct 2018 10:17:07 -0700 (PDT) Received: from localhost ([::1]:60923 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gHu78-0002UT-38 for importer@patchew.org; Wed, 31 Oct 2018 13:17:06 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53955) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gHter-0006j8-8Q for qemu-devel@nongnu.org; Wed, 31 Oct 2018 12:47:57 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gHten-0004di-6n for qemu-devel@nongnu.org; Wed, 31 Oct 2018 12:47:53 -0400 Received: from relay.sw.ru ([185.231.240.75]:55778) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1gHtem-0004d5-P7; Wed, 31 Oct 2018 12:47:49 -0400 Received: from [172.16.25.136] (helo=localhost.sw.ru) by relay.sw.ru with esmtp (Exim 4.90_1) (envelope-from ) id 1gHteQ-0005jc-Qn; Wed, 31 Oct 2018 19:47:26 +0300 From: Andrey Shinkevich To: qemu-devel@nongnu.org, qemu-block@nongnu.org Date: Wed, 31 Oct 2018 19:47:20 +0300 Message-Id: <1541004440-182262-3-git-send-email-andrey.shinkevich@virtuozzo.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1541004440-182262-1-git-send-email-andrey.shinkevich@virtuozzo.com> References: <1541004440-182262-1-git-send-email-andrey.shinkevich@virtuozzo.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x X-Received-From: 185.231.240.75 X-Mailman-Approved-At: Wed, 31 Oct 2018 13:12:12 -0400 Subject: [Qemu-devel] [PATCH 2/2] Discard blocks while copy-on-read X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kwolf@redhat.com, vsementsov@virtuozzo.com, jcody@redhat.com, armbru@redhat.com, dgilbert@redhat.com, andrey.shinkevich@virtuozzo.com, den@openvz.org, mreitz@redhat.com Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Discards the block duplicated in an intermediate backing file after the block have been copied into the active layer during QMP block-stream operation. It saves the disk space while merging snapshots. Signed-off-by: Andrey Shinkevich --- block/stream.c | 400 +++++++++++++++++++++++++++++++++++++++++++++++++++++= ++-- 1 file changed, 389 insertions(+), 11 deletions(-) diff --git a/block/stream.c b/block/stream.c index db81df4..0adceb4 100644 --- a/block/stream.c +++ b/block/stream.c @@ -12,6 +12,7 @@ */ =20 #include "qemu/osdep.h" +#include "qemu/cutils.h" #include "trace.h" #include "block/block_int.h" #include "block/blockjob_int.h" @@ -35,9 +36,62 @@ typedef struct StreamBlockJob { BlockdevOnError on_error; char *backing_file_str; int bs_flags; + bool discard; + BlockDriverState *stream_top_bs; + GSList *im_nodes; } StreamBlockJob; =20 -static int coroutine_fn stream_populate(BlockBackend *blk, +typedef struct IntermediateNode { + BlockBackend *blk; + int flags; +} IntermediateNode; + +static inline void restore_all_im_nodes(StreamBlockJob *s) +{ + GSList *l; + BlockDriverState *bs_active; + BlockDriverState *bs_im; + IntermediateNode *im_node; + BlockReopenQueue *queue =3D NULL; + Error *local_err =3D NULL; + + assert(s->stream_top_bs && s->stream_top_bs->backing && + s->stream_top_bs->backing->bs); + bs_active =3D backing_bs(s->stream_top_bs); + assert(backing_bs(bs_active)); + + bdrv_subtree_drained_begin(backing_bs(bs_active)); + + for (l =3D s->im_nodes; l; l =3D l->next) { + im_node =3D l->data; + if (im_node->blk) { + bs_im =3D blk_bs(im_node->blk); + + if (im_node->flags !