On Sunday 27 January 2008 08:07, Graham Adams wrote:
> On 1/23/08, Graham Adams <gadams@xxxxxxxxxxx> wrote:
> > On 1/23/08, Len Brown <lenb@xxxxxxxxxx> wrote:
> > > On Tuesday 22 January 2008 12:56, you wrote:
> > > > On 1/22/08, Len Brown <lenb@xxxxxxxxxx> wrote:
> > > > > On Monday 21 January 2008 12:28, Graham Adams wrote:
> > > > > > On 1/19/08, Len Brown <lenb@xxxxxxxxxx> wrote:
> > > > > > >
> > > > > > > > Please find attached the output of dmidecode after booting with
> > > > > > > > acpi_osi=!Linux as requested in dmesg output. The system seems to be
> > > > > > > > working OK apart from problems with STAC9221 Line In (which has never
> > > > > > > > worked under Linux).
> > > > > > >
> > > > > > > > The system is a Dell Dimension 5150.
> > > > > > >
> > > > > > > > System Information
> > > > > > > > Manufacturer: Dell Inc.
> > > > > > > > Product Name: Dell DM051
> > > > > > > > Version: Not Specified
> > > > > > > > Serial Number: 6X8B32J
> > > > > > > > UUID: 44454C4C-5800-1038-8042-B6C04F33324A
> > > > > > > > Wake-up Type: Power Switch
> > > > > > > >
> > > > > > > > Handle 0x0200, DMI type 2, 8 bytes
> > > > > > > > Base Board Information
> > > > > > > > Manufacturer: Dell Inc.
> > > > > > > > Product Name: 0WG261
> > > > >
> > > > > This is "different"....
> > > > >
> > > > > some magic happens at initialization and at wake from suspend-to-disk
> > > > > to PCI device 00:1d -- which is apparently USB.
> > > > >
> > > > > The code reads as if Dell really did intend to do something special for Linux -
> > > > > at least the version of Linux they were testing...
> > > > >
> > > > > for non-Linux, ISLI does some SMI magic, for Linux the SMI magic is skipped.
> > > > >
> > > > > Does USB work properly in both cases, both before and after suspend to disk?
> > > > >
> > > > > thanks,
> > > > > -Len
> > > > >
> > > > > Method (ISLI, 0, NotSerialized)
> > > > > {
> > > > > If (CondRefOf (_OSI, Local0))
> > > > > {
> > > > > If (_OSI ("Linux"))
> > > > > {
> > > > > Return (One)
> > > > > }
> > > > > Else
> > > > > {
> > > > > Return (Zero)
> > > > > }
> > > > > }
> > > > > Else
> > > > > {
> > > > > Return (Zero)
> > > > > }
> > > > > }
> > > > >
> > > > > ...
> > > > > Method (GUSB, 0, NotSerialized)
> > > > > {
> > > > > If (ISLI ()) # the only reference to ISLI is this one
> > > > > {
> > > > > Return (Zero)
> > > > > # return valie of GUSB is ignored,
> > > > > # so this is the NO-OP path
> > > > > }
> > > > > Else
> > > > > {
> > > > > Return (SMI2 (0xBA))
> > > > > }
> > > > > }
> > > > > ...
> > > > > # surely this is SMI command/SMI Data method:
> > > > > # no telling what the actual command 0xBA is, though...
> > > > >
> > > > > Method (SMI2, 1, NotSerialized)
> > > > > {
> > > > > Acquire (SMIM, 0xFFFF)
> > > > > Store (Arg0, SMIC)
> > > > > Store (SMIC, Local0)
> > > > > Store (SMID, Local0)
> > > > > ShiftLeft (Local0, 0x08, Local0)
> > > > > Add (SMIC, Local0, Local0)
> > > > > Release (SMIM)
> > > > > Return (Local0)
> > > > > }
> > > > > ...
> > > > > Method (_WAK, 1, NotSerialized)
> > > > > {
> > > > > If (LEqual (Arg0, 0x04))
> > > > > {
> > > > > # we come here on wakeup from suspend-to-disk
> > > > >
> > > > > GUSB ()
> > > > > }
> > > > >
> > > > > If (Or (And (PMS2, One), Or (And (LNot (HACK ()),
> > > > > And (GLBT, 0x11)), And (LNot (HACK ()), And (THRP, 0x18
> > > > > )))))
> > > > > {
> > > > > Notify (\_SB.VBTN, 0x02)
> > > > > }
> > > > >
> > > > > Return (Zero)
> > > > > }
> > > > > ...
> > > > > Scope (_SB.PCI0)
> > > > > {
> > > > > Device (USB0)
> > > > > {
> > > > > Name (_ADR, 0x001D0000)
> > > > > Name (_UID, 0x05)
> > > > > Name (_PRW, Package (0x02)
> > > > > {
> > > > > 0x03,
> > > > > 0x03
> > > > > })
> > > > > OperationRegion (UPC1, PCI_Config, 0xC1, One)
> > > > > Field (UPC1, ByteAcc, NoLock, Preserve)
> > > > > {
> > > > > LEGK, 8
> > > > > }
> > > > >
> > > > > Method (_INI, 0, NotSerialized)
> > > > > {
> > > > > If (HACK ()) {}
> > > > > Else
> > > > > {
> > > > > Store (LEGK, Local0)
> > > > > And (Local0, 0x60, Local0)
> > > > > Or (Local0, 0x20, Local0)
> > > > > Store (Local0, LEGK)
> > > > > }
> > > > >
> > > > > GUSB ()
> > > > > }
> > > > >
> > > > > ...
