Skip to contentSkip to conversation

x86, HPET: mark HPET PCI BARs as IORESOURCE_PCI_FIXED

PATCH Superseded Bjorn Helgaas 3 participants 7 messages Sep 23, 2010 – Sep 25, 2010

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86, HPET: mark HPET PCI BARs as IORESOURCE_PCI_FIXED

We often discover the HPET early, via the static ACPI HPET table, before enumerating PCI devices. If the HPET is implemented as a PCI function, we will discover it again during PCI device enumeration. We must make sure we never move the PCI function out from under the driver.

Doing this for *any* PCI BAR that matches a previously discovered HPET should be more generic than the previous ATI-specific quirk.

This is for a regression from 2.6.34, but the reporter has been unable to test it yet.

Reference: https://bugzilla.kernel.org/show_bug.cgi?id=18482

---

 arch/x86/kernel/quirks.c |   19 +++++++++++++++++++
 arch/x86/pci/fixup.c     |   28 ----------------------------
 2 files changed, 19 insertions(+), 28 deletions(-)
arch/x86/kernel/quirks.c +19 -0
index 939b9e9..8acd667 100644
@@ -507,6 +507,25 @@ static void force_disable_hpet_msi(struct pci_dev *unused)
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS,
 			 force_disable_hpet_msi);
 
+static void mark_pci_hpet_fixed(struct pci_dev *dev)
+{
+	int i;
+
+	if (!hpet_address)
+		return;
+
+	for (i = 0; i < 6; i++) {
+		struct resource *res = &dev->resource[i];
+
+		if (resource_type(res) == IORESOURCE_MEM &&
+		    res->start == hpet_address) {
+			dev_info(&dev->dev, "BAR %d: %pR is an HPET we found earlier; making it immovable\n",
+				 i, res);
+			res->flags |= IORESOURCE_PCI_FIXED;
+		}
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, mark_pci_hpet_fixed);
 #endif
 
 #if defined(CONFIG_PCI) && defined(CONFIG_NUMA)
arch/x86/pci/fixup.c +0 -28
index 6dd8955..08eba69 100644
@@ -493,31 +493,3 @@ static void __devinit pci_siemens_interrupt_controller(struct pci_dev *dev)
 }
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_SIEMENS, 0x0015,
 			  pci_siemens_interrupt_controller);
-
-/*
- * SB600: Disable BAR1 on device 14.0 to avoid HPET resources from
- * confusing the PCI engine:
- */
-static void sb600_disable_hpet_bar(struct pci_dev *dev)
-{
-	u8 val;
-
-	/*
-	 * The SB600 and SB700 both share the same device
-	 * ID, but the PM register 0x55 does something different
-	 * for the SB700, so make sure we are dealing with the
-	 * SB600 before touching the bit:
-	 */
-
-	pci_read_config_byte(dev, 0x08, &val);
-
-	if (val < 0x2F) {
-		outb(0x55, 0xCD6);
-		val = inb(0xCD7);
-
-		/* Set bit 7 in PM register 0x55 */
-		outb(0x55, 0xCD6);
-		outb(val | 0x80, 0xCD7);
-	}
-}
-DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_ATI, 0x4385, sb600_disable_hpet_bar);
#1 lore raw

Sorry for the duplicate; I tried to include Frédéric's proper name, but stg mail got horribly confused and made a mess of the cc: list. This one should be better.

#2 lore raw
In reply to Bjorn Helgaas

Sorry Yinghai, I forgot to copy you on the initial post. Any comments on this? If not, we should try to get it into linux-next soon.

Quoted 90 lines · Bjorn Helgaas

We often discover the HPET early, via the static ACPI HPET table, before enumerating PCI devices. If the HPET is implemented as a PCI function, we will discover it again during PCI device enumeration. We must make sure we never move the PCI function out from under the driver.

Doing this for *any* PCI BAR that matches a previously discovered HPET should be more generic than the previous ATI-specific quirk.

This is for a regression from 2.6.34, but the reporter has been unable to test it yet.

