[Mesa-dev] [PATCH 1/2] nir: Use alloca instead of variable length arrays.
kenneth at whitecape.org
Fri Feb 27 11:51:01 PST 2015
On Friday, February 27, 2015 02:04:47 PM Jose Fonseca wrote:
> On 26/02/15 18:07, Brian Paul wrote:
> > AFAIK, there's no zeroing version of alloca().
> Yes, I also searched. And unfortunately inline functions can't be used
> due to alloca semantics. I'm not sure if there's any C-preprocessor magic.
> Anyway, depending on the maximum size of these alloca arrays, we might
> actually need to use malloc, as Windows stack is 1MB by default, plus
> the OpenGL driver only gets to use whatever stack is left from the
> If using C++ on NIR is acceptable, std::vector<> would be a nice
> solution overall: practically the same syntax/semantics as C99 VLAs,
> plus no risk of stack overflow.
We're hoping to keep it C99 - malloc and freeing them is probably the
best plan. Thanks again!
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 819 bytes
Desc: This is a digitally signed message part.
More information about the mesa-dev