=3D bdrv_get_flags(bs_im) && bs_im) { + queue =3D bdrv_reopen_queue(queue, bs_im, NULL, im_node->f= lags); + } + /* Give up write permissions before making it read-only */ + blk_set_perm(im_node->blk, 0, BLK_PERM_ALL, &error_abort); + blk_unref(im_node->blk); + bdrv_unref(bs_im); + } + g_free(im_node); + } + g_slist_free(s->im_nodes); + s->im_nodes =3D NULL; + + if (queue) { + bdrv_reopen_multiple(bdrv_get_aio_context(bs_active), queue, + &local_err); + if (local_err !=3D NULL) { + error_report_err(local_err); + } + } + + bdrv_subtree_drained_end(backing_bs(bs_active)); +} + +static int coroutine_fn stream_populate(const StreamBlockJob *s, int64_t offset, uint64_t bytes, void *buf) { @@ -46,12 +100,83 @@ static int coroutine_fn stream_populate(BlockBackend *= blk, .iov_len =3D bytes, }; QEMUIOVector qiov; + GSList *l; + IntermediateNode *im_node; + int ret; =20 + assert(s); assert(bytes < SIZE_MAX); qemu_iovec_init_external(&qiov, &iov, 1); =20 /* Copy-on-read the unallocated clusters */ - return blk_co_preadv(blk, offset, qiov.size, &qiov, BDRV_REQ_COPY_ON_R= EAD); + ret =3D blk_co_preadv(s->common.blk, offset, qiov.size, &qiov, + BDRV_REQ_COPY_ON_READ); + + if (ret < 0 || !s->discard) { + return ret; + } + + for (l =3D s->im_nodes; l; l =3D l->next) { + im_node =3D l->data; + blk_co_pdiscard(im_node->blk, offset, bytes); + } + + return ret; +} + +static int stream_exit_discard(StreamBlockJob *s) +{ + BlockJob *bjob =3D &s->common; + BlockDriverState *bs =3D backing_bs(s->stream_top_bs); + BlockDriverState *base =3D s->base; + Error *local_err =3D NULL; + int ret =3D 0; + + /* Make sure that the BDS doesn't go away during bdrv_replace_node, + * before we can call bdrv_drained_end */ + bdrv_ref(s->stream_top_bs); + /* Reopen intermediate images back in read-only mode */ + restore_all_im_nodes(s); + /* Hold a guest back from writing until we remove the filter */ + bdrv_drained_begin(bs); + /* Dropping WRITE is required before changing the backing file. */ + bdrv_child_try_set_perm(s->stream_top_bs->backing, 0, BLK_PERM_ALL, + &error_abort); + if (bs->backing) { + const char *base_id =3D NULL, *base_fmt =3D NULL; + if (base) { + base_id =3D s->backing_file_str; + if (base->drv) { + base_fmt =3D base->drv->format_name; + } + } + ret =3D bdrv_change_backing_file(bs, base_id, base_fmt); + bdrv_set_backing_hd(bs, base, &local_err); + if (local_err) { + error_report_err(local_err); + ret =3D -EPERM; + } + } + /* Remove the filter driver from the graph. Before this, get rid of + * the blockers on the intermediate nodes so that the resulting state = is + * valid. Also give up permissions on stream_top_bs->backing, which mi= ght + * block the removal. */ + block_job_remove_all_bdrv(bjob); + bdrv_child_try_set_perm(s->stream_top_bs->backing, 0, BLK_PERM_ALL, + &error_abort); + bdrv_replace_node(s->stream_top_bs, backing_bs(s->stream_top_bs), + &error_abort); + /* We just changed the BDS the job BB refers to (with either or both o= f the + * bdrv_replace_node() calls), so switch the BB back so the cleanup do= es + * the right thing. We don't need any permissions any more now. */ + blk_remove_bs(bjob->blk); + blk_set_perm(bjob->blk, 0, BLK_PERM_ALL, &error_abort); + blk_insert_bs(bjob->blk, s->stream_top_bs, &error_abort); + + bdrv_drained_end(bs); + bdrv_unref(s->stream_top_bs); + + return ret; } =20 static int stream_prepare(Job *job) @@ -63,6 +188,10 @@ static int stream_prepare(Job *job) Error *local_err =3D NULL; int ret =3D 0; =20 + if (s->discard) { + return