> > > > > This is different too....
> > > > > Method (_INI, 0, NotSerialized)
> > > > > {
> > > > > If (CondRefOf (_OSI, Local0))
> > > > > {
> > > > > If (_OSI ("Windows 2001"))
> > > > > {
> > > > > Store (0x04, MSOS)
> > > > > # yes, Linux does come here
> > > > > # MSOS is effectively "XP" or newer
> > > > > }
> > > > > }
> > > > > }
> > > > >
> > > > > ...
> > > > > Method (HACK, 0, NotSerialized)
> > > > > {
> > > > > If (LEqual (MSOS, 0x04))
> > > > > {
> > > > > Store (CMRD (0x6F), Local0)
> > > > > Or (Local0, 0x04, Local0)
> > > > > CMWR (0x6F, Local0)
> > > > > # Linux come shere b/c of OSI(Microsoft Windows 2001)
> > > > > Return (One)
> > > > > }
> > > > > Else
> > > > > {
> > > > > Store (CMRD (0x6F), Local0)
> > > > > And (Local0, 0xFB, Local0)
> > > > > CMWR (0x6F, Local0)
> > > > > }
> > > > >
> > > > > Add (SizeOf (_OS), One, Local0)
> > > > > Store (Local0, Local1)
> > > > > Name (BUF0, Buffer (Local0) {})
> > > > > Name (BUF1, Buffer (Local0) {})
> > > > > Name (OSNT, One)
> > > > > Name (OS98, One)
> > > > > Store (_OS, BUF0)
> > > > > Store ("Microsoft Windows NT", BUF1)
> > > > > While (Local0)
> > > > > {
> > > > > # an in-line strcmp()...
> > > > > Decrement (Local0)
> > > > > If (LEqual (DerefOf (Index (BUF0, Local0)), DerefOf (Index (BUF1,
> > > > > Local0)))) {}
> > > > > Else
> > > > > {
> > > > > Store (Zero, OSNT)
> > > > > }
> > > > > }
> > > > >
> > > > > If (OSNT)
> > > > > {
> > > > > Store (0x03, MSOS)
> > > > > Return (One)
> > > > > }
> > > > > Else
> > > > > {
> > > > > Store ("Microsoft Windows", BUF1)
> > > > > While (Local1)
> > > > > {
> > > > > Decrement (Local1)
> > > > > If (LEqual (DerefOf (Index (BUF0, Local1)), DerefOf (Index (BUF1,
> > > > > Local1)))) {}
> > > > > Else
> > > > > {
> > > > > Store (Zero, OS98)
> > > > > }
> > > > > }
> > > > >
> > > > > If (OS98)
> > > > > {
> > > > > Store (One, MSOS)
> > > > > Return (Zero)
> > > > > }
> > > > > Else
> > > > > {
> > > > > Store (0x02, MSOS)
> > > > > Return (Zero)
> > > > > }
> > > > > }
> > > > > }
> > > > >
> > > >
> > > >
> > > >
> > > > Len,
> > > >
> > > > I had to learn how to configure my desktop to suspend to disk.
> > > >
> > > > Using acpi_osi=Linux it does not resume from suspend, I get an error
> > > > "Could not stat resume device file"
> > > > The resume device is /dev/sda5 specified on the kermel command line
> > > > via the grub menu.
> > > > Using acpi_osi=!Linux the resume works (and USB devices work).
> > >
> > > Hmmm, so you've to 2.6.22 or earlier where the default is acpi_osi=Linux
> > > and suspend to disk fails; but when you add "acpi_osi=!Linux" suspend
> > > to disk works?!
> > >
> > > I guess it is important that we stay Windows compatible
> > > on this box and trigger whatever SMI Dell has in that path.
> > >
> > > thanks,
> > > -Len
> > >
> >
> > I'll have another look at this today if I get a chance - it's possible
> > I messed something up with the suspend as it's the first time I've set
> > it up - a normal reboot (not after suspend) failed with
> > resume=/dev/sda5 and acpi_osi=!Linux. I don't know if it always tries
> > a resume if you give a resume= kernel argument or looks for a
> > signature and if not found does a normal reboot.
> >
> > --
> > Graham Adams
> > gadams@xxxxxxxxxxx
> >
>
> Len,
>
> My problems were due I think to switching to initramfs from initrd so
> I could use suspend. Initramfs seems to detect devices in a different
> order and it couldn't find the root filesystem, strange it worked
> sometimes though. I'm sorry about the confusion.
>
> Anyway, the upshot is that I haven't found any functional difference
> after all with acpi_osi=!Linux or acpi_osi=Linux on a normal boot or a
> resume after suspend.
> USB seems to work OK in each case (hotplugged disk, also keyboard and
> mouse are USB anyway).
>
>
> Let me know if you want me to try anything else.
Thanks for testing -- I've removed the comment that OSI(Linux) breaks hibernate.
I'll keep OSI(Linux) disabled on this box since it is a risk and has not
proven to help anything.
-Len
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
