Subject: kern/20245: 'sa_switch: not enough stacks' when running smms
To: None <gnats-bugs@gnats.netbsd.org>
From: None <dogcow@babymeat.com>
List: netbsd-bugs
Date: 02/07/2003 15:25:34
>Number:         20245
>Category:       kern
>Synopsis:       'sa_switch: not enough stacks' when running smms
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Feb 07 15:27:00 PST 2003
>Closed-Date:
>Last-Modified:
>Originator:     Tom Spindler
>Release:        NetBSD 1.6N
>Organization:
	
>Environment:
	
	
System: NetBSD beefcake.babymeat.com 1.6N NetBSD 1.6N (BEEFCAKE) #0: Thu Feb 6 19:57:47 PST 2003 dogcow@x1-6-00-48-54-6c-9b-29:/media/src/src/sys/arch/i386/compile/BEEFCAKE i386
Architecture: i386
Machine: i386
>Description:
	
when running xmms, every once in a while I get the following message on
the console:

sa_switch(24802.8 flag 900004): Not enough stacks.
(or from another invocation)
sa_switch(16903.9 flag 900004): Not enough stacks.

when this happens, the audio stops and won't return until xmms is restarted.

here's the p s-s and ps/l output:
p3 dogcow@beefcake: ~ 2586 % ps -swww16903
 UID   PID PPID CPU LID NLWP PRI NI  VSZ  RSS WCHAN    STAT TT    TIME COMMAND
9107 16903  273   0   9    9  38  5 3352 5424 sawait   S-   p4 5:51.18 xmms -e
9107 16903  273   0   8    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   7    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   2    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   4    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   5    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   3    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   6    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
9107 16903  273   0   1    9  10  5 3352 5424 sastacks S-   p4 5:51.18 xmms -e
 
and hand-transcribed from the ddb screen:
 
LID FLAGS    STRUCT LWP*        UAREA* WAIT
1   0xe00084 d3c3e008           d4055000
2   0xe00084 d3c3e588           d408f000
3   0xe00084 d3c3e288           d3f92000
4   0xe00084 d3c3ed88           d4039000
5   0xe00084 d3c3e188           d4147000
6   0xe00084 d3c3eb88           d4051000
7   0xe00084 d3c3ea88           d40a3000
8   0xe00084 d3c3e988           d4087000
9   0x800084 d3f48698           d409b000 

I'll probably be recompiling the kernel with options DEBUG real soon.

 
>How-To-Repeat:
Run xmms, possibly hitting the disk with other programs
	
>Fix:
	
>Release-Note:
>Audit-Trail:
>Unformatted: