The following changes since commit 42f6c9179be4401974dd3a75ee72defd16b5092d:

Merge tag 'pull-ppc-20211112' of https://github.com/legoater/qemu into staging (2021-11-12 12:28:25 +0100)

are available in the Git repository at:

https://gitlab.com/hreitz/qemu.git tags/pull-block-2021-11-16

for you to fetch changes up to 5dbd0ce115c7720268e653fe928bb6a0c1314a80:

file-posix: Fix alignment after reopen changing O_DIRECT (2021-11-16 11:30:29 +0100)

- Note that because I had to modify Kevin's pull request, I did not want

to merge it partially (with a merge commit), but instead decided to

apply all patches from the pull request mails (including their message

IDs)

----------------------------------------------------------------

Hanna Reitz (10):

stream: Traverse graph after modification

block: Manipulate children list in .attach/.detach

block: Unite remove_empty_child and child_free

block: Drop detached child from ignore list

block: Pass BdrvChild ** to replace_child_noperm

block: Restructure remove_file_or_backing_child()

transactions: Invoke clean() after everything else

block: Let replace_child_tran keep indirect pointer

block: Let replace_child_noperm free children

iotests/030: Unthrottle parallel jobs in reverse

Kevin Wolf (2):

docs: Deprecate incorrectly typed device_add arguments

file-posix: Fix alignment after reopen changing O_DIRECT

Stefan Hajnoczi (1):

softmmu/qdev-monitor: fix use-after-free in qdev_set_id()

docs/about/deprecated.rst | 14 +++

include/qemu/transactions.h | 3 +

block.c | 233 +++++++++++++++++++++++++++---------

block/file-posix.c | 20 +++-

block/stream.c | 7 +-

softmmu/qdev-monitor.c | 2 +-

util/transactions.c | 8 +-

tests/qemu-iotests/030 | 11 +-

tests/qemu-iotests/142 | 29 +++++

tests/qemu-iotests/142.out | 18 +++

10 files changed, 279 insertions(+), 66 deletions(-)

2.33.1

bdrv_cor_filter_drop() modifies the block graph. That means that other

parties can also modify the block graph before it returns. Therefore,

we cannot assume that the result of a graph traversal we did before

remains valid afterwards.

We should thus fetch `base` and `unfiltered_base` afterwards instead of

before.

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Reviewed-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Message-Id: <20211111120829.81329-2-hreitz@redhat.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-2-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

block/stream.c | 7 +++++--

1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/block/stream.c b/block/stream.c

--- a/block/stream.c

+++ b/block/stream.c

@@ -XXX,XX +XXX,XX @@ static int stream_prepare(Job *job)

{

StreamBlockJob *s = container_of(job, StreamBlockJob, common.job);

BlockDriverState *unfiltered_bs = bdrv_skip_filters(s->target_bs);

- BlockDriverState *base = bdrv_filter_or_cow_bs(s->above_base);

- BlockDriverState *unfiltered_base = bdrv_skip_filters(base);

+ BlockDriverState *base;

+ BlockDriverState *unfiltered_base;

Error *local_err = NULL;

int ret = 0;

@@ -XXX,XX +XXX,XX @@ static int stream_prepare(Job *job)

bdrv_cor_filter_drop(s->cor_filter_bs);

s->cor_filter_bs = NULL;

+ base = bdrv_filter_or_cow_bs(s->above_base);

+ unfiltered_base = bdrv_skip_filters(base);

+

if (bdrv_cow_child(unfiltered_bs)) {

const char *base_id = NULL, *base_fmt = NULL;

if (unfiltered_base) {

2.33.1

The children list is specific to BDS parents. We should not modify it

in the general children modification code, but let BDS parents deal with

it in their .attach() and .detach() methods.

This also has the advantage that a BdrvChild is removed from the

children list before its .bs pointer can become NULL. BDS parents

generally assume that their children's .bs pointer is never NULL, so

this is actually a bug fix.

