[Qemu-devel] [PATCH v3 0/9] block: Add auto-read-only option

[Qemu-devel] [PATCH v3 0/9] block: Add auto-read-only option

[Kevin Wolf]

See patch 2 for an explanation of the motivation.

v3:
- Clarified QAPI schema documentation that auto-read-only can only
  degrade read-write to read-only, not the other way round [Eric]
- Don't refuse to set copy-on-read=on and auto-read-only=on at the same
  time; only complain when actually trying to degrade to read-only
- Let bdrv_apply_auto_read_only() return -EACCESS on all errors
- Fixed file-posix and gluster implementations [Eric, Niels]
- Added a patch to make auto-read-only=on the default for human user
  interfaces (-drive/-hda/...)

v2:
- Turn bdrv_set_read_only() into bdrv_apply_auto_read_only()
- Support the option in a lot more block drivers

Kevin Wolf (9):
  block: Update flags in bdrv_set_read_only()
  block: Add auto-read-only option
  block: Require auto-read-only for existing fallbacks
  nbd: Support auto-read-only option
  file-posix: Support auto-read-only option
  curl: Support auto-read-only option
  gluster: Support auto-read-only option
  iscsi: Support auto-read-only option
  block: Make auto-read-only=on default for -drive

 qapi/block-core.json  |  7 ++++++
 include/block/block.h |  5 +++-
 block.c               | 54 +++++++++++++++++++++++++++++++++++--------
 block/bochs.c         | 17 +++++---------
 block/cloop.c         | 16 ++++---------
 block/curl.c          |  8 +++----
 block/dmg.c           | 16 ++++---------
 block/file-posix.c    | 19 ++++++++++++---
 block/gluster.c       | 12 ++++++++--
 block/iscsi.c         |  8 ++++---
 block/nbd-client.c    | 10 ++++----
 block/rbd.c           | 14 ++++-------
 block/vvfat.c         | 11 +++------
 blockdev.c            |  1 +
 14 files changed, 120 insertions(+), 78 deletions(-)

-- 
2.19.1

[Qemu-devel] [PATCH v3 1/9] block: Update flags in bdrv_set_read_only()

[Kevin Wolf]

To fully change the read-only state of a node, we must not only change
bs->read_only, but also update bs->open_flags.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
---
 block.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/block.c b/block.c
index 0d6e5f1a76..d7bd6d29b4 100644
--- a/block.c
+++ b/block.c
@@ -281,6 +281,13 @@ int bdrv_set_read_only(BlockDriverState *bs, bool read_only, Error **errp)
     }
 
     bs->read_only = read_only;
+
+    if (read_only) {
+        bs->open_flags &= ~BDRV_O_RDWR;
+    } else {
+        bs->open_flags |= BDRV_O_RDWR;
+    }
+
     return 0;
 }
 
-- 
2.19.1

[Qemu-devel] [PATCH v3 2/9] block: Add auto-read-only option

[Kevin Wolf]

If a management application builds the block graph node by node, the
protocol layer doesn't inherit its read-only option from the format
layer any more, so it must be set explicitly.

Backing files should work on read-only storage, but at the same time, a
block job like commit should be able to reopen them read-write if they
are on read-write storage. However, without option inheritance, reopen
only changes the read-only option for the root node (typically the
format layer), but not the protocol layer, so reopening fails (the
format layer wants to get write permissions, but the protocol layer is
still read-only).

A simple workaround for the problem in the management tool would be to
open the protocol layer always read-write and to make only the format
layer read-only for backing files. However, sometimes the file is
actually stored on read-only storage and we don't know whether the image
can be opened read-write (for example, for NBD it depends on the server
we're trying to connect to). This adds an option that makes QEMU try to
open the image read-write, but allows it to degrade to a read-only mode
without returning an error.

The documentation for this option is consciously phrased in a way that
allows QEMU to switch to a better model eventually: Instead of trying
when the image is first opened, making the read-only flag dynamic and
changing it automatically whenever the first BLK_PERM_WRITE user is
attached or the last one is detached would be much more useful
behaviour.

Unfortunately, this more useful behaviour is also a lot harder to
implement, and libvirt needs a solution now before it can switch to
-blockdev, so let's start with this easier approach for now.

