All of lore.kernel.org
 help / color / mirror / Atom feed
From: Kevin Wolf <kwolf@redhat.com>
To: qemu-block@nongnu.org
Cc: kwolf@redhat.com, peter.maydell@linaro.org, qemu-devel@nongnu.org
Subject: [PULL 51/64] blockdev: Fix active commit choice
Date: Mon,  7 Sep 2020 13:09:23 +0200	[thread overview]
Message-ID: <20200907110936.261684-52-kwolf@redhat.com> (raw)
In-Reply-To: <20200907110936.261684-1-kwolf@redhat.com>

From: Max Reitz <mreitz@redhat.com>

We have to perform an active commit whenever the top node has a parent
that has taken the WRITE permission on it.

This means that block-commit's @backing-file parameter is no longer
allowed for such nodes, and that users will have to issue a
block-job-complete command.  Neither should pose a problem in practice,
because this case was basically just broken until now.

(Since this commit already touches block-commit's documentation, it also
moves up the chunk explaining general block-commit behavior that for
some reason was situated under @backing-file.)

Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 qapi/block-core.json | 39 ++++++++++++++++++++-------------------
 blockdev.c           | 35 ++++++++++++++++++++++++++++++-----
 2 files changed, 50 insertions(+), 24 deletions(-)

diff --git a/qapi/block-core.json b/qapi/block-core.json
index e34796d98f..0345f6f2d2 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1569,6 +1569,18 @@
 # Live commit of data from overlay image nodes into backing nodes - i.e.,
 # writes data between 'top' and 'base' into 'base'.
 #
+# If top == base, that is an error.
+# If top has no overlays on top of it, or if it is in use by a writer,
+# the job will not be completed by itself.  The user needs to complete
+# the job with the block-job-complete command after getting the ready
+# event. (Since 2.0)
+#
+# If the base image is smaller than top, then the base image will be
+# resized to be the same size as top.  If top is smaller than the base
+# image, the base will not be truncated.  If you want the base image
+# size to match the size of the smaller top, you can safely truncate
+# it yourself once the commit operation successfully completes.
+#
 # @job-id: identifier for the newly-created block job. If
 #          omitted, the device name will be used. (Since 2.7)
 #
@@ -1593,14 +1605,15 @@
 #       accepted
 #
 # @backing-file: The backing file string to write into the overlay
-#                image of 'top'.  If 'top' is the active layer,
-#                specifying a backing file string is an error. This
-#                filename is not validated.
+#                image of 'top'.  If 'top' does not have an overlay
+#                image, or if 'top' is in use by a writer, specifying
+#                a backing file string is an error.
 #
-#                If a pathname string is such that it cannot be
-#                resolved by QEMU, that means that subsequent QMP or
-#                HMP commands must use node-names for the image in
-#                question, as filename lookup methods will fail.
+#                This filename is not validated.  If a pathname string
+#                is such that it cannot be resolved by QEMU, that
+#                means that subsequent QMP or HMP commands must use
+#                node-names for the image in question, as filename
+#                lookup methods will fail.
 #
 #                If not specified, QEMU will automatically determine
 #                the backing file string to use, or error out if
@@ -1609,18 +1622,6 @@
 #                filename or protocol.
 #                (Since 2.1)
 #
-#                If top == base, that is an error.
-#                If top == active, the job will not be completed by itself,
-#                user needs to complete the job with the block-job-complete
-#                command after getting the ready event. (Since 2.0)
-#
-#                If the base image is smaller than top, then the base image
-#                will be resized to be the same size as top.  If top is
-#                smaller than the base image, the base will not be
-#                truncated.  If you want the base image size to match the
-#                size of the smaller top, you can safely truncate it
-#                yourself once the commit operation successfully completes.
-#
 # @speed: the maximum speed, in bytes per second
 #
 # @on-error: the action to take on an error. 'ignore' means that the request
diff --git a/blockdev.c b/blockdev.c
index f308adc9aa..7f2561081e 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -2602,6 +2602,7 @@ void qmp_block_commit(bool has_job_id, const char *job_id, const char *device,
     AioContext *aio_context;
     Error *local_err = NULL;
     int job_flags = JOB_DEFAULT;
+    uint64_t top_perm, top_shared;
 
     if (!has_speed) {
         speed = 0;
@@ -2717,14 +2718,38 @@ void qmp_block_commit(bool has_job_id, const char *job_id, const char *device,
         goto out;
     }
 
