qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Kevin Wolf <kwolf@redhat.com>
To: qemu-block@nongnu.org
Cc: kwolf@redhat.com, qemu-devel@nongnu.org
Subject: [Qemu-devel] [PULL 09/27] block: Handle permission errors in change_parent_backing_link()
Date: Tue,  7 Mar 2017 16:40:33 +0100	[thread overview]
Message-ID: <1488901251-16214-10-git-send-email-kwolf@redhat.com> (raw)
In-Reply-To: <1488901251-16214-1-git-send-email-kwolf@redhat.com>

Instead of just trying to change parents by parent over to reference @to
instead of @from, and abort()ing whenever the permissions don't allow
this, do proper permission checking beforehand and pass any error to the
callers.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
---
 block.c | 50 ++++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 44 insertions(+), 6 deletions(-)

diff --git a/block.c b/block.c
index a7b09d3..a310132 100644
--- a/block.c
+++ b/block.c
@@ -2933,21 +2933,53 @@ static bool should_update_child(BdrvChild *c, BlockDriverState *to)
 }
 
 static void change_parent_backing_link(BlockDriverState *from,
-                                       BlockDriverState *to)
+                                       BlockDriverState *to, Error **errp)
 {
     BdrvChild *c, *next;
+    GSList *list = NULL, *p;
+    uint64_t old_perm, old_shared;
+    uint64_t perm = 0, shared = BLK_PERM_ALL;
+    int ret;
+
+    /* Make sure that @from doesn't go away until we have successfully attached
+     * all of its parents to @to. */
+    bdrv_ref(from);
 
+    /* Put all parents into @list and calculate their cumulative permissions */
     QLIST_FOREACH_SAFE(c, &from->parents, next_parent, next) {
         if (!should_update_child(c, to)) {
             continue;
         }
+        list = g_slist_prepend(list, c);
+        perm |= c->perm;
+        shared &= c->shared_perm;
+    }
+
+    /* Check whether the required permissions can be granted on @to, ignoring
+     * all BdrvChild in @list so that they can't block themselves. */
+    ret = bdrv_check_update_perm(to, perm, shared, list, errp);
+    if (ret < 0) {
+        bdrv_abort_perm_update(to);
+        goto out;
+    }
+
+    /* Now actually perform the change. We performed the permission check for
+     * all elements of @list at once, so set the permissions all at once at the
+     * very end. */
+    for (p = list; p != NULL; p = p->next) {
+        c = p->data;
 
         bdrv_ref(to);
-        /* FIXME Are we sure that bdrv_replace_child() can't run into
-         * &error_abort because of permissions? */
-        bdrv_replace_child(c, to, true);
+        bdrv_replace_child_noperm(c, to);
         bdrv_unref(from);
     }
+
+    bdrv_get_cumulative_perm(to, &old_perm, &old_shared);
+    bdrv_set_perm(to, old_perm | perm, old_shared | shared);
+
+out:
+    g_slist_free(list);
+    bdrv_unref(from);
 }
 
 /*
@@ -2980,7 +3012,12 @@ void bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top,
         goto out;
     }
 
-    change_parent_backing_link(bs_top, bs_new);
+    change_parent_backing_link(bs_top, bs_new, &local_err);
+    if (local_err) {
+        error_propagate(errp, local_err);
+        bdrv_set_backing_hd(bs_new, NULL, &error_abort);
+        goto out;
+    }
 
     /* bs_new is now referenced by its new parents, we don't need the
      * additional reference any more. */
@@ -2995,7 +3032,8 @@ void bdrv_replace_in_backing_chain(BlockDriverState *old, BlockDriverState *new)
 
     bdrv_ref(old);
 
-    change_parent_backing_link(old, new);
+    /* FIXME Proper error handling */
+    change_parent_backing_link(old, new, &error_abort);
 
     bdrv_unref(old);
 }
-- 
1.8.3.1

  parent reply	other threads:[~2017-03-07 15:41 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-07 15:40 [Qemu-devel] [PULL 00/27] Block layer fixes for 2.9.0-rc0 Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 01/27] commit: Fix error handling Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 02/27] mirror: Fix permission problem with 'replaces' Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 03/27] mirror: Fix permissions for removing mirror_top_bs Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 04/27] mirror: Fix error path for dirty bitmap creation Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 05/27] block: Fix blockdev-snapshot error handling Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 06/27] block: Factor out should_update_child() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 07/27] block: Factor out bdrv_replace_child_noperm() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 08/27] block: Ignore multiple children in bdrv_check_update_perm() Kevin Wolf
2017-03-07 15:40 ` Kevin Wolf [this message]
2017-03-07 15:40 ` [Qemu-devel] [PULL 10/27] block: Fix error handling in bdrv_replace_in_backing_chain() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 11/27] sheepdog: Defuse time bomb in sd_open() error handling Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 12/27] sheepdog: Fix error handling in sd_snapshot_delete() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 13/27] sheepdog: Fix error handling sd_create() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 14/27] sheepdog: Mark sd_snapshot_delete() lossage FIXME Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 15/27] sheepdog: Fix snapshot ID parsing in _open(), _create, _goto() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 16/27] sheepdog: Don't truncate long VDI name in _open(), _create() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 17/27] sheepdog: Report errors in pseudo-filename more usefully Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 18/27] sheepdog: Use SocketAddress and socket_connect() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 19/27] sheepdog: Implement bdrv_parse_filename() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 20/27] gluster: Drop assumptions on SocketTransport names Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 21/27] gluster: Don't duplicate qapi-util.c's qapi_enum_parse() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 22/27] gluster: Plug memory leaks in qemu_gluster_parse_json() Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 23/27] qapi-schema: Rename GlusterServer to SocketAddressFlat Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 24/27] qapi-schema: Rename SocketAddressFlat's variant tcp to inet Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 25/27] sheepdog: Support blockdev-add Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 26/27] block: Don't use error_abort in blk_new_open Kevin Wolf
2017-03-07 15:40 ` [Qemu-devel] [PULL 27/27] commit: Don't use error_abort in commit_start Kevin Wolf
2017-03-08 14:49 ` [Qemu-devel] [PULL 00/27] Block layer fixes for 2.9.0-rc0 Peter Maydell

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=1488901251-16214-10-git-send-email-kwolf@redhat.com \
    --to=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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).