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]

Re: PATCH: new GC implementation

To: ak at muc dot de
Subject: Re: PATCH: new GC implementation
From: Mark Mitchell <mark at codesourcery dot com>
Date: Wed, 22 Sep 1999 08:03:30 -0700
Cc: samuel at codesourcery dot com, gcc-patches at egcs dot cygnus dot com
Organization: CodeSourcery, LLC
References: <14312.36444.325569.263933@coriander.orconet.com><m34sgnqpyz.fsf@fred.muc.de>

>>>>> "Andi" == Andi Kleen <ak@muc.de> writes:

    Andi> Erm. Of course Linux zeroes all pages it hands to user
    Andi> processes. Otherwise that would be a gapping security
    Andi> hole. So mapping /dev/zero and anonymous mappings should be
    Andi> equivalent, both will zero lazily. You can probably remove
    Andi> the special case for Linux.

Trust us; we ran profilers.  It makes a very significant difference.
I suspect that Linux is clever enough to hand back pages most recently
used by processes with the same UID or some such; there is definitely
a significant win.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

Follow-Ups:
- Re: PATCH: new GC implementation
  - From: Andi Kleen

References:
- PATCH: new GC implementation
  - From: Alex Samuel
- Re: PATCH: new GC implementation
  - From: Andi Kleen

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