Instead of adding a new auto-read-only option, turning the existing
read-only into an enum (with a bool alternate for compatibility) was
considered, but it complicated the implementation to the point that it
didn't seem to be worth it.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 qapi/block-core.json  |  7 +++++++
 include/block/block.h |  2 ++
 block.c               | 17 +++++++++++++++++
 block/vvfat.c         |  1 +
 4 files changed, 27 insertions(+)

diff --git a/qapi/block-core.json b/qapi/block-core.json
index cfb37f8c1d..2ac4e69e68 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -3651,6 +3651,12 @@
 #                 either generally or in certain configurations. In this case,
 #                 the default value does not work and the option must be
 #                 specified explicitly.
+# @auto-read-only: if true and @read-only is false, QEMU may automatically
+#                  decide not to open the image read-write as requested, but
+#                  fall back to read-only instead (and switch between the modes
+#                  later), e.g. depending on whether the image file is writable
+#                  or whether a writing user is attached to the node
+#                  (default: false, since 3.1)
 # @detect-zeroes: detect and optimize zero writes (Since 2.1)
 #                 (default: off)
 # @force-share:   force share all permission on added nodes.
@@ -3666,6 +3672,7 @@
             '*discard': 'BlockdevDiscardOptions',
             '*cache': 'BlockdevCacheOptions',
             '*read-only': 'bool',
+            '*auto-read-only': 'bool',
             '*force-share': 'bool',
             '*detect-zeroes': 'BlockdevDetectZeroesOptions' },
   'discriminator': 'driver',
diff --git a/include/block/block.h b/include/block/block.h
index b189cf422e..580b3716c3 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -115,6 +115,7 @@ typedef struct HDGeometry {
                                       select an appropriate protocol driver,
                                       ignoring the format layer */
 #define BDRV_O_NO_IO       0x10000 /* don't initialize for I/O */
+#define BDRV_O_AUTO_RDONLY 0x20000 /* degrade to read-only if opening read-write fails */
 
 #define BDRV_O_CACHE_MASK  (BDRV_O_NOCACHE | BDRV_O_NO_FLUSH)
 
@@ -125,6 +126,7 @@ typedef struct HDGeometry {
 #define BDRV_OPT_CACHE_DIRECT   "cache.direct"
 #define BDRV_OPT_CACHE_NO_FLUSH "cache.no-flush"
 #define BDRV_OPT_READ_ONLY      "read-only"
+#define BDRV_OPT_AUTO_READ_ONLY "auto-read-only"
 #define BDRV_OPT_DISCARD        "discard"
 #define BDRV_OPT_FORCE_SHARE    "force-share"
 
diff --git a/block.c b/block.c
index d7bd6d29b4..ed73522cc6 100644
--- a/block.c
+++ b/block.c
@@ -930,6 +930,7 @@ static void bdrv_inherited_options(int *child_flags, QDict *child_options,
 
     /* Inherit the read-only option from the parent if it's not set */
     qdict_copy_default(child_options, parent_options, BDRV_OPT_READ_ONLY);
+    qdict_copy_default(child_options, parent_options, BDRV_OPT_AUTO_READ_ONLY);
 
     /* Our block drivers take care to send flushes and respect unmap policy,
      * so we can default to enable both on lower layers regardless of the
@@ -1053,6 +1054,7 @@ static void bdrv_backing_options(int *child_flags, QDict *child_options,
 
     /* backing files always opened read-only */
     qdict_set_default_str(child_options, BDRV_OPT_READ_ONLY, "on");
+    qdict_set_default_str(child_options, BDRV_OPT_AUTO_READ_ONLY, "off");
     flags &= ~BDRV_O_COPY_ON_READ;
 
     /* snapshot=on is handled on the top layer */
@@ -1142,6 +1144,10 @@ static void update_flags_from_options(int *flags, QemuOpts *opts)
         *flags |= BDRV_O_RDWR;
     }
 
+    assert(qemu_opt_find(opts, BDRV_OPT_AUTO_READ_ONLY));
+    if (qemu_opt_get_bool_del(opts, BDRV_OPT_AUTO_READ_ONLY, false)) {
+        *flags |= BDRV_O_AUTO_RDONLY;
+    }
 }
 
 static void update_options_from_flags(QDict *options, int flags)
@@ -1156,6 +1162,10 @@ static void update_options_from_flags(QDict *options, int flags)
     if (!qdict_haskey(options, BDRV_OPT_READ_ONLY)) {
         qdict_put_bool(options, BDRV_OPT_READ_ONLY, !(flags & BDRV_O_RDWR));
     }
+    if (!qdict_haskey(options, BDRV_OPT_AUTO_READ_ONLY)) {
+        qdict_put_bool(options, BDRV_OPT_AUTO_READ_ONLY,
+                       flags & BDRV_O_AUTO_RDONLY);
+    }
 }
 
