powerpc/64s: Un-blacklist system_call() from kprobes
authorNaveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Thu, 29 Jun 2017 17:49:18 +0000 (23:19 +0530)
committerMichael Ellerman <mpe@ellerman.id.au>
Mon, 3 Jul 2017 13:12:01 +0000 (23:12 +1000)
It is actually safe to probe system_call() in entry_64.S, but only till
we unset MSR_RI. To allow this, add a new symbol system_call_exit()
after the mtmsrd and blacklist that.

Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
arch/powerpc/kernel/entry_64.S

index e5cf98d..ed45e86 100644 (file)
@@ -188,6 +188,18 @@ system_call:                       /* label this so stack traces look sane */
        andi.   r10,r8,MSR_RI
        beq-    unrecov_restore
 #endif
+
+/*
+ * This is a few instructions into the actual syscall exit path (which actually
+ * starts at .Lsyscall_exit) to cater to kprobe blacklisting and to reduce the
+ * number of visible symbols for profiling purposes.
+ *
+ * We can probe from system_call until this point as MSR_RI is set. But once it
+ * is cleared below, we won't be able to take a trap.
+ *
+ * This is blacklisted from kprobes further below with _ASM_NOKPROBE_SYMBOL().
+ */
+system_call_exit:
        /*
         * Disable interrupts so current_thread_info()->flags can't change,
         * and so that we don't get interrupted after loading SRR0/1.
@@ -414,7 +426,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
        b       .       /* prevent speculative execution */
 #endif
 _ASM_NOKPROBE_SYMBOL(system_call_common);
-_ASM_NOKPROBE_SYMBOL(system_call);
+_ASM_NOKPROBE_SYMBOL(system_call_exit);
 
 /* Save non-volatile GPRs, if not already saved. */
 _GLOBAL(save_nvgprs)