Skip to contentSkip to conversation

bpf: Fix the build on BPF_SYSCALL= y && !CONFIG_TRACING kernels, make it more configurable

tip:perf/core Alexei Starovoitov 2 participants 2 messages Apr 2, 2015 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Alexei Starovoitov

author Acked-by
#1 lore raw
Quoted 4 lines · tip-bot for Ingo Molnar
Committer:  Ingo Molnar <[email protected]>
CommitDate: Thu, 2 Apr 2015 16:28:06 +0200

bpf: Fix the build on BPF_SYSCALL=y && !CONFIG_TRACING kernels, make it more configurable

Thanks for the fix.
I've tried this combination and it was fine, since whole kernel/trace/ directory is not compiled when !CONFIG_TRACING, but I missed the fact that CONFIG_RING_BUFFER=y also enables kernel/trace/ which this ia64 .config exploited :( I'll add it to my set of configs. Thanks again for the fix. It makes dependency clear.

#2 lore raw
Quoted 11 lines · Alexei Starovoitov
Quoted 4 lines · tip-bot for Ingo Molnar
Committer:  Ingo Molnar <[email protected]>
CommitDate: Thu, 2 Apr 2015 16:28:06 +0200

bpf: Fix the build on BPF_SYSCALL=y && !CONFIG_TRACING kernels, make it more configurable

Thanks for the fix.
I've tried this combination and it was fine, since whole kernel/trace/ directory is not compiled when !CONFIG_TRACING, but I missed the fact that CONFIG_RING_BUFFER=y also enables kernel/trace/ which this ia64 .config exploited :( I'll add it to my set of configs.

I saw this on x86 as well, but yeah, it was a rare thing.

Thanks,

Ingo