Subject: Re: ld: symbol __PROCEDURE_LINKAGE_TABLE_ remains undefined
To: None <>
From: Matt Thomas <>
List: tech-toolchain
Date: 08/27/1998 15:51:22
At 03:25 PM 8/27/98 , Matt Thomas wrote:
>At 03:18 PM 8/27/98 , Perry E. Metzger wrote:

>>Wacky question:
>>As we are trying to terminate use of the current toolchain at some
>>point, and would like to eventually migrate to ELF, why not make the
>>vax use ELF/ELF shared libs, instead of hacking on a new a.out shared
>>lib thing, when we have the glory (on the vax) of not needing to be
>>backward compatible with an old shared library format?
>The difference between writing ~200-300 lines of code and 4000-5000
>(if not more) lines.  I don't have the time for the latter and the
>former I can do while at the IETF.

The gas VAX assembler is not a BFD assembler so I'd need to convert
it to BFD (a requirement for doing ELF).  Looking at the i386 support
for BFD, this could be a lot work.

I then would need to write a elf-vax.[ch] and add VAX-specific BFD
relocation types.  A good thing would be that I could add 
specific relotypes for JMP/JSB and CALLS/CALLG opcodes (since they
need different plt types).  The bad thing is that this is a lot of
work.  And I hate working on the BFD code.

And then you need to hack ld to do the "right" thing.

I agree the above would all be a great step forward but I just don't
have the time to do it.  I have binutils 2.9.1 gas to emit PIC code
for VAX.  I have a 2.9.1 disassembler for VAX (well, most of one).
Once the a.out shared library diffs get merged, it's about an hour
to merge in the changes required for VAX support.  

So the question is, you can either have a.out VAX shared libs in a 
week and eventually have ELF VAX shared libs someday or wait 
(possibly forever) for the ELF VAX shared libs to appear.

Considering the limited memory and general slowness of most VAXen,
the sooner shared libraries can be used the better (even if they 
are a.out).

Matt Thomas               Internet:
3am Software Foundry      WWW URL:
Sunnyvale, CA             Disclaimer: I avow all knowledge of this message