[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH RFC v0 5/6] target-ppc: add vsrv instruction
From: |
Nikunj A Dadhania |
Subject: |
Re: [Qemu-devel] [PATCH RFC v0 5/6] target-ppc: add vsrv instruction |
Date: |
Wed, 27 Jul 2016 12:27:27 +0530 |
User-agent: |
Notmuch/0.21 (https://notmuchmail.org) Emacs/25.0.94.1 (x86_64-redhat-linux-gnu) |
David Gibson <address@hidden> writes:
> [ Unknown signature status ]
> On Wed, Jul 27, 2016 at 12:01:33PM +0530, Nikunj A Dadhania wrote:
>> David Gibson <address@hidden> writes:
>>
>> > [ Unknown signature status ]
>> > On Wed, Jul 27, 2016 at 12:56:57AM +0530, Nikunj A Dadhania wrote:
>> >> From: Vivek Andrew Sha <address@hidden>
>> >>
>> >> Adds Vector Shift Right Variable instruction.
>> >>
>> >> Signed-off-by: Vivek Andrew Sha <address@hidden>
>> >> Signed-off-by: Nikunj A Dadhania <address@hidden>
>> >> ---
>> >> target-ppc/helper.h | 1 +
>> >> target-ppc/int_helper.c | 17 +++++++++++++++++
>> >> target-ppc/translate.c | 2 ++
>> >> 3 files changed, 20 insertions(+)
>> >>
>> >> diff --git a/target-ppc/helper.h b/target-ppc/helper.h
>> >> index 9703f85..8eada2f 100644
>> >> --- a/target-ppc/helper.h
>> >> +++ b/target-ppc/helper.h
>> >> @@ -211,6 +211,7 @@ DEF_HELPER_3(vslw, void, avr, avr, avr)
>> >> DEF_HELPER_3(vsld, void, avr, avr, avr)
>> >> DEF_HELPER_3(vslo, void, avr, avr, avr)
>> >> DEF_HELPER_3(vsro, void, avr, avr, avr)
>> >> +DEF_HELPER_3(vsrv, void, avr, avr, avr)
>> >> DEF_HELPER_3(vslv, void, avr, avr, avr)
>> >> DEF_HELPER_3(vaddcuw, void, avr, avr, avr)
>> >> DEF_HELPER_3(vsubcuw, void, avr, avr, avr)
>> >> diff --git a/target-ppc/int_helper.c b/target-ppc/int_helper.c
>> >> index 412398f..f4776f0 100644
>> >> --- a/target-ppc/int_helper.c
>> >> +++ b/target-ppc/int_helper.c
>> >> @@ -1722,6 +1722,23 @@ void helper_vslv(ppc_avr_t *r, ppc_avr_t *a,
>> >> ppc_avr_t *b)
>> >> }
>> >> }
>> >>
>> >> +void helper_vsrv(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
>> >> +{
>> >> + int i;
>> >> + unsigned int shift, bytes, src[ARRAY_SIZE(r->u8) + 1];
>> >> +
>> >> + src[0] = 0;
>> >> + for (i = 0; i < ARRAY_SIZE(r->u8); i++) {
>> >> + src[i + 1] = a->u8[i];
>> >> + }
>> >> +
>> >> + for (i = 0; i < ARRAY_SIZE(r->u8); i++) {
>> >> + shift = b->u8[i] & 0x7; /* extract shift value */
>> >> + bytes = (src[i] << 8) + src[i + 1]; /* extract adjacent
>> >> bytes */
>> >
>> > I think you should be able to construct bytes on the fly without
>> > pre-generating teh whole of src, as you already did for vslv.
>>
>> Had done that, but that introduces a bug like this, for eg:
>>
>> vslv vra,vra,vrb
>>
>> So modified vra->u8[i] is used during subsequent operation as input.
>>
>> Assuming I take care or special casing "0":
>>
>> bytes = ((vra->u8[i - 1] << 8) | (vra->u8[i]))
>> vra->u8[i] = (bytes >> shift) & 0xFF;
>>
>> when i = 1, bytes will ((vra->u8[0] << 8) | (vra->u8[1])). But vra->u8[0],
>> was changed in the previous operation.
>
> Ah, good point.
>
>> Thats the reason src[] is needed
>
> It's probably possible to avoid generating all of src by instead
> generating the bytes one loop iteration ahead, but that sounds fiddly,
> so the current approach is fine for now.
Or do the operation in the reverse: starting from size to 0. Let me try
that.
Regards
Nikunj
- Re: [Qemu-devel] [PATCH RFC v0 2/6] target-ppc: add vabsdu[b, h, w] instructions, (continued)
[Qemu-devel] [PATCH RFC v0 6/6] target-ppc: add extswsli[.] instruction, Nikunj A Dadhania, 2016/07/26
Re: [Qemu-devel] [PATCH RFC v0 0/6] POWER9 TCG enablements - part2, David Gibson, 2016/07/28