egcs-1.1.2-pre3 sh-hms c-torture failure

Toshiyasu Morita tm@netcom.com
Wed Mar 31 23:54:00 GMT 1999


This bug appears to be in the regmove pass.

[root@robin test]# sh-hms-gcc -O2 -fomit-frame-pointer -finline-functions 
-g struct-ret-1.c
[root@robin test]# sh-hms-run a.out
X f(B,char,double,B):({0.1,{2,3,0}},'',0.3,{0.2,{5,4,3}}) <- a.i[x] wrong
X f(B,char,double,B):({0.1,{1,2,3}},'',0.3,{0.2,{5,4,3}}) (strings different)

[root@robin test]# sh-hms-gcc -O2 -fno-regmove -fomit-frame-pointer 
-finline-functions -g struct-ret-1.c
[root@robin test]# sh-hms-run a.out
X f(B,char,double,B):({0.1,{1,2,3}},'',0.3,{0.2,{5,4,3}}) <- a.i[x] ok
X f(B,char,double,B):({0.1,{1,2,3}},'',0.3,{0.2,{5,4,3}}) (strings identical)

Toshi

> 
> version: egcs-1.1.2-pre3
>    host: i386-linux (Red Hat 4.2)
>  target: sh-hms
> 
> I see one failure on c-torture-1.45 with the cross-compiler:
> 
> ERROR: struct-ret-1.c: compiled program got signal 45 with: 
> -DSTACK_SIZE=16 -O2 -fomit-frame-pointer -finline-functions
> 
> Toshi
> 
> 




More information about the Gcc-bugs mailing list