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: dlltool -N [PATCH]


Shaun Jackman wrote:

> <http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/binutils/dlltool.c?rev=1.53&content-type=text/x-cvsweb-markup&cvsroot=src>
> 
>>at asm_prefix(), note that ARM specifically does not have any "ASM
>>prefix," unlike i386.  Is this code inaccurate?  Or does the APCS-32 ABI
>>call for an underscore when arm-pe, in general, does not?  I was unable
>>to find any authoritative documentation on this.  Does someone know?
> 
> I don't know what the ABI says, but gcc uses uses an underscore asm_prefix:

For the record, I don't have any objection with this patch, provided it
also does the right thing for the __declspec(dllimport) __imp_ names.
(This patch doesn't add an underscore to those names, but it it might
need to--I don't know.)

I think the GCC used here is doing the wrong thing, because it doesn't
match the ABI used by binutils, which seems to be right, because its
also what the official Microsoft toolchain does.  I beleive there is a
single macro that can be used to toggle whether GCC emits the leading
underscore.

However, since obviously there are already some compilers that are doing
the wrong thing as far as the ABI is concerned, it seems like a good
thing to have a 'quick fix' option like this, similar to --kill-at and
similar.  -N might be useful in other situations.

I think that if a maintainer does not have some other sort of problem
with this patch, it should be approved.

Aaron W. LaFramboise


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