This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Question about the .idata$ sections and their missing symbols
2009/1/5 Dave Korn <dave.korn.cygwin@googlemail.com>:
> Kai Tietz wrote:
>> 2009/1/4 Dave Korn:
>>> Kai Tietz wrote:
>>>
>>>> The issue is that those idata$<n> sections aren't necessarily in a
>>>> writable section. E.g. on WM 61 OS expects that IAT (idata$5) is
>>>> placed within the .data section and the rest is put into the .rdata
>>>> section.
>>> Won't --enable-auto-import fix that?
>>>
>> No, because WM 6.1 expects that the IAT is in .data section, but the
>> rest have to be in read-only section. Otherwise it simply rejects the
>> start. Auto-importing doesn't helps here.
>
> Ah, well then maybe you need a variant linker script that puts idata$5 into
> .data and puts the remaining idata$ sections into rdata?
>
> cheers,
> DaveK
>
No, the script isn't the problem. But I would like to get an anwser to
my question
"The point I am interested in is in in bfd/peXXigen.c. There is much code
about data directory " (the PE(+) Directory Entries) ... "filling
which references idata$<n> sections in " ... (function) ..." post link
script. This code seems to be dead for most pe(+) coff targets IIUC. I
came to this, while adding proper value for mingw of
the IAT directory. There are some comments, that there are pending
patches for this.
Maybe somebody know which patches these are, and if there is still a
developer pushing them?"
Cheers,
Kai
--
| (\_/) This is Bunny. Copy and paste
| (='.'=) Bunny into your signature to help
| (")_(") him gain world domination