[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
CVS: cvs.openbsd.org: src
- To: source-changes_(_at_)_cvs_(_dot_)_openbsd_(_dot_)_org
- Subject: CVS: cvs.openbsd.org: src
- From: "Constantine A. Murenin" <cnst_(_at_)_cvs_(_dot_)_openbsd_(_dot_)_org>
- Date: Tue, 12 Jun 2007 10:33:27 -0600 (MDT)
CVSROOT: /cvs
Module name: src
Changes by: cnst_(_at_)_cvs_(_dot_)_openbsd_(_dot_)_org 2007/06/12 10:33:27
Modified files:
usr.bin/make : job.c
Log message:
Fix a 10-year-old pointer-arithmetic bug in make,
introduced on 1997-12-16, make/job.c revision 1.8.
This should fix the CPU spinout bug when running make with a -j option --
from now on, `make -j24` successfully completes the kernel build on my
Core 2 Duo, instead of halting the build and eating one CPU until ^C.
Found and debugged by myself, with the help of malloc.conf options.
"definitely a bug, nice catch" beck@
"let's also fix a [potential] memory leak in realloc" by millert@
"should go in" deraadt@
Visit your host, monkey.org