Re: [PR89528] reset debug uses of return value when dropping dead RTL call

2019-04-11 Thread Richard Biener
ted to the trick you used in pr89892.c, of using a more complex >expression to get UNSUPPORTED results instead of FAILs when the call >and >thus the variable are optimized out. How's this? Looks good! Thanks, Richard. > >[PR89528] reset debug uses of return value when droppin

Re: [PR89528] reset debug uses of return value when dropping dead RTL call

2019-04-11 Thread Alexandre Oliva
instead of FAILs when the call and thus the variable are optimized out. How's this? [PR89528] reset debug uses of return value when dropping dead RTL call When we remove an RTL call, we wouldn't clean up references to the return value of the call in debug insns. Make it so that we do.

Re: [PR89528] reset debug uses of return value when dropping dead RTL call

2019-04-05 Thread Richard Sandiford
Alexandre Oliva writes: > When we remove an RTL call, we wouldn't clean up references to the > return value of the call in debug insns. Make it so that we do. > > Regstrapped on x86_64- and i686-linux-gnu. Ok to install? > > > for gcc/ChangeLog > > PR debug/89528 > * valtrack.c (dea

Re: [PR89528] reset debug uses of return value when dropping dead RTL call

2019-04-05 Thread Richard Biener
On Fri, Apr 5, 2019 at 9:05 AM Alexandre Oliva wrote: > > When we remove an RTL call, we wouldn't clean up references to the > return value of the call in debug insns. Make it so that we do. > > Regstrapped on x86_64- and i686-linux-gnu. Ok to install? Looks OK but can you adjust the testcase t

[PR89528] reset debug uses of return value when dropping dead RTL call

2019-04-05 Thread Alexandre Oliva
When we remove an RTL call, we wouldn't clean up references to the return value of the call in debug insns. Make it so that we do. Regstrapped on x86_64- and i686-linux-gnu. Ok to install? for gcc/ChangeLog PR debug/89528 * valtrack.c (dead_debug_insert_temp): Reset debug ref