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Message-Id: <20211111120829.81329-3-hreitz@redhat.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-3-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

---

block.c | 14 +++++---------

1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/block.c b/block.c

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_attach(BdrvChild *child)

{

BlockDriverState *bs = child->opaque;

+ QLIST_INSERT_HEAD(&bs->children, child, next);

if (child->role & BDRV_CHILD_COW) {

bdrv_backing_attach(child);

}

@@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_detach(BdrvChild *child)

}

bdrv_unapply_subtree_drain(child, bs);

+ QLIST_REMOVE(child, next);

}

static int bdrv_child_cb_update_filename(BdrvChild *c, BlockDriverState *base,

@@ -XXX,XX +XXX,XX @@ static void bdrv_child_free(void *opaque)

static void bdrv_remove_empty_child(BdrvChild *child)

{

assert(!child->bs);

- QLIST_SAFE_REMOVE(child, next);

+ assert(!child->next.le_prev); /* not in children list */

bdrv_child_free(child);

}

@@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_noperm(BlockDriverState *parent_bs,

return ret;

- QLIST_INSERT_HEAD(&parent_bs->children, *child, next);

- /*

- * child is removed in bdrv_attach_child_common_abort(), so don't care to

- * abort this change separately.

- */

return 0;

}

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_filter_or_cow_child_abort(void *opaque)

BdrvRemoveFilterOrCowChild *s = opaque;

BlockDriverState *parent_bs = s->child->opaque;

- QLIST_INSERT_HEAD(&parent_bs->children, s->child, next);

if (s->is_backing) {

parent_bs->backing = s->child;

} else {

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

};

tran_add(tran, &bdrv_remove_filter_or_cow_child_drv, s);

- QLIST_SAFE_REMOVE(child, next);

if (s->is_backing) {

bs->backing = NULL;

} else {

2.33.1

Now that bdrv_remove_empty_child() no longer removes the child from the

parent's children list but only checks that it is not in such a list, it

is only a wrapper around bdrv_child_free() that checks that the child is

empty and unused. That should apply to all children that we free, so

put those checks into bdrv_child_free() and drop

bdrv_remove_empty_child().

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Reviewed-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Message-Id: <20211111120829.81329-4-hreitz@redhat.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-4-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

---

block.c | 26 +++++++++++++-------------

1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/block.c b/block.c

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild *child,

-static void bdrv_child_free(void *opaque)

-{

- BdrvChild *c = opaque;

-

- g_free(c->name);

- g_free(c);

-}

-

-static void bdrv_remove_empty_child(BdrvChild *child)

+/**

+ * Free the given @child.

+ *

+ * The child must be empty (i.e. `child->bs == NULL`) and it must be

+ * unused (i.e. not in a children list).

+ */

+static void bdrv_child_free(BdrvChild *child)

{

assert(!child->bs);

assert(!child->next.le_prev); /* not in children list */

- bdrv_child_free(child);

+

+ g_free(child->name);

+ g_free(child);

}

typedef struct BdrvAttachChildCommonState {

@@ -XXX,XX +XXX,XX @@ static void bdrv_attach_child_common_abort(void *opaque)

bdrv_unref(bs);

- bdrv_remove_empty_child(child);

+ bdrv_child_free(child);

*s->child = NULL;

}

@@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_common(BlockDriverState *child_bs,

if (ret < 0) {

error_propagate(errp, local_err);

- bdrv_remove_empty_child(new_child);

+ bdrv_child_free(new_child);

return ret;

}

}

@@ -XXX,XX +XXX,XX @@ static void bdrv_detach_child(BdrvChild *child)

BlockDriverState *old_bs = child->bs;

bdrv_replace_child_noperm(child, NULL);

- bdrv_remove_empty_child(child);

+ bdrv_child_free(child);

if (old_bs) {

/*

2.33.1

bdrv_attach_child_common_abort() restores the parent's AioContext. To

do so, the child (which was supposed to be attached, but is now detached

again by this abort handler) is added to the ignore list for the

AioContext changing functions.

However, since we modify a BDS's children list in the BdrvChildClass's

.attach and .detach handlers, the child is already effectively detached

from the parent by this point. We do not need to put it into the ignore

list.

