On Tue, 2013-09-10 at 13:58 +0800, Alex Shi wrote:
Could you like to send the bad config to reproduce the fixing issue?
This occurs with the Versatile Express config used in Linaro, i.e. the output created by...
ARCH=arm scripts/kconfig/merge_config.sh \ linaro/configs/linaro-base.conf \ linaro/configs/distribution.conf \ linaro/configs/big-LITTLE-MP.conf \ linaro/configs/big-LITTLE-IKS.conf \ linaro/configs/vexpress.conf
Though to get the section missmatch from secondary_startup you will also need to disable CONFIG_THUMB2_KERNEL and section mismatch checking doesn't work full for Thumb kernels.
Oh, and you will need CONFIG_DEBUG_SECTION_MISMATCH=y as well. I always have that on my commandline when building kernels.
On 09/06/2013 11:25 PM, Mark Hambleton wrote:
diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S index 8bac553..2725c87 100644 --- a/arch/arm/kernel/head.S +++ b/arch/arm/kernel/head.S @@ -342,7 +342,6 @@ __turn_mmu_on_loc: .long __turn_mmu_on_end #if defined(CONFIG_SMP)
- __CPUINIT ENTRY(secondary_startup)
I just has few concern of the solution of upstream kernel. the commit 2449189bb7c73b5fe55a18bc0d289e39bdcd4998 in Linus' tree did similar fixing.
That fixes the other commit you mention below and doesn't work properly without it...
#if defined(CONFIG_SMP)
.text/* * Common entry point for secondary CPUs. diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c index 84d55da..a666911 100644 --- a/arch/arm/kernel/smp.c +++ b/arch/arm/kernel/smp.c @@ -57,7 +57,7 @@ struct secondary_data secondary_data;
- control for which core is the next to come out of the secondary
- boot "holding pen"
*/ -volatile int __cpuinitdata pen_release = -1; +volatile int pen_release = -1;
and commit 8bd26e3a7e49af2697449bbc, includes this change on Linus' tree too. I saw Greg also picked up the commit.
That is the ARM part of the whole cpuinit feature removal, so if that's being backported to the 3.10 stable kernel then none of my patch is needed and we can just ignore it and get on with other things :-)