This is the mail archive of the gcc-help@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: Compiler options for PowerPC


Hi,

Segher Boessenkool <segher@kernel.crashing.org> wrote:

[...]


See Ian's answer.

Ian is right. I just looked at the disassembly of the binary a little closer and saw that the dcbt instruction occurs in the strncmp function of the glibc. I looked at this function in glibc source and it turned out strncmp is implemented in assembler code and the dcbt instruction is used.

It probably would be best for you to get the simulator fixed
anyway; since the simulator doesn't seem to simulate the caches
at all (or it would definitely have to support the dcbX insns!),
it can just "nop" that instruction.  Should be really easy to
add to the simulator.

Your right too. The best way would be to fix the simulator. But I'm currently just investigating if the simulator is suitable for my needs and I don't want to put too much effort into rewriting the simulator at the moment. So may be I'll just replace that instruction by a nop, either in the binary itself or in the strncmp source or in the decoding unit of the simulator.

You helped me a lot tracking down the problem. Thanks!

Best regards
Jakob



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]