Fix 2 moderately simple sparc PRs

David S. Miller davem@redhat.com
Sun Apr 28 21:07:00 GMT 2002


   From: Mark Mitchell <mark@codesourcery.com>
   Date: Sun, 28 Apr 2002 11:02:38 -0700
   
   Make it easier for me: tell me whether or not these are regressions.
   
   The changes look sensible, but we'll never stop here if we just keep
   making sensible changes.  Are their programs for which we used to get
   good performance, but now get bad performance?

I belive both are regressions.

Especially for the prefetch case.  I sent last night an email to the
person who added the Sparc prefetch support in and he replied to me
"Oh crap, it is indeed broken, sorry"

It looks like Jakub has duplicated my work on one these bugs
on the mainline since I posted these patches :(



More information about the Gcc-patches mailing list