[PATCH 0/4] LLVMLinux: Patches to enable the kernel to be compiled with clang/LLVM

[Olof Johansson]

On Thu, Jul 31, 2014 at 8:48 AM, Will Deacon <will.deacon at arm.com> wrote:
> On Thu, Jul 31, 2014 at 04:39:55PM +0100, Olof Johansson wrote:
>> On Thu, Jul 31, 2014 at 3:33 AM, Will Deacon <will.deacon at arm.com> wrote:
>> > On Thu, Jul 31, 2014 at 12:57:25AM +0100, behanw at converseincode.com wrote:
>> >> From: Behan Webster <behanw at converseincode.com>
>> >>
>> >> This patch set moves from using locally defined named registers to access the
>> >> stack pointer to using a globally defined named register. This allows the code
>> >> to work both with gcc and clang.
>> >>
>> >> The LLVMLinux project aims to fully build the Linux kernel using both gcc and
>> >> clang (the C front end for the LLVM compiler infrastructure project).
>> >>
>> >> Behan Webster (4):
>> >>   arm64: LLVMLinux: Add current_stack_pointer() for arm64
>> >>   arm64: LLVMLinux: Use current_stack_pointer in save_stack_trace_tsk
>> >>   arm64: LLVMLinux: Calculate current_thread_info from
>> >>     current_stack_pointer
>> >>   arm64: LLVMLinux: Use current_stack_pointer in kernel/traps.c
>> >
>> > Once Andreas's comments have been addressed:
>> >
>> >   Acked-by: Will Deacon <will.deacon at arm.com>
>> >
>> > Please can you send a new series after the merge window?
>>
>> Given that the ARM64 KVM guys are still actively breaking -next, and
>> this is considerably smaller and lower risk than that, I think you can
>> take it for 3.17?
>
> Hey, those breakages are in the kvm tree not the arm64 tree! I'd really
> rather wait on these as I don't see the rush to get them in for 3.17 and
> it's not beyond the realms of possibility that they could cause problems
> for a particular version of GCC (at the very least, I'd need to re-run
> all the testing I've been doing).


Yeah, it was somewhat tongue in cheek that made not have made it
across the wire.

3.18 is fine with me too even though I find the concerns to be quite unlikely.


-Olof