Use this opportunity to clean up the empty line structure: Keep setting

the ignore list, invoking the AioContext function, and freeing the

ignore list in blocks separated by empty lines.

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Reviewed-by: Kevin Wolf <kwolf@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Message-Id: <20211111120829.81329-5-hreitz@redhat.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-5-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

block.c | 8 +++++---

1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/block.c b/block.c

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static void bdrv_attach_child_common_abort(void *opaque)

if (bdrv_child_get_parent_aio_context(child) != s->old_parent_ctx) {

- GSList *ignore = g_slist_prepend(NULL, child);

+ GSList *ignore;

+ /* No need to ignore `child`, because it has been detached already */

+ ignore = NULL;

child->klass->can_set_aio_ctx(child, s->old_parent_ctx, &ignore,

&error_abort);

g_slist_free(ignore);

- ignore = g_slist_prepend(NULL, child);

- child->klass->set_aio_ctx(child, s->old_parent_ctx, &ignore);

+ ignore = NULL;

+ child->klass->set_aio_ctx(child, s->old_parent_ctx, &ignore);

g_slist_free(ignore);

2.33.1

bdrv_replace_child_noperm() modifies BdrvChild.bs, and can potentially

set it to NULL. That is dangerous, because BDS parents generally assume

that their children's .bs pointer is never NULL. We therefore want to

let bdrv_replace_child_noperm() set the corresponding BdrvChild pointer

to NULL, too.

This patch lays the foundation for it by passing a BdrvChild ** pointer

to bdrv_replace_child_noperm() so that it can later use it to NULL the

BdrvChild pointer immediately after setting BdrvChild.bs to NULL.

(We will still need to undertake some intermediate steps, though.)

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Message-Id: <20211111120829.81329-6-hreitz@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-6-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

block.c | 23 ++++++++++++-----------

1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/block.c b/block.c

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_open_inherit(const char *filename,

static bool bdrv_recurse_has_child(BlockDriverState *bs,

BlockDriverState *child);

-static void bdrv_replace_child_noperm(BdrvChild *child,

+static void bdrv_replace_child_noperm(BdrvChild **child,

BlockDriverState *new_bs);

static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

BdrvChild *child,

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_abort(void *opaque)

BlockDriverState *new_bs = s->child->bs;

/* old_bs reference is transparently moved from @s to @s->child */

- bdrv_replace_child_noperm(s->child, s->old_bs);

+ bdrv_replace_child_noperm(&s->child, s->old_bs);

bdrv_unref(new_bs);

}

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_tran(BdrvChild *child, BlockDriverState *new_bs,

if (new_bs) {

bdrv_ref(new_bs);

}

- bdrv_replace_child_noperm(child, new_bs);

+ bdrv_replace_child_noperm(&child, new_bs);

/* old_bs reference is transparently moved from @child to @s */

}

@@ -XXX,XX +XXX,XX @@ uint64_t bdrv_qapi_perm_to_blk_perm(BlockPermission qapi_perm)

return permissions[qapi_perm];

}

-static void bdrv_replace_child_noperm(BdrvChild *child,

+static void bdrv_replace_child_noperm(BdrvChild **childp,

BlockDriverState *new_bs)

{

+ BdrvChild *child = *childp;

BlockDriverState *old_bs = child->bs;

int new_bs_quiesce_counter;

int drain_saldo;

@@ -XXX,XX +XXX,XX @@ static void bdrv_attach_child_common_abort(void *opaque)

BdrvChild *child = *s->child;

BlockDriverState *bs = child->bs;

- bdrv_replace_child_noperm(child, NULL);

+ bdrv_replace_child_noperm(s->child, NULL);

if (bdrv_get_aio_context(bs) != s->old_child_ctx) {

bdrv_try_set_aio_context(bs, s->old_child_ctx, &error_abort);

@@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_common(BlockDriverState *child_bs,

}

bdrv_ref(child_bs);

- bdrv_replace_child_noperm(new_child, child_bs);

+ bdrv_replace_child_noperm(&new_child, child_bs);

*child = new_child;

@@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_noperm(BlockDriverState *parent_bs,

return 0;