-    if (top_bs == bs) {
+    /*
+     * Active commit is required if and only if someone has taken a
+     * WRITE permission on the top node.  Historically, we have always
+     * used active commit for top nodes, so continue that practice
+     * lest we possibly break clients that rely on this behavior, e.g.
+     * to later attach this node to a writing parent.
+     * (Active commit is never really wrong.)
+     */
+    bdrv_get_cumulative_perm(top_bs, &top_perm, &top_shared);
+    if (top_perm & BLK_PERM_WRITE ||
+        bdrv_skip_filters(top_bs) == bdrv_skip_filters(bs))
+    {
         if (has_backing_file) {
-            error_setg(errp, "'backing-file' specified,"
-                             " but 'top' is the active layer");
+            if (bdrv_skip_filters(top_bs) == bdrv_skip_filters(bs)) {
+                error_setg(errp, "'backing-file' specified,"
+                                 " but 'top' is the active layer");
+            } else {
+                error_setg(errp, "'backing-file' specified, but 'top' has a "
+                                 "writer on it");
+            }
             goto out;
         }
-        commit_active_start(has_job_id ? job_id : NULL, bs, base_bs,
-                            job_flags, speed, on_error,
+        if (!has_job_id) {
+            /*
+             * Emulate here what block_job_create() does, because it
+             * is possible that @bs != @top_bs (the block job should
+             * be named after @bs, even if @top_bs is the actual
+             * source)
+             */
+            job_id = bdrv_get_device_name(bs);
+        }
+        commit_active_start(job_id, top_bs, base_bs, job_flags, speed, on_error,
                             filter_node_name, NULL, NULL, false, &local_err);
     } else {
         BlockDriverState *overlay_bs = bdrv_find_overlay(bs, top_bs);
-- 
2.25.4



