[GOLD] Got_offset_list: addend field

[Alan Modra]

On Thu, Sep 02, 2021 at 06:13:57PM -0700, Cary Coutant wrote:
> > This is the first in a series of patches aimed at supporting GOT
> > entries against symbol plus addend generally for PowerPC64 rather than
> > just section symbol plus addend as gold has currently.
> >
> > This patch adds an addend field to Got_offset_list, so that both local
> > and global symbols can have GOT entries with addend.  Note the FIXME
> > in incremental.cc, which will need attention if powerpc is to ever
> > support incremental linking.
> 
> Do you think you'd ever want incremental linking on powerpc?

No, I don't think so.  And I'll argue against it if someone tries to
talk me into implementing it.  :-)

> Frankly,
> the effort for just the one target platform was pretty high, the
> maintenance on it is burdensome, and I'm tempted to deprecate it and
> rip it out at some point in the future.
> 
> > As per the recommendation in
> > https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#Rf-default-args
> > the changes to symtab.h Symbol methods could be done with what looks
> > like a good use of default arguments but they seem to be against the
> > gold maintainers' coding practice.  So instead we get more overloads.
> 
> For the record, I've got nothing against default parameter values, as
> long as they're obvious default values that shouldn't ever need to be
> changed. I'm not sure what Ian's position was,

There's this:
https://sourceware.org/pipermail/binutils/2010-October/069026.html

> but he did use a
> default parameter value in at least one place (Descriptors::open in
> descriptors.h).

Heh.

> OK, with or without your suggestion to use a default parameter value
> for has_got_offset.

Thanks.  I wrote the patch initially with default parameter values so
it will be easy to go back to that.

-- 
Alan Modra
Australia Development Lab, IBM