}

-static void bdrv_detach_child(BdrvChild *child)

+static void bdrv_detach_child(BdrvChild **childp)

{

- BlockDriverState *old_bs = child->bs;

+ BlockDriverState *old_bs = (*childp)->bs;

- bdrv_replace_child_noperm(child, NULL);

- bdrv_child_free(child);

+ bdrv_replace_child_noperm(childp, NULL);

+ bdrv_child_free(*childp);

if (old_bs) {

/*

@@ -XXX,XX +XXX,XX @@ void bdrv_root_unref_child(BdrvChild *child)

BlockDriverState *child_bs;

child_bs = child->bs;

- bdrv_detach_child(child);

+ bdrv_detach_child(&child);

bdrv_unref(child_bs);

}

2.33.1

As of a future patch, bdrv_replace_child_tran() will take a BdrvChild **

pointer. Prepare for that by getting such a pointer and using it where

applicable, and (dereferenced) as a parameter for

bdrv_replace_child_tran().

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Message-Id: <20211111120829.81329-7-hreitz@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-7-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

---

block.c | 21 ++++++++++++---------

1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/block.c b/block.c

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

BdrvChild *child,

Transaction *tran)

+ BdrvChild **childp;

BdrvRemoveFilterOrCowChild *s;

- assert(child == bs->backing || child == bs->file);

-

if (!child) {

return;

+ if (child == bs->backing) {

+ childp = &bs->backing;

+ } else if (child == bs->file) {

+ childp = &bs->file;

+ g_assert_not_reached();

if (child->bs) {

- bdrv_replace_child_tran(child, NULL, tran);

+ bdrv_replace_child_tran(*childp, NULL, tran);

s = g_new(BdrvRemoveFilterOrCowChild, 1);

*s = (BdrvRemoveFilterOrCowChild) {

.child = child,

- .is_backing = (child == bs->backing),

+ .is_backing = (childp == &bs->backing),

};

tran_add(tran, &bdrv_remove_filter_or_cow_child_drv, s);

- if (s->is_backing) {

- bs->backing = NULL;

- } else {

- bs->file = NULL;

- }

+ *childp = NULL;

/*

2.33.1

Invoke the transaction drivers' .clean() methods only after all

.commit() or .abort() handlers are done.

This makes it easier to have nested transactions where the top-level

transactions pass objects to lower transactions that the latter can

still use throughout their commit/abort phases, while the top-level

transaction keeps a reference that is released in its .clean() method.

(Before this commit, that is also possible, but the top-level

transaction would need to take care to invoke tran_add() before the

lower-level transaction does. This commit makes the ordering

irrelevant, which is just a bit nicer.)

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Message-Id: <20211111120829.81329-8-hreitz@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-8-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

include/qemu/transactions.h | 3 +++

util/transactions.c | 8 ++++++--

2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/include/qemu/transactions.h b/include/qemu/transactions.h

--- a/include/qemu/transactions.h

+++ b/include/qemu/transactions.h

@@ -XXX,XX +XXX,XX @@

* tran_create(), call your "prepare" functions on it, and finally call

* tran_abort() or tran_commit() to finalize the transaction by corresponding

* finalization actions in reverse order.

+ *

+ * The clean() functions registered by the drivers in a transaction are called

+ * last, after all abort() or commit() functions have been called.

*/

#ifndef QEMU_TRANSACTIONS_H

diff --git a/util/transactions.c b/util/transactions.c

index XXXXXXX..XXXXXXX 100644

--- a/util/transactions.c

+++ b/util/transactions.c

@@ -XXX,XX +XXX,XX @@ void tran_abort(Transaction *tran)

{

TransactionAction *act, *next;

- QSLIST_FOREACH_SAFE(act, &tran->actions, entry, next) {

+ QSLIST_FOREACH(act, &tran->actions, entry) {

if (act->drv->abort) {

act->drv->abort(act->opaque);

+ }

+ QSLIST_FOREACH_SAFE(act, &tran->actions, entry, next) {

if (act->drv->clean) {

act->drv->clean(act->opaque);

