[Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[Nathan Froyd]

This patch series adds support for integer Altivec instructions to QEMU,
including element-wise loads and stores.  It's rather long (42
individual patches), so to avoid patch-bombing the list, I've placed the
patch series at:

  http://www.method-combination.net/qemu/altivec-patches/

For easier viewing of the whole series, there's also one comprehensive
patch:

  http://www.method-combination.net/qemu/altivec-patches/0000-BIG-PATCH.patch

diffstat says:

 target-ppc/cpu.h       |   10 +
 target-ppc/helper.h    |  127 +++++++
 target-ppc/op_helper.c |  862 ++++++++++++++++++++++++++++++++++++++++++++++++
 target-ppc/translate.c |  416 +++++++++++++++++++++++
 4 files changed, 1415 insertions(+), 0 deletions(-)

Why only the integer instructions?  I originally wrote support for the
whole instruction set, but I did it in the days of dyngen.  So a
straight forward-port was out of the question.  The original patch also
sloppily used native float, rather than float32, and used C99 math
functions to implement some of the more exotic Altivec instructions.
Both of these decisions mean that some care has to be taken in porting
the floating-point instructions.

I figured it'd be better to push out the integer instructions now and
the floating-point instructions later, rather than waiting for some
unspecified time for full support.  (After all, target-ppc has done just
fine with only {l,st}vx{,l}, almost anything additional is an
improvement!)

Comments and/or commits on my behalf welcome. :)

-Nathan

Re: [Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[C.W. Betts]

[-- Attachment #1: Type: text/plain, Size: 1761 bytes --]

I get a 403 forbidden error when trying to access  http://www.method-combination.net/qemu/altivec-patches/
>
>
On Dec 13, 2008, at 12:08 PM, Nathan Froyd wrote:

> This patch series adds support for integer Altivec instructions to  
> QEMU,
> including element-wise loads and stores.  It's rather long (42
> individual patches), so to avoid patch-bombing the list, I've placed  
> the
> patch series at:
>
>  http://www.method-combination.net/qemu/altivec-patches/
>
> For easier viewing of the whole series, there's also one comprehensive
> patch:
>
>  http://www.method-combination.net/qemu/altivec-patches/0000-BIG-PATCH.patch
>
> diffstat says:
>
> target-ppc/cpu.h       |   10 +
> target-ppc/helper.h    |  127 +++++++
> target-ppc/op_helper.c |  862 +++++++++++++++++++++++++++++++++++++++ 
> +++++++++
> target-ppc/translate.c |  416 +++++++++++++++++++++++
> 4 files changed, 1415 insertions(+), 0 deletions(-)
>
> Why only the integer instructions?  I originally wrote support for the
> whole instruction set, but I did it in the days of dyngen.  So a
> straight forward-port was out of the question.  The original patch  
> also
> sloppily used native float, rather than float32, and used C99 math
> functions to implement some of the more exotic Altivec instructions.
> Both of these decisions mean that some care has to be taken in porting
> the floating-point instructions.
>
> I figured it'd be better to push out the integer instructions now and
> the floating-point instructions later, rather than waiting for some
> unspecified time for full support.  (After all, target-ppc has done  
> just
> fine with only {l,st}vx{,l}, almost anything additional is an
> improvement!)
>
> Comments and/or commits on my behalf welcome. :)
>
> -Nathan
>
>


[-- Attachment #2: Type: text/html, Size: 2422 bytes --]

Re: [Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[C.W. Betts]

[-- Attachment #1: Type: text/plain, Size: 1761 bytes --]

I get a 403 forbidden error when trying to access  http://www.method-combination.net/qemu/altivec-patches/
>
>
On Dec 13, 2008, at 12:08 PM, Nathan Froyd wrote:

> This patch series adds support for integer Altivec instructions to  
> QEMU,
> including element-wise loads and stores.  It's rather long (42
> individual patches), so to avoid patch-bombing the list, I've placed  
> the
> patch series at:
>
>  http://www.method-combination.net/qemu/altivec-patches/
>
> For easier viewing of the whole series, there's also one comprehensive
> patch:
>
>  http://www.method-combination.net/qemu/altivec-patches/0000-BIG-PATCH.patch
>
> diffstat says:
>
> target-ppc/cpu.h       |   10 +
> target-ppc/helper.h    |  127 +++++++
> target-ppc/op_helper.c |  862 +++++++++++++++++++++++++++++++++++++++ 
> +++++++++
> target-ppc/translate.c |  416 +++++++++++++++++++++++
> 4 files changed, 1415 insertions(+), 0 deletions(-)
>
> Why only the integer instructions?  I originally wrote support for the
> whole instruction set, but I did it in the days of dyngen.  So a
> straight forward-port was out of the question.  The original patch  
> also
> sloppily used native float, rather than float32, and used C99 math
> functions to implement some of the more exotic Altivec instructions.
> Both of these decisions mean that some care has to be taken in porting
> the floating-point instructions.
>
> I figured it'd be better to push out the integer instructions now and
> the floating-point instructions later, rather than waiting for some
> unspecified time for full support.  (After all, target-ppc has done  
> just
> fine with only {l,st}vx{,l}, almost anything additional is an
> improvement!)
>
> Comments and/or commits on my behalf welcome. :)
>
> -Nathan
>
>


[-- Attachment #2: Type: text/html, Size: 2422 bytes --]

Re: [Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[Nathan Froyd]

On Sat, Dec 13, 2008 at 04:37:41PM -0700, C.W. Betts wrote:
> I get a 403 forbidden error when trying to access  
> http://www.method-combination.net/qemu/altivec-patches/

Whoops, sorry about that.  Fixed now.  (Accessing files by name worked,
of course, but that only helped for the big patch.)

-Nathan

Re: [Qemu-devel] [PATCH] target-ppc: add integer Altivec instructions

[Thiemo Seufer]

Nathan Froyd wrote:
> This patch series adds support for integer Altivec instructions to QEMU,
> including element-wise loads and stores.  It's rather long (42
> individual patches), so to avoid patch-bombing the list, I've placed the
> patch series at:
> 
>   http://www.method-combination.net/qemu/altivec-patches/

Please patch-bomb instead, so we have patch and related discussion
(and signed-off lines) in the same mail archive.


Thiemo