This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
RE: [committed] MIPS16/GAS: Fix delay slot filling across frags
- From: "Maciej W. Rozycki" <macro at imgtec dot com>
- To: Matthew Fortune <Matthew dot Fortune at imgtec dot com>
- Cc: "binutils at sourceware dot org" <binutils at sourceware dot org>
- Date: Thu, 30 Jun 2016 15:44:18 +0100
- Subject: RE: [committed] MIPS16/GAS: Fix delay slot filling across frags
- Authentication-results: sourceware.org; auth=none
- References: <6D39441BF12EF246A7ABCE6654B023537E448CF7 at HHMAIL01 dot hh dot imgtec dot org> <alpine dot DEB dot 2 dot 00 dot 1606211958240 dot 7182 at tp dot orcam dot me dot uk> <6D39441BF12EF246A7ABCE6654B023537E46634C at HHMAIL01 dot hh dot imgtec dot org>
On Thu, 30 Jun 2016, Matthew Fortune wrote:
> > I have committed the change below then, with just a minor update to
> > your proposed code change to have the declaration and statements
> > separated from each other in the piece of code affected. I decided your
> > test case was insufficient though, as I think we want to make sure not
> > only that code assembles successfully, but also that the same code is
> > produced with and without a listing requested as well, and that both 16-
> > bit and 32-bit
> > MIPS16 jump instructions are handled correctly across frags.
>
> Thanks Maciej. The extended test cases are definitely a good idea, this
> area seems fragile but fewer special cases in the code will be better.
If you check how this code looked like as at 1e91584932ef^, then you
may actually understand why whoever wrote this originally decided to
keep the MIPS16 case separate and not to bother wiring it in to the
general case. It was a very good clean-up indeed made by Richard there.
Maciej