Re: [testsuite]: Move pr45830.c to gcc.dg/torture and make it pass for AVR

2011-12-12 Thread Mike Stump
On Dec 7, 2011, at 3:59 AM, Georg-Johann Lay wrote: > In principle, test case pr45830.c works for target avr, but there is an issue > with the -ftree-switch-conversion optimization activated at higher > optimization > levels: It transforms code size into .data usage and thus exceeds AVRs' RAM > si

Ping #1: [testsuite]: Move pr45830.c to gcc.dg/torture and make it pass for AVR

2011-12-12 Thread Georg-Johann Lay
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00474.html Georg-Johann Lay wrote: > In principle, test case pr45830.c works for target avr, but there is an issue > with the -ftree-switch-conversion optimization activated at higher > optimization > levels: It transforms code size into .data usage an

[testsuite]: Move pr45830.c to gcc.dg/torture and make it pass for AVR

2011-12-07 Thread Georg-Johann Lay
In principle, test case pr45830.c works for target avr, but there is an issue with the -ftree-switch-conversion optimization activated at higher optimization levels: It transforms code size into .data usage and thus exceeds AVRs' RAM size because of big CSWTCH lookup tables located in RAM. The pat