This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH,rs6000] fix interrupt safety issue on E500 targets
On Fri, Oct 05, 2007 at 06:00:44PM +0930, Alan Modra wrote:
> So, another fix for the SPE problem is to use frame_reg_rtx in the SPE
> support code instead of inventing another rtx for r11. That's what I
> set out to do in the following patch, but then noticed a few other
> * config/rs6000/rs6000.c (rs6000_emit_epilogue): Correct
> altivec sp_offset. Rearrange sp_offset assignments to
> correspond to stack adjustments. Use frame_reg_rtx for
> SPE register restores. Correct SPE stack adjustment.
> A powerpc-linux bootstrap/regtest is still in progress. I don't have
> access to SPE hardware to properly test the SPE changes, so would
> someone mind regression testing this for me on SPE?
Bootstrap has proceeded without problems, regtesting underway.