i386/136008: [acpi] Dell Vostro 1310 will not shutdown (Requires user intervention)

[Ilya Bakulin]

The following reply was made to PR i386/136008; it has been noted by GNATS.

From: Ilya Bakulin <webmaster at kibab.com>
To: bug-followup at FreeBSD.org, dan at dburkland.com
Cc:  
Subject: Re: i386/136008: [acpi] Dell Vostro 1310 will not shutdown
 (Requires user intervention)
Date: Mon, 28 Sep 2009 20:27:18 +0400

 --Sig_/EAzJ16sc/d0hEZlOsUlyXqA
 Content-Type: multipart/mixed; boundary="MP_/CraSF8uEy=eJAif63J7.+Ca"
 
 --MP_/CraSF8uEy=eJAif63J7.+Ca
 Content-Type: text/plain; charset=US-ASCII
 Content-Transfer-Encoding: quoted-printable
 Content-Disposition: inline
 
 I have exactly the same laptop model with similar symptoms.
 It seems that EC cannot work in polled mode at all. Since FreeBSD ACPI code=
  disables GPEs on shutdown, it is not able to receive answer from EC, and r=
 eboots machine.
 
 Atached patch adds new sysctl "debug.acpi.ec.disable_polling", setting it t=
 o "1" resolves all described symptoms.
 
 --=20
 Regards,
 Ilya Bakulin
 http://kibab.com
 xmpp://kibab612@jabber.ru
 
 --MP_/CraSF8uEy=eJAif63J7.+Ca
 Content-Type: text/x-patch
 Content-Transfer-Encoding: quoted-printable
 Content-Disposition: attachment; filename=acpi_ec_addon.diff
 
 --- sys/dev/acpica/acpi_ec.c	2009-06-05 22:44:36.418313000 +0400
 +++ sys/dev/acpica/acpi_ec.c	2009-09-04 15:49:54.000000000 +0400
 @@ -197,6 +197,10 @@
  SYSCTL_INT(_debug_acpi_ec, OID_AUTO, timeout, CTLFLAG_RW, &ec_timeout,
      EC_TIMEOUT, "Total time spent waiting for a response (poll+sleep)");
 =20
 +static int	ec_disable_polling;
 +SYSCTL_INT(_debug_acpi_ec, OID_AUTO, disable_polling, CTLFLAG_RW, &ec_disa=
 ble_polling, 0,
 +    "Totally disable use of polled mode (Dell EC problem workaround)");
 +
  static ACPI_STATUS
  EcLock(struct acpi_ec_softc *sc)
  {
 @@ -577,7 +581,9 @@
 =20
      /* Disable the GPE so we don't get EC events during shutdown. */
      sc =3D device_get_softc(dev);
 -    AcpiDisableGpe(sc->ec_gpehandle, sc->ec_gpebit, ACPI_NOT_ISR);
 +    if (!ec_disable_polling) {
 +	AcpiDisableGpe(sc->ec_gpehandle, sc->ec_gpebit, ACPI_NOT_ISR);
 +    }
      return (0);
  }
 =20
 @@ -814,7 +820,7 @@
 =20
      ACPI_SERIAL_ASSERT(ec);
      Status =3D AE_NO_HARDWARE_RESPONSE;
 -    int need_poll =3D cold || rebooting || ec_polled_mode || sc->ec_suspen=
 ding;
 +    int need_poll =3D (cold || rebooting || ec_polled_mode || sc->ec_suspe=
 nding )  &&  !ec_disable_polling;
      /*
       * The main CPU should be much faster than the EC.  So the status shou=
 ld
       * be "not ready" when we start waiting.  But if the main CPU is really
 @@ -894,7 +900,11 @@
  	    device_printf(sc->ec_dev,
  		"wait timed out (%sresponse), forcing polled mode\n",
  		Status =3D=3D AE_OK ? "" : "no ");
 -	    ec_polled_mode =3D TRUE;
 +	    if (ec_disable_polling) {
 +		device_printf(sc->ec_dev, "Polling explicitly disabled! Continue waiting=
  for generated GPEs...\n");
 +	    } else {
 +		ec_polled_mode =3D TRUE;
 +	    }
  	}
      }
      if (Status !=3D AE_OK)
 
 --MP_/CraSF8uEy=eJAif63J7.+Ca--
 
 --Sig_/EAzJ16sc/d0hEZlOsUlyXqA
 Content-Type: application/pgp-signature; name=signature.asc
 Content-Disposition: attachment; filename=signature.asc
 
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.12 (FreeBSD)
 
 iEYEARECAAYFAkrA4+sACgkQo9vlj1oadwg8YwCfT2LuqsTHf77i3ObhIpV89YJd
 k2oAniZlLrJfVQSmFVf478sMnOA98xSJ
 =kZw9
 -----END PGP SIGNATURE-----
 
 --Sig_/EAzJ16sc/d0hEZlOsUlyXqA--