Reference: https://bugzilla.kernel.org/show_bug.cgi?id=18482

---

 arch/x86/kernel/quirks.c |   19 +++++++++++++++++++
 arch/x86/pci/fixup.c     |   28 ----------------------------
 2 files changed, 19 insertions(+), 28 deletions(-)
arch/x86/kernel/quirks.c +19 -0
index 939b9e9..8acd667 100644
@@ -507,6 +507,25 @@ static void force_disable_hpet_msi(struct pci_dev *unused)
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS,
 			 force_disable_hpet_msi);
 
+static void mark_pci_hpet_fixed(struct pci_dev *dev)
+{
+	int i;
+
+	if (!hpet_address)
+		return;
+
+	for (i = 0; i < 6; i++) {
+		struct resource *res = &dev->resource[i];
+
+		if (resource_type(res) == IORESOURCE_MEM &&
+		    res->start == hpet_address) {
+			dev_info(&dev->dev, "BAR %d: %pR is an HPET we found earlier; making it immovable\n",
+				 i, res);
+			res->flags |= IORESOURCE_PCI_FIXED;
+		}
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, mark_pci_hpet_fixed);
 #endif
 
 #if defined(CONFIG_PCI) && defined(CONFIG_NUMA)
arch/x86/pci/fixup.c +0 -28
index 6dd8955..08eba69 100644
@@ -493,31 +493,3 @@ static void __devinit pci_siemens_interrupt_controller(struct pci_dev *dev)
 }
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_SIEMENS, 0x0015,
 			  pci_siemens_interrupt_controller);
-
-/*
- * SB600: Disable BAR1 on device 14.0 to avoid HPET resources from
- * confusing the PCI engine:
- */
-static void sb600_disable_hpet_bar(struct pci_dev *dev)
-{
-	u8 val;
-
-	/*
-	 * The SB600 and SB700 both share the same device
-	 * ID, but the PM register 0x55 does something different
-	 * for the SB700, so make sure we are dealing with the
-	 * SB600 before touching the bit:
-	 */
-
-	pci_read_config_byte(dev, 0x08, &val);
-
-	if (val < 0x2F) {
-		outb(0x55, 0xCD6);
-		val = inb(0xCD7);
-
-		/* Set bit 7 in PM register 0x55 */
-		outb(0x55, 0xCD6);
-		outb(val | 0x80, 0xCD7);
-	}
-}
-DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_ATI, 0x4385, sb600_disable_hpet_bar);
#3 lore raw
Quoted 3 lines · Bjorn Helgaas

Sorry Yinghai, I forgot to copy you on the initial post.  Any comments on this?  If not, we should try to get it into linux-next soon.

should be good.

only one thing, that will punish all other vendors... and run that quirk for every pci device.

please check if you can have +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_ANY_ID, mark_pci_hpet_fixed);

AND _AMD maybe, if AMD dump _ATI brand later.

Thanks

Yinghai

#4 lore raw
Quoted 7 lines · Yinghai Lu
Quoted 3 lines · Bjorn Helgaas

Sorry Yinghai, I forgot to copy you on the initial post. Any comments on this? If not, we should try to get it into linux-next soon.

should be good.

Quoted 8 lines

only one thing, that will punish all other vendors... and run that quirk for every pci device.

please check if you can have +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_ANY_ID, mark_pci_hpet_fixed);

AND _AMD maybe, if AMD dump _ATI brand later.

No, because the problem is known to happen on multiple vendors.

-hpa

#5 lore raw
Quoted 11 lines · H. Peter Anvin
Quoted 10 lines · Yinghai Lu
Quoted 3 lines · Bjorn Helgaas

Sorry Yinghai, I forgot to copy you on the initial post. Any comments on this? If not, we should try to get it into linux-next soon.

should be good.

only one thing, that will punish all other vendors... and run that quirk for every pci device.

Quoted 8 lines
Quoted 5 lines

please check if you can have +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_ANY_ID, mark_pci_hpet_fixed);

AND _AMD maybe, if AMD dump _ATI brand later.

No, because the problem is known to happen on multiple vendors.

Do you have URL to them?

INTEL ?
ServerWorks?
Via?

Yinghai

#6 lore raw
Quoted 9 lines · Yinghai Lu
Quoted 3 lines

No, because the problem is known to happen on multiple vendors.

Do you have URL to them?

INTEL ?
ServerWorks?
Via?

It'll happen on any platform where the HPET is a BAR. There is no point in enumerating it by vendor, since there is a static test for it.

-hpa