------- Comment #2 from rguenther at suse dot de 2009-04-08 16:21 ------- Subject: Re: IPA should compute if a function may recurse
On Wed, 8 Apr 2009, hubicka at ucw dot cz wrote: > ------- Comment #1 from hubicka at ucw dot cz 2009-04-08 16:10 ------- > Subject: Re: New: IPA should compute if a function may recurse > > While testing for possible recurssion can definitly be added (and it > will return true on everything that might lead to call outsie of visible > unit), there is problem with using it on local statics. > > Inlining of function can distribute use of local static pretty much > everywhere, so value of origin of local static for optimization is quite > zero. Of course one can track what functions reffers to the given > variable and then one needs to track if function A referring to it might > call to any of the others referring to it direclty or indirectly. Hmm, but will the inlined function not still be visible as such in the cycle? So what I want to know is whether the local static is referred to in any function the function calls (thus, a more complete analysis would be performed by IPA mod-ref anyway). > Isn't this however exactly same as asking ipa-pureconst bit? No. The called function may read and clobber memory, I just want to be able to check if a local static variable "escaped" even if its address was not taken (thus, we recursed). Richard. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39689