@@ -XXX,XX +XXX,XX @@ void tran_commit(Transaction *tran)

{

TransactionAction *act, *next;

- QSLIST_FOREACH_SAFE(act, &tran->actions, entry, next) {

+ QSLIST_FOREACH(act, &tran->actions, entry) {

if (act->drv->commit) {

act->drv->commit(act->opaque);

+ }

+ QSLIST_FOREACH_SAFE(act, &tran->actions, entry, next) {

if (act->drv->clean) {

act->drv->clean(act->opaque);

2.33.1

As of a future commit, bdrv_replace_child_noperm() will clear the

indirect BdrvChild pointer passed to it if the new child BDS is NULL.

bdrv_replace_child_tran() will want to let it do that, but revert this

change in its abort handler. For that, we need to have it receive a

BdrvChild ** pointer, too, and keep it stored in the

BdrvReplaceChildState object that we attach to the transaction.

Note that we do not need to store it in the BdrvReplaceChildState when

new_bs is not NULL, because then there is nothing to revert. This is

important so that bdrv_replace_node_noperm() can pass a pointer to a

loop-local variable to bdrv_replace_child_tran() without worrying that

this pointer will outlive one loop iteration.

(Of course, for that to work, bdrv_replace_node_noperm() and in turn

bdrv_replace_node() and its relatives may not be called with a NULL @to

node. Luckily, they already are not, but now we should assert this.)

bdrv_remove_file_or_backing_child() on the other hand needs to ensure

that the indirect pointer it passes will stay valid for the duration of

the transaction. Ensure this by keeping a strong reference to the BDS

whose &bs->backing or &bs->file it passes to bdrv_replace_child_tran(),

and giving up that reference only in the transaction .clean() handler.

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Message-Id: <20211111120829.81329-9-hreitz@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-9-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

block.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++++++-------

1 file changed, 73 insertions(+), 10 deletions(-)

diff --git a/block.c b/block.c

index XXXXXXX..XXXXXXX 100644

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static int bdrv_drv_set_perm(BlockDriverState *bs, uint64_t perm,

typedef struct BdrvReplaceChildState {

BdrvChild *child;

+ BdrvChild **childp;

BlockDriverState *old_bs;

} BdrvReplaceChildState;

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_abort(void *opaque)

BdrvReplaceChildState *s = opaque;

BlockDriverState *new_bs = s->child->bs;

- /* old_bs reference is transparently moved from @s to @s->child */

+ /*

+ * old_bs reference is transparently moved from @s to s->child.

+ *

+ * Pass &s->child here instead of s->childp, because:

+ * (1) s->old_bs must be non-NULL, so bdrv_replace_child_noperm() will not

+ * modify the BdrvChild * pointer we indirectly pass to it, i.e. it

+ * will not modify s->child. From that perspective, it does not matter

+ * whether we pass s->childp or &s->child.

+ * (TODO: Right now, bdrv_replace_child_noperm() never modifies that

+ * pointer anyway (though it will in the future), so at this point it

+ * absolutely does not matter whether we pass s->childp or &s->child.)

+ * (2) If new_bs is not NULL, s->childp will be NULL. We then cannot use

+ * it here.

+ * (3) If new_bs is NULL, *s->childp will have been NULLed by

+ * bdrv_replace_child_tran()'s bdrv_replace_child_noperm() call, and we

+ * must not pass a NULL *s->childp here.

+ * (TODO: In its current state, bdrv_replace_child_noperm() will not

+ * have NULLed *s->childp, so this does not apply yet. It will in the

+ * future.)

+ *

+ * So whether new_bs was NULL or not, we cannot pass s->childp here; and in

+ * any case, there is no reason to pass it anyway.

+ */

bdrv_replace_child_noperm(&s->child, s->old_bs);

bdrv_unref(new_bs);

}

@@ -XXX,XX +XXX,XX @@ static TransactionActionDrv bdrv_replace_child_drv = {

* Note: real unref of old_bs is done only on commit.

*

* The function doesn't update permissions, caller is responsible for this.

