-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/06/11 12:39, Steven Bosscher wrote: > On Wed, Apr 6, 2011 at 8:37 PM, Steven Bosscher <stevenb....@gmail.com> wrote: >> On Wed, Apr 6, 2011 at 8:15 PM, Jeff Law <l...@redhat.com> wrote: >> >>>> AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only >>>> affects RTL. >>> But the elimination of the PHI results in creating RTL that is inserted >>> on a CFG edge. >> >> Yes, but gimple_expand_cfg() calls find_many_sub_basic_blocks(), and >> that should be enough, no?? >> >> /me goes back to trying to understand this code :-) > > Could you please add an explanation to the PR about how that PHI > results in a loop on an edge? My fantasy is not big enough to > visualize any case where that can happen! Presumably it's the vector initialization. THere's a fair amount of backend goop that comes into play. Peek at expand_set_or_movmem_via_loop.
jeff -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJNnL/yAAoJEBRtltQi2kC7LfoH/jWh+8b77bjuLQnVyBrB8Naj UWOPmPQHzs6DT1THXz8ef+pS3Bvuhbm+RxnHWhUTJQ9qBYwCf2oXzZDgytadfGo7 PKEtRpJTe4z7dwTGvp6UUX16TEI29OLHeNyyiDdEQ2ryCHJaSYB1MC8PEANilaHW uYAvTOkLbk6ORjx06pleVGy0IJW1UwLeQoJ2ggZvvmPZz8NghAWuvdfVkoX409wo vjOJ/EnDI603zKazh8yLI0c1K+jZNjnqqlxM8kC3GSt1lJt0LSO5vKW47H0E4zf0 rXhfj5WTiNJ0b0QjGGreIbKjUT8HLPjTJe3gurqgS25R2NtWCv9zxkYFu7sCGPo= =2tG9 -----END PGP SIGNATURE-----
