perf/x86: Only allow rdpmc if a perf_event is mapped
authorAndy Lutomirski <luto@amacapital.net>
Fri, 24 Oct 2014 22:58:12 +0000 (15:58 -0700)
committerIngo Molnar <mingo@kernel.org>
Wed, 4 Feb 2015 11:10:47 +0000 (12:10 +0100)
commit7911d3f7af14a614617e38245fedf98a724e46a9
tree8a8d163aee11daadeda7ecd9e835b86c2eb5890e
parentc1317ec2b906442930318d9d6e51425c5a69e9cb
perf/x86: Only allow rdpmc if a perf_event is mapped

We currently allow any process to use rdpmc.  This significantly
weakens the protection offered by PR_TSC_DISABLED, and it could be
helpful to users attempting to exploit timing attacks.

Since we can't enable access to individual counters, use a very
coarse heuristic to limit access to rdpmc: allow access only when
a perf_event is mmapped.  This protects seccomp sandboxes.

There is plenty of room to further tighen these restrictions.  For
example, this allows rdpmc for any x86_pmu event, but it's only
useful for self-monitoring tasks.

As a side effect, cap_user_rdpmc will now be false for AMD uncore
events.  This isn't a real regression, since .event_idx is disabled
for these events anyway for the time being.  Whenever that gets
re-added, the cap_user_rdpmc code can be adjusted or refactored
accordingly.

Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Vince Weaver <vince@deater.net>
Cc: "hillf.zj" <hillf.zj@alibaba-inc.com>
Cc: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/a2bdb3cf3a1d70c26980d7c6dddfbaa69f3182bf.1414190806.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/include/asm/mmu.h
arch/x86/include/asm/mmu_context.h
arch/x86/kernel/cpu/perf_event.c
arch/x86/kernel/cpu/perf_event.h