arm64: Use the reserved TTBR0 if context switching to the init_mm
authorCatalin Marinas <catalin.marinas@arm.com>
Mon, 23 Mar 2015 15:06:50 +0000 (15:06 +0000)
committerCatalin Marinas <catalin.marinas@arm.com>
Mon, 23 Mar 2015 18:42:23 +0000 (18:42 +0000)
The idle_task_exit() function may call switch_mm() with next ==
&init_mm. On arm64, init_mm.pgd cannot be used for user mappings, so
this patch simply sets the reserved TTBR0.

Cc: <stable@vger.kernel.org>
Reported-by: Jon Medhurst (Tixy) <tixy@linaro.org>
Tested-by: Jon Medhurst (Tixy) <tixy@linaro.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
arch/arm64/include/asm/mmu_context.h

index a9eee33dfa62dc031ab8262c275eba79f8609bac..101a42bde728a8b9547bca989b696d473dcd7e42 100644 (file)
@@ -151,6 +151,15 @@ switch_mm(struct mm_struct *prev, struct mm_struct *next,
 {
        unsigned int cpu = smp_processor_id();
 
+       /*
+        * init_mm.pgd does not contain any user mappings and it is always
+        * active for kernel addresses in TTBR1. Just set the reserved TTBR0.
+        */
+       if (next == &init_mm) {
+               cpu_set_reserved_ttbr0();
+               return;
+       }
+
        if (!cpumask_test_and_set_cpu(cpu, mm_cpumask(next)) || prev != next)
                check_and_switch_context(next, tsk);
 }