All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
To: qemu-block@nongnu.org
Cc: qemu-devel@nongnu.org, kwolf@redhat.com, hreitz@redhat.com,
	vsementsov@yandex-team.ru
Subject: [PATCH v6 09/15] Revert "block: Let replace_child_noperm free children"
Date: Sat, 25 Jun 2022 00:28:24 +0300	[thread overview]
Message-ID: <20220624212830.316919-10-vsementsov@yandex-team.ru> (raw)
In-Reply-To: <20220624212830.316919-1-vsementsov@yandex-team.ru>

We are going to reimplement this behavior (clear bs->file / bs->backing
pointers automatically when child->bs is cleared) in a nicer way, see
further commit
"block: Manipulate bs->file / bs->backing pointers in .attach/.detach".

With this revert we bring back a problem that was fixed by b0a9f6fed3d8.
Still the problem was mostly theoretical, we don't have concrete bugs
fixed by b0a9f6fed3d8, we don't have a specific test. Probably some
accidental failures of iotests are related.

Alternatively, we may merge this and following three reverts into final
"block: Manipulate ..." to avoid any kind of regression. But seems that
in this case having separate clear revert commits is better.

This reverts commit b0a9f6fed3d80de610dcd04a7e66f9f30a04174f.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---
 block.c | 102 +++++++++++++-------------------------------------------
 1 file changed, 23 insertions(+), 79 deletions(-)

diff --git a/block.c b/block.c
index 883b1bb242..34ca046470 100644
--- a/block.c
+++ b/block.c
@@ -90,10 +90,8 @@ 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,
-                                      bool free_empty_child);
+                                      BlockDriverState *new_bs);
 static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,
                                               BdrvChild *child,
                                               Transaction *tran);
@@ -2338,7 +2336,6 @@ typedef struct BdrvReplaceChildState {
     BdrvChild *child;
     BdrvChild **childp;
     BlockDriverState *old_bs;
-    bool free_empty_child;
 } BdrvReplaceChildState;
 
 static void bdrv_replace_child_commit(void *opaque)
@@ -2346,9 +2343,6 @@ static void bdrv_replace_child_commit(void *opaque)
     BdrvReplaceChildState *s = opaque;
     GLOBAL_STATE_CODE();
 
-    if (s->free_empty_child && !s->child->bs) {
-        bdrv_child_free(s->child);
-    }
     bdrv_unref(s->old_bs);
 }
 
@@ -2366,26 +2360,22 @@ 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, 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_replace_child_noperm(&s->child, s->old_bs);
     bdrv_unref(new_bs);
 }
 
@@ -2402,44 +2392,30 @@ 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.
  *
- * 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).
+ * (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 **childp,
                                     BlockDriverState *new_bs,
-                                    Transaction *tran,
-                                    bool free_empty_child)
+                                    Transaction *tran)
 {
     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);
     }
-    /*
-     * 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);
+    bdrv_replace_child_noperm(childp, new_bs);
     /* old_bs reference is transparently moved from *childp to @s */
 }
 
@@ -2821,22 +2797,8 @@ 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,
-                                      bool free_empty_child)
+                                      BlockDriverState *new_bs)
 {
     BdrvChild *child = *childp;
     BlockDriverState *old_bs = child->bs;
@@ -2875,9 +2837,6 @@ static void bdrv_replace_child_noperm(BdrvChild **childp,
     }
 
     child->bs = new_bs;
-    if (!new_bs) {
-        *childp = NULL;
-    }
 
     if (new_bs) {
         assert_bdrv_graph_writable(new_bs);
@@ -2908,10 +2867,6 @@ static void bdrv_replace_child_noperm(BdrvChild **childp,
         bdrv_parent_drained_end_single(child);
         drain_saldo++;
     }
-
-    if (free_empty_child && !child->bs) {
-        bdrv_child_free(child);
-    }
 }
 
 /**
@@ -2943,14 +2898,7 @@ static void bdrv_attach_child_common_abort(void *opaque)
     BlockDriverState *bs = child->bs;
 
