Support: Don't set RLIMIT_AS on child processes when applying a memory limit

It doesn't seem relevant to set an address space limit - this isn't
important in any sense that I'm aware & it gets in the way of things
that use a lot of address space, like llvm-symbolizer.

This came up when I realized that bugpoint regression tests were much
slower with -gsplit-dwarf than plain -g. Turned out that bugpoint
subprocesses (opt, etc) were crashing and doing symbolization - but
bugpoint runs those subprocesses with a 400MB memory limit. So with
plain -g, mmaping the opt binary would exceed the memory limit, fail,
and thus be really fast - no symbolization occurred. Whereas with
-gsplit-dwarf, comically, having less to map in, it would succeed and
then spend lots of time symbolizing.

I've fixed at least the critical part of bugpoint's perf problem there
by adding an option to allow bugpoint to disable symbolization. Thus
improving the perfromance for -gsplit-dwarf and making the -g-esque
speed available without this quirk/accidental benefit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@305242 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Support/Unix/Program.inc b/lib/Support/Unix/Program.inc
index 7d3537e20727e275466d339bc6083451f140cef1..2df0eaff47e52a507e3087aa28deb372d7b91612 100644 (file)
--- a/lib/Support/Unix/Program.inc
+++ b/lib/Support/Unix/Program.inc
@@ -163,16 +163,6 @@ static void SetMemoryLimits (unsigned size)
   r.rlim_cur = limit;
   setrlimit (RLIMIT_RSS, &r);
 #endif
-#ifdef RLIMIT_AS  // e.g. NetBSD doesn't have it.
-  // Don't set virtual memory limit if built with any Sanitizer. They need 80Tb
-  // of virtual memory for shadow memory mapping.
-#if !LLVM_MEMORY_SANITIZER_BUILD && !LLVM_ADDRESS_SANITIZER_BUILD
-  // Virtual memory.
-  getrlimit (RLIMIT_AS, &r);
-  r.rlim_cur = limit;
-  setrlimit (RLIMIT_AS, &r);
-#endif
-#endif
 #endif
 }