Am 27.08.2013 um 18:07 hat Stefan Weil geschrieben: > Am 27.08.2013 17:56, schrieb Kevin Wolf: > > The failing condition is checked immediately before the assertion, so > > keeping the assertion is kind of redundant. > > Are you sure that it is not the check which is redundant? > Is file == NULL allowed in this 'else' branch?
Allowed in the sense of it's possible to get there, yes. The check that errors out on the condition isn't dead code after reverting 8afaefb8 (which enables the recent work towards blockdev-add): $ x86_64-softmmu/qemu-system-x86_64 -drive file.driver=qcow2,file=/tmp/test.qcow2 qemu-system-x86_64: -drive file.driver=qcow2,file=/tmp/test.qcow2: Can't use 'qcow2' as a block driver for the protocol level qemu-system-x86_64: -drive file.driver=qcow2,file=/tmp/test.qcow2: could not open disk image /tmp/test.qcow2: Invalid argument Kevin > > Signed-off-by: Kevin Wolf <kw...@redhat.com> > > --- > > block.c | 1 - > > 1 file changed, 1 deletion(-) > > > > diff --git a/block.c b/block.c > > index a387c1a..26639e8 100644 > > --- a/block.c > > +++ b/block.c > > @@ -743,7 +743,6 @@ static int bdrv_open_common(BlockDriverState *bs, > > BlockDriverState *file, > > ret = -EINVAL; > > goto free_and_fail; > > } > > - assert(file != NULL); > > bs->file = file; > > ret = drv->bdrv_open(bs, options, open_flags); > > } >