 static void bdrv_assign_node_name(BlockDriverState *bs,
@@ -1328,6 +1338,11 @@ QemuOptsList bdrv_runtime_opts = {
             .type = QEMU_OPT_BOOL,
             .help = "Node is opened in read-only mode",
         },
+        {
+            .name = BDRV_OPT_AUTO_READ_ONLY,
+            .type = QEMU_OPT_BOOL,
+            .help = "Node can become read-only if opening read-write fails",
+        },
         {
             .name = "detect-zeroes",
             .type = QEMU_OPT_STRING,
@@ -2486,6 +2501,8 @@ BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp)
         qdict_set_default_str(qdict, BDRV_OPT_CACHE_DIRECT, "off");
         qdict_set_default_str(qdict, BDRV_OPT_CACHE_NO_FLUSH, "off");
         qdict_set_default_str(qdict, BDRV_OPT_READ_ONLY, "off");
+        qdict_set_default_str(qdict, BDRV_OPT_AUTO_READ_ONLY, "off");
+
     }
 
     bs = bdrv_open_inherit(NULL, reference, qdict, 0, NULL, NULL, errp);
diff --git a/block/vvfat.c b/block/vvfat.c
index f2e7d501cf..98ba5e2bac 100644
--- a/block/vvfat.c
+++ b/block/vvfat.c
@@ -3130,6 +3130,7 @@ static void vvfat_qcow_options(int *child_flags, QDict *child_options,
                                int parent_flags, QDict *parent_options)
 {
     qdict_set_default_str(child_options, BDRV_OPT_READ_ONLY, "off");
+    qdict_set_default_str(child_options, BDRV_OPT_AUTO_READ_ONLY, "off");
     qdict_set_default_str(child_options, BDRV_OPT_CACHE_NO_FLUSH, "on");
 }
 
-- 
2.19.1

Re: [Qemu-devel] [PATCH v3 2/9] block: Add auto-read-only option

[Eric Blake]

On 10/17/18 11:41 AM, Kevin Wolf wrote:
> If a management application builds the block graph node by node, the
> protocol layer doesn't inherit its read-only option from the format
> layer any more, so it must be set explicitly.
> 

> The documentation for this option is consciously phrased in a way that
> allows QEMU to switch to a better model eventually: Instead of trying
> when the image is first opened, making the read-only flag dynamic and
> changing it automatically whenever the first BLK_PERM_WRITE user is
> attached or the last one is detached would be much more useful
> behaviour.
> 

> +++ b/qapi/block-core.json
> @@ -3651,6 +3651,12 @@
>   #                 either generally or in certain configurations. In this case,
>   #                 the default value does not work and the option must be
>   #                 specified explicitly.
> +# @auto-read-only: if true and @read-only is false, QEMU may automatically
> +#                  decide not to open the image read-write as requested, but
> +#                  fall back to read-only instead (and switch between the modes
> +#                  later), e.g. depending on whether the image file is writable
> +#                  or whether a writing user is attached to the node
> +#                  (default: false, since 3.1)

Definitely better wording compared to v2.

Reviewed-by: Eric Blake <eblake@redhat.com>

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

[Qemu-devel] [PATCH v3 3/9] block: Require auto-read-only for existing fallbacks

[Kevin Wolf]

Some block drivers have traditionally changed their node to read-only
mode without asking the user. This behaviour has been marked deprecated
since 2.11, expecting users to provide an explicit read-only=on option.

Now that we have auto-read-only=on, enable these drivers to make use of
the option.

This is the only use of bdrv_set_read_only(), so we can make it a bit
more specific and turn it into a bdrv_apply_auto_read_only() that is
more convenient for drivers to use.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 include/block/block.h |  3 ++-
 block.c               | 42 +++++++++++++++++++++++++++---------------
 block/bochs.c         | 17 ++++++-----------
 block/cloop.c         | 16 +++++-----------
 block/dmg.c           | 16 +++++-----------
 block/rbd.c           | 14 ++++----------
 block/vvfat.c         | 10 ++--------
 7 files changed, 51 insertions(+), 67 deletions(-)

diff --git a/include/block/block.h b/include/block/block.h
index 580b3716c3..7f5453b45b 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -438,7 +438,8 @@ int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base,
 bool bdrv_is_read_only(BlockDriverState *bs);
 int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only,
                            bool ignore_allow_rdw, Error **errp);
-int bdrv_set_read_only(BlockDriverState *bs, bool read_only, Error **errp);
+int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg,
+                              Error **errp);
 bool bdrv_is_writable(BlockDriverState *bs);
 bool bdrv_is_sg(BlockDriverState *bs);
 bool bdrv_is_inserted(BlockDriverState *bs);
diff --git a/block.c b/block.c
index ed73522cc6..a97e324f39 100644
--- a/block.c
+++ b/block.c
@@ -266,29 +266,41 @@ int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only,
     return 0;
 }
 
