All of lore.kernel.org
 help / color / mirror / Atom feed
From: Max Reitz <mreitz@redhat.com>
To: qemu-block@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>,
	qemu-devel@nongnu.org, Max Reitz <mreitz@redhat.com>
Subject: [PATCH for-5.0 15/31] block: Split bdrv_default_perms_for_storage()
Date: Wed, 27 Nov 2019 14:16:08 +0100	[thread overview]
Message-ID: <20191127131624.1062403-16-mreitz@redhat.com> (raw)
In-Reply-To: <20191127131624.1062403-1-mreitz@redhat.com>

We can be less restrictive about pure data children than those with
metadata on them.

For bdrv_format_default_perms(), we keep the safe option of
bdrv_default_perms_for_metadata() (until we drop
bdrv_format_default_perms() altogether).

That means that bdrv_default_perms_for_data() is unused at this point.
We will use it for all children that have the DATA role, but not the
METADATA role.  So far, no child has any role, so we do not use it, but
that will change.

Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block.c | 53 +++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 43 insertions(+), 10 deletions(-)

diff --git a/block.c b/block.c
index 4d4ccbacdf..33abd7f64e 100644
--- a/block.c
+++ b/block.c
@@ -2294,18 +2294,17 @@ static void bdrv_default_perms_for_backing(BlockDriverState *bs, BdrvChild *c,
     *nshared = shared;
 }
 
-static void bdrv_default_perms_for_storage(BlockDriverState *bs, BdrvChild *c,
-                                           const BdrvChildClass *child_class,
-                                           BdrvChildRole role,
-                                           BlockReopenQueue *reopen_queue,
-                                           uint64_t perm, uint64_t shared,
-                                           uint64_t *nperm, uint64_t *nshared)
+static void bdrv_default_perms_for_metadata(BlockDriverState *bs, BdrvChild *c,
+                                            const BdrvChildClass *child_class,
+                                            BdrvChildRole role,
+                                            BlockReopenQueue *reopen_queue,
+                                            uint64_t perm, uint64_t shared,
+                                            uint64_t *nperm, uint64_t *nshared)
 {
     int flags;
 
     assert(child_class == &child_file ||
-           (child_class == &child_of_bds &&
-            (role & (BDRV_CHILD_METADATA | BDRV_CHILD_DATA))));
+           (child_class == &child_of_bds && (role & BDRV_CHILD_METADATA)));
 
     flags = bdrv_reopen_get_flags(reopen_queue, bs);
 
@@ -2338,6 +2337,40 @@ static void bdrv_default_perms_for_storage(BlockDriverState *bs, BdrvChild *c,
     *nshared = shared;
 }
 
+/* TODO: Use */
+static void __attribute__((unused))
+bdrv_default_perms_for_data(BlockDriverState *bs, BdrvChild *c,
+                            const BdrvChildClass *child_class,
+                            BdrvChildRole role,
+                            BlockReopenQueue *reopen_queue,
+                            uint64_t perm, uint64_t shared,
+                            uint64_t *nperm, uint64_t *nshared)
+{
+    assert(child_class == &child_of_bds && (role & BDRV_CHILD_DATA));
+
+    /*
+     * Apart from the modifications below, the same permissions are
+     * forwarded and left alone as for filters
+     */
+    bdrv_filter_default_perms(bs, c, child_class, role, reopen_queue,
+                              perm, shared, &perm, &shared);
+
+    /*
+     * We cannot allow other users to resize the file because the
+     * format driver might have some assumptions about the size
+     * (e.g. because it is stored in metadata, or because the file is
+     * split into fixed-size data files).
+     */
+    shared &= ~BLK_PERM_RESIZE;
+
+    if (bs->open_flags & BDRV_O_INACTIVE) {
+        shared |= BLK_PERM_WRITE | BLK_PERM_RESIZE;
+    }
+
+    *nperm = perm;
+    *nshared = shared;
+}
+
 void bdrv_format_default_perms(BlockDriverState *bs, BdrvChild *c,
                                const BdrvChildClass *child_class,
                                BdrvChildRole role,
@@ -2349,8 +2382,8 @@ void bdrv_format_default_perms(BlockDriverState *bs, BdrvChild *c,
     assert(child_class == &child_backing || child_class == &child_file);
 
     if (!backing) {
-        bdrv_default_perms_for_storage(bs, c, child_class, role, reopen_queue,
-                                       perm, shared, nperm, nshared);
+        bdrv_default_perms_for_metadata(bs, c, child_class, role, reopen_queue,
+                                        perm, shared, nperm, nshared);
     } else {
         bdrv_default_perms_for_backing(bs, c, child_class, role, reopen_queue,
                                        perm, shared, nperm, nshared);
-- 
2.23.0



  parent reply	other threads:[~2019-11-27 13:28 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-27 13:15 [PATCH for-5.0 00/31] block: Introduce real BdrvChildRole Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 01/31] block: Rename BdrvChildRole to BdrvChildClass Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 02/31] block: Add BdrvChildRole Max Reitz
2019-11-28  9:31   ` Vladimir Sementsov-Ogievskiy
2019-11-28 10:52     ` Max Reitz
2019-11-28 11:24       ` Vladimir Sementsov-Ogievskiy
2019-11-28 14:00         ` Max Reitz
2019-11-28 14:12   ` Kevin Wolf
2019-11-28 16:36     ` Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 03/31] block: Add BdrvChildRole to BdrvChild Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 04/31] block: Pass BdrvChildRole to bdrv_child_perm() Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 05/31] block: Drop BdrvChildClass.stay_at_node Max Reitz
2019-11-27 13:15 ` [PATCH for-5.0 06/31] block: Keep BDRV_O_NO_IO in *inherited_fmt_options Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 07/31] block: Pass BdrvChildRole to .inherit_options() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 08/31] block: Unify bdrv_*inherited_options() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 09/31] block: Unify bdrv_child_cb_attach() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 10/31] block: Unify bdrv_child_cb_detach() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 11/31] block: Add child_of_bds Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 12/31] block: Distinguish paths in *_format_default_perms Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 13/31] block: Pull out bdrv_default_perms_for_backing() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 14/31] block: Pull out bdrv_default_perms_for_storage() Max Reitz
2019-11-27 13:16 ` Max Reitz [this message]
2019-11-27 13:16 ` [PATCH for-5.0 16/31] block: Add bdrv_default_perms() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 17/31] raw-format: Split raw_read_options() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 18/31] block: Switch child_format users to child_of_bds Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 19/31] block: Drop child_format Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 20/31] block: Make backing files child_of_bds children Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 21/31] block: Drop child_backing Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 22/31] block: Make format drivers use child_of_bds Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 23/31] block: Make filter " Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 24/31] block: Use child_of_bds in remaining places Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 25/31] tests: Use child_of_bds instead of child_file Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 26/31] block: Use bdrv_default_perms() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 27/31] block: Make bdrv_filter_default_perms() static Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 28/31] block: Drop bdrv_format_default_perms() Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 29/31] block: Drop child_file Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 30/31] block: Pass BdrvChildRole in remaining cases Max Reitz
2019-11-27 13:16 ` [PATCH for-5.0 31/31] block: Drop @child_class from bdrv_child_perm() 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=20191127131624.1062403-16-mreitz@redhat.com \
    --to=mreitz@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.