stream_exit_discard(s); + } + if (bs->backing) { const char *base_id =3D NULL, *base_fmt =3D NULL; if (base) { @@ -102,7 +231,7 @@ static int coroutine_fn stream_run(Job *job, Error **er= rp) { StreamBlockJob *s =3D container_of(job, StreamBlockJob, common.job); BlockBackend *blk =3D s->common.blk; - BlockDriverState *bs =3D blk_bs(blk); + BlockDriverState *bs; BlockDriverState *base =3D s->base; int64_t len; int64_t offset =3D 0; @@ -112,6 +241,12 @@ static int coroutine_fn stream_run(Job *job, Error **e= rrp) int64_t n =3D 0; /* bytes */ void *buf; =20 + if (s->discard) { + bs =3D backing_bs(s->stream_top_bs); + } else { + bs =3D blk_bs(blk); + } + if (!bs->backing) { goto out; } @@ -165,7 +300,7 @@ static int coroutine_fn stream_run(Job *job, Error **er= rp) } trace_stream_one_iteration(s, offset, n, ret); if (copy) { - ret =3D stream_populate(blk, offset, n, buf); + ret =3D stream_populate(s, offset, n, buf); } if (ret < 0) { BlockErrorAction action =3D @@ -206,6 +341,209 @@ out: return ret; } =20 +static int coroutine_fn bdrv_stream_top_preadv(BlockDriverState *bs, + uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags) +{ + return bdrv_co_preadv(bs->backing, offset, bytes, qiov, flags); +} + +static int coroutine_fn bdrv_stream_top_pwritev(BlockDriverState *bs, + uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags) +{ + return bdrv_co_pwritev(bs->backing, offset, bytes, qiov, flags); +} + +static int coroutine_fn bdrv_stream_top_flush(BlockDriverState *bs) +{ + if (bs->backing =3D=3D NULL) { + /* we can be here after failed bdrv_append in stream_start */ + return 0; + } + return bdrv_co_flush(bs->backing->bs); +} + +static int coroutine_fn bdrv_stream_top_pwrite_zeroes(BlockDriverState *bs, + int64_t offset, int bytes, BdrvRequestFlags flags) +{ + return bdrv_co_pwrite_zeroes(bs->backing, offset, bytes, flags); +} + +static int coroutine_fn bdrv_stream_top_pdiscard(BlockDriverState *bs, + int64_t offset, int bytes) +{ + return bdrv_co_pdiscard(bs->backing, offset, bytes); +} + +static int bdrv_stream_top_get_info(BlockDriverState *bs, BlockDriverInfo = *bdi) +{ + return bdrv_get_info(bs->backing->bs, bdi); +} + +static void bdrv_stream_top_refresh_filename(BlockDriverState *bs, QDict *= opts) +{ + if (bs->backing =3D=3D NULL) { + /* we can be here after failed bdrv_attach_child in + * bdrv_set_backing_hd */ + return; + } + bdrv_refresh_filename(bs->backing->bs); + pstrcpy(bs->exact_filename, sizeof(bs->exact_filename), + bs->backing->bs->filename); +} + +static void bdrv_stream_top_child_perm(BlockDriverState *bs, BdrvChild *c, + const BdrvChildRole *role, + BlockReopenQueue *reopen_queue, + uint64_t perm, uint64_t shared, + uint64_t *nperm, uint64_t *nshared) +{ + /* Must be able to forward guest writes to the real image */ + *nperm =3D 0; + if (perm & BLK_PERM_WRITE) { + *nperm |=3D BLK_PERM_WRITE; + } + + *nshared =3D BLK_PERM_ALL; +} + +/* Dummy node that provides consistent read to its users without requiring= it + * from its backing file and that allows writes on the backing file chain.