-/* TODO Remove (deprecated since 2.11)
- * Block drivers are not supposed to automatically change bs->read_only.
- * Instead, they should just check whether they can provide what the user
- * explicitly requested and error out if read-write is requested, but they can
- * only provide read-only access. */
-int bdrv_set_read_only(BlockDriverState *bs, bool read_only, Error **errp)
+/*
+ * Called by a driver that can only provide a read-only image.
+ *
+ * Returns 0 if the node is already read-only or it could switch the node to
+ * read-only because BDRV_O_AUTO_RDONLY is set.
+ *
+ * Returns -EACCES if the node is read-write and BDRV_O_AUTO_RDONLY is not set
+ * or bdrv_can_set_read_only() forbids making the node read-only. If @errmsg
+ * is not NULL, it is used as the error message for the Error object.
+ */
+int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg,
+                              Error **errp)
 {
     int ret = 0;
 
-    ret = bdrv_can_set_read_only(bs, read_only, false, errp);
-    if (ret < 0) {
-        return ret;
+    if (!(bs->open_flags & BDRV_O_RDWR)) {
+        return 0;
+    }
+    if (!(bs->open_flags & BDRV_O_AUTO_RDONLY)) {
+        goto fail;
     }
 
-    bs->read_only = read_only;
-
-    if (read_only) {
-        bs->open_flags &= ~BDRV_O_RDWR;
-    } else {
-        bs->open_flags |= BDRV_O_RDWR;
+    ret = bdrv_can_set_read_only(bs, true, false, NULL);
+    if (ret < 0) {
+        goto fail;
     }
 
+    bs->read_only = true;
+    bs->open_flags &= ~BDRV_O_RDWR;
+
     return 0;
+
+fail:
+    error_setg(errp, "%s", errmsg ?: "Image is read-only");
+    return -EACCES;
 }
 
 void bdrv_get_full_backing_filename_from_filename(const char *backed,
diff --git a/block/bochs.c b/block/bochs.c
index 50c630047b..22e7d44211 100644
--- a/block/bochs.c
+++ b/block/bochs.c
@@ -105,23 +105,18 @@ static int bochs_open(BlockDriverState *bs, QDict *options, int flags,
     struct bochs_header bochs;
     int ret;
 
+    /* No write support yet */
+    ret = bdrv_apply_auto_read_only(bs, NULL, errp);
+    if (ret < 0) {
+        return ret;
+    }
+
     bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file,
                                false, errp);
     if (!bs->file) {
         return -EINVAL;
     }
 
-    if (!bdrv_is_read_only(bs)) {
-        error_report("Opening bochs images without an explicit read-only=on "
-                     "option is deprecated. Future versions will refuse to "
-                     "open the image instead of automatically marking the "
-                     "image read-only.");
-        ret = bdrv_set_read_only(bs, true, errp); /* no write support yet */
-        if (ret < 0) {
-            return ret;
-        }
-    }
-
     ret = bdrv_pread(bs->file, 0, &bochs, sizeof(bochs));
     if (ret < 0) {
         return ret;
diff --git a/block/cloop.c b/block/cloop.c
index 2be68987bd..df2b85f723 100644
--- a/block/cloop.c
+++ b/block/cloop.c
@@ -67,23 +67,17 @@ static int cloop_open(BlockDriverState *bs, QDict *options, int flags,
     uint32_t offsets_size, max_compressed_block_size = 1, i;
     int ret;
 
+    ret = bdrv_apply_auto_read_only(bs, NULL, errp);
+    if (ret < 0) {
+        return ret;
+    }
+
     bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file,
                                false, errp);
     if (!bs->file) {
         return -EINVAL;
     }
 
