Tom Stellard <t...@stellard.net> writes:

> Hi,
>
> This series improves clover's kernel argument handling by allowing drivers to
> specify an alignment for the input buffer and also by sign/zero extending the
> arguments when required.  In order to test these patches on r600g and 
> radeonsi,
> you will need to also apply the LLVM patches which were just sent to the list.
>
> -Tom

Hi Tom,

can you have a look at this branch? [1] It addresses some concerns I had
about this series.  I've dropped patches 1, 4 and 5, and rewritten 6 and
7.

I've removed the KERNEL_ARG_ALIGNMENT cap and pad arguments, instead all
the alignment information is obtained from the compiler, which
simplifies things considerably but I'm not really sure if it's going to
work for you.

The "noext" argument extension type seemed redundant so I dropped it
too.

I've also added support for byte-swapping for the case where host and
device have different endianness, and fixed all argument subclasses so
they can deal with arbitrary data sizes -- Letting the compiler specify
argument::__size isn't enough, we also needed some way to transform the
base data type to that size.

Thanks.

[1] http://cgit.freedesktop.org/~currojerez/mesa/log/

Attachment: pgplgfxui3w6p.pgp
Description: PGP signature

_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev

Reply via email to