Opened 16 years ago

Closed 16 years ago

Last modified 16 years ago

#232 closed defect (fixed)

Trap in ACPI.PSD when testing _OSI multiple times

Reported by: Chuck McKinnis Owned by: pasha
Priority: blocker Milestone:
Component: ACPI PSD Version: 3.08
Keywords: Cc:

Description

_OSI Method very often used to identify OS version for example http://www.microsoft.com/whdc/archive/_OSI-method.mspx

Usually this block included in _INI method and looks like:

If (CondRefOf (_OSI, Local0))
{
   If (_OSI ("LINX"))
   {
      Store (0x0001, OSYS)
   }
...
   If (_OSI ("Windows 2006"))
   {
      Store (0x07D6, OSYS)
   }
}


Very often this code cause system trap in ACPI.PSD. I made some tests and found that ACPI.PSD ALWAYS trap if

If (_OSI ("Some string"))

executed mode then once. For example:

If (CondRefOf (_OSI, Local0))
{
   Store (0x0001, OSYS)
} else
{
   Store (0x0002, OSYS)
}

Doesn't cause trap.

If (CondRefOf (_OSI, Local0))
{
   If (_OSI ("LINX"))
   {
      Store (0x0001, OSYS)
   }
}

doesn't cause trap. while

   If (_OSI ("System 1"))
   {
      Store (0x0001, OSYS)
   }
   If (_OSI ("System 2"))
   {
      Store (0x0002, OSYS)
   }

cause the trap. I made a search of internet and didn't found any information about traps is such cases on other OS, like linux or FreeBSD. Sometimes users of such systems comment out CondRefOf?() blocks, leaving only one case statement (no way to change _OSI value in linux?), when they have problems with batteries or other ACPI related devices, but never due to the trap in this place. It seems that this but is OS/2 specific, as trap itself happend after cleaning up in UtRemoveReference?() function, when reference couldn't be found, something wrong with stack?

Apropriated logs and trapdumps can be provided if necessary.

Attachments (2)

acpi-3.08a.log (15.7 KB) - added by Chuck McKinnis 16 years ago.
acpi-3.08a-pci.log (29.7 KB) - added by Chuck McKinnis 16 years ago.

Download all attachments as: .zip

Change History (14)

comment:1 Changed 16 years ago by Chuck McKinnis

Priority: majorblocker

comment:2 Changed 16 years ago by Chuck McKinnis

Can you please send an email to rwklein at mensys.nl I would like to be in touch with you about this ACPI bug. Thanks

Roderick Klein Mensys

comment:3 Changed 16 years ago by Chuck McKinnis

e-co:

"No troubles in other OSes.."

because other OSes use different Intel ACPI CA code?

  • Pasha was talking that Linux is using ACPI, 2005,

eComStation uses ACPI, 2008/03. So.. different Intel ACPI CA machines.

  • froloff: FreeBSD 7.0 includes Intel ACPI-CA 20070320.

comment:4 Changed 16 years ago by pasha

Resolution: invalid
Status: newclosed

_OSI for special use, don't for playing

comment:5 Changed 16 years ago by Chuck McKinnis

Resolution: invalid
Status: closedreopened

Nobody playing games here. We are looking for the way how to improve the ACPI component. It is clear what for _OSI must be used. Please reread Rüdiger Ihle message with fear from 06.11.2007. Real situations from real hardware presented here and the way how to model it in order to examine more precise was shown.

comment:6 Changed 16 years ago by pasha

Resolution: invalid
Status: reopenedclosed

Read docs and do as normal user or going to acpi-dev mail list.

comment:7 Changed 16 years ago by Chuck McKinnis

Resolution: invalid
Status: closedreopened

I don't see what is wrong with ticket open on this topic.

Roderick Klein Mensys

comment:8 Changed 16 years ago by pasha

/OSI is removed from acpi.psd. Need remove /OS too? I can't support this options for any string. Ticket is't wrong, user is wrong, which use this options without consulting.

comment:9 Changed 16 years ago by Chuck McKinnis

Last build works fine, where all previos trapped.

Thanks.

Changed 16 years ago by Chuck McKinnis

Attachment: acpi-3.08a.log added

comment:10 Changed 16 years ago by pasha

No problem, now you can try any combination of /OSi=xxxx, this options is absent.

comment:11 Changed 16 years ago by eco

Resolution: fixed
Status: reopenedclosed

OK, there is progress!

We should invite Acer users to test new version.

Changed 16 years ago by Chuck McKinnis

Attachment: acpi-3.08a-pci.log added

comment:12 Changed 16 years ago by Chuck McKinnis

Seems fixed with latest ACPI build and ticket #211 and # 234.

Note: See TracTickets for help on using tickets.