qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH] TCG: Add preprocessor guards for optional tcg o


From: Alexander Graf
Subject: Re: [Qemu-devel] [PATCH] TCG: Add preprocessor guards for optional tcg ops
Date: Thu, 11 Aug 2011 14:36:52 +0200

On 11.08.2011, at 14:28, Avi Kivity wrote:

> On 08/11/2011 02:40 PM, Alexander Graf wrote:
>> On 11.08.2011, at 13:24, malc wrote:
>> 
>> >  On Thu, 11 Aug 2011, Alexander Graf wrote:
>> >
>> >>  While compiling current HEAD on a ppc64 box, I was confronted with the
>> >>  following compile errors:
>> >>
>> >>   tcg/optimize.c: In function ?tcg_constant_folding?:
>> >>   tcg/optimize.c:546: error: ?INDEX_op_not_i32? undeclared (first use in 
>> >> this function)
>> >>   tcg/optimize.c:546: error: (Each undeclared identifier is reported only 
>> >> once
>> >>   tcg/optimize.c:546: error: for each function it appears in.)
>> >>   tcg/optimize.c:546: error: ?INDEX_op_not_i64? undeclared (first use in 
>> >> this function)
>> >>   tcg/optimize.c:573: error: ?INDEX_op_ext32u_i64? undeclared (first use 
>> >> in this function)
>> >>   make[1]: *** [tcg/optimize.o] Error 1
>> >>
>> >>  Obviously, the optimize.c tries to use TCG opcode constants that are 
>> >> optional
>> >>  and thus not defined in some targets, such as ppc64.
>> >>
>> >>  This patch guards them with the proper #ifdefs, so compilation works 
>> >> again.
>> >>
>> >>  Signed-off-by: Alexander Graf<address@hidden>
>> >>  ---
>> >>  tcg/optimize.c |    8 +++++++-
>> >>  1 files changed, 7 insertions(+), 1 deletions(-)
>> >>
>> >>  diff --git a/tcg/optimize.c b/tcg/optimize.c
>> >>  index 7eb5eb1..7b4954c 100644
>> >>  --- a/tcg/optimize.c
>> >>  +++ b/tcg/optimize.c
>> >>  @@ -543,7 +543,11 @@ static TCGArg *tcg_constant_folding(TCGContext *s, 
>> >> uint16_t *tcg_opc_ptr,
>> >>              gen_args += 2;
>> >>              args += 2;
>> >>              break;
>> >>  +#if ((TCG_TARGET_REG_BITS == 32)&&  defined(TCG_TARGET_HAS_not_i32)) || 
>> >> \
>> >>  +    ((TCG_TARGET_REG_BITS == 64)&&  defined(TCG_TARGET_HAS_not_i32)&&  \
>> >>  +     defined(TCG_TARGET_HAS_not_i64))
>> >>          CASE_OP_32_64(not):
>> >>  +#endif
>> >>  #ifdef TCG_TARGET_HAS_ext8s_i32
>> >>          case INDEX_op_ext8s_i32:
>> >>  #endif
>> >>  @@ -568,8 +572,10 @@ static TCGArg *tcg_constant_folding(TCGContext *s, 
>> >> uint16_t *tcg_opc_ptr,
>> >>  #ifdef TCG_TARGET_HAS_ext16u_i64
>> >>          case INDEX_op_ext16u_i64:
>> >>  #endif
>> >>  -#if TCG_TARGET_REG_BITS == 64
>> >>  +#if (TCG_TARGET_REG_BITS == 64)&&  defined(TCG_TARGET_HAS_ext32s_i64)
>> >>          case INDEX_op_ext32s_i64:
>> >>  +#endif
>> >>  +#if (TCG_TARGET_REG_BITS == 64)&&  defined(TCG_TARGET_HAS_ext32u_i64)
>> >>          case INDEX_op_ext32u_i64:
>> >>  #endif
>> >>              if (temps[args[1]].state == TCG_TEMP_CONST) {
>> >>
>> >
>> >  Or (not even compile tested):
>> 
>> Sure, but then the next target would pop up with optionals not implemented. 
>> The real alternative would be to make these ops non-optional :)
>> 
> 
> Or to have automatic generation of the optionals based on the primitives, if 
> the optionals are not present.

That's what's happening in the background already, no? The line is towards tcg 
users though, not it tcg internal code.


Alex




reply via email to

[Prev in Thread] Current Thread [Next in Thread]