  parent reply	other threads:[~2020-09-07 11:34 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-07 11:08 [PULL 00/64] Block layer patches Kevin Wolf
2020-09-07 11:08 ` [PULL 01/64] block: Raise an error when backing file parameter is an empty string Kevin Wolf
2020-09-07 11:08 ` [PULL 02/64] block/nvme: Replace magic value by SCALE_MS definition Kevin Wolf
2020-09-07 11:08 ` [PULL 03/64] block/nvme: Avoid further processing if trace event not enabled Kevin Wolf
2020-09-07 11:08 ` [PULL 04/64] block/nvme: Let nvme_create_queue_pair() fail gracefully Kevin Wolf
2020-09-07 11:08 ` [PULL 05/64] block/nvme: Define INDEX macros to ease code review Kevin Wolf
2020-09-07 11:08 ` [PULL 06/64] block/nvme: Improve error message when IO queue creation failed Kevin Wolf
2020-09-07 11:08 ` [PULL 07/64] block/nvme: Use common error path in nvme_add_io_queue() Kevin Wolf
2020-09-07 11:08 ` [PULL 08/64] block/nvme: Rename local variable Kevin Wolf
2020-09-07 11:08 ` [PULL 09/64] block/nvme: Use union of NvmeIdCtrl / NvmeIdNs structures Kevin Wolf
2020-09-07 11:08 ` [PULL 10/64] block/nvme: Replace qemu_try_blockalign0 by qemu_try_blockalign/memset Kevin Wolf
2020-09-07 11:08 ` [PULL 11/64] block/nvme: Replace qemu_try_blockalign(bs) by qemu_try_memalign(pg_sz) Kevin Wolf
2020-09-07 11:08 ` [PULL 12/64] block/nvme: Simplify nvme_init_queue() arguments Kevin Wolf
2020-09-07 11:08 ` [PULL 13/64] block/nvme: Replace BDRV_POLL_WHILE by AIO_WAIT_WHILE Kevin Wolf
2020-09-07 11:08 ` [PULL 14/64] block/nvme: Simplify nvme_create_queue_pair() arguments Kevin Wolf
2020-09-07 11:08 ` [PULL 15/64] block/nvme: Extract nvme_poll_queue() Kevin Wolf
2020-09-07 11:08 ` [PULL 16/64] block/nvme: Use an array of EventNotifier Kevin Wolf
2020-09-07 11:08 ` [PULL 17/64] block: Add child access functions Kevin Wolf
2020-09-07 11:08 ` [PULL 18/64] block: Add chain helper functions Kevin Wolf
2020-09-07 11:08 ` [PULL 19/64] block: bdrv_cow_child() for bdrv_has_zero_init() Kevin Wolf
2020-09-07 11:08 ` [PULL 20/64] block: bdrv_set_backing_hd() is about bs->backing Kevin Wolf
2020-09-07 11:08 ` [PULL 21/64] block: Include filters when freezing backing chain Kevin Wolf
2020-09-07 11:08 ` [PULL 22/64] block: Drop bdrv_is_encrypted() Kevin Wolf
2020-09-07 11:08 ` [PULL 23/64] block: Add bdrv_supports_compressed_writes() Kevin Wolf
2020-09-07 11:08 ` [PULL 24/64] throttle: Support compressed writes Kevin Wolf
2020-09-07 11:08 ` [PULL 25/64] copy-on-read: " Kevin Wolf
2020-09-07 11:08 ` [PULL 26/64] block: Use bdrv_filter_(bs|child) where obvious Kevin Wolf
2020-09-07 11:08 ` [PULL 27/64] block: Use CAFs in block status functions Kevin Wolf
2020-09-07 11:09 ` [PULL 28/64] stream: Deal with filters Kevin Wolf
2020-09-07 11:09 ` [PULL 29/64] block: Use CAFs when working with backing chains Kevin Wolf
2020-09-07 11:09 ` [PULL 30/64] block: Use bdrv_cow_child() in bdrv_co_truncate() Kevin Wolf
2020-09-07 11:09 ` [PULL 31/64] block: Re-evaluate backing file handling in reopen Kevin Wolf
2020-09-07 11:09 ` [PULL 32/64] block: Flush all children in generic code Kevin Wolf
2020-09-07 11:09 ` [PULL 33/64] vmdk: Drop vmdk_co_flush() Kevin Wolf
2020-09-07 11:09 ` [PULL 34/64] block: Iterate over children in refresh_limits Kevin Wolf
2020-09-07 11:09 ` [PULL 35/64] block: Use CAFs in bdrv_refresh_filename() Kevin Wolf
2020-09-07 11:09 ` [PULL 36/64] block: Use CAF in bdrv_co_rw_vmstate() Kevin Wolf
2020-09-07 11:09 ` [PULL 37/64] block/snapshot: Fix fallback Kevin Wolf
2021-04-30 22:30   ` Peter Maydell
2021-05-03  9:40     ` Kevin Wolf
2021-05-03  9:45       ` Max Reitz
2021-05-03 10:17         ` Kevin Wolf
2020-09-07 11:09 ` [PULL 38/64] block: Use CAFs for debug breakpoints Kevin Wolf
2020-09-07 11:09 ` [PULL 39/64] block: Improve get_allocated_file_size's default Kevin Wolf
2020-09-07 11:09 ` [PULL 40/64] block/null: Implement bdrv_get_allocated_file_size Kevin Wolf
2020-09-07 11:09 ` [PULL 41/64] blockdev: Use CAF in external_snapshot_prepare() Kevin Wolf
2020-09-07 11:09 ` [PULL 42/64] block: Report data child for query-blockstats Kevin Wolf
2020-09-07 11:09 ` [PULL 43/64] block: Use child access functions for QAPI queries Kevin Wolf
2020-09-07 11:09 ` [PULL 44/64] block-copy: Use CAF to find sync=top base Kevin Wolf
2020-09-07 11:09 ` [PULL 45/64] mirror: Deal with filters Kevin Wolf
2020-09-07 11:09 ` [PULL 46/64] backup: " Kevin Wolf
2020-09-07 11:09 ` [PULL 47/64] commit: " Kevin Wolf
2020-09-07 11:09 ` [PULL 48/64] nbd: Use CAF when looking for dirty bitmap Kevin Wolf
2020-09-07 11:09 ` [PULL 49/64] qemu-img: Use child access functions Kevin Wolf
2020-09-07 11:09 ` [PULL 50/64] block: Drop backing_bs() Kevin Wolf
2020-09-07 11:09 ` Kevin Wolf [this message]
2020-09-07 11:09 ` [PULL 52/64] block: Inline bdrv_co_block_status_from_*() Kevin Wolf
2020-09-07 11:09 ` [PULL 53/64] block: Leave BDS.backing_{file,format} constant Kevin Wolf
2020-09-07 11:09 ` [PULL 54/64] iotests: Test that qcow2's data-file is flushed Kevin Wolf
2020-09-07 11:09 ` [PULL 55/64] iotests: Let complete_and_wait() work with commit Kevin Wolf
2020-09-07 11:09 ` [PULL 56/64] iotests: Add filter commit test cases Kevin Wolf
2020-09-07 11:09 ` [PULL 57/64] iotests: Add filter mirror " Kevin Wolf
2020-09-07 11:09 ` [PULL 58/64] iotests: Add test for commit in sub directory Kevin Wolf
2020-09-07 11:09 ` [PULL 59/64] iotests: Test committing to overridden backing Kevin Wolf
2020-09-07 11:09 ` [PULL 60/64] iotests: Allow running from different directory Kevin Wolf
2020-09-07 11:09 ` [PULL 61/64] file-win32: Fix "locking" option Kevin Wolf
2020-09-07 11:09 ` [PULL 62/64] block/nvme: Group controller registers in NVMeRegs structure Kevin Wolf
2020-09-07 11:09 ` [PULL 63/64] block/nvme: Use generic NvmeBar structure Kevin Wolf
2020-09-07 11:09 ` [PULL 64/64] block/nvme: Pair doorbell registers Kevin Wolf
2020-09-07 20:22 ` [PULL 00/64] Block layer patches Peter Maydell
2020-09-08  7:01   ` Kevin Wolf
2020-09-08  9:01     ` Max 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=20200907110936.261684-52-kwolf@redhat.com \
    --to=kwolf@redhat.com \
    --cc=peter.maydell@linaro.org \
    --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.