This is the mail archive of the mailing list for the GDB 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: [patch] Set calling convention of methods

>>>>> "Jonas" == Jonas Maebe <> writes:

Tom> Yeah, this is a bit gross.  AFAICT the calling convention field only
Tom> takes values from dwarf.  So, I suppose the question is how you plan
Tom> to use this information in later patches.  If you check the calling
Tom> convention in general code then I suppose we will need some generic
Tom> set of values here.

Jonas> It is already used in general code (well, in non-DWARF-specific code),
Jonas> see the use of DW_CC_GNU_renesas_sh in sh-tdep.c

Yeah, I saw that.  I suspect that this means that the SH maintainer
knows that DWARF is the only debug format in use on the SH.

"dwarf2.h" is not included in many places in gdb, and not by any
header.  So, assuming your uses of this field will end up in truly
generic code, I think the thing to do is introduce a new enum, and
then translate values in the DWARF reader.  It is tempting to simply
reuse the DWARF enum internally, but I think that goes against the GDB


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