Conference turris::digital_unix

9043.0. "URGENT - 4.0b patch work for 4.0a???" by NETRIX::"[email protected]" (Kim Roy) Wed Mar 05 1997 15:02

Hi there,

I have a customer whose 4100 running dunix 4.0a is crashing with the following
panic string:  vm_map_swapout: negative resident count.  His stack trace is as
follows:


>  0 stop_secondary_cpu(do_lwc = 0x0)
["../../../../src/kernel/arch/alpha/cpu.c":493, 0xfffffc00004f6f74]
   1 panic(s = 0xfffffc0000668de8 = "event_timeout: panic request")
["../../../../src/kernel/bsd/subr_prf.c":703, 0xfffffc0000290ae0]
   2 event_timeout(func = 0xfffffc0000290d10, arg = 0xfffffc00006ce560,
timeout
= 0x5) ["../../../../src/kernel/arch/alpha/cpu.c":987, 0xfffffc00004f7d58]
   3 xcpu_puts(s = 0xfffffc000069f690, prfbufp = 0xfffffc00006ce560)
["../../../../src/kernel/bsd/subr_prf.c":844, 0xfffffc0000290d74]
   4 printf(va_alist = 0xfffffc0000627188)
["../../../../src/kernel/bsd/subr_prf.c":379, 0xfffffc0000290170]
   5 panic(s = 0xfffffc00006673a0 = "vm_map_swapout: negative resident count")
["../../../../src/kernel/bsd/subr_prf.c":753, 0xfffffc0000290c38]
   6 vm_map_swapout(0x1, 0xfffffc00067dbec0, 0x1, 0xfffffc00086a9600,
0xfffffc0008def560) ["../../../../src/kernel/vm/vm_map.c":777,
0xfffffc00004db2e4]
   7 task_swapout(0xfffffc00006cf500, 0x2, 0xfffffc0008def500, 0x0, 0x0)
["../../../../src/kernel/kern/task_swap.c":466, 0xfffffc00002c4940]
   8 task_swapout_thread(0xfffffc00019af080, 0x0, 0x0, 0x0,
0xfffffc003fe102c0)
["../../../../src/kernel/kern/task_swap.c":666, 0xfffffc00002c4e34]
End Trace for machine_slot[paniccpu].cpu_panic_thread:

There is a patch for this in v4.0b, but not in v4.0a.  Can I safely apply 
this patch to a v4.0a system?   THe patch details follow:

New PatchID:  22.00
PATCH ID: OSF410-400127     SUBSET(s): OSFBIN410 

******************************************************************************
*

PROBLEM:        (49818) (Patch ID: OSF410-400127)
********
System panics with messgae: "vm_map_swapout: negative resident count".

The crash will most likely occur on a multi-processor system during
a heavy paging load.

A typical stack trace for the panic will be as follows:

  panic("vm_map_swapout: negative resident count")
   
   vm_map_swapout()
   task_swapout()
   task_swapout_thread()


FILE(s):

/sys/BINARY/vm.mod              subset OSFBIN410
CHECKSUM: 54019    424

Thanks for any advice.
[Posted by WWW Notes gateway]

    Kim,
    
    please read note 8919 and use CANASTA when faced with crash situations.
    
    If there is no RULE in CANASTA and the SIMILAR_CASES information does
    not point you to an IPMT, which allows you to find a solution or
    workaround, raise an IPMT yourself.
    
    USEG is committed to using CANASTA to provide rules about solved problems.
    If they don't sufficiently document their solution, they will have to
    live with the IPMTs...
    
    Volker.

Hi!

I sent the crash to canasta as suggested, but no rule matched this case.
Anybody have any ideas what I can do about this crash?

Thanks in advance.

kim
[Posted by WWW Notes gateway]

[Posted by WWW Notes gateway]

    Kim,
    
    the FIND_SIMILAR_CASES function of CANASTA told you about another crash
    in V4.0A with the SAME FOOTPRINT, raised as IPMT HPAQ1107Q !
    
    Did you check the status of that IPMT ?
    
    Volker.