-    if (!bdrv_is_read_only(bs)) {
-        error_report("Opening cloop images without an explicit read-only=on "
-                     "option is deprecated. Future versions will refuse to "
-                     "open the image instead of automatically marking the "
-                     "image read-only.");
-        ret = bdrv_set_read_only(bs, true, errp);
-        if (ret < 0) {
-            return ret;
-        }
-    }
-
     /* read header */
     ret = bdrv_pread(bs->file, 128, &s->block_size, 4);
     if (ret < 0) {
diff --git a/block/dmg.c b/block/dmg.c
index c9b3c519c4..1d9283ba2f 100644
--- a/block/dmg.c
+++ b/block/dmg.c
@@ -413,23 +413,17 @@ static int dmg_open(BlockDriverState *bs, QDict *options, int flags,
     int64_t offset;
     int ret;
 
+    ret = bdrv_apply_auto_read_only(bs, NULL, errp);
+    if (ret < 0) {
+        return ret;
+    }
+
     bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file,
                                false, errp);
     if (!bs->file) {
         return -EINVAL;
     }
 
-    if (!bdrv_is_read_only(bs)) {
-        error_report("Opening dmg images without an explicit read-only=on "
-                     "option is deprecated. Future versions will refuse to "
-                     "open the image instead of automatically marking the "
-                     "image read-only.");
-        ret = bdrv_set_read_only(bs, true, errp);
-        if (ret < 0) {
-            return ret;
-        }
-    }
-
     block_module_load_one("dmg-bz2");
 
     s->n_chunks = 0;
diff --git a/block/rbd.c b/block/rbd.c
index 014c68d629..ee0b4a6941 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -780,16 +780,10 @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags,
     /* If we are using an rbd snapshot, we must be r/o, otherwise
      * leave as-is */
     if (s->snap != NULL) {
-        if (!bdrv_is_read_only(bs)) {
-            error_report("Opening rbd snapshots without an explicit "
-                         "read-only=on option is deprecated. Future versions "
-                         "will refuse to open the image instead of "
-                         "automatically marking the image read-only.");
-            r = bdrv_set_read_only(bs, true, &local_err);
-            if (r < 0) {
-                error_propagate(errp, local_err);
-                goto failed_open;
-            }
+        r = bdrv_apply_auto_read_only(bs, "rbd snapshots are read-only", errp);
+        if (r < 0) {
+            rbd_close(s->image);
+            goto failed_open;
         }
     }
 
diff --git a/block/vvfat.c b/block/vvfat.c
index 98ba5e2bac..e4df255d58 100644
--- a/block/vvfat.c
+++ b/block/vvfat.c
@@ -1262,15 +1262,9 @@ static int vvfat_open(BlockDriverState *bs, QDict *options, int flags,
                        "Unable to set VVFAT to 'rw' when drive is read-only");
             goto fail;
         }
-    } else  if (!bdrv_is_read_only(bs)) {
-        error_report("Opening non-rw vvfat images without an explicit "
-                     "read-only=on option is deprecated. Future versions "
-                     "will refuse to open the image instead of "
-                     "automatically marking the image read-only.");
-        /* read only is the default for safety */
-        ret = bdrv_set_read_only(bs, true, &local_err);
+    } else {
+        ret = bdrv_apply_auto_read_only(bs, NULL, errp);
         if (ret < 0) {
-            error_propagate(errp, local_err);
             goto fail;
         }
     }
-- 
2.19.1

Re: [Qemu-devel] [PATCH v3 3/9] block: Require auto-read-only for existing fallbacks

[Eric Blake]

On 10/17/18 11:41 AM, Kevin Wolf wrote:
> Some block drivers have traditionally changed their node to read-only
> mode without asking the user. This behaviour has been marked deprecated
> since 2.11, expecting users to provide an explicit read-only=on option.
> 
> Now that we have auto-read-only=on, enable these drivers to make use of
> the option.
> 
> This is the only use of bdrv_set_read_only(), so we can make it a bit
> more specific and turn it into a bdrv_apply_auto_read_only() that is
> more convenient for drivers to use.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---

> +/*
> + * Called by a driver that can only provide a read-only image.
> + *
> + * Returns 0 if the node is already read-only or it could switch the node to
> + * read-only because BDRV_O_AUTO_RDONLY is set.
> + *
> + * Returns -EACCES if the node is read-write and BDRV_O_AUTO_RDONLY is not set
> + * or bdrv_can_set_read_only() forbids making the node read-only. If @errmsg
> + * is not NULL, it is used as the error message for the Error object.

Works for me.

