[libvirt] [PATCH] util: storage: Fix build after 25924dec0f9329d429aadae14e273602307e2214
The commit referenced above changed function arguments of virStorageFileGetMetadataFromBuf() but didn't tweak the ATTRIBUTE_NONNULL tied to them. This was caught by coverity as it actually obeys them. We disabled them for GCC and thus it didn't show up. Additionally in commit 3ea661deeabadc3c114dfb6f662b9fd17d714a01 I passed NULL to the backingFormat argument which was also marked as nonnull. Use a dummy int's address when the argument isn't supplied so that the code doesn't need to change much. --- src/util/virstoragefile.c | 4 ++++ src/util/virstoragefile.h | 3 +-- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c index 505b652..de0e27a 100644 --- a/src/util/virstoragefile.c +++ b/src/util/virstoragefile.c @@ -955,6 +955,10 @@ virStorageFileGetMetadataFromBuf(const char *path, int *backingFormat) { virStorageSourcePtr ret = NULL; + int dummy; + + if (!backingFormat) + backingFormat = &dummy; if (!(ret = virStorageFileMetadataNew(path, format))) return NULL; diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h index 744a6ba..528cfad 100644 --- a/src/util/virstoragefile.h +++ b/src/util/virstoragefile.h @@ -301,8 +301,7 @@ virStorageSourcePtr virStorageFileGetMetadataFromBuf(const char *path, size_t len, int format, int *backingFormat) - ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) ATTRIBUTE_NONNULL(4) - ATTRIBUTE_NONNULL(5); + ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); int virStorageFileChainGetBroken(virStorageSourcePtr chain, char **broken_file); -- 2.0.0
On 09.07.2014 10:05, Peter Krempa wrote:
The commit referenced above changed function arguments of virStorageFileGetMetadataFromBuf() but didn't tweak the ATTRIBUTE_NONNULL tied to them. This was caught by coverity as it actually obeys them. We disabled them for GCC and thus it didn't show up.
Additionally in commit 3ea661deeabadc3c114dfb6f662b9fd17d714a01 I passed NULL to the backingFormat argument which was also marked as nonnull. Use a dummy int's address when the argument isn't supplied so that the code doesn't need to change much. --- src/util/virstoragefile.c | 4 ++++ src/util/virstoragefile.h | 3 +-- 2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c index 505b652..de0e27a 100644 --- a/src/util/virstoragefile.c +++ b/src/util/virstoragefile.c @@ -955,6 +955,10 @@ virStorageFileGetMetadataFromBuf(const char *path, int *backingFormat) { virStorageSourcePtr ret = NULL; + int dummy; + + if (!backingFormat) + backingFormat = &dummy;
if (!(ret = virStorageFileMetadataNew(path, format))) return NULL; diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h index 744a6ba..528cfad 100644 --- a/src/util/virstoragefile.h +++ b/src/util/virstoragefile.h @@ -301,8 +301,7 @@ virStorageSourcePtr virStorageFileGetMetadataFromBuf(const char *path, size_t len, int format, int *backingFormat) - ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) ATTRIBUTE_NONNULL(4) - ATTRIBUTE_NONNULL(5); + ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); int virStorageFileChainGetBroken(virStorageSourcePtr chain, char **broken_file);
So if we need to cripple passing NULL, why do we even have ATTRIBUTE_NONNULL? Wouldn't it be simpler if virStorageFileGetMetadataFromBuf would accept NULL? Michal
On 07/09/14 10:12, Michal Privoznik wrote:
On 09.07.2014 10:05, Peter Krempa wrote:
The commit referenced above changed function arguments of virStorageFileGetMetadataFromBuf() but didn't tweak the ATTRIBUTE_NONNULL tied to them. This was caught by coverity as it actually obeys them. We disabled them for GCC and thus it didn't show up.
Additionally in commit 3ea661deeabadc3c114dfb6f662b9fd17d714a01 I passed NULL to the backingFormat argument which was also marked as nonnull. Use a dummy int's address when the argument isn't supplied so that the code doesn't need to change much. --- src/util/virstoragefile.c | 4 ++++ src/util/virstoragefile.h | 3 +-- 2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h index 744a6ba..528cfad 100644 --- a/src/util/virstoragefile.h +++ b/src/util/virstoragefile.h @@ -301,8 +301,7 @@ virStorageSourcePtr virStorageFileGetMetadataFromBuf(const char *path, size_t len, int format, int *backingFormat) - ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) ATTRIBUTE_NONNULL(4) - ATTRIBUTE_NONNULL(5); + ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); int virStorageFileChainGetBroken(virStorageSourcePtr chain, char **broken_file);
So if we need to cripple passing NULL, why do we even have ATTRIBUTE_NONNULL? Wouldn't it be simpler if virStorageFileGetMetadataFromBuf would accept NULL?
This patch actually removes the ATTRIBUTE_NONNULL from two places. Argument 4, as it's now an int and the attribute doesn't make sense there. And argument 5 as we are passing NULL in one place now. Therefore I also added the null tolerance code as the underlying code doesn't tolerate NULL. In virStorageFileGetMetadataFromFD we do the same stuff. Peter
On 07/09/2014 02:05 AM, Peter Krempa wrote:
The commit referenced above changed function arguments of virStorageFileGetMetadataFromBuf() but didn't tweak the ATTRIBUTE_NONNULL tied to them. This was caught by coverity as it actually obeys them. We disabled them for GCC and thus it didn't show up.
Yeah, because gcc 4.8 still misbehaves on the attribute (silently "optimizing" code without flagging invalid uses). There's been talk on the gcc list about fixing it, although probably no sooner than 4.10 timeframe.
Additionally in commit 3ea661deeabadc3c114dfb6f662b9fd17d714a01 I passed NULL to the backingFormat argument which was also marked as nonnull. Use a dummy int's address when the argument isn't supplied so that the code doesn't need to change much. --- src/util/virstoragefile.c | 4 ++++ src/util/virstoragefile.h | 3 +-- 2 files changed, 5 insertions(+), 2 deletions(-)
ACK. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
On 07/09/14 15:01, Eric Blake wrote:
On 07/09/2014 02:05 AM, Peter Krempa wrote:
The commit referenced above changed function arguments of virStorageFileGetMetadataFromBuf() but didn't tweak the ATTRIBUTE_NONNULL tied to them. This was caught by coverity as it actually obeys them. We disabled them for GCC and thus it didn't show up.
Yeah, because gcc 4.8 still misbehaves on the attribute (silently "optimizing" code without flagging invalid uses). There's been talk on the gcc list about fixing it, although probably no sooner than 4.10 timeframe.
Additionally in commit 3ea661deeabadc3c114dfb6f662b9fd17d714a01 I passed NULL to the backingFormat argument which was also marked as nonnull. Use a dummy int's address when the argument isn't supplied so that the code doesn't need to change much. --- src/util/virstoragefile.c | 4 ++++ src/util/virstoragefile.h | 3 +-- 2 files changed, 5 insertions(+), 2 deletions(-)
ACK.
Pushed; Thanks. Peter
participants (3)
-
Eric Blake -
Michal Privoznik -
Peter Krempa