This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: MIPS gas is broken now.


Daniel Jacobowitz wrote:
> On Mon, Sep 02, 2002 at 04:17:58PM +0200, Thiemo Seufer wrote:
> > Maciej W. Rozycki wrote:
> > > On Mon, 2 Sep 2002, Thiemo Seufer wrote:
> > > 
> > > > AFAICS the embedded PIC tests shouldn't be checked for mips*-linux
> > > > targets. Maybe they should apply only for mips*-elf targets, but
> > > > I'm not sure about this. Chris, what's your opinion about this idea?
> > > 
> > >  They are handy for checking if a generic MIPS change doesn't break
> > > embedded targets.
> > 
> > It wasn't a generic change but a change to mips*-linux targets.
> > It doesn't affect toolchains for embedded targets.
> 
> ... and (if I remember correctly) the bug actually affected more than
> just empic; quite a bit more.

I haven't seen such a case yet (but maybe haven't looked close enough).
It would mean the 32 bit only ELF MIPS targets are broken without
--enable-64-bit-bfd. Do you have a testcase?

> It's just that it didn't show up in the
> testcase otherwise.  64-bit bfd_vma's are necessary for much of the
> MIPS stuff to work.

I don't believe 32 bit only code handling actually needs 64 bit wide
values. Things like signedness errors may lead to wrong code with 32
bit BFD, but these are IMHO bugs which should be fixed.


Thiemo


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]