[PATCH] ARM: Add workaround for I-Cache line size mismatch between CPU cores

[Marek Szyprowski]

Some big.LITTLE systems have I-Cache line size mismatch between
LITTLE and big cores. This patch adds a workaround for proper I-Cache
support on such systems. Without it, some class of the userspace code
(typically self-modifying) might suffer from random SIGILL failures.

Similar workaround already exists for ARM64 architecture. I has been
added by commit 116c81f427ff ("arm64: Work around systems with mismatched
cache line sizes").

Signed-off-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
---
This workaround is needed on all supported Exynos big.LITTLE SoCs: 5420,
5422 and 5800.

Resend reason: removed RFC tag as there are no comments, I will upload
this patch to the patch tracking system
---
 arch/arm/configs/exynos_defconfig |  1 +
 arch/arm/include/asm/cacheflush.h |  7 +++++++
 arch/arm/kernel/smp.c             |  1 +
 arch/arm/mm/Kconfig               |  8 ++++++++
 arch/arm/mm/cache-v7.S            | 13 +++++++++++++
 arch/arm/mm/init.c                | 16 ++++++++++++++++
 arch/arm/mm/mm.h                  |  2 ++
 7 files changed, 48 insertions(+)

diff --git a/arch/arm/configs/exynos_defconfig b/arch/arm/configs/exynos_defconfig
index c95c54284da2..9b959afaaa12 100644
--- a/arch/arm/configs/exynos_defconfig
+++ b/arch/arm/configs/exynos_defconfig
@@ -9,6 +9,7 @@ CONFIG_MODULE_UNLOAD=y
 CONFIG_PARTITION_ADVANCED=y
 CONFIG_ARCH_EXYNOS=y
 CONFIG_ARCH_EXYNOS3=y
+CONFIG_CPU_ICACHE_MISMATCH_WORKAROUND=y
 CONFIG_SMP=y
 CONFIG_BIG_LITTLE=y
 CONFIG_NR_CPUS=8
diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h
index ec1a5fd0d294..ec4fd2e2dd60 100644
--- a/arch/arm/include/asm/cacheflush.h
+++ b/arch/arm/include/asm/cacheflush.h
@@ -479,4 +479,11 @@ static inline void __sync_cache_range_r(volatile void *p, size_t size)
 void flush_uprobe_xol_access(struct page *page, unsigned long uaddr,
 			     void *kaddr, unsigned long len);
 
+
+#ifdef CONFIG_CPU_ICACHE_MISMATCH_WORKAROUND
+void check_cpu_icache_size(int cpuid);
+#else
+static inline void check_cpu_icache_size(int cpuid) { }
+#endif
+
 #endif
diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 8687d619260f..261be0e5bc03 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -373,6 +373,7 @@ static void smp_store_cpu_info(unsigned int cpuid)
 	cpu_info->cpuid = read_cpuid_id();
 
 	store_cpu_topology(cpuid);
+	check_cpu_icache_size(cpuid);
 }
 
 /*
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig
index b169e580bf82..d4733e086f2b 100644
--- a/arch/arm/mm/Kconfig
+++ b/arch/arm/mm/Kconfig
@@ -780,6 +780,14 @@ config CPU_ICACHE_DISABLE
 	  Say Y here to disable the processor instruction cache. Unless
 	  you have a reason not to or are unsure, say N.
 
+config CPU_ICACHE_MISMATCH_WORKAROUND
+	bool "Workaround for I-Cache line size mismatch between CPU cores"
+	depends on SMP && CPU_V7
+	help
+	  Some big.LITTLE systems have I-Cache line size mismatch between
+	  LITTLE and big cores.  Say Y here to enable a workaround for
+	  proper I-Cache support on such systems.  If unsure, say N.
+
 config CPU_DCACHE_DISABLE
 	bool "Disable D-Cache (C-bit)"
 	depends on (CPU_CP15 && !SMP) || CPU_V7M
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index 2149b47a0c5a..db3986708c8a 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -19,6 +19,14 @@
 
 #include "proc-macros.S"
 
+#ifdef CONFIG_CPU_ICACHE_MISMATCH_WORKAROUND
+.globl icache_size
+	.data
+	.align	2
+icache_size:
+	.long	64
+	.text
+#endif
 /*
  * The secondary kernel init calls v7_flush_dcache_all before it enables
  * the L1; however, the L1 comes out of reset in an undefined state, so
@@ -284,7 +292,12 @@ ENTRY(v7_coherent_user_range)
 	cmp	r12, r1
 	blo	1b
 	dsb	ishst
+#ifdef CONFIG_CPU_ICACHE_MISMATCH_WORKAROUND
+	ldr	r3, =icache_size
+	ldr	r2, [r3, #0]
+#else
 	icache_line_size r2, r3
+#endif
 	sub	r3, r2, #1
 	bic	r12, r0, r3
 2:
diff --git a/arch/arm/mm/init.c b/arch/arm/mm/init.c
index be0b42937888..1a66af5bd259 100644
--- a/arch/arm/mm/init.c
+++ b/arch/arm/mm/init.c
@@ -242,6 +242,22 @@ static void __init arm_initrd_init(void)
 #endif
 }
 
+#ifdef CONFIG_CPU_ICACHE_MISMATCH_WORKAROUND
+void check_cpu_icache_size(int cpuid)
+{
+	u32 size, ctr;
+
+	asm("mrc p15, 0, %0, c0, c0, 1" : "=r" (ctr));
+
+	size = 1 << ((ctr & 0xf) + 2);
+	if (cpuid != 0 && icache_size != size)
+		pr_info("CPU%u: detected I-Cache line size mismatch, workaround enabled\n",
+			cpuid);
+	if (icache_size > size)
+		icache_size = size;
+}
+#endif
+
 void __init arm_memblock_init(const struct machine_desc *mdesc)
 {
 	/* Register the kernel text, kernel data and initrd with memblock. */
diff --git a/arch/arm/mm/mm.h b/arch/arm/mm/mm.h
index 6b045c6653ea..941356d95a67 100644
--- a/arch/arm/mm/mm.h
+++ b/arch/arm/mm/mm.h
@@ -8,6 +8,8 @@
 /* the upper-most page table pointer */
 extern pmd_t *top_pmd;
 
+extern int icache_size;
+
 /*
  * 0xffff8000 to 0xffffffff is reserved for any ARM architecture
  * specific hacks for copying pages efficiently, while 0xffff4000
-- 
2.17.1