On Wed, Aug 16, 2017 at 7:51 PM, Stephen Rothwell <s...@canb.auug.org.au> wrote:
> Hi James,
>
> After merging the security tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
>
> In file included from samples/seccomp/bpf-fancy.c:12:0:
> samples/seccomp/bpf-fancy.c: In function 'main':
> samples/seccomp/bpf-helper.h:47:26: error: 'SECCOMP_RET_KILL_THREAD'
> undeclared (first use in this function)
> BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_KILL_THREAD)
> ^
> ./usr/include/linux/filter.h:48:59: note: in definition of macro 'BPF_STMT'
> #define BPF_STMT(code, k) { (unsigned short)(code), 0, 0, k }
> ^
> samples/seccomp/bpf-fancy.c:41:3: note: in expansion of macro 'DENY'
> DENY, /* Don't passthrough into a label */
> ^~~~
> samples/seccomp/bpf-helper.h:47:26: note: each undeclared identifier is
> reported only once for each function it appears in
> BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_KILL_THREAD)
> ^
> ./usr/include/linux/filter.h:48:59: note: in definition of macro 'BPF_STMT'
> #define BPF_STMT(code, k) { (unsigned short)(code), 0, 0, k }
> ^
> samples/seccomp/bpf-fancy.c:41:3: note: in expansion of macro 'DENY'
> DENY, /* Don't passthrough into a label */
> ^~~~
>
> [Note: this is a cross build, if that is relevant ...]
>
> Presumably caused by commit
>
> fd76875ca289 ("seccomp: Rename SECCOMP_RET_KILL to SECCOMP_RET_KILL_THREAD")
>
> I have used the security tree from next-20170816 for today.
Hmmm, I think we've had problems like this before due to samples being
built before the headers have been installed. Regardless, I'll
un-rename that macro in the samples...
-Kees
--
Kees Cook
Pixel Security
