boost/build
2
0
Fork 0
mirror of https://github.com/boostorg/build.git synced 2026-02-13 12:22:17 +00:00
Code Issues Projects Releases Wiki Activity

Fix a bug Chris Cambly reported with the timeout code on

Browse Source 
AIX.  Apparently AIX doesn't permit a forked process to 
reference (set) memory in the parent's address space.  No 
other system seems to object to this practice but it taught
me a lesson!

The fix was to move the call to get the child process start 
time directly before calling vfork.  This isn't really fair 
to the forked process as we start counting time against the
child process that we haven't even forked (we count the 
vfork/exec call overhead against the child process).

Tested Rene's test.sh script on Sun, Linux, AIX, and Sgi.



[SVN r39399]
This commit is contained in:
K. Noel Belcourt
2007-09-20 04:11:11 +00:00
parent 733d202e9d
commit e50a11396b
1 changed files with 18 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

33
src/engine/execunix.c
View File

@@ -63,7 +63,6 @@ static clock_t tps = 0;
static struct timeval tv;
static int intr = 0;
static int cmdsrunning = 0;
static void (*istat)( int );
#define OUT 0
#define ERR 1
@@ -161,8 +160,6 @@ execcmd(
argv[3] = 0;
}
if (tps == 0) tps = sysconf(_SC_CLK_TCK);
/* increment jobs running */
++cmdsrunning;
@@ -186,6 +183,24 @@ execcmd(
/* Start the command */
if (0 < globs.timeout) {
struct tms buf;
/*
* handle hung processes using different mechanism
* manually track elapsed time and signal process
* when time limit expires
*
* could use this approach for both consuming too
* much cpu and hung processes, but it never hurts
* to have backup
*/
cmdtab[ slot ].start_time = times(&buf);
/* make a global, only do this once */
if (tps == 0) tps = sysconf(_SC_CLK_TCK);
}
if ((cmdtab[slot].pid = vfork()) == 0)
{
close(out[0]);
@@ -204,7 +219,6 @@ execcmd(
/* terminate processes only if timeout is positive */
if (0 < globs.timeout) {
struct rlimit rl;
struct tms buf;
/*
* set hard and soft resource limits for cpu usage
@@ -213,17 +227,6 @@ execcmd(
rl.rlim_cur = globs.timeout;
rl.rlim_max = globs.timeout;
setrlimit(RLIMIT_CPU, &rl);
/*
* handle hung processes using different mechanism
* manually track elapsed time and signal process
* when time limit expires
*
* could use this approach for both consuming too
* much cpu and hung processes, but it never hurts
* to have backup
*/
cmdtab[ slot ].start_time = times(&buf);
}
execvp( argv[0], argv );