This is the mail archive of the gcc-patches@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]

Re: [tree-ssa] Fix bootstrap failure caused by tree-ssa-loop.c


In message <95BA529E-2C53-11D8-8CD9-000A95DA505C@dberlin.org>, Daniel Berlin wr
ites:
 >>
 >>
 >>>> You can see this lossage in action if you have the loop code enabled 
 >>>> and
 >>>> do a checking bootstrap with -O1.
 >>>
 >>> ??? It passed without any problems for me.  Interesting, I will have 
 >>> to
 >>> check whether I don't bootstrap with checking disabled, but I would be
 >>> very surprised if it was the case.
 >> Hmm, I may have mis-spoken.  Try with -O2 -fno-tree-pre, checking 
 >> enabled.
 >
 >Let me guess, PRE's rebuilding of the dominator tree masks this?
If PRE rebuilds the dominator tree, then you're probably right.  It was
about 2am when I finally figured out that the dominator tree didn't match
reality.  Needless to say I didn't investigate much further other than to
note it was something done in the loop discovery code.

jeff




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