+ *

+ * Note that if new_bs == NULL, @childp is stored in a state object attached

+ * to @tran, so that the old child can be reinstated in the abort handler.

+ * Therefore, if @new_bs can be NULL, @childp must stay valid until the

+ * transaction is committed or aborted.

+ *

+ * (TODO: The reinstating does not happen yet, but it will once

+ * bdrv_replace_child_noperm() NULLs *childp when new_bs is NULL.)

*/

-static void bdrv_replace_child_tran(BdrvChild *child, BlockDriverState *new_bs,

+static void bdrv_replace_child_tran(BdrvChild **childp,

+ BlockDriverState *new_bs,

Transaction *tran)

{

BdrvReplaceChildState *s = g_new(BdrvReplaceChildState, 1);

*s = (BdrvReplaceChildState) {

- .child = child,

- .old_bs = child->bs,

+ .child = *childp,

+ .childp = new_bs == NULL ? childp : NULL,

+ .old_bs = (*childp)->bs,

};

tran_add(tran, &bdrv_replace_child_drv, s);

if (new_bs) {

bdrv_ref(new_bs);

}

- bdrv_replace_child_noperm(&child, new_bs);

- /* old_bs reference is transparently moved from @child to @s */

+ bdrv_replace_child_noperm(childp, new_bs);

+ /* old_bs reference is transparently moved from *childp to @s */

}

/*

@@ -XXX,XX +XXX,XX @@ static bool should_update_child(BdrvChild *c, BlockDriverState *to)

typedef struct BdrvRemoveFilterOrCowChild {

BdrvChild *child;

+ BlockDriverState *bs;

bool is_backing;

} BdrvRemoveFilterOrCowChild;

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_filter_or_cow_child_commit(void *opaque)

bdrv_child_free(s->child);

}

+static void bdrv_remove_filter_or_cow_child_clean(void *opaque)

+{

+ BdrvRemoveFilterOrCowChild *s = opaque;

+

+ /* Drop the bs reference after the transaction is done */

+ bdrv_unref(s->bs);

+ g_free(s);

+}

+

static TransactionActionDrv bdrv_remove_filter_or_cow_child_drv = {

.abort = bdrv_remove_filter_or_cow_child_abort,

.commit = bdrv_remove_filter_or_cow_child_commit,

- .clean = g_free,

+ .clean = bdrv_remove_filter_or_cow_child_clean,

};

/*

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

return;

}

+ /*

+ * Keep a reference to @bs so @childp will stay valid throughout the

+ * transaction (required by bdrv_replace_child_tran())

+ */

+ bdrv_ref(bs);

if (child == bs->backing) {

childp = &bs->backing;

} else if (child == bs->file) {

@@ -XXX,XX +XXX,XX @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

}

if (child->bs) {

- bdrv_replace_child_tran(*childp, NULL, tran);

+ bdrv_replace_child_tran(childp, NULL, tran);

}

s = g_new(BdrvRemoveFilterOrCowChild, 1);

*s = (BdrvRemoveFilterOrCowChild) {

.child = child,

+ .bs = bs,

.is_backing = (childp == &bs->backing),

};

tran_add(tran, &bdrv_remove_filter_or_cow_child_drv, s);

@@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_noperm(BlockDriverState *from,

{

BdrvChild *c, *next;

+ assert(to != NULL);

+

QLIST_FOREACH_SAFE(c, &from->parents, next_parent, next) {

assert(c->bs == from);

if (!should_update_child(c, to)) {

@@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_noperm(BlockDriverState *from,

c->name, from->node_name);

return -EPERM;

}

- bdrv_replace_child_tran(c, to, tran);

+

+ /*

+ * Passing a pointer to the local variable @c is fine here, because

+ * @to is not NULL, and so &c will not be attached to the transaction.

+ */

+ bdrv_replace_child_tran(&c, to, tran);

}

return 0;

@@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_noperm(BlockDriverState *from,

*

* With @detach_subchain=true @to must be in a backing chain of @from. In this

* case backing link of the cow-parent of @to is removed.

+ *

+ * @to must not be NULL.

