This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: reload problem:

I meant gcc does not look at the condition of nameless insn patterns.

If it does, why reload fails then to satisfy "0" ?


Alexandre Oliva wrote:

>On Jan 22, 2002, Dmitry <diwil@mail.ru> wrote:
>
>>I missed a fact, that during instruction recognition, gcc does not look at the condition.
>>
>
>It does.  It doesn't look at constraints, though.  Is this what you
>meant?  I'll assume so henceforth.
>
>>So, during reload pass it faces with a condition which cannot be satisfied.
>>
>
>The `0' constraint can be satisfied by reload: it suffices to save the
>input register elsewhere, unless it's dead after that instruction, and
>then store the output register in the intended output location.
>
Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]