> +++ b/block/rbd.c
> @@ -780,16 +780,10 @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags,
>       /* If we are using an rbd snapshot, we must be r/o, otherwise
>        * leave as-is */
>       if (s->snap != NULL) {
> -        if (!bdrv_is_read_only(bs)) {
> -            error_report("Opening rbd snapshots without an explicit "
> -                         "read-only=on option is deprecated. Future versions "
> -                         "will refuse to open the image instead of "
> -                         "automatically marking the image read-only.");
> -            r = bdrv_set_read_only(bs, true, &local_err);
> -            if (r < 0) {
> -                error_propagate(errp, local_err);
> -                goto failed_open;
> -            }
> +        r = bdrv_apply_auto_read_only(bs, "rbd snapshots are read-only", errp);
> +        if (r < 0) {
> +            rbd_close(s->image);
> +            goto failed_open;

That rbd_close() is an independent bugfix.  Should probably be split to 
a separate commit, or at a minimum called out in the commit message as 
intentional.

Actually, is it really needed to prevent a leak, or does the existing 
rados_shutdown() in failed_open already implicitly cover the actions of 
rbd_close()?

With the RBD issue touched up,
Reviewed-by: Eric Blake <eblake@redhat.com>

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

Re: [Qemu-devel] [PATCH v3 3/9] block: Require auto-read-only for existing fallbacks

[Kevin Wolf]

Am 17.10.2018 um 20:53 hat Eric Blake geschrieben:
> On 10/17/18 11:41 AM, Kevin Wolf wrote:
> > Some block drivers have traditionally changed their node to read-only
> > mode without asking the user. This behaviour has been marked deprecated
> > since 2.11, expecting users to provide an explicit read-only=on option.
> > 
> > Now that we have auto-read-only=on, enable these drivers to make use of
> > the option.
> > 
> > This is the only use of bdrv_set_read_only(), so we can make it a bit
> > more specific and turn it into a bdrv_apply_auto_read_only() that is
> > more convenient for drivers to use.
> > 
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>

> > +++ b/block/rbd.c
> > @@ -780,16 +780,10 @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags,
> >       /* If we are using an rbd snapshot, we must be r/o, otherwise
> >        * leave as-is */
> >       if (s->snap != NULL) {
> > -        if (!bdrv_is_read_only(bs)) {
> > -            error_report("Opening rbd snapshots without an explicit "
> > -                         "read-only=on option is deprecated. Future versions "
> > -                         "will refuse to open the image instead of "
> > -                         "automatically marking the image read-only.");
> > -            r = bdrv_set_read_only(bs, true, &local_err);
> > -            if (r < 0) {
> > -                error_propagate(errp, local_err);
> > -                goto failed_open;
> > -            }
> > +        r = bdrv_apply_auto_read_only(bs, "rbd snapshots are read-only", errp);
> > +        if (r < 0) {
> > +            rbd_close(s->image);
> > +            goto failed_open;
> 
> That rbd_close() is an independent bugfix.  Should probably be split to a
> separate commit, or at a minimum called out in the commit message as
> intentional.

Okay, I'll split it.

> Actually, is it really needed to prevent a leak, or does the existing
> rados_shutdown() in failed_open already implicitly cover the actions of
> rbd_close()?

I don't know if rados_shutdown() would implicitly do that as well, but
the normal .bdrv_close implementation calls both, too, so I suppose the
safe option is to let the error path do the same.

Kevin

[Qemu-devel] [PATCH v3 4/9] nbd: Support auto-read-only option

[Kevin Wolf]

If read-only=off, but auto-read-only=on is given, open a read-write NBD
connection if the server provides a read-write export, but instead of
erroring out for read-only exports, just degrade to read-only.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
---
 block/nbd-client.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/block/nbd-client.c b/block/nbd-client.c
index 9686ecbd5e..76e9ca3abe 100644
--- a/block/nbd-client.c
+++ b/block/nbd-client.c
@@ -992,11 +992,11 @@ int nbd_client_init(BlockDriverState *bs,
         logout("Failed to negotiate with the NBD server\n");
         return ret;
     }
-    if (client->info.flags & NBD_FLAG_READ_ONLY &&
-        !bdrv_is_read_only(bs)) {
-        error_setg(errp,
-                   "request for write access conflicts with read-only export");
-        return -EACCES;
+    if (client->info.flags & NBD_FLAG_READ_ONLY) {
+        ret = bdrv_apply_auto_read_only(bs, "NBD export is read-only", errp);
+        if (ret < 0) {
+            return ret;
+        }
     }
     if (client->info.flags & NBD_FLAG_SEND_FUA) {
         bs->supported_write_flags = BDRV_REQ_FUA;
-- 
2.19.1