*/

static int bdrv_replace_node_common(BlockDriverState *from,

BlockDriverState *to,

@@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_common(BlockDriverState *from,

BlockDriverState *to_cow_parent = NULL;

int ret;

+ assert(to != NULL);

+

if (detach_subchain) {

assert(bdrv_chain_contains(from, to));

assert(from != to);

@@ -XXX,XX +XXX,XX @@ out:

return ret;

}

+/**

+ * Replace node @from by @to (where neither may be NULL).

+ */

int bdrv_replace_node(BlockDriverState *from, BlockDriverState *to,

Error **errp)

{

@@ -XXX,XX +XXX,XX @@ int bdrv_replace_child_bs(BdrvChild *child, BlockDriverState *new_bs,

bdrv_drained_begin(old_bs);

bdrv_drained_begin(new_bs);

- bdrv_replace_child_tran(child, new_bs, tran);

+ bdrv_replace_child_tran(&child, new_bs, tran);

found = g_hash_table_new(NULL, NULL);

refresh_list = bdrv_topological_dfs(refresh_list, found, old_bs);

2.33.1

In most of the block layer, especially when traversing down from other

BlockDriverStates, we assume that BdrvChild.bs can never be NULL. When

it becomes NULL, it is expected that the corresponding BdrvChild pointer

also becomes NULL and the BdrvChild object is freed.

Therefore, once bdrv_replace_child_noperm() sets the BdrvChild.bs

pointer to NULL, it should also immediately set the corresponding

BdrvChild pointer (like bs->file or bs->backing) to NULL.

In that context, it also makes sense for this function to free the

child. Sometimes we cannot do so, though, because it is called in a

transactional context where the caller might still want to reinstate the

child in the abort branch (and free it only on commit), so this behavior

has to remain optional.

In bdrv_replace_child_tran()'s abort handler, we now rely on the fact

that the BdrvChild passed to bdrv_replace_child_tran() must have had a

non-NULL .bs pointer initially. Make a note of that and assert it.

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

Message-Id: <20211111120829.81329-10-hreitz@redhat.com>

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Message-Id: <20211115145409.176785-10-kwolf@redhat.com>

Signed-off-by: Hanna Reitz <hreitz@redhat.com>

block.c | 102 +++++++++++++++++++++++++++++++++++++++++++-------------

1 file changed, 79 insertions(+), 23 deletions(-)

diff --git a/block.c b/block.c

index XXXXXXX..XXXXXXX 100644

--- a/block.c

+++ b/block.c

@@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_open_inherit(const char *filename,

static bool bdrv_recurse_has_child(BlockDriverState *bs,

BlockDriverState *child);

+static void bdrv_child_free(BdrvChild *child);

static void bdrv_replace_child_noperm(BdrvChild **child,

- BlockDriverState *new_bs);

+ BlockDriverState *new_bs,

+ bool free_empty_child);

static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,

BdrvChild *child,

Transaction *tran);

@@ -XXX,XX +XXX,XX @@ typedef struct BdrvReplaceChildState {

BdrvChild *child;

BdrvChild **childp;

BlockDriverState *old_bs;

+ bool free_empty_child;

} BdrvReplaceChildState;

static void bdrv_replace_child_commit(void *opaque)

{

BdrvReplaceChildState *s = opaque;

+ if (s->free_empty_child && !s->child->bs) {

+ bdrv_child_free(s->child);

+ }

bdrv_unref(s->old_bs);

}

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_abort(void *opaque)

* modify the BdrvChild * pointer we indirectly pass to it, i.e. it

* will not modify s->child. From that perspective, it does not matter

* whether we pass s->childp or &s->child.

- * (TODO: Right now, bdrv_replace_child_noperm() never modifies that

- * pointer anyway (though it will in the future), so at this point it

- * absolutely does not matter whether we pass s->childp or &s->child.)

* (2) If new_bs is not NULL, s->childp will be NULL. We then cannot use

* it here.

* (3) If new_bs is NULL, *s->childp will have been NULLed by

* bdrv_replace_child_tran()'s bdrv_replace_child_noperm() call, and we

* must not pass a NULL *s->childp here.