     GLOBAL_STATE_CODE();
-    /*
-     * Pass free_empty_child=false, because we still need the child
-     * for the AioContext operations on the parent below; those
-     * BdrvChildClass methods all work on a BdrvChild object, so we
-     * need to keep it as an empty shell (after this function, it will
-     * not be attached to any parent, and it will not have a .bs).
-     */
-    bdrv_replace_child_noperm(s->child, NULL, false);
+    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);
@@ -2972,6 +2920,7 @@ static void bdrv_attach_child_common_abort(void *opaque)
 
     bdrv_unref(bs);
     bdrv_child_free(child);
+    *s->child = NULL;
 }
 
 static TransactionActionDrv bdrv_attach_child_common_drv = {
@@ -3050,9 +2999,7 @@ static int bdrv_attach_child_common(BlockDriverState *child_bs,
     }
 
     bdrv_ref(child_bs);
-    bdrv_replace_child_noperm(&new_child, child_bs, true);
-    /* child_bs was non-NULL, so new_child must not have been freed */
-    assert(new_child != NULL);
+    bdrv_replace_child_noperm(&new_child, child_bs);
 
     *child = new_child;
 
@@ -3113,7 +3060,8 @@ static void bdrv_detach_child(BdrvChild **childp)
     BlockDriverState *old_bs = (*childp)->bs;
 
     GLOBAL_STATE_CODE();
-    bdrv_replace_child_noperm(childp, NULL, true);
+    bdrv_replace_child_noperm(childp, NULL);
+    bdrv_child_free(*childp);
 
     if (old_bs) {
         /*
@@ -5167,11 +5115,7 @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,
     }
 
     if (child->bs) {
-        /*
-         * Pass free_empty_child=false, we will free the child in
-         * bdrv_remove_filter_or_cow_child_commit()
-         */
-        bdrv_replace_child_tran(childp, NULL, tran, false);
+        bdrv_replace_child_tran(childp, NULL, tran);
     }
 
     s = g_new(BdrvRemoveFilterOrCowChild, 1);
@@ -5181,6 +5125,8 @@ static void bdrv_remove_file_or_backing_child(BlockDriverState *bs,
         .is_backing = (childp == &bs->backing),
     };
     tran_add(tran, &bdrv_remove_filter_or_cow_child_drv, s);
+
+    *childp = NULL;
 }
 
 /*
@@ -5224,7 +5170,7 @@ static int bdrv_replace_node_noperm(BlockDriverState *from,
          * 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, true);
+        bdrv_replace_child_tran(&c, to, tran);
     }
 
     return 0;
@@ -5389,9 +5335,7 @@ 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, true);
-    /* @new_bs must have been non-NULL, so @child must not have been freed */
-    assert(child != NULL);
+    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);
-- 
2.25.1



  parent reply	other threads:[~2022-06-24 21:39 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-24 21:28 [PATCH v6 00/15] block: cleanup backing and file handling Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 01/15] block: BlockDriver: add .filtered_child_is_backing field Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 02/15] block: introduce bdrv_open_file_child() helper Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 03/15] block/blklogwrites: don't care to remove bs->file child on failure Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 04/15] test-bdrv-graph-mod: update test_parallel_perm_update test case Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 05/15] tests-bdrv-drain: bdrv_replace_test driver: declare supports_backing Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 06/15] test-bdrv-graph-mod: fix filters to be filters Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 07/15] block: document connection between child roles and bs->backing/bs->file Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 08/15] block/snapshot: stress that we fallback to primary child Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` Vladimir Sementsov-Ogievskiy [this message]
2022-06-24 21:28 ` [PATCH v6 10/15] Revert "block: Let replace_child_tran keep indirect pointer" Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 11/15] Revert "block: Restructure remove_file_or_backing_child()" Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 12/15] Revert "block: Pass BdrvChild ** to replace_child_noperm" Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 13/15] block: Manipulate bs->file / bs->backing pointers in .attach/.detach Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 14/15] block/snapshot: drop indirection around bdrv_snapshot_fallback_ptr Vladimir Sementsov-Ogievskiy
2022-06-24 21:28 ` [PATCH v6 15/15] block: refactor bdrv_remove_file_or_backing_child to bdrv_remove_child Vladimir Sementsov-Ogievskiy
2022-06-30 13:18 ` [PATCH v6 00/15] block: cleanup backing and file handling Hanna Reitz

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220624212830.316919-10-vsementsov@yandex-team.ru \
    --to=vsementsov@yandex-team.ru \
    --cc=hreitz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.