linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
* build failure with VDSO on non-glibc based systems
@ 2015-06-23 10:35 Szabolcs Nagy
  2015-06-23 15:16 ` Nathan Lynch
  0 siblings, 1 reply; 2+ messages in thread
From: Szabolcs Nagy @ 2015-06-23 10:35 UTC (permalink / raw)
  To: linux-arm-kernel

if the host toolchain is not glibc based then the arm kernel build fails with

 arch/arm/vdso/vdsomunge.c:53:19: fatal error: error.h: No such file or directory

error.h is a glibc only header (ie not available in musl, newlib and bsd libcs).

i attached a patch that fixes the build failure for me.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-fix-arm-vdsomunge-not-to-depend-on-glibc-specific-er.patch
Type: text/x-patch
Size: 4168 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150623/81f37be6/attachment.bin>

^ permalink raw reply	[flat|nested] 2+ messages in thread

* build failure with VDSO on non-glibc based systems
  2015-06-23 10:35 build failure with VDSO on non-glibc based systems Szabolcs Nagy
@ 2015-06-23 15:16 ` Nathan Lynch
  0 siblings, 0 replies; 2+ messages in thread
From: Nathan Lynch @ 2015-06-23 15:16 UTC (permalink / raw)
  To: linux-arm-kernel

On 06/23/2015 05:35 AM, Szabolcs Nagy wrote:
> if the host toolchain is not glibc based then the arm kernel build fails with
> 
>  arch/arm/vdso/vdsomunge.c:53:19: fatal error: error.h: No such file or directory
> 
> error.h is a glibc only header (ie not available in musl, newlib and bsd libcs).

Thanks for the report and the patch.  I agree it is inappropriate to
unnecessarily depend on glibc features in host programs.


> i attached a patch that fixes the build failure for me.

A few issues with the patch prevent me from accepting it as-is.

The patch lacks the required Signed-off-by: line; see part 11 in
Documentation/SubmittingPatches.

The body of your email would suffice as the commit log; please include
it in the patch itself.


>  
> +#define fail(...) ( \
> +	failed = 1, \
> +	fprintf(stderr, "%s: ", argv0), \
> +	fprintf(stderr, __VA_ARGS__), \
> +	exit(EXIT_FAILURE))
> +

It would be more in keeping with Linux coding style to write this as a
do { ... } while (0) block (Chapter 12 of Documentation/CodingStyle).

Writing it as a function, if possible, would be best.

Please attend to these issues and resubmit.  Alternatively, I have a
musl installation I can test with and develop an equivalent fix
(crediting you with the initial report, of course).

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2015-06-23 15:16 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-06-23 10:35 build failure with VDSO on non-glibc based systems Szabolcs Nagy
2015-06-23 15:16 ` Nathan Lynch

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).