qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: John Snow <jsnow@redhat.com>
To: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>,
	qemu-block@nongnu.org, qemu-devel@nongnu.org
Cc: kwolf@redhat.com, famz@redhat.com, armbru@redhat.com,
	mreitz@redhat.com, stefanha@redhat.com, pbonzini@redhat.com,
	den@openvz.org
Subject: Re: [Qemu-devel] [PATCH v20 10/30] block/dirty-bitmap: add readonly field to BdrvDirtyBitmap
Date: Fri, 2 Jun 2017 17:02:47 -0400	[thread overview]
Message-ID: <ccd143a1-e6dd-a08c-5747-cd46bbbf0429@redhat.com> (raw)
In-Reply-To: <20170602112158.232757-11-vsementsov@virtuozzo.com>



On 06/02/2017 07:21 AM, Vladimir Sementsov-Ogievskiy wrote:
> It will be needed in following commits for persistent bitmaps.
> If bitmap is loaded from read-only storage (and we can't mark it
> "in use" in this storage) corresponding BdrvDirtyBitmap should be
> read-only.
> 
> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> ---
>  block/dirty-bitmap.c         | 32 ++++++++++++++++++++++++++++++++
>  block/io.c                   |  8 ++++++++
>  blockdev.c                   |  6 ++++++
>  include/block/dirty-bitmap.h |  4 ++++
>  4 files changed, 50 insertions(+)
> 
> diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c
> index f25428868c..1c9ffb292a 100644
> --- a/block/dirty-bitmap.c
> +++ b/block/dirty-bitmap.c
> @@ -45,6 +45,12 @@ struct BdrvDirtyBitmap {
>      bool disabled;              /* Bitmap is disabled. It skips all writes to
>                                     the device */
>      int active_iterators;       /* How many iterators are active */
> +    bool readonly;              /* Bitmap is read-only and may be changed only
> +                                   by deserialize* functions. This field blocks

In what way do the deserialize functions change the bitmaps, again?

> +                                   any changing operations on owning image
> +                                   (writes and discards), if bitmap is readonly
> +                                   such operations must fail and not change
> +                                   image or this bitmap */
>      QLIST_ENTRY(BdrvDirtyBitmap) list;
>  };
>  
> @@ -437,6 +443,7 @@ void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
>                             int64_t cur_sector, int64_t nr_sectors)
>  {
>      assert(bdrv_dirty_bitmap_enabled(bitmap));
> +    assert(!bdrv_dirty_bitmap_readonly(bitmap));

I still feel as if bdrv_dirty_bitmap_enabled() can return false if
bdrv_dirty_bitmap_readonly is true, and you wouldn't have to edit these
parts, but I don't care enough to press the issue.

>      hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
>  }
>  
> @@ -444,12 +451,14 @@ void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
>                               int64_t cur_sector, int64_t nr_sectors)
>  {
>      assert(bdrv_dirty_bitmap_enabled(bitmap));
> +    assert(!bdrv_dirty_bitmap_readonly(bitmap));
>      hbitmap_reset(bitmap->bitmap, cur_sector, nr_sectors);
>  }
>  
>  void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out)
>  {
>      assert(bdrv_dirty_bitmap_enabled(bitmap));
> +    assert(!bdrv_dirty_bitmap_readonly(bitmap));
>      if (!out) {
>          hbitmap_reset_all(bitmap->bitmap);
>      } else {
> @@ -520,6 +529,7 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
>          if (!bdrv_dirty_bitmap_enabled(bitmap)) {
>              continue;
>          }
> +        assert(!bdrv_dirty_bitmap_readonly(bitmap));

Highlighting the difference in strictness between "disabled" and "readonly."

