On Thu, Dec 01, 2011 at 03:51:00PM +0100, Daniel Lezcano wrote:
commit 549158d2ab01e8370d2773044fe09738a26f7086 Author: Nicolas Pitre nicolas.pitre@linaro.org Date: Thu Aug 25 00:35:59 2011 -0400
ARM: move iotable mappings within the vmalloc regionIn order to remove the build time variation between different SOCs with regards to VMALLOC_END, the iotable mappings are now allocated inside the vmalloc region. This allows for VMALLOC_END to be identical across all machines.
I recently reported an issue with this patch on s3c64xx which I'm avoiding with the below change, I believe Nicolas folded this in to his code but it's not propagated into -next yet.
From d53e2ce3fb18e097678b324932591044eb80c0f1 Mon Sep 17 00:00:00 2001
From: Mark Brown broonie@opensource.wolfsonmicro.com Date: Thu, 24 Nov 2011 12:46:04 +0000 Subject: [PATCH] ARM: Handle empty maps in iotable_init()
Some CPUs (such as the S3C6410) have been relying on being able to call iotable_init() with no io_descs in order to simplify passing through machine-specific io_descs. The changes in "ARM: move iotable mappings within the vmalloc region" broke this by adding an early_alloc_aligned() for an array of vm_structs. Fix this by returning early if no descriptors have been passed.
I'm not sure if this is the most tasteful fix but it preserves existing behaviour and allows boot to proceed on my system.
Signed-off-by: Mark Brown broonie@opensource.wolfsonmicro.com --- arch/arm/mm/mmu.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c index 06e2aef..94c5a0c 100644 --- a/arch/arm/mm/mmu.c +++ b/arch/arm/mm/mmu.c @@ -763,6 +763,9 @@ void __init iotable_init(struct map_desc *io_desc, int nr) struct map_desc *md; struct vm_struct *vm;
+ if (!nr) + return; + vm = early_alloc_aligned(sizeof(*vm) * nr, __alignof__(*vm));
for (md = io_desc; nr; md++, nr--) {