This is the mail archive of the
mailing list for the binutils project.
Re: nop in load-delay-slot (on MIPS o32 PIC code)
At Tue, 20 Apr 2004 18:06:56 +0200 (CEST), Maciej W. Rozycki wrote:
> Fixing it shouldn't actually be that tough these days.
Cool. I'm glad to hear that it's gotten easier.
> The current way of handling variant frags (thanks, Richard) makes such
> changes to them quite easy to handle. That may mean your patch requires a
> major update, though.
Undoubtedly. As noted, we've already discarded it.
It sounds like it might be better for interested parties to just do it
over from scratch. "just look for nop." 8-)
> Hmm, I guess it's just a matter of adjusting exiting test cases for the
> -march=mips2 case. If there is no test case for a particular variation,
> then it should probably be added whether changing the code involved or
> not. ;-)
Well, really what you need is to make sure that there are both
with-nops and without-nops test cases, and then you should run them
for all architectures like some of the tests do.
But, this touches a **lot** of tests. E.g., there are a lot of tests
which load addresses off of GP, then use the results. That includes
the load-delay nop, and variants with and without should be tested.
Agreed about "should be." And it's nice when people improve the
testsuite w/o provocation. But unfortunately, (IMO) the onus is on
patch submitters to include appropriate/complete tests for their
changes. In this case, due to the existing state of the testsuite,
that may be substantial.