llvm dev - Mar 2016 - [PATCH] D15965: Add support for dumping relocations in non-relocatable files

It sounds like what you’re asking is, rather that universally calling
RelocationRef::getOffset inside llvm-objdump.cpp

I should:

*        Check if Obj “isRelocatableObject”

*        If it is, call RelocationRef::getOffset()

*        If it isn’t

o   Call RelocationRef::getAddress()

o   Build an ordered map of all sections and their bounds, check if the
relocation lands within a section

o   Subtract the section address from the relocation address

 

Since getAddress isn’t in RelocationRef at the moment, this is going to have to
be getAddress for MachO and COFF.

 

COFFObjectFile::getRelocationOffset returns R->VirtualAddress, so inside here
I’m going to have to find the section it belongs to, add it to the result so it
can later be subtracted off by llvm-objdump

 

MachOObjectFile::getRelocationOffset also returns getAnyRelocationAddress so
I’ll have to also find its containing section and add this in just so
llvm-objdump can later subtract it.

 

Can you detail why this is a better plan?

 

From: Rafael Espíndola [mailto:rafael.espindola at gmail.com] 
Sent: Thursday, March 24, 2016 3:13 PM
To: Colin LeMahieu <colinl at codeaurora.org>
Cc: llvm-commits <llvm-commits at lists.llvm.org>; Shankar Easwaran
<shankare at codeaurora.org>; Hemant Kulkarni <khemant at
codeaurora.org>; reviews+D15965+public+319b19c852e3c4e3 at reviews.llvm.org
Subject: RE: [PATCH] D15965: Add support for dumping relocations in
non-relocatable files

 


On Mar 24, 2016 4:08 PM, "Colin LeMahieu" <colinl at codeaurora.org
<mailto:colinl at codeaurora.org> > wrote:
>
> llvm-objdump wants to print the relocationOffset, so it calls
RelocationRef::relocationOffset which calls ELFObjectFile::getRelocationOffset
>
> Since objdump wants the relocation offset, and this pipes down to
ELFObjectFile::getRelocationOffset, at what point is the address supposed to be
asked for?
>
In llvm-objdump. It can then search the sections to see if one happens to
include that address.
> -----Original Message-----
> From: Rafael Espíndola [mailto:rafael.espindola at gmail.com
<mailto:rafael.espindola at gmail.com> ]
> Sent: Thursday, March 24, 2016 2:45 PM
> To: reviews+D15965+public+319b19c852e3c4e3 at reviews.llvm.org
<mailto:reviews%2BD15965%2Bpublic%2B319b19c852e3c4e3 at reviews.llvm.org>
> Cc: colinl at codeaurora.org <mailto:colinl at codeaurora.org> ;
Hemant Kulkarni <khemant at codeaurora.org <mailto:khemant at
codeaurora.org> >; Shankar Easwaran <shankare at codeaurora.org
<mailto:shankare at codeaurora.org> >; llvm-commits <llvm-commits at
lists.llvm.org <mailto:llvm-commits at lists.llvm.org> >
> Subject: Re: [PATCH] D15965: Add support for dumping relocations in
non-relocatable files
>
> On 24 March 2016 at 15:35, Shankar Easwaran <shankare at codeaurora.org
<mailto:shankare at codeaurora.org> > wrote:
> > shankare added a reviewer: shankare.
> > shankare added a comment.
> >
> > http://www.sco.com/developers/gabi/latest/ch4.reloc.html
> >
> >   r_offset
> >   This member gives the location at which to apply the relocation
action. For a relocatable file, the value is the byte offset from the beginning
of the section to the storage unit affected by the relocation. For an executable
file or a shared object, the value is the virtual address of the storage unit
affected by the relocation.
> >
> > This shows that the assertion is incorrect, as the ELF ABI suggests
reading relocation records is valid in Executables as well.
> >
> > While relocation offset is not looked up for shared libraries from
> > .rela.dyn and .rela.plt sections, there is a usecase where relocation
> > sections are generated with the linker switch **--emit-relocs**
>
> And those records have addresses, not offsets. Therefore the assertion is
correct and you may *NOT* remove it.
>
> CHeers,
> Rafael
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160324/328587eb/attachment.html>

On 24 March 2016 at 16:35, Colin LeMahieu <colinl at codeaurora.org>
wrote:
> It sounds like what you’re asking is, rather that universally calling
> RelocationRef::getOffset inside llvm-objdump.cpp
>
> I should:
>
> ·        Check if Obj “isRelocatableObject”
>
> ·        If it is, call RelocationRef::getOffset()
>
> ·        If it isn’t
>
> o   Call RelocationRef::getAddress()
>
> o   Build an ordered map of all sections and their bounds, check if the
> relocation lands within a section
>
> o   Subtract the section address from the relocation address
>
That is the overall structure, yes. You do have to handle the case of
there being no section that covers the address.

I will take a look at the updated patch.

Cheers,
Rafael