Package: ppp Version: 2.4.3-20050321+2 Severity: normal
Pppd doesn't close the pty slave in the forked pty process/on exec, which in case of rp-pppoe (and in principle others, too, I guess) can cause it to hang in a write() to fd 1 (that's what I've seen, but read() on fd 0 would be possible just as well, I guess) when its pppd is long gone, waiting for itself to read the data from the slave fd, which, of course, won't happen. I didn't really look at the source, but I guess this doesn't allow any other explanation: $ ps axfu [...] root 30742 0.0 0.6 1500 432 ? S Sep18 0:00 /usr/sbin/pppoe -I eth1 -T 12 -M 00:00:CB:22:15:75 root 17109 0.0 0.6 1500 432 ? S Sep18 0:00 /usr/sbin/pppoe -I eth1 -T 12 -M 00:00:CB:22:15:77 root 17117 0.0 0.6 1500 432 ? S Sep18 0:00 /usr/sbin/pppoe -I eth1 -T 12 -M 00:00:CB:22:15:75 [...] $ ls -l /proc/30742/fd/ total 0 lrwx------ 1 root root 64 Sep 28 04:41 0 -> /dev/ptmx lrwx------ 1 root root 64 Sep 28 04:41 1 -> /dev/ptmx lrwx------ 1 root root 64 Sep 28 04:41 10 -> /dev/pts/22 lrwx------ 1 root root 64 Sep 28 04:41 11 -> /dev/pts/34 lrwx------ 1 root root 64 Sep 28 04:41 12 -> /dev/pts/44 lrwx------ 1 root root 64 Sep 28 04:41 13 -> /dev/pts/50 lrwx------ 1 root root 64 Sep 28 04:41 14 -> /dev/pts/59 lrwx------ 1 root root 64 Sep 28 04:41 15 -> /dev/pts/63 l-wx------ 1 root root 64 Sep 28 04:41 2 -> /var/log/ppp-connect-errors lrwx------ 1 root root 64 Sep 28 04:41 3 -> socket:[142223127] lrwx------ 1 root root 64 Sep 28 04:41 4 -> socket:[142223128] lrwx------ 1 root root 64 Sep 28 04:41 5 -> socket:[142223240] lrwx------ 1 root root 64 Sep 28 04:41 6 -> socket:[142223241] lrwx------ 1 root root 64 Sep 28 04:41 9 -> /dev/pts/46 $ strace -p 30742 Process 30742 attached - interrupt to quit write(1, "~\377}#} W`} } } } X}&4 }!})h} }!} "..., 181 (pressed ^C after a while)<unfinished ...> Process 30742 detached -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
