Change in binutils-gdb[master]: [gdb] Fix gdb.dwarf2/amd64-entry-value-param.exp with -fPIE/-pie

Change in binutils-gdb[master]: [gdb] Fix gdb.dwarf2/amd64-entry-value-param.exp with -fPIE/-pie

[Tom de Vries (Code Review)]

Tom de Vries has posted comments on this change.

Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35
......................................................................


Patch Set 1:

Patch submitted here: https://sourceware.org/ml/gdb-patches/2019-08/msg00406.html


-- 
To view, visit https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35
To unsubscribe, or for help writing mail filters, visit https://gnutoolchain-gerrit.osci.io/r/settings

[review] [gdb] Fix gdb.dwarf2/amd64-entry-value-param.exp with -fPIE/-pie

[Tom Tromey (Code Review)]

Tom Tromey has posted comments on this change.

Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35
......................................................................


Patch Set 1:

(3 comments)

Thank you for the patch.

I had a few comments, see below.

https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1//COMMIT_MSG 
Commit Message:

https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1//COMMIT_MSG@34 
PS1, Line 34: relocation offset at the use sites in call_site_for_pc and
This seems like a good idea to me, and also the direction
I'd like us to head more generally, because it means that
data can be shared across objfiles (eventually).


https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1/gdb/dwarf2read.c 
File gdb/dwarf2read.c:

https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1/gdb/dwarf2read.c@14123 
PS1, Line 14123: 	    SET_FIELD_PHYSADDR (call_site->target, lowpc);
Other spots in dwarf2read do this weird thing where they
relocate the value, then call gdbarch_adjust_dwarf2_addr,
and then un-relocate the value.  Perhaps this should too.

... though in the long run, ideally, we would just change
gdbarch to make it clear that this can be called with
unrelocated addresses.


https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1/gdb/objfiles.c 
File gdb/objfiles.c:

https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/35/1/gdb/objfiles.c@1502 
PS1, Line 1502: find_objfile_section_for_unrelocated_addr (struct objfile *objfile,
This seems somewhat questionable to me.  I guess on ELF it
should work ok, but is there really a guarantee that the
sections cannot overlap in the unrelocated object file?

The current code just uses SECT_OFF_TEXT unconditionally, so
maybe just doing things that way would be fine, and avoid this
question.  (Or maybe the answer is "yes" and we don't have to
worry -- I do not know.)