This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR91126
- From: Jeff Law <law at redhat dot com>
- To: Richard Biener <rguenther at suse dot de>, gcc-patches at gcc dot gnu dot org
- Date: Wed, 10 Jul 2019 08:11:28 -0600
- Subject: Re: [PATCH] Fix PR91126
- References: <alpine.LSU.2.20.1907101313280.2976@zhemvz.fhfr.qr>
On 7/10/19 5:15 AM, Richard Biener wrote:
>
> The following patch fixes an old (but now manifesting in a testsuite
> fail) issue with value-numbering on big-endian machines. I've
> checked the VN results after the patch on aarch64 with -mbig-endian.
>
> Bootstrap and regtest running on x86_64-unknown-linux-gnu.
>
> I'll commit after this succeeded.
>
> Richard.
>
> 2019-07-10 Richard Biener <rguenther@suse.de>
>
> PR tree-optimization/91126
> * tree-ssa-sccvn.c (n_walk_cb_data::push_partial_def): Adjust
> native encoding offset for BYTES_BIG_ENDIAN.
> (vn_reference_lookup_3): Likewise.
>
> * gcc.dg/torture/pr91126.c: New testcase.
FWIW, the easiest big endian system to test is ppc64 as there's a
machine in the build farm. It gets a bootstrap & regression cycle
daily. Including the glibc & kernel build it takes ~4hrs.
My tester will also bootstrap aarch64_be, m68k and armeb, but I think
the ppc64 tester is the most effective and you can debug natively if
something goes wrong. Debugging gcc via the qemu stub is, umm, painful.
jeff