Hi Greg,

I was able to reproduce the kernel crash on 4.14, 4.19 and 4.20 kernels. Will 
apply your patches and re-test.

Thanks,
Nelson

-----Original Message-----
From: Greg Kroah-Hartman [mailto:[email protected]] 
Sent: Monday, March 18, 2019 2:26 AM
To: [email protected]
Cc: Greg Kroah-Hartman <[email protected]>; [email protected]; 
Tony Jones <[email protected]>; DSouza, Nelson <[email protected]>; Peter 
Zijlstra (Intel) <[email protected]>; Thomas Gleixner <[email protected]>; 
[email protected]; [email protected]; [email protected]
Subject: [PATCH 4.14 31/34] perf/x86/intel: Fix memory corruption

4.14-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Peter Zijlstra <[email protected]>

commit ede271b059463731cbd6dffe55ffd70d7dbe8392 upstream.

Through:

  validate_event()
    x86_pmu.get_event_constraints(.idx=-1)
      tfa_get_event_constraints()
        dyn_constraint()

cpuc->constraint_list[-1] is used, which is an obvious out-of-bound access.

In this case, simply skip the TFA constraint code, there is no event constraint 
with just PMC3, therefore the code will never result in the empty set.

Fixes: 400816f60c54 ("perf/x86/intel: Implement support for TSX Force Abort")
Reported-by: Tony Jones <[email protected]>
Reported-by: "DSouza, Nelson" <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Tested-by: Tony Jones <[email protected]>
Tested-by: "DSouza, Nelson" <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Link: https://lkml.kernel.org/r/[email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>

---
 arch/x86/events/intel/core.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -3262,7 +3262,7 @@ tfa_get_event_constraints(struct cpu_hw_
        /*
         * Without TFA we must not use PMC3.
         */
-       if (!allow_tsx_force_abort && test_bit(3, c->idxmsk)) {
+       if (!allow_tsx_force_abort && test_bit(3, c->idxmsk) && idx >= 0) {
                c = dyn_constraint(cpuc, c, idx);
                c->idxmsk64 &= ~(1ULL << 3);
                c->weight--;


Reply via email to