[Qemu-devel] [PATCH v3 5/9] file-posix: Support auto-read-only option

[Kevin Wolf]

If read-only=off, but auto-read-only=on is given, open the file
read-write if we have the permissions, but instead of erroring out for
read-only files, just degrade to read-only.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/file-posix.c | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/block/file-posix.c b/block/file-posix.c
index 2da3a76355..0c1b81ce4b 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -527,9 +527,22 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
 
     s->fd = -1;
     fd = qemu_open(filename, s->open_flags, 0644);
-    if (fd < 0) {
-        ret = -errno;
-        error_setg_errno(errp, errno, "Could not open '%s'", filename);
+    ret = fd < 0 ? -errno : 0;
+
+    if (ret == -EACCES || ret == -EROFS) {
+        /* Try to degrade to read-only, but if it doesn't work, still use the
+         * normal error message. */
+        if (bdrv_apply_auto_read_only(bs, NULL, NULL) == 0) {
+            bdrv_flags &= ~BDRV_O_RDWR;
+            raw_parse_flags(bdrv_flags, &s->open_flags);
+            assert(!(s->open_flags & O_CREAT));
+            fd = qemu_open(filename, s->open_flags);
+            ret = fd < 0 ? -errno : 0;
+        }
+    }
+
+    if (ret < 0) {
+        error_setg_errno(errp, -ret, "Could not open '%s'", filename);
         if (ret == -EROFS) {
             ret = -EACCES;
         }
-- 
2.19.1

Re: [Qemu-devel] [PATCH v3 5/9] file-posix: Support auto-read-only option

[Eric Blake]

On 10/17/18 11:41 AM, Kevin Wolf wrote:
> If read-only=off, but auto-read-only=on is given, open the file
> read-write if we have the permissions, but instead of erroring out for
> read-only files, just degrade to read-only.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   block/file-posix.c | 19 ++++++++++++++++---
>   1 file changed, 16 insertions(+), 3 deletions(-)
> 

Reviewed-by: Eric Blake <eblake@redhat.com>

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

[Qemu-devel] [PATCH v3 6/9] curl: Support auto-read-only option

[Kevin Wolf]

If read-only=off, but auto-read-only=on is given, just degrade to
read-only.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
---
 block/curl.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/block/curl.c b/block/curl.c
index fabb2b4da7..db5d2bd8ef 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -684,10 +684,10 @@ static int curl_open(BlockDriverState *bs, QDict *options, int flags,
     const char *protocol_delimiter;
     int ret;
 
-
-    if (flags & BDRV_O_RDWR) {
-        error_setg(errp, "curl block device does not support writes");
-        return -EROFS;
+    ret = bdrv_apply_auto_read_only(bs, "curl driver does not support writes",
+                                    errp);
+    if (ret < 0) {
+        return ret;
     }
 
     if (!libcurl_initialized) {
-- 
2.19.1

[Qemu-devel] [PATCH v3 7/9] gluster: Support auto-read-only option

[Kevin Wolf]

If read-only=off, but auto-read-only=on is given, open the file
read-write if we have the permissions, but instead of erroring out for
read-only files, just degrade to read-only.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/gluster.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/block/gluster.c b/block/gluster.c
index 4fd55a9cc5..5e300c96c8 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -849,8 +849,16 @@ static int qemu_gluster_open(BlockDriverState *bs,  QDict *options,
     qemu_gluster_parse_flags(bdrv_flags, &open_flags);
 
     s->fd = glfs_open(s->glfs, gconf->path, open_flags);
-    if (!s->fd) {
-        ret = -errno;
+    ret = s->fd ? 0 : -errno;
+
+    if (ret == -EACCES || ret == -EROFS) {
+        /* Try to degrade to read-only, but if it doesn't work, still use the
+         * normal error message. */
+        if (bdrv_apply_auto_read_only(bs, NULL, NULL) == 0) {
+            open_flags = (open_flags & ~O_RDWR) | O_RDONLY;
+            s->fd = glfs_open(s->glfs, gconf->path, open_flags);
+            ret = s->fd ? 0 : -errno;
+        }
     }
 
     s->supports_seek_data = qemu_gluster_test_seek(s->fd);
-- 
2.19.1

Re: [Qemu-devel] [Qemu-block] [PATCH v3 7/9] gluster: Support auto-read-only option

[Niels de Vos]

