[patch, mips, stabs] Fix stabs failures in GCC testsuite
Maciej W. Rozycki
macro@codesourcery.com
Tue May 7 12:52:00 GMT 2013
On Sat, 4 May 2013, Richard Sandiford wrote:
> >> because I think all this is really papering over the problem of forgetting
> >> the label state across a section switch. This happens implicitly for
> >> these .stab directives (that are also seldom used these days), but I
> >> reckon I saw a case where it mattered for explicit section switches and
> >> had to be worked around somehow. It would be good to have some thought as
> >> to how to do this properly some day.
> >
> > However after some further thinking I believe a good less-effort solution
> > for GCC might be to produce a .insn directive after any label considered a
> > code entry point, where GAS supports the pseudo-op. This would avoid
> > surprises in case something else turned out to clear the instruction mode
> > again. What do you think?
>
> I think that's going in the wrong direction. If we find a specific
> "real-world" case that isn't being handled correctly, we should fix
> it in GAS, so that it works with both compiler and hand-written code.
Maybe you're right after all.
The problem with section switches dropping the labels is a bit painful
though -- I have now recalled the case where .insn had to be used to adapt
previously working code for microMIPS assembly and work this limitation
around was the Linux kernel.
Maciej
More information about the Binutils
mailing list