Skip to contentSkip to conversation

Force enable HPET on (some?) ICH9 boards

PATCH In review Alistair John Strachan 3 participants 4 messages Jan 27, 2008 – Jan 31, 2008

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Force enable HPET on (some?) ICH9 boards

Some consumer ICH9 boards (such as the Abit IP35 Pro) do not provide a BIOS option for enabling the HPET. The same ICH workaround used for 6,7,8 can be applied to 9. Here I enable the only PCI id that was visible on my system.

I have confirmed the HPETs work both from userspace and as a clocksource for the running kernel (2.6.24 here) after applying this patch.

Force enabled HPET at base address 0xfed00000
hpet clockevent registered
hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0, 0
hpet0: 4 64-bit timers, 14318180 Hz

---
 arch/x86/kernel/quirks.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
arch/x86/kernel/quirks.c +2 -0
index fab30e1..150ba29 100644
@@ -162,6 +162,8 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH7_31,
 			 ich_force_enable_hpet);
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_1,
 			 ich_force_enable_hpet);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH9_7,
+			 ich_force_enable_hpet);
 
 
 static struct pci_dev *cached_dev;
Signature · 4 lines
Cheers,
Alistair.

137/1 Warrender Park Road, Edinburgh, UK.
Acked-by on v1 1/1
#1 lore raw

Patch looks good.
If BIOS does not report HPET on more of such systems we may have to add other chipsets in ICH9 family (ICH9_8, ...) as well.

Quoted 53 lines

-----Original Message-----
From: Alistair John Strachan [mailto:[email protected]]
Sent: Sunday, January 27, 2008 6:33 AM
To: Pallipadi, Venkatesh
Cc: Ingo Molnar; Linux Kernel Mailing List; Alistair John Strachan
Subject: [PATCH] Force enable HPET on (some?) ICH9 boards

Some consumer ICH9 boards (such as the Abit IP35 Pro) do not provide a BIOS
option for enabling the HPET. The same ICH workaround used for 6,7,8 can be
applied to 9. Here I enable the only PCI id that was visible on my system.

I have confirmed the HPETs work both from userspace and as a clocksource for
the running kernel (2.6.24 here) after applying this patch.

Force enabled HPET at base address 0xfed00000
hpet clockevent registered
hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0, 0
hpet0: 4 64-bit timers, 14318180 Hz

---
arch/x86/kernel/quirks.c | 2 ++

1 files changed, 2 insertions(+), 0 deletions(-)
arch/x86/kernel/quirks.c +1 -0
index fab30e1..150ba29 100644
@@ -162,6 +162,8 @@ 
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 
 PCI_DEVICE_ID_INTEL_ICH7_31,
 			 ich_force_enable_hpet);
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 
 PCI_DEVICE_ID_INTEL_ICH8_1,
 			 ich_force_enable_hpet);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 

PCI_DEVICE_ID_INTEL_ICH9_7,

arch/x86/kernel/quirks.c +1 -0
+			 ich_force_enable_hpet);

static struct pci_dev *cached_dev;

Signature · 4 lines
Cheers,
Alistair.

137/1 Warrender Park Road, Edinburgh, UK.
#2 lore raw
Quoted 5 lines · Venkatesh

Patch looks good.
If BIOS does not report HPET on more of such systems we may have to add other chipsets in ICH9 family (ICH9_8, ...) as well.

Hi Ingo,

Are you going to pick this patch up via the x86 tree, or shall I forward it to Andrew or trivial@?

Do you want me to re-send with Venki's ack?

Signature · 4 lines
Cheers,
Alistair.

137/1 Warrender Park Road, Edinburgh, UK.
#3 lore raw
Quoted 4 lines · Alistair John Strachan

Are you going to pick this patch up via the x86 tree, or shall I forward it to Andrew or trivial@?

Do you want me to re-send with Venki's ack?

no need, we merged your improvement via x86.git and it is now in Linus's latest tree:

commit dff244af014144e4ac87dfc5b6e450dc8832710e
Author: Alistair John Strachan <[email protected]>
Date: Wed Jan 30 13:33:39 2008 +0100

x86: force enable HPET on (some?) ICH9 boards

it should first show up in v2.6.25-rc1. Thanks,

Ingo