On Wed, Oct 17, 2018 at 06:41:58PM +0200, Kevin Wolf wrote:
> If read-only=off, but auto-read-only=on is given, open the file
> read-write if we have the permissions, but instead of erroring out for
> read-only files, just degrade to read-only.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>  block/gluster.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/block/gluster.c b/block/gluster.c
> index 4fd55a9cc5..5e300c96c8 100644
> --- a/block/gluster.c
> +++ b/block/gluster.c
> @@ -849,8 +849,16 @@ static int qemu_gluster_open(BlockDriverState *bs,  QDict *options,
>      qemu_gluster_parse_flags(bdrv_flags, &open_flags);
>  
>      s->fd = glfs_open(s->glfs, gconf->path, open_flags);
> -    if (!s->fd) {
> -        ret = -errno;
> +    ret = s->fd ? 0 : -errno;
> +
> +    if (ret == -EACCES || ret == -EROFS) {
> +        /* Try to degrade to read-only, but if it doesn't work, still use the
> +         * normal error message. */
> +        if (bdrv_apply_auto_read_only(bs, NULL, NULL) == 0) {
> +            open_flags = (open_flags & ~O_RDWR) | O_RDONLY;
> +            s->fd = glfs_open(s->glfs, gconf->path, open_flags);
> +            ret = s->fd ? 0 : -errno;
> +        }
>      }
>  
>      s->supports_seek_data = qemu_gluster_test_seek(s->fd);
> -- 

Looks good to me, thanks.

Reviewed-by: Niels de Vos <ndevos@redhat.com>

[Qemu-devel] [PATCH v3 8/9] iscsi: Support auto-read-only option

[Kevin Wolf]

If read-only=off, but auto-read-only=on is given, open the volume
read-write if we have the permissions, but instead of erroring out for
read-only volumes, just degrade to read-only.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
---
 block/iscsi.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/block/iscsi.c b/block/iscsi.c
index bb69faf34a..130ae26f5f 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -1878,9 +1878,11 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,
     /* Check the write protect flag of the LUN if we want to write */
     if (iscsilun->type == TYPE_DISK && (flags & BDRV_O_RDWR) &&
         iscsilun->write_protected) {
-        error_setg(errp, "Cannot open a write protected LUN as read-write");
-        ret = -EACCES;
-        goto out;
+        ret = bdrv_apply_auto_read_only(bs, "LUN is write protected", errp);
+        if (ret < 0) {
+            goto out;
+        }
+        flags &= ~BDRV_O_RDWR;
     }
 
     iscsi_readcapacity_sync(iscsilun, &local_err);
-- 
2.19.1

[Qemu-devel] [PATCH v3 9/9] block: Make auto-read-only=on default for -drive

[Kevin Wolf]

While we want machine interfaces like -blockdev and QMP blockdev-add to
add as little auto-detection as possible so that management tools are
explicit about their needs, -drive is a convenience option for human
users. Enabling auto-read-only=on by default there enables users to use
read-only images for read-only guest devices without having to specify
read-only=on explicitly. If they try to attach the image to a read-write
device, they will still get an error message.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 blockdev.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/blockdev.c b/blockdev.c
index a8755bd908..ab57a8a53a 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -590,6 +590,7 @@ static BlockBackend *blockdev_init(const char *file, QDict *bs_opts,
         qdict_set_default_str(bs_opts, BDRV_OPT_CACHE_NO_FLUSH, "off");
         qdict_set_default_str(bs_opts, BDRV_OPT_READ_ONLY,
                               read_only ? "on" : "off");
+        qdict_set_default_str(bs_opts, BDRV_OPT_AUTO_READ_ONLY, "on");
         assert((bdrv_flags & BDRV_O_CACHE_MASK) == 0);
 
         if (runstate_check(RUN_STATE_INMIGRATE)) {
-- 
2.19.1

Re: [Qemu-devel] [PATCH v3 9/9] block: Make auto-read-only=on default for -drive

[Eric Blake]

On 10/17/18 11:42 AM, Kevin Wolf wrote:
> While we want machine interfaces like -blockdev and QMP blockdev-add to
> add as little auto-detection as possible so that management tools are
> explicit about their needs, -drive is a convenience option for human
> users. Enabling auto-read-only=on by default there enables users to use
> read-only images for read-only guest devices without having to specify
> read-only=on explicitly. If they try to attach the image to a read-write
> device, they will still get an error message.
> 
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   blockdev.c | 1 +
>   1 file changed, 1 insertion(+)
> 

Reviewed-by: Eric Blake <eblake@redhat.com>

However, the obligatory: "iotests coverage?"

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org