[PATCH v3 4/4] ARM: reinsert ARCH_MULTI_V4 Kconfig option

[PATCH v3 4/4] ARM: reinsert ARCH_MULTI_V4 Kconfig option

[Jonas Jensen]

CPU_FA526 is thumbless and adding under e.g. ARCH_MULTI_V4T instead means inclusion of CPU_ARM920T
which in turn selects CPU_ABRT_EV4T, CPU_CACHE_V4WT, CPU_COPY_V4WB, CPU_TLB_V4WBI, which in turn can
mean trouble when the kernel eventually tries to exec init:

[    2.580000] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000004
[    2.580000]
[    2.580000] CPU: 0 PID: 1 Comm: init Not tainted 3.10.0-next-20130703+ #3
[    2.580000] [<c000cc64>] (unwind_backtrace+0x0/0xf0) from [<c000b524>] (show_stack+0x10/0x14)
[    2.580000] [<c000b524>] (show_stack+0x10/0x14) from [<c021d6dc>] (panic+0x90/0x1e0)
[    2.580000] [<c021d6dc>] (panic+0x90/0x1e0) from [<c0019664>] (do_exit+0x724/0x834)
[    2.580000] [<c0019664>] (do_exit+0x724/0x834) from [<c00197d8>] (do_group_exit+0x3c/0xb0)
[    2.580000] [<c00197d8>] (do_group_exit+0x3c/0xb0) from [<c0024c34>] (get_signal_to_deliver+0x2ec/0x598)
[    2.580000] [<c0024c34>] (get_signal_to_deliver+0x2ec/0x598) from [<c021d2a8>] (do_signal+0x9c/0x3cc)
[    2.580000] [<c021d2a8>] (do_signal+0x9c/0x3cc) from [<c000b110>] (do_work_pending+0xc4/0xe0)
[    2.580000] [<c000b110>] (do_work_pending+0xc4/0xe0) from [<c0009334>] (work_pending+0xc/0x20)

Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com>
---
 arch/arm/Kconfig | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index d84161c..87e93ab 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -878,6 +878,11 @@ menu "Multiple platform selection"
 
 comment "CPU Core family selection"
 
+config ARCH_MULTI_V4
+	bool "ARMv4 based platforms (FA526, StrongARM)"
+	depends on !ARCH_MULTI_V6_V7
+	select ARCH_MULTI_V4_V5
+
 config ARCH_MULTI_V4T
 	bool "ARMv4T based platforms (ARM720T, ARM920T, ...)"
 	depends on !ARCH_MULTI_V6_V7
-- 
1.8.2.1

[PATCH v3 4/4] ARM: reinsert ARCH_MULTI_V4 Kconfig option

[Jonas Jensen]

Arnd suggested removing one (or many) of the options:

CPU_ABRT_EV4T
CPU_CACHE_V4WT
CPU_COPY_V4WB
CPU_TLB_V4WBI

However boot is still broken after removing all including CPU_32v4T.

Selecting CPU_SA1100 instead, it boots normally and init works.

The idea is, If you can help figure out what exactly breaks boot, the
patch can be dropped.

>From these tests, I have created two branches ( arch/arm/mm have
replacements for calls to no longer compiled in ):

http://code.google.com/p/linux-next-20130703-moxart/source/detail?r=a88f33ad22d3f9a31fe126f1c95568e5b2e8f32d&name=multi_V4_test_arm920t
http://code.google.com/p/linux-next-20130703-moxart/source/detail?r=9f4c21f1123a122c51792cd374217ce6e426e74a&name=multi_V4_test_sa1100

Best regards,
Jonas

[PATCH v3 4/4] ARM: reinsert ARCH_MULTI_V4 Kconfig option

[Arnd Bergmann]

On Monday 08 July 2013, Jonas Jensen wrote:
> Arnd suggested removing one (or many) of the options:
> 
> CPU_ABRT_EV4T
> CPU_CACHE_V4WT
> CPU_COPY_V4WB
> CPU_TLB_V4WBI
> 
> However boot is still broken after removing all including CPU_32v4T.
> 
> Selecting CPU_SA1100 instead, it boots normally and init works.
> 
> The idea is, If you can help figure out what exactly breaks boot, the
> patch can be dropped.
> 
> From these tests, I have created two branches ( arch/arm/mm have
> replacements for calls to no longer compiled in ):
> 
> http://code.google.com/p/linux-next-20130703-moxart/source/detail?r=a88f33ad22d3f9a31fe126f1c95568e5b2e8f32d&name=multi_V4_test_arm920t
> http://code.google.com/p/linux-next-20130703-moxart/source/detail?r=9f4c21f1123a122c51792cd374217ce6e426e74a&name=multi_V4_test_sa1100

It looks like you select CPU_ARM926T but change the entry for CPU_ARM920T
instead, which would have no effect then.

	Arnd