Bug 7214

Summary: [PBST] Namespace lookup failure, AE_NOT_FOUND
Product: ACPI Reporter: paolo (oopla)
Component: Config-TablesAssignee: Robert Moore (Robert.Moore)
Status: REJECTED INVALID    
Severity: normal    
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.17.13 Subsystem:
Regression: --- Bisected commit-id:
Attachments: DSDT of Toshiba A110-262

Description paolo 2006-09-27 08:43:08 UTC
Most recent kernel where this bug did not occur:
Distribution: Debian Etch
Hardware Environment: Toshiba A110-262
Software Environment: GNU/Linux
Problem Description: koog flood 

Steps to reproduce: just some apps checking battery status

I'm seeing many msgs like:

2006-09-27_14:27:51.96853 kern.warn: ACPI Error (psargs-0353): [PBST] Namespace
lookup failure, AE_NOT_FOUND
2006-09-27_14:27:51.96858 kern.warn: ACPI Error (psparse-0517): Method parse/exe
cution failed [\_SB_.PCI0.LPCB.BAT1._BST] (Node df606c68), AE_NOT_FOUND
2006-09-27_14:27:51.96860 kern.warn: acpi_battery-0205 [02] battery_get_status
  : Error evaluating _BST
Comment 1 Robert Moore 2006-09-28 13:28:02 UTC
Please post the acpidump for this machine
Comment 2 paolo 2006-09-28 14:27:30 UTC
Created attachment 9122 [details]
DSDT of Toshiba A110-262
Comment 3 Robert Moore 2006-09-28 14:48:41 UTC
The DSDT contains a method that creates a namespace object, but is not 
serialized. This will cause various race conditions with multiple threads. Use 
acpi_serialize configuration option to force serialization, this should fix 
the problem.

Manufacturer should be notified.

Method (_BST, 0, NotSerialized)
{
    Name (PBST, Package (0x04)
    {
        Zero, 
        Ones, 
        Ones, 
        0x2B5C
    })
    Store (0x2B5C, Local3)
    If (LAnd (ECOK (), LEqual (ECDY, Zero)))
    {
        Sleep (0x14)
        Store (^^EC0.BST0, BST1)
        Sleep (0x14)
        Store (^^EC0.GAU0, BGU1)
        Sleep (0x14)
    }
...
    Return (PBST)
}
Comment 4 paolo 2006-09-28 23:27:33 UTC
thanks, I'll put acpi_serialize as boot param and see what happen after 1st
chance to reboot that (remote) machine, and leave it run 1 day. Will be back
with results. Then I'll put up a notice on Toshiba's Linux website.
Comment 5 paolo 2006-10-07 00:51:29 UTC
ok, no more such messages in logs. Seems that acpi_serialize solved the problem.
So bug should be closed, though I'm unsure which 'resolution' code applies here.
Comment 6 Robert Moore 2006-10-11 11:24:23 UTC
This is a BIOS bug, not an ACPI bug