[Bug lto/87434] LTO corrupts code

2018-10-03 Thread kelvin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87434 kelvin at gcc dot gnu.org changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Resolution|-

[Bug lto/87434] LTO corrupts code

2018-09-26 Thread kelvin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87434 --- Comment #2 from kelvin at gcc dot gnu.org --- Thanks for this suggestion. I'll investigate further. (My intent was to advise the compiler that this new memory address expression computes the same memory location as the original memory addres

[Bug lto/87434] LTO corrupts code

2018-09-26 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87434 --- Comment #1 from Richard Biener --- + rtx new_mem = gen_rtx_MEM (GET_MODE (mem), derived_ptr_reg); + MEM_COPY_ATTRIBUTES (new_mem, mem); I think it's dangerous to use old MEMs attributes this way, esp. MEM_EXPR and fri