On 05/13/2015 02:15 PM, Peter Maydell wrote:
> On 13 May 2015 at 22:08, Richard Henderson wrote:
>>> Are unaligned accesses always slow-path, by the way?
>
> Would it be in theory possible to fast-path MO_UNALN
> unaligned accesses if the host CPU supported unaligned
> accesses for whatever load/
On 13 May 2015 at 22:08, Richard Henderson wrote:
>> Are unaligned accesses always slow-path, by the way?
Would it be in theory possible to fast-path MO_UNALN
unaligned accesses if the host CPU supported unaligned
accesses for whatever load/store insn we use in the TLB
hit case?
-- PMM
> /* MO_UNALN accesses are never checked for alignment; MO_ALIGN
> * accesses will result in a call to the CPU's do_unaligned_access
> * hook if the guest address is not aligned. The default depends
> * on whether the target CPU defines ALIGNED_ONLY.
> */
Fair enough.
> Ar
On 13 May 2015 at 20:10, Richard Henderson wrote:
> These modifiers control, on a per-memory-op basis, whether
> unaligned memory accesses are allowed. The default setting
> reflects the target's definition of ALIGNED_ONLY.
>
> Signed-off-by: Richard Henderson
> diff --git a/tcg/tcg.h b/tcg/tcg
These modifiers control, on a per-memory-op basis, whether
unaligned memory accesses are allowed. The default setting
reflects the target's definition of ALIGNED_ONLY.
Signed-off-by: Richard Henderson
---
softmmu_template.h | 72 --
tcg/tcg.h