[PATCH][AArch64] Adjust writeback in non-zero memset

[Wilco Dijkstra]

Hi Siddhesh,

>On 14/11/18 8:15 PM, Wilco Dijkstra wrote:
>> v2: Also bias dst in the zva_other code to avoid issues with zva sizes >= 256.
>> 
>> This fixes an ineffiency in the non-zero memset.  Delaying the writeback
>> until the end of the loop is slightly faster on some cores - this shows
>> ~5% performance gain on Cortex-A53 when doing large non-zero memsets.
>> 
>> Tested against the GLIBC testsuite, OK for commit?
>
> Can you please also summarize the performance results for other processors?

I ran it on Cortex-A72, and there isn't a performance difference on the memset bench
beyond measurement errors. On any out-of-order core address increments are
executed in parallel with loads/stores, ie. they have no measurable latency and so
their exact placement is irrelevant.

Wilco