>          hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
>      }
>  }
> @@ -541,3 +551,25 @@ int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap)
>  {
>      return hbitmap_count(bitmap->meta);
>  }
> +
> +bool bdrv_dirty_bitmap_readonly(const BdrvDirtyBitmap *bitmap)
> +{
> +    return bitmap->readonly;
> +}
> +
> +void bdrv_dirty_bitmap_set_readonly(BdrvDirtyBitmap *bitmap, bool value)
> +{
> +    bitmap->readonly = value;
> +}
> +
> +bool bdrv_has_readonly_bitmaps(BlockDriverState *bs)
> +{
> +    BdrvDirtyBitmap *bm;
> +    QLIST_FOREACH(bm, &bs->dirty_bitmaps, list) {
> +        if (bm->readonly) {
> +            return true;
> +        }
> +    }
> +
> +    return false;
> +}
> diff --git a/block/io.c b/block/io.c
> index fdd7485c22..0e28a1f595 100644
> --- a/block/io.c
> +++ b/block/io.c
> @@ -1349,6 +1349,10 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child,
>      uint64_t bytes_remaining = bytes;
>      int max_transfer;
>  
> +    if (bdrv_has_readonly_bitmaps(bs)) {
> +        return -EPERM;
> +    }
> +

Should this be a dynamic error, or an assertion? We should probably
endeavor to never actually hit this circumstance (we should not have
readonly bitmaps on a RW node.)

>      assert(is_power_of_2(align));
>      assert((offset & (align - 1)) == 0);
>      assert((bytes & (align - 1)) == 0);
> @@ -2437,6 +2441,10 @@ int coroutine_fn bdrv_co_pdiscard(BlockDriverState *bs, int64_t offset,
>          return -ENOMEDIUM;
>      }
>  
> +    if (bdrv_has_readonly_bitmaps(bs)) {
> +        return -EPERM;
> +    }
> +
>      ret = bdrv_check_byte_request(bs, offset, count);
>      if (ret < 0) {
>          return ret;
> diff --git a/blockdev.c b/blockdev.c
> index c63f4e82c7..2b397abf66 100644
> --- a/blockdev.c
> +++ b/blockdev.c
> @@ -2033,6 +2033,9 @@ static void block_dirty_bitmap_clear_prepare(BlkActionState *common,
>      } else if (!bdrv_dirty_bitmap_enabled(state->bitmap)) {
>          error_setg(errp, "Cannot clear a disabled bitmap");
>          return;
> +    } else if (bdrv_dirty_bitmap_readonly(state->bitmap)) {
> +        error_setg(errp, "Cannot clear a readonly bitmap");
> +        return;
>      }

Probably getting close to easier to specify what state we DO allow
bitmaps to be cleared in (enabled/active, not frozen, disabled or readonly.)

>  
>      bdrv_clear_dirty_bitmap(state->bitmap, &state->backup);
> @@ -2813,6 +2816,9 @@ void qmp_block_dirty_bitmap_clear(const char *node, const char *name,
>                     "Bitmap '%s' is currently disabled and cannot be cleared",
>                     name);
>          goto out;
> +    } else if (bdrv_dirty_bitmap_readonly(bitmap)) {
> +        error_setg(errp, "Bitmap '%s' is readonly and cannot be cleared", name);
> +        goto out;
>      }

Random thought, maybe qmp_block_dirty_bitmap_clear should utilize the
transactional action core to perform this action instead of
reimplementing it. This has nothing to do with you, though.

>  
>      bdrv_clear_dirty_bitmap(bitmap, NULL);
> diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h
> index 1e17729ac2..aa6d47ee00 100644
> --- a/include/block/dirty-bitmap.h
> +++ b/include/block/dirty-bitmap.h
> @@ -75,4 +75,8 @@ void bdrv_dirty_bitmap_deserialize_ones(BdrvDirtyBitmap *bitmap,
>                                          bool finish);
>  void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap);
>  
> +bool bdrv_dirty_bitmap_readonly(const BdrvDirtyBitmap *bitmap);
> +void bdrv_dirty_bitmap_set_readonly(BdrvDirtyBitmap *bitmap, bool value);
> +bool bdrv_has_readonly_bitmaps(BlockDriverState *bs);
> +
>  #endif
> 

