Subject: Re: After newlock2 merge: Different pthread behavior for userland programs?
To: None <current-users@netbsd.org>
From: Sverre Froyen <sverre@viewmark.com>
List: current-users
Date: 03/23/2007 10:53:58
On Wednesday 21 March 2007, Markus W Kilbinger wrote:
> I just resynced with your latest changes several minutes ago and can
> confirm that now perl's 'make test' passes the ext/threads part
> flawlessly, too! Thanks! (Close pr/36008?)
While system stability has improved tremendously over the last few weeks, I
still see hangs when using Apache with Perl CGI scripts. Apache accesses to
plain HTML pages appear to work fine. Here's a top output example:
load averages: 0.07, 0.05, 0.04 up 0 days, 1:34
10:31:17
59 processes: 58 sleeping, 1 on processor
CPU states: 0.0% user, 0.0% nice, 0.0% system, 0.0% interrupt, 100% idle
Memory: 136M Act, 4188K Wired, 10M Exec, 104M File, 303M Free
Swap: 1000M Total, 1000M Free
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
16 root 18 0 0K 12M syncer 0:05 0.00% 0.00% [ioflush]
2175 sverre 28 0 240K 1160K CPU 0:00 0.00% 0.00% top
417 www 29 0 4176K 1556K tstile 0:00 0.00% 0.00% httpd
1229 www 29 0 4176K 1068K tstile 0:00 0.00% 0.00% httpd
1367 www 28 0 4176K 1556K tstile 0:00 0.00% 0.00% httpd
1908 www 28 0 4176K 1068K tstile 0:00 0.00% 0.00% httpd
1440 root 28 0 244K 920K tstile 0:00 0.00% 0.00% cron
11 root 10 0 0K 12M cardslot 0:00 0.00% 0.00% [cardslot0]
12 root 10 0 0K 12M cardslot 0:00 0.00% 0.00% [cardslot1]
3 root 10 0 0K 12M apmev 0:00 0.00% 0.00% [apm0]
4 root 10 0 0K 12M pmsreset 0:00 0.00% 0.00% [pms0]
5 root 10 0 0K 12M usbevt 0:00 0.00% 0.00% [usb0]
6 root 10 0 0K 12M usbtsk 0:00 0.00% 0.00%
[usbtask-hc]
7 root 10 0 0K 12M usbtsk 0:00 0.00% 0.00%
[usbtask-dr]
8 root 10 0 0K 12M usbevt 0:00 0.00% 0.00% [usb1]
9 root 10 0 0K 12M usbevt 0:00 0.00% 0.00% [usb2]
10 root 10 0 0K 12M usbevt 0:00 0.00% 0.00% [usb3]
2025 root 3 0 56K 844K ttyin 0:00 0.00% 0.00% getty
1182 root 3 0 56K 844K ttyin 0:00 0.00% 0.00% getty
306 root 2 0 3928K 5552K select 0:00 0.00% 0.00% cupsd
1117 mysql 2 0 23M 3968K select 0:00 0.00% 0.00% mysqld
Notice the processes in state 'tstile'. What can I do to help track this
problem down?
System is a Thinkpad T42 i386 NetBSD 4.99.16, basically LAPTOP_GENERIC from
22 March.
apache-2.2.4nb3 Apache HTTP (Web) server, version 2
perl-5.8.8nb4 Practical Extraction and Report Language
Both are compiled with threading option and are from around 7 March.
Thanks,
Sverre