[Qemu-devel] [PATCH] qcow2: Fix Coverity warning when calculating the refcount cache size

[Qemu-devel] [PATCH] qcow2: Fix Coverity warning when calculating the refcount cache size

[Alberto Garcia]

MIN_REFCOUNT_CACHE_SIZE is 4 and the cluster size is guaranteed to be
at most 2MB, so the minimum refcount cache size (in bytes) is always
going to fit in a 32-bit integer.

Coverity doesn't know that, and since we're storing the result in a
uint64_t (*refcount_cache_size) it thinks that we need the 64 bits and
that we probably want to do a 64-bit multiplication to prevent the
result from being truncated.

This is a false positive in this case, but it's a fair warning.
We could do a 64-bit multiplication to get rid of it, but since we
know that a 32-bit variable is enough to store this value let's simply
reuse min_refcount_cache, make it a normal int and stop doing casts.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
---
 block/qcow2.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/block/qcow2.c b/block/qcow2.c
index 6d532470a8..a007dc4246 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -768,6 +768,7 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts,
     BDRVQcow2State *s = bs->opaque;
     uint64_t combined_cache_size;
     bool l2_cache_size_set, refcount_cache_size_set, combined_cache_size_set;
+    int min_refcount_cache = MIN_REFCOUNT_CACHE_SIZE * s->cluster_size;
 
     combined_cache_size_set = qemu_opt_get(opts, QCOW2_OPT_CACHE_SIZE);
     l2_cache_size_set = qemu_opt_get(opts, QCOW2_OPT_L2_CACHE_SIZE);
@@ -804,8 +805,6 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts,
         } else {
             uint64_t virtual_disk_size = bs->total_sectors * BDRV_SECTOR_SIZE;
             uint64_t max_l2_cache = virtual_disk_size / (s->cluster_size / 8);
-            uint64_t min_refcount_cache =
-                (uint64_t) MIN_REFCOUNT_CACHE_SIZE * s->cluster_size;
 
             /* Assign as much memory as possible to the L2 cache, and
              * use the remainder for the refcount cache */
@@ -825,7 +824,7 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts,
                                  * s->cluster_size);
         }
         if (!refcount_cache_size_set) {
-            *refcount_cache_size = MIN_REFCOUNT_CACHE_SIZE * s->cluster_size;
+            *refcount_cache_size = min_refcount_cache;
         }
     }
 
-- 
2.11.0

Re: [Qemu-devel] [PATCH] qcow2: Fix Coverity warning when calculating the refcount cache size

[Eric Blake]

On 05/28/2018 10:01 AM, Alberto Garcia wrote:
> MIN_REFCOUNT_CACHE_SIZE is 4 and the cluster size is guaranteed to be
> at most 2MB, so the minimum refcount cache size (in bytes) is always
> going to fit in a 32-bit integer.
> 
> Coverity doesn't know that, and since we're storing the result in a
> uint64_t (*refcount_cache_size) it thinks that we need the 64 bits and
> that we probably want to do a 64-bit multiplication to prevent the
> result from being truncated.
> 
> This is a false positive in this case, but it's a fair warning.
> We could do a 64-bit multiplication to get rid of it, but since we
> know that a 32-bit variable is enough to store this value let's simply
> reuse min_refcount_cache, make it a normal int and stop doing casts.
> 
> Signed-off-by: Alberto Garcia <berto@igalia.com>
> Reported-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>   block/qcow2.c | 5 ++---
>   1 file changed, 2 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

Re: [Qemu-devel] [PATCH] qcow2: Fix Coverity warning when calculating the refcount cache size

[Kevin Wolf]

Am 28.05.2018 um 17:01 hat Alberto Garcia geschrieben:
> MIN_REFCOUNT_CACHE_SIZE is 4 and the cluster size is guaranteed to be
> at most 2MB, so the minimum refcount cache size (in bytes) is always
> going to fit in a 32-bit integer.
> 
> Coverity doesn't know that, and since we're storing the result in a
> uint64_t (*refcount_cache_size) it thinks that we need the 64 bits and
> that we probably want to do a 64-bit multiplication to prevent the
> result from being truncated.
> 
> This is a false positive in this case, but it's a fair warning.
> We could do a 64-bit multiplication to get rid of it, but since we
> know that a 32-bit variable is enough to store this value let's simply
> reuse min_refcount_cache, make it a normal int and stop doing casts.
> 
> Signed-off-by: Alberto Garcia <berto@igalia.com>
> Reported-by: Peter Maydell <peter.maydell@linaro.org>

Thanks, applied to the block branch.

Kevin