This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils 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: PowerPC gas -many


I understand all this. I guess my question is as a developer trying what is the rule of thumb with regards to -many?

For example, which mftb should work with just 'as -many', the classic or book-e form?
Also, for disassembly does AltiVec or SPE or the macc* take precedence?


- kumar

On Tuesday, September 2, 2003, at 09:32 AM, Alan Modra wrote:

On Tue, Sep 02, 2003 at 08:38:47AM -0500, Kumar Gala wrote:
I'm also concerned about the use of -many. For example, between Book-E
and classic PPC mftb changes encoding. I'm not sure what expectation
users have today with -many, but it may be more confusion when
conflicts arise that the 'default' behavior is not what they expect
based on their focus.

With the patch I posted, you can set the default. eg. as -m440 -many.


Also, disassembly is a nightmare with -many.

No one is forcing you to use -Many. It's a new option for the disassembler.

I'm pretty sure between AltiVec / SPE / macc* there are over lapping
opcodes and which one should get precedences?

Yes, it's true there are overlapping opcodes. If you know you are disassembling SPE code, use -Me500. Or -Me500 -Many to get SPE plus others, with SPE taking precedence.

--
Alan Modra
IBM OzLabs - Linux Technology Centre


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