ARM: 7559/1: smp: switch away from the idmap before updating init_mm.mm_count
authorWill Deacon <will.deacon@arm.com>
Fri, 19 Oct 2012 16:53:01 +0000 (17:53 +0100)
committerRussell King <rmk+kernel@arm.linux.org.uk>
Mon, 22 Oct 2012 14:33:51 +0000 (15:33 +0100)
When booting a secondary CPU, the primary CPU hands two sets of page
tables via the secondary_data struct:

(1) swapper_pg_dir: a normal, cacheable, shared (if SMP) mapping
    of the kernel image (i.e. the tables used by init_mm).

(2) idmap_pgd: an uncached mapping of the .idmap.text ELF
    section.

The idmap is generally used when enabling and disabling the MMU, which
includes early CPU boot. In this case, the secondary CPU switches to
swapper as soon as it enters C code:

struct mm_struct *mm = &init_mm;
unsigned int cpu = smp_processor_id();

/*
 * All kernel threads share the same mm context; grab a
 * reference and switch to it.
 */
atomic_inc(&mm->mm_count);
current->active_mm = mm;
cpumask_set_cpu(cpu, mm_cpumask(mm));
cpu_switch_mm(mm->pgd, mm);

This causes a problem on ARMv7, where the identity mapping is treated as
strongly-ordered leading to architecturally UNPREDICTABLE behaviour of
exclusive accesses, such as those used by atomic_inc.

This patch re-orders the secondary_start_kernel function so that we
switch to swapper before performing any exclusive accesses.

Cc: <stable@vger.kernel.org>
Cc: David McKay <david.mckay@st.com>
Reported-by: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
arch/arm/kernel/smp.c

index 8e20754dd31d5946f5297aae2023ae2c8d5cd572..fbc8b2623d82f08c8c2926375a7b8a31f943361b 100644 (file)
@@ -294,18 +294,24 @@ static void percpu_timer_setup(void);
 asmlinkage void __cpuinit secondary_start_kernel(void)
 {
        struct mm_struct *mm = &init_mm;
-       unsigned int cpu = smp_processor_id();
+       unsigned int cpu;
+
+       /*
+        * The identity mapping is uncached (strongly ordered), so
+        * switch away from it before attempting any exclusive accesses.
+        */
+       cpu_switch_mm(mm->pgd, mm);
+       enter_lazy_tlb(mm, current);
+       local_flush_tlb_all();
 
        /*
         * All kernel threads share the same mm context; grab a
         * reference and switch to it.
         */
+       cpu = smp_processor_id();
        atomic_inc(&mm->mm_count);
        current->active_mm = mm;
        cpumask_set_cpu(cpu, mm_cpumask(mm));
-       cpu_switch_mm(mm->pgd, mm);
-       enter_lazy_tlb(mm, current);
-       local_flush_tlb_all();
 
        printk("CPU%u: Booted secondary processor\n", cpu);