Re: [Qemu-devel] [RFC PATCH 08/11] softmmu: make do_unaligned_access extern

["Andreas Färber" <afaerber@suse.de>]

Am 08.05.2014 18:09, schrieb Paolo Bonzini:
> We will reference it from more files in the next patch.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  target-alpha/cpu.h         | 3 +++
>  target-alpha/mem_helper.c  | 4 ++--
>  target-mips/cpu.h          | 3 +++
>  target-mips/op_helper.c    | 8 ++------
>  target-sparc/cpu.h         | 4 ++++
>  target-sparc/ldst_helper.c | 9 +++------
>  target-xtensa/cpu.h        | 3 +++
>  target-xtensa/op_helper.c  | 5 +----
>  8 files changed, 21 insertions(+), 18 deletions(-)
> 
> diff --git a/target-alpha/cpu.h b/target-alpha/cpu.h
> index d9b861f..242068f 100644
> --- a/target-alpha/cpu.h
> +++ b/target-alpha/cpu.h
> @@ -498,6 +498,9 @@ static inline void cpu_get_tb_cpu_state(CPUAlphaState *env, target_ulong *pc,
>      *pflags = flags;
>  }
>  
> +void do_unaligned_access(CPUAlphaState *env, target_ulong addr,
> +                         int is_write, int is_user, uintptr_t retaddr);
> +
>  #include "exec/exec-all.h"
>  
>  #endif /* !defined (__CPU_ALPHA_H__) */
[snip]

Have you considered turning this into CPUClass hooks instead, as done
with do_unassigned_access?

Before your patch, this was a static function used only by
softmmu_template.h - making it global is not exactly helping with
multi-target support. Any solution to remedy regressions appreciated.

Regards,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg