diff options
author | Masami Hiramatsu <mhiramat@kernel.org> | 2019-11-27 14:56:52 +0900 |
---|---|---|
committer | Ingo Molnar <mingo@kernel.org> | 2019-11-27 07:44:25 +0100 |
commit | 285a54efe3861976af9d15e85ff8c91a78d1407b (patch) | |
tree | fd65cb9e79d9ae242702ade625024bacc8b97058 | |
parent | 76ffa7204b1ad7d321ac3a0292fdf3975d14866b (diff) | |
download | linux-285a54efe3861976af9d15e85ff8c91a78d1407b.tar.bz2 |
x86/alternatives: Sync bp_patching update for avoiding NULL pointer exception
ftracetest multiple_kprobes.tc testcase hits the following NULL pointer
exception:
BUG: kernel NULL pointer dereference, address: 0000000000000000
PGD 800000007bf60067 P4D 800000007bf60067 PUD 7bf5f067 PMD 0
Oops: 0000 [#1] PREEMPT SMP PTI
RIP: 0010:poke_int3_handler+0x39/0x100
Call Trace:
<IRQ>
do_int3+0xd/0xf0
int3+0x42/0x50
RIP: 0010:sched_clock+0x6/0x10
poke_int3_handler+0x39 was alternatives:958:
static inline void *text_poke_addr(struct text_poke_loc *tp)
{
return _stext + tp->rel_addr; <------ Here is line #958
}
This seems to be caused by tp (bp_patching.vec) being NULL but
bp_patching.nr_entries != 0. There is a small chance for this
to happen, because we have no synchronization between the zeroing
of bp_patching.nr_entries and before clearing bp_patching.vec.
Steve suggested we could fix this by adding sync_core(), because int3
is done with interrupts disabled, and the on_each_cpu() requires
all CPUs to have had their interrupts enabled.
[ mingo: Edited the comments and the changelog. ]
Suggested-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: bristot@redhat.com
Fixes: c0213b0ac03c ("x86/alternative: Batch of patch operations")
Link: https://lkml.kernel.org/r/157483421229.25881.15314414408559963162.stgit@devnote2
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-rw-r--r-- | arch/x86/kernel/alternative.c | 8 |
1 files changed, 7 insertions, 1 deletions
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 4552795a8df4..30e86730655c 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -1134,8 +1134,14 @@ static void text_poke_bp_batch(struct text_poke_loc *tp, unsigned int nr_entries * sync_core() implies an smp_mb() and orders this store against * the writing of the new instruction. */ - bp_patching.vec = NULL; bp_patching.nr_entries = 0; + /* + * This sync_core () call ensures that all INT3 handlers in progress + * have finished. This allows poke_int3_handler() after this to + * avoid touching bp_paching.vec by checking nr_entries == 0. + */ + text_poke_sync(); + bp_patching.vec = NULL; } void text_poke_loc_init(struct text_poke_loc *tp, void *addr, |