[ltp] T42: slow (2-minute) resumes from S3 sleep

[Marius Gedminas]

--ZoaI/ZTpAVc4A5k6
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hi,

I run Ubuntu Breezy on my T42.  I use a 2.4.12 kernel patched with the
Radeon ACPI power drain patch, so that my battery doesn't die overnight
while suspended.

Sometimes (not always), when I press Fn-F4 to suspend, the resume takes
literally 2 wall-clock minutes.  The kernel just sits there and does
nothing for about 100 seconds (I will include the excerpt of dmesg with
timestamps at the end of the email).  If I just run the exact same
/etc/acpi/sleep.sh script directly (say, from the gnome Run dialog, with
gksudo), the resume is fast.

This pause doesn't always appear while resuming, but when it starts
appearing, it keeps on appearing on ever resume until next reboot.

I have not seen such pauses with a 2.4.10 kernel.

Any ideas or similar experiences?

Here's the excerpt from dmesg:

[4295604.733000] Stopping tasks: =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D|

  (I have trimmed the number of '=3D' signs a bit here)

[4295604.748000] GTM info 78,3c,0,0,13
[4295604.760000] GTM info 78,14,0,0,13
[4295605.749000] ACPI: PCI interrupt for device 0000:02:00.1 disabled
[4295605.749000] ACPI: PCI interrupt for device 0000:02:00.0 disabled
[4295605.749000] radeonfb (0000:01:00.0): suspending to state: 3...
[4295605.749000] radeonfb (0000:01:00.0): switching to D2 state...
[4295605.886000] ACPI: PCI interrupt for device 0000:01:00.0 disabled
[4295606.381000] ACPI: PCI interrupt for device 0000:00:1f.5 disabled
[4295606.393000] Back to C!

  That was the suspend part.  Resume starts here -- witness the large
  difference in timestamps.

[4320370.060000] ACPI: PCI Interrupt 0000:00:1f.1[A] -> Link [LNKC] -> GSI =
11 (level, low) -> IRQ 11
[4320370.060000] ACPI: PCI Interrupt 0000:00:1f.5[B] -> Link [LNKB] -> GSI =
11 (level, low) -> IRQ 11
[4320370.060000] PCI: Setting latency timer of device 0000:00:1f.5 to 64
[4320370.361000] radeonfb (0000:01:00.0): resuming from state: 3...
[4320370.361000] ACPI: PCI Interrupt 0000:01:00.0[A] -> Link [LNKA] -> GSI =
11 (level, low) -> IRQ 11
[4320370.361000] radeonfb (0000:01:00.0): switching to D0 state...
[4320372.803000] ACPI: PCI Interrupt 0000:02:00.0[A] -> Link [LNKA] -> GSI =
11 (level, low) -> IRQ 11
[4320372.803000] ACPI: PCI Interrupt 0000:02:00.1[B] -> Link [LNKB] -> GSI =
11 (level, low) -> IRQ 11

  Here is a 8-second pause that I see almost always.

[4320380.605000] Start STM
[4320380.606000] start GTF
[4320380.606000] data 2,0,0,0,0,a0,ef
[4320380.606000] data 0,0,0,0,0,a0,f5
[4320380.606000] data 3,45,0,0,0,a0,ef
[4320380.606000] data 3,c,0,0,0,a0,ef
[4320380.708000] Start STM
[4320381.308000] start GTF
[4320381.308000] data 3,42,0,0,0,a0,ef
[4320381.308000] data 3,c,0,0,0,a0,ef
[4320381.308000] data 0,10,0,0,0,a0,e3
[4320381.308000] data 0,3,0,0,0,a0,e3

  Here is a 100-second pause that annoys me to no end.  It does not
  happen when I do not use an ACPI event to suspend.

[4320479.417000] Restarting tasks... done

  The rest goes fast.

Marius Gedminas
--=20
"Wipe Info uses hexadecimal values to wipe files. This provides more securi=
ty
than wiping with decimal values."
	-- Norton SystemWorks 2002 Manual

--ZoaI/ZTpAVc4A5k6
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFDVexPkVdEXeem148RAmK+AJ9aLW3/Uk4L4vzmCCEoY3YpM00nnACfWGZ8
sjwsqX7rdxGWSH5wwdQfdX0=
=er6b
-----END PGP SIGNATURE-----

--ZoaI/ZTpAVc4A5k6--