This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug middle-end/29519] [4.2 Regression] Bad code on MIPS with -fnon-call-exceptions



------- Comment #3 from pinskia at gcc dot gnu dot org  2006-10-19 23:16 -------
The difference between with and without -fnon-call-exceptions is:
insn 10 9 11 3 (set (reg/f:SI 196)
        (mem/s/f/j:SI (reg/v/f:SI 194 [ t ]) [0 <variable>.f13+0 S4 A32])) -1
(nil)
    (nil))

vs:

(insn 10 9 11 3 (set (reg:SI 196)
        (mem/s/f/j:SI (reg/v/f:SI 194 [ t ]) [0 <variable>.f13+0 S4 A32])) -1
(nil)
    (nil))


Note in the case where it is set, we record 196 as a pointer so we know that
pointers don't wrap so we decide that pointer+1 can never be 0 in CSE.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29519


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]