Date: Mon, 14 Jan 2008 11:11:48 -0800 From: Nate Lawson <nate@root.org> To: stable@freebsd.org Cc: current@FreeBSD.org Subject: Re: Fix for GPE livelock on HPs Message-ID: <478BB3F4.7050600@root.org> In-Reply-To: <4789B267.1000505@root.org> References: <47893F8B.7010805@root.org> <4789756B.8060203@vwsoft.com> <4789B267.1000505@root.org>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------070609030008030803090405
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Nate Lawson wrote:
> Volker wrote:
>> On 12/23/-58 20:59, Nate Lawson wrote:
>>> I've committed the below patch and want to MFC it to 7.0. To do this, I
>>> need people to test this quickly. It probably has no effect in 6.x and
>>> probably doesn't apply cleanly there.
>>>
>>> Please try this patch if you have a laptop and 7.x. If you have
>>> -current, just cvsup. I'd like to make sure there is no regression.
>>> I'm already aware that it fixes things for some HP users.
>> Nate,
>>
>> can you be a bit specific for a) what GPE is, b) what the problem is,
>> c) what to look for (any test procedures?) and d) which HP laptop
>> models might be affected?
>>
>> I do have an Omnibook vt6200 (P-IV 1.8G) running 6-STABLE and a new HP
>> 6715b (Tur-X2 TL-60) running 7-RC1 (currently installing on this, OS
>> is not yet fully set up).
>>
>> If I knew what to look for, I might test your patches (at least on the
>> 7-RC1 version).
>
> A GPE is an interrupt of sorts. I'm looking for any bad behavior the
> patch might cause. I'm certain it fixes lockups some HPs had during
> thermal zone events (i.e. fan switching on when it gets hot). Pretty
> much anyone with a laptop that locks up and you suspect acpi should test
> it. And anyone who is willing to test it on another brand laptop to be
> sure the patch doesn't break anything more would be welcome.
>
> You should be able to do "sysctl hw.acpi" and see the temperature and
> "apm" to see battery status without any new problems after applying the
> patch.
I've added a patch for 6-stable also (attached). Please test if you
have a laptop and 6.x. See -current for the 7.x patch if that's
relevant to you.
-Nate
--------------070609030008030803090405
Content-Type: text/x-patch;
name="defer_gpe_6x.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="defer_gpe_6x.diff"
Index: sys/dev/acpica/Osd/OsdSchedule.c
===================================================================
RCS file: /home/ncvs/src/sys/dev/acpica/Osd/OsdSchedule.c,v
retrieving revision 1.32.2.6
diff -u -r1.32.2.6 OsdSchedule.c
--- sys/dev/acpica/Osd/OsdSchedule.c 6 Jul 2006 08:32:49 -0000 1.32.2.6
+++ sys/dev/acpica/Osd/OsdSchedule.c 14 Jan 2008 19:06:27 -0000
@@ -106,10 +106,8 @@
at->at_context = Context;
switch (Priority) {
case OSD_PRIORITY_GPE:
- pri = 4;
- break;
case OSD_PRIORITY_HIGH:
- pri = 3;
+ pri = 4;
break;
case OSD_PRIORITY_MED:
pri = 2;
Index: sys/contrib/dev/acpica/evgpe.c
===================================================================
RCS file: /home/ncvs/src/sys/contrib/dev/acpica/evgpe.c,v
retrieving revision 1.1.1.9.2.1
diff -u -r1.1.1.9.2.1 evgpe.c
--- sys/contrib/dev/acpica/evgpe.c 7 Nov 2005 09:55:48 -0000 1.1.1.9.2.1
+++ sys/contrib/dev/acpica/evgpe.c 14 Jan 2008 19:02:15 -0000
@@ -587,6 +587,10 @@
*
******************************************************************************/
+static void
+AcpiEvAsynchEnableGpe (
+ void *Context);
+
static void ACPI_SYSTEM_XFACE
AcpiEvAsynchExecuteGpeMethod (
void *Context)
@@ -656,13 +660,26 @@
}
}
- if ((LocalGpeEventInfo.Flags & ACPI_GPE_XRUPT_TYPE_MASK) == ACPI_GPE_LEVEL_TRIGGERED)
+ /* Defer enabling of GPE until all notify handlers are done */
+ AcpiOsExecute(OSL_NOTIFY_HANDLER, AcpiEvAsynchEnableGpe, GpeEventInfo);
+ return_VOID;
+}
+
+static void
+AcpiEvAsynchEnableGpe (
+ void *Context)
+{
+ ACPI_GPE_EVENT_INFO *GpeEventInfo = (void *) Context;
+ ACPI_STATUS Status;
+
+ if ((GpeEventInfo->Flags & ACPI_GPE_XRUPT_TYPE_MASK) ==
+ ACPI_GPE_LEVEL_TRIGGERED)
{
/*
* GPE is level-triggered, we clear the GPE status bit after
* handling the event.
*/
- Status = AcpiHwClearGpe (&LocalGpeEventInfo);
+ Status = AcpiHwClearGpe (GpeEventInfo);
if (ACPI_FAILURE (Status))
{
return_VOID;
@@ -671,7 +688,7 @@
/* Enable this GPE */
- (void) AcpiHwWriteGpeEnableReg (&LocalGpeEventInfo);
+ (void) AcpiHwWriteGpeEnableReg (GpeEventInfo);
return_VOID;
}
--------------070609030008030803090405--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?478BB3F4.7050600>