[Bug general/24498] 0.176: isn't LTO ready

["mark at klomp dot org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=24498

Mark Wielaard <mark at klomp dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

--- Comment #12 from Mark Wielaard <mark at klomp dot org> ---
(In reply to Martin Liška from comment #9)
> (In reply to Mark Wielaard from comment #8)
> > I proposed a patch to use the new gcc 10 symver attribute to define symbol
> > versioning:
> > 
> > https://sourceware.org/pipermail/elfutils-devel/2020q2/002606.html
> > 
> > This should help with using LTO since it makes the function symbols and
> > versions visible to the compiler (before they were "hidden" inside asm
> > statements).
> 
> There's an ongoing effort that will extend .symver GAS syntax:
> https://sourceware.org/bugzilla/show_bug.cgi?id=25295
> 
> Later then, the GCC will utilize the new syntax.

Interesting, thanks for the reference. I also added a comment to the GCC bug
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=48200#c43 because the current
syntax (not supporting triple @ variant) makes things a bit cumbersome.

This is partly because the interaction between linker version script and
assembler .symver isn't really documented fully. The assembler .symver should
override what is in the linker version script, but it doesn't fully. In
particular, as shown by this hack, the linker might still add a (default)
version to a symbol, even if the assembler already added a version.

-- 
You are receiving this mail because:
You are on the CC list for the bug.