Reviewed-by: John Snow <jsnow@redhat.com>

  reply	other threads:[~2017-06-02 21:03 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-06-02 11:21 [Qemu-devel] [PATCH v20 00/30] qcow2: persistent dirty bitmaps Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 01/30] specs/qcow2: fix bitmap granularity qemu-specific note Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 02/30] specs/qcow2: do not use wording 'bitmap header' Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 03/30] hbitmap: improve dirty iter Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 04/30] tests: add hbitmap iter test Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 05/30] block: fix bdrv_dirty_bitmap_granularity signature Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 06/30] block/dirty-bitmap: add deserialize_ones func Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 07/30] qcow2-refcount: rename inc_refcounts() and make it public Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 08/30] qcow2: add bitmaps extension Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 09/30] block/dirty-bitmap: fix comment for BlockDirtyBitmap.disabled field Vladimir Sementsov-Ogievskiy
2017-06-02 21:02   ` John Snow
2017-06-09 12:29     ` Max Reitz
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 10/30] block/dirty-bitmap: add readonly field to BdrvDirtyBitmap Vladimir Sementsov-Ogievskiy
2017-06-02 21:02   ` John Snow [this message]
2017-06-03 17:19     ` Sementsov-Ogievskiy Vladimir
2017-06-09 12:56   ` Max Reitz
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 11/30] qcow2: autoloading dirty bitmaps Vladimir Sementsov-Ogievskiy
2017-06-02 21:48   ` John Snow
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 12/30] block: refactor bdrv_reopen_commit Vladimir Sementsov-Ogievskiy
2017-06-02 21:57   ` John Snow
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 13/30] block: new bdrv_reopen_bitmaps_rw interface Vladimir Sementsov-Ogievskiy
2017-06-02 22:17   ` John Snow
2017-06-03 16:53     ` Sementsov-Ogievskiy Vladimir
2017-06-09 13:27   ` Max Reitz
2017-06-13 10:25     ` Vladimir Sementsov-Ogievskiy
2017-06-13 15:28       ` Max Reitz
2017-06-14  9:03         ` Vladimir Sementsov-Ogievskiy
2017-06-14 12:04           ` Max Reitz
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 14/30] qcow2: support .bdrv_reopen_bitmaps_rw Vladimir Sementsov-Ogievskiy
2017-06-02 23:09   ` John Snow
2017-06-09 13:38   ` Max Reitz
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 15/30] block/dirty-bitmap: add autoload field to BdrvDirtyBitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 16/30] block: bdrv_close: release bitmaps after drv->bdrv_close Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 17/30] block: introduce persistent dirty bitmaps Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 18/30] block/dirty-bitmap: add bdrv_dirty_bitmap_next() Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 19/30] qcow2: add persistent dirty bitmaps support Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 20/30] qcow2: store bitmaps on reopening image as read-only Vladimir Sementsov-Ogievskiy
2017-06-09 13:48   ` Max Reitz
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 21/30] block: add bdrv_can_store_new_dirty_bitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 22/30] qcow2: add .bdrv_can_store_new_dirty_bitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 23/30] qmp: add persistent flag to block-dirty-bitmap-add Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 24/30] qmp: add autoload parameter " Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 25/30] qmp: add x-debug-block-dirty-bitmap-sha256 Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 26/30] iotests: test qcow2 persistent dirty bitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 27/30] block/dirty-bitmap: add bdrv_remove_persistent_dirty_bitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 28/30] qcow2: add .bdrv_remove_persistent_dirty_bitmap Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 29/30] qmp: block-dirty-bitmap-remove: remove persistent Vladimir Sementsov-Ogievskiy
2017-06-02 11:21 ` [Qemu-devel] [PATCH v20 30/30] block: release persistent bitmaps on inactivate Vladimir Sementsov-Ogievskiy
2017-06-09 14:01   ` 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=ccd143a1-e6dd-a08c-5747-cd46bbbf0429@redhat.com \
    --to=jsnow@redhat.com \
    --cc=armbru@redhat.com \
    --cc=den@openvz.org \
    --cc=famz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=mreitz@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=vsementsov@virtuozzo.com \
    /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).