= */ +static BlockDriver bdrv_stream_top =3D { + .format_name =3D "stream_top", + .bdrv_co_preadv =3D bdrv_stream_top_preadv, + .bdrv_co_pwritev =3D bdrv_stream_top_pwritev, + .bdrv_co_pwrite_zeroes =3D bdrv_stream_top_pwrite_zeroes, + .bdrv_co_pdiscard =3D bdrv_stream_top_pdiscard, + .bdrv_get_info =3D bdrv_stream_top_get_info, + .bdrv_co_flush =3D bdrv_stream_top_flush, + .bdrv_co_block_status =3D bdrv_co_block_status_from_backing, + .bdrv_refresh_filename =3D bdrv_stream_top_refresh_filename, + .bdrv_child_perm =3D bdrv_stream_top_child_perm, +}; + +/* In the case of block discard, add a dummy driver + * to make the backing chain writable. */ +static BlockDriverState *insert_filter(BlockDriverState *bs, Error **errp) +{ + const char *filter_node_name =3D NULL; + BlockDriverState *stream_top_bs; + Error *local_err =3D NULL; + + stream_top_bs =3D bdrv_new_open_driver(&bdrv_stream_top, filter_node_n= ame, + BDRV_O_RDWR, errp); + if (stream_top_bs =3D=3D NULL) { + return NULL; + } + if (!filter_node_name) { + stream_top_bs->implicit =3D true; + } + + stream_top_bs->total_sectors =3D bs->total_sectors; + stream_top_bs->supported_write_flags =3D BDRV_REQ_WRITE_UNCHANGED; + stream_top_bs->supported_zero_flags =3D BDRV_REQ_WRITE_UNCHANGED; + bdrv_set_aio_context(stream_top_bs, bdrv_get_aio_context(bs)); + + /* bdrv_append takes ownership of the stream_top_bs reference, need to= keep + * it alive until block_job_create() succeeds even if bs has no parent= . */ + bdrv_ref(stream_top_bs); + bdrv_drained_begin(bs); + bdrv_append(stream_top_bs, bs, &local_err); + bdrv_drained_end(bs); + + if (local_err) { + bdrv_unref(stream_top_bs); + error_propagate(errp, local_err); + return NULL; + } + + return stream_top_bs; +} + +/* Makes intermediate block chain writable */ +static int init_intermediate_nodes(StreamBlockJob *s, + BlockDriverState *bs, + BlockDriverState *base, Error **errp) +{ + BlockDriverState *iter; + int backing_bs_flags; + IntermediateNode *im_node; + BlockBackend *blk; + BlockReopenQueue *queue =3D NULL; + Error *local_err =3D NULL; + int ret; + + /* Sanity check */ + if (!backing_bs(bs)) { + error_setg(errp, "Top BDS does not have a backing file."); + return -EINVAL; + } + if (base && !bdrv_chain_contains(bs, base)) { + error_setg(errp, "The backing chain does not contain the base file= ."); + return -EINVAL; + } + + /* Reopen intermediate images in read-write mode */ + bdrv_subtree_drained_begin(backing_bs(bs)); + + for (iter =3D backing_bs(bs); iter && iter !=3D base; iter =3D backing= _bs(iter)) { + /* Keep the intermediate backing chain with BDRV original flags */ + backing_bs_flags =3D bdrv_get_flags(iter); + im_node =3D g_new0(IntermediateNode, 1); + im_node->blk =3D NULL; + im_node->flags =3D backing_bs_flags; + bdrv_ref(iter); + s->im_nodes =3D g_slist_prepend(s->im_nodes, im_node); + + if ((backing_bs_flags & BDRV_O_RDWR) =3D=3D 0) { + queue =3D bdrv_reopen_queue(queue, iter, NULL, + backing_bs_flags | BDRV_O_RDWR); + } + } + + if (queue) { + ret =3D bdrv_reopen_multiple(bdrv_get_aio_context(bs), queue, &loc= al_err); + if (local_err !=3D NULL) { + error_propagate(errp, local_err); + bdrv_subtree_drained_end(backing_bs(bs)); + restore_all_im_nodes(s); + return -1; + } + } + + bdrv_subtree_drained_end(backing_bs(bs)); + + s->im_nodes =3D g_slist_reverse(s->im_nodes); + GSList *l =3D s->im_nodes; + + for (iter =3D backing_bs(bs); iter && iter !=3D base; iter =3D backing= _bs(iter)) { + blk =3D blk_new(BLK_PERM_WRITE, BLK_PERM_CONSISTENT_READ | + BLK_PERM_WRITE | BLK_PERM_WRITE_UNCHANGED | + BLK_PERM_GRAPH_MOD); + if (!blk) { + error_setg(errp, + "Block Stream: failed to create new Block Backend."