On 4/12/07, Philippe Gerum <[EMAIL PROTECTED]> wrote: > Saving I0-I3 in __mcount "seems" to fix the current issue, but I'm not > sure this is the proper way to address this problem. Specifically, > should the mcount call be emitted before any scratch reg is initialized > instead, or is mcount() an exception to the common ABI rules (in which > case, are there any other scratch regs to save here)? > Yes. It's the proper way. GCC doesn't know there is a call to _mcount in the inserted instrument code when -pg. So _mcount need save and restore caller saved registers besides callee saved ones.
Jie _______________________________________________ Uclinux-dist-devel mailing list [EMAIL PROTECTED] http://blackfin.uclinux.org/mailman/listinfo/uclinux-dist-devel
