Re: [Qemu-devel] [PATCH] block: Fix bdrv_is_allocated() for short backing files

[Kevin Wolf <kwolf@redhat.com>]

Am 06.05.2014 um 21:53 hat Max Reitz geschrieben:
> On 06.05.2014 15:30, Kevin Wolf wrote:
> >bdrv_is_allocated() shouldn't return true for sectors that are
> >unallocated, but after the end of a short backing file, even though
> >such sectors are (correctly) marked as containing zeros.
> >
> >Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> >---
> >  block.c               |  8 +++++---
> >  include/block/block.h | 11 +++++++----
> >  2 files changed, 12 insertions(+), 7 deletions(-)
> >
> >diff --git a/block.c b/block.c
> >index c90c71a..d3a9906 100644
> >--- a/block.c
> >+++ b/block.c
> >@@ -3883,6 +3883,10 @@ static int64_t coroutine_fn bdrv_co_get_block_status(BlockDriverState *bs,
> >                                       *pnum, pnum);
> >      }
> >+    if (ret & (BDRV_BLOCK_DATA | BDRV_BLOCK_ZERO)) {
> >+        ret |= BDRV_BLOCK_ALLOCATED;
> >+    }
> >+
> 
> Shouldn't BDRV_BLOCK_ALLOCATED be set in the
> !bs->drv->bdrv_co_get_block_status case as well?

It should. Thanks, I'll send a v2.

> >      if (!(ret & BDRV_BLOCK_DATA) && !(ret & BDRV_BLOCK_ZERO)) {
> >          if (bdrv_unallocated_blocks_are_zero(bs)) {
> >              ret |= BDRV_BLOCK_ZERO;
> >@@ -3959,9 +3963,7 @@ int coroutine_fn bdrv_is_allocated(BlockDriverState *bs, int64_t sector_num,
> >      if (ret < 0) {
> >          return ret;
> >      }
> >-    return
> >-        (ret & BDRV_BLOCK_DATA) ||
> >-        ((ret & BDRV_BLOCK_ZERO) && !bdrv_has_zero_init(bs));
> >+    return (ret & BDRV_BLOCK_ALLOCATED);
> >  }
> >  /*
> >diff --git a/include/block/block.h b/include/block/block.h
> >index 2fda81c..ad4c7e8 100644
> >--- a/include/block/block.h
> >+++ b/include/block/block.h
> >@@ -116,6 +116,8 @@ typedef enum {
> >  /* BDRV_BLOCK_DATA: data is read from bs->file or another file
> >   * BDRV_BLOCK_ZERO: sectors read as zero
> >   * BDRV_BLOCK_OFFSET_VALID: sector stored in bs->file as raw data
> >+ * BDRV_BLOCK_ALLOCATED: the content of the block is determined by this
> >+ *                       layer (as opposed to the backing file)
> 
> I guess this is above BDRV_BLOCK_RAW (albeit having a greater value)
> because it is not only used internally? (to pick up on the topic of
> OCD :-P)

Because it felt right. This may or may not be equivalent.

Kevin