= ); + goto fail; + } + + ret =3D blk_insert_bs(blk, iter, errp); + if (ret < 0) { + goto fail; + } + + assert(l); + im_node =3D l->data; + im_node->blk =3D blk; + l =3D l->next; + } + + return 0; + +fail: + restore_all_im_nodes(s); + + return -1; +} + static const BlockJobDriver stream_job_driver =3D { .job_driver =3D { .instance_size =3D sizeof(StreamBlockJob), @@ -224,9 +562,12 @@ void stream_start(const char *job_id, BlockDriverState= *bs, int creation_flags, int64_t speed, bool discard, BlockdevOnError on_error, Error **errp) { - StreamBlockJob *s; + StreamBlockJob *s =3D NULL; BlockDriverState *iter; int orig_bs_flags; + BlockDriverState *stream_top_bs; + int node_shared_flags =3D BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UN= CHANGED; + int ret; =20 /* Make sure that the image is opened in read-write mode */ orig_bs_flags =3D bdrv_get_flags(bs); @@ -236,10 +577,19 @@ void stream_start(const char *job_id, BlockDriverStat= e *bs, } } =20 + if (discard) { + node_shared_flags |=3D BLK_PERM_WRITE; + stream_top_bs =3D insert_filter(bs, errp); + if (stream_top_bs =3D=3D NULL) { + goto fail; + } + } else { + stream_top_bs =3D bs; + } /* Prevent concurrent jobs trying to modify the graph structure here, = we * already have our own plans. Also don't allow resize as the image si= ze is * queried only at the job start and then cached. */ - s =3D block_job_create(job_id, &stream_job_driver, NULL, bs, + s =3D block_job_create(job_id, &stream_job_driver, NULL, stream_top_bs, BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UNCHANG= ED | BLK_PERM_GRAPH_MOD, BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UNCHANG= ED | @@ -251,18 +601,28 @@ void stream_start(const char *job_id, BlockDriverStat= e *bs, =20 /* Block all intermediate nodes between bs and base, because they will * disappear from the chain after this operation. The streaming job re= ads - * every block only once, assuming that it doesn't change, so block wr= ites - * and resizes. */ + * every block only once, assuming that it doesn't change, so forbid w= rites + * and resizes. Allow writing in case of discard. */ for (iter =3D backing_bs(bs); iter && iter !=3D base; iter =3D backing= _bs(iter)) { block_job_add_bdrv(&s->common, "intermediate node", iter, 0, - BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UNCHA= NGED, - &error_abort); + node_shared_flags, &error_abort); + } + + if (discard) { + s->stream_top_bs =3D stream_top_bs; + /* The block job now has a reference to this node */ + bdrv_unref(stream_top_bs); + + ret =3D init_intermediate_nodes(s, bs, base, errp); + if (ret < 0) { + goto fail; + } } =20 s->base =3D base; s->backing_file_str =3D g_strdup(backing_file_str); s->bs_flags =3D orig_bs_flags; - + s->discard =3D discard; s->on_error =3D on_error; trace_stream_start(bs, base, s); job_start(&s->common.job); @@ -272,4 +632,22 @@ fail: if (orig_bs_flags !=3D bdrv_get_flags(bs)) { bdrv_reopen(bs, orig_bs_flags, NULL); } + if (!discard) { + return; + } + if (s) { + /* Make sure this BDS does not go away until we have completed the= graph + * changes below */ + bdrv_ref(stream_top_bs); + job_early_fail(&s->common.job); + } + if (stream_top_bs) { + bdrv_drained_begin(bs); + bdrv_child_try_set_perm(stream_top_bs->backing, 0, BLK_PERM_ALL, + &error_abort); + bdrv_replace_node(stream_top_bs, backing_bs(stream_top_bs), + &error_abort); + bdrv_drained_end(bs); + bdrv_unref(stream_top_bs); + } } --=20 1.8.3.1