Re: [Mesa-dev] [PATCH] nv50/ir: Check for valid insn instead of defs size

2016-03-09 Thread Pierre Moreau
On 06:10 PM - Mar 08 2016, Ilia Mirkin wrote: > Patch is fine, description is wrong (or at least inaccurate). > > The real issue is that function arguments have defs, but no defining > instruction. As a result, there's nothing to do when allocating > registers. This has nothing to do with $r0,

Re: [Mesa-dev] [PATCH] nv50/ir: Check for valid insn instead of defs size

2016-03-08 Thread Ilia Mirkin
Patch is fine, description is wrong (or at least inaccurate). The real issue is that function arguments have defs, but no defining instruction. As a result, there's nothing to do when allocating registers. This has nothing to do with $r0, but it does have something to do with the fact that nv50

Re: [Mesa-dev] [PATCH] nv50/ir: Check for valid insn instead of defs size

2016-03-08 Thread Samuel Pitoiset
This seems like correct. Reviewed-by: Samuel Pitoiset On 02/25/2016 02:03 AM, Pierre Moreau wrote: On Tesla cards, the first register $r0 contains the thread id; later generations use a specialised register for it. In order to prevent the register from being given

[Mesa-dev] [PATCH] nv50/ir: Check for valid insn instead of defs size

2016-02-24 Thread Pierre Moreau
On Tesla cards, the first register $r0 contains the thread id; later generations use a specialised register for it. In order to prevent the register from being given to anyone, and thus lose the thread id information, an lvalue is created to represent $r0 and is passed as an argument to the `main`