On Sat, Nov 12, 2016 at 10:14:49PM +0000, Sergei Trofimovich wrote:
> > > Basically gcc tries to say us BUILD_ASSERT is not visible.
> > >
> > > BUILD_ASSERT lives in kerncompat.h which this change adds.
> >
> > I think including the kerncompat.h is too intrusive here, I've fixed by
> > providing an empty macro if it's not defined. I'll release 4.8.2 soon.
>
> Apologies. I did not test your fix right afterwards. Seems now header is incomplete
> due to missing NULL (gcc-6):
>
> btrfs-progs-v4.8.3 $ gcc -c ioctl.h -o /tmp/a.o
> ioctl.h: In function 'btrfs_err_str':
> ioctl.h:711:11: error: 'NULL' undeclared (first use in this function)
> return NULL;
> ^~~~
> ioctl.h:711:11: note: each undeclared identifier is reported only once for each function it appears in
The ioctl.h file can be included in both C and C++ code, I'd rahter
avoid to ifdef the right definition of NULL, so s/NULL/0/ seems as a
best fix to me.
Compiling with g++ shows other type errors like
ioctl.h:709:5: warning: ISO C++ forbids converting a string constant to ‘char*’ [-Wwrite-strings]
"in progress";
so we'd have to change the type to 'const char*'. I'm not sure how much
breakage this could cause in programs using the header.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