- * (TODO: In its current state, bdrv_replace_child_noperm() will not

- * have NULLed *s->childp, so this does not apply yet. It will in the

- * future.)

*

* So whether new_bs was NULL or not, we cannot pass s->childp here; and in

* any case, there is no reason to pass it anyway.

*/

- bdrv_replace_child_noperm(&s->child, s->old_bs);

+ bdrv_replace_child_noperm(&s->child, s->old_bs, true);

+ /*

+ * The child was pre-existing, so s->old_bs must be non-NULL, and

+ * s->child thus must not have been freed

+ */

+ assert(s->child != NULL);

+ if (!new_bs) {

+ /* As described above, *s->childp was cleared, so restore it */

+ assert(s->childp != NULL);

+ *s->childp = s->child;

+ }

bdrv_unref(new_bs);

}

@@ -XXX,XX +XXX,XX @@ static TransactionActionDrv bdrv_replace_child_drv = {

*

* The function doesn't update permissions, caller is responsible for this.

*

+ * (*childp)->bs must not be NULL.

+ *

* Note that if new_bs == NULL, @childp is stored in a state object attached

* to @tran, so that the old child can be reinstated in the abort handler.

* Therefore, if @new_bs can be NULL, @childp must stay valid until the

* transaction is committed or aborted.

*

- * (TODO: The reinstating does not happen yet, but it will once

- * bdrv_replace_child_noperm() NULLs *childp when new_bs is NULL.)

+ * If @free_empty_child is true and @new_bs is NULL, the BdrvChild is

+ * freed (on commit). @free_empty_child should only be false if the

+ * caller will free the BDrvChild themselves (which may be important

+ * if this is in turn called in another transactional context).

*/

static void bdrv_replace_child_tran(BdrvChild **childp,

BlockDriverState *new_bs,

- Transaction *tran)

+ Transaction *tran,

+ bool free_empty_child)

{

BdrvReplaceChildState *s = g_new(BdrvReplaceChildState, 1);

*s = (BdrvReplaceChildState) {

.child = *childp,

.childp = new_bs == NULL ? childp : NULL,

.old_bs = (*childp)->bs,

+ .free_empty_child = free_empty_child,

};

tran_add(tran, &bdrv_replace_child_drv, s);

+ /* The abort handler relies on this */

+ assert(s->old_bs != NULL);

if (new_bs) {

bdrv_ref(new_bs);

}

- bdrv_replace_child_noperm(childp, new_bs);

+ /*

+ * Pass free_empty_child=false, we will free the child (if

+ * necessary) in bdrv_replace_child_commit() (if our

+ * @free_empty_child parameter was true).

+ */

+ bdrv_replace_child_noperm(childp, new_bs, false);

/* old_bs reference is transparently moved from *childp to @s */

}

@@ -XXX,XX +XXX,XX @@ uint64_t bdrv_qapi_perm_to_blk_perm(BlockPermission qapi_perm)

return permissions[qapi_perm];

}

+/**

+ * Replace (*childp)->bs by @new_bs.

+ *

+ * If @new_bs is NULL, *childp will be set to NULL, too: BDS parents

+ * generally cannot handle a BdrvChild with .bs == NULL, so clearing

+ * BdrvChild.bs should generally immediately be followed by the

+ * BdrvChild pointer being cleared as well.

+ *

+ * If @free_empty_child is true and @new_bs is NULL, the BdrvChild is

+ * freed. @free_empty_child should only be false if the caller will

+ * free the BdrvChild themselves (this may be important in a

+ * transactional context, where it may only be freed on commit).

+ */

static void bdrv_replace_child_noperm(BdrvChild **childp,

- BlockDriverState *new_bs)

+ BlockDriverState *new_bs,

+ bool free_empty_child)

{

BdrvChild *child = *childp;

BlockDriverState *old_bs = child->bs;

@@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild **childp,

}

child->bs = new_bs;

+ if (!new_bs) {

+ *childp = NULL;

+ }

if (new_bs) {

QLIST_INSERT_HEAD(&new_bs->parents, child, next_parent);