summaryrefslogtreecommitdiffstats
path: root/src/lib/gssapi/generic/gssapi.h
diff options
context:
space:
mode:
authorTheodore Tso <tytso@mit.edu>1995-04-18 21:51:11 +0000
committerTheodore Tso <tytso@mit.edu>1995-04-18 21:51:11 +0000
commitca05cace316acdb0ea7677a6ddc83610697f2b24 (patch)
tree40a451eb8760024fc2911f1f27e3121dc1c18a6a /src/lib/gssapi/generic/gssapi.h
parent8cb7facb419fdbe46ad337b8b62d359a7a87c6a4 (diff)
downloadkrb5-ca05cace316acdb0ea7677a6ddc83610697f2b24.tar.gz
krb5-ca05cace316acdb0ea7677a6ddc83610697f2b24.tar.xz
krb5-ca05cace316acdb0ea7677a6ddc83610697f2b24.zip
kpropd uses fork instead of vfork, because it's closing file
descriptors in the child. Print out the pid of the child process if debugging is enabled. Use _exit() instead exit() after a fork(), so that child process doesn't cause the stdio buffers don't get flushed twice. When debugging is enabled and kpropd can't bind to the port, it should try binding again after setting SO_REUSEADDR on the socket. This avoids problems when debugging modifications to kprop/kpropd or its setup, when kpropd is run twice quickly in succession. kpropd shouldn't signal(SIGCHLD, SIG_IGN) in the parent. Instead, it should wait() until the child exits. This was causing problems because setting SIGCHLD to SIG_IGN under SYSV causes child processes to get reaped automatically, so the wait() for kdb5_edit was failing. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@5368 dc483132-0cff-0310-8789-dd5450dbe970
Diffstat (limited to 'src/lib/gssapi/generic/gssapi.h')
0 files changed, 0 insertions, 0 deletions