mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Suresh Siddha <suresh.b.siddha@intel.com>
To: Jeff Garzik <jgarzik@pobox.com>
Cc: Zwane Mwaikambo <zwane@fsmlabs.com>,
	Suresh Siddha <suresh.b.siddha@intel.com>,
	linux-kernel@vger.kernel.org, asit.k.mallick@intel.com
Subject: Re: [Patch 0/2] Disable SW irqbalance/irqaffinity for E7520/E7320/E7525
Date: Fri, 24 Sep 2004 15:20:27 -0700	[thread overview]
Message-ID: <20040924152026.A25742@unix-os.sc.intel.com> (raw)
In-Reply-To: <4154828F.6090205@pobox.com>; from jgarzik@pobox.com on Fri, Sep 24, 2004 at 04:24:47PM -0400

On Fri, Sep 24, 2004 at 04:24:47PM -0400, Jeff Garzik wrote:
> Zwane Mwaikambo wrote:
> > On Thu, 23 Sep 2004, Suresh Siddha wrote:
> > 
> > 
> >>As part of the workaround for the "Interrupt message re-ordering across
> >>hub interface" errata (page #16 in
> >>http://developer.intel.com/design/chipsets/specupdt/30288402.pdf),
> >>BIOS may enable hardware IRQ balancing for Lindenhurst/Tumwater based chipset
> >>platforms. Software based irq_balance/irq_affinity should be disabled if
> >>hardware IRQ balancing is enabled.
> > 
> > 
> > Can we avoid those tests? Like not starting the irq balancer code at all 
> > with those chipsets?
> 
> 
> That's my preference.
> 

Ok. How about this patch?

Add pci quirks to disable irqbalance/affinity for E7520/E7320/E7525
with revision ID 0x09 and below.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>

diff -Nru linux-2.6.9-rc2/arch/i386/kernel/io_apic.c linux-irq/arch/i386/kernel/io_apic.c
--- linux-2.6.9-rc2/arch/i386/kernel/io_apic.c	2004-09-12 22:32:00.000000000 -0700
+++ linux-irq/arch/i386/kernel/io_apic.c	2004-09-04 12:12:42.499670496 -0700
@@ -31,6 +31,7 @@
 #include <linux/mc146818rtc.h>
 #include <linux/compiler.h>
 #include <linux/acpi.h>
+#include <linux/pci.h>
 
 #include <linux/sysdev.h>
 #include <asm/io.h>
@@ -260,7 +261,7 @@
 cpumask_t __cacheline_aligned pending_irq_balance_cpumask[NR_IRQS];
 
 #define IRQBALANCE_CHECK_ARCH -999
-static int irqbalance_disabled = IRQBALANCE_CHECK_ARCH;
+int irqbalance_disabled = IRQBALANCE_CHECK_ARCH;
 static int physical_balance = 0;
 
 struct irq_cpu_info {
@@ -582,6 +583,30 @@
 	return 0;
 }
 
+/* HW irqbalance may be enabled for E7520/E7320/E7525 with revision ID 
+ * 0x09 and below. Disable sw irqbalance in that case
+ */
+static void check_hw_irqbalance(void)
+{
+	struct pci_dev *dev = NULL;
+	extern void quirk_intel_irqbalance(struct pci_dev *dev);
+
+	dev = pci_find_device(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_SMCH, 
+			      dev);
+	if (!dev)
+		dev = pci_find_device(PCI_VENDOR_ID_INTEL, 
+				      PCI_DEVICE_ID_INTEL_E7320_MCH, dev);
+	if (!dev)
+		dev = pci_find_device(PCI_VENDOR_ID_INTEL, 
+				      PCI_DEVICE_ID_INTEL_E7525_MCH, dev);
+	if (!dev)
+		return;
+
+	quirk_intel_irqbalance(dev);
+
+	return;	
+}
+
 static int __init balanced_irq_init(void)
 {
 	int i;
@@ -590,6 +615,8 @@
 
 	cpus_shift_right(tmp, cpu_online_map, 2);
         c = &boot_cpu_data;
+
+	check_hw_irqbalance();
 	/* When not overwritten by the command line ask subarchitecture. */
 	if (irqbalance_disabled == IRQBALANCE_CHECK_ARCH)
 		irqbalance_disabled = NO_BALANCE_IRQ;
@@ -636,7 +663,7 @@
 	return 0;
 }
 
-static int __init irqbalance_disable(char *str)
+int __init irqbalance_disable(char *str)
 {
 	irqbalance_disabled = 1;
 	return 0;
diff -Nru linux-2.6.9-rc2/arch/i386/kernel/irq.c linux-irq/arch/i386/kernel/irq.c
--- linux-2.6.9-rc2/arch/i386/kernel/irq.c	2004-09-12 22:31:30.000000000 -0700
+++ linux-irq/arch/i386/kernel/irq.c	2004-09-04 11:48:21.000000000 -0700
@@ -272,7 +272,7 @@
 
 static int noirqdebug;
 
-static int __init noirqdebug_setup(char *str)
+int __init noirqdebug_setup(char *str)
 {
 	noirqdebug = 1;
 	printk("IRQ lockup detection disabled\n");
@@ -997,13 +997,15 @@
 	return len;
 }
 
+int no_irq_affinity;
+
 static int irq_affinity_write_proc(struct file *file, const char __user *buffer,
 					unsigned long count, void *data)
 {
 	int irq = (long)data, full_count = count, err;
 	cpumask_t new_value, tmp;
 
-	if (!irq_desc[irq].handler->set_affinity)
+	if (!irq_desc[irq].handler->set_affinity || no_irq_affinity)
 		return -EIO;
 
 	err = cpumask_parse(buffer, count, new_value);
diff -Nru linux-2.6.9-rc2/arch/x86_64/kernel/irq.c linux-irq/arch/x86_64/kernel/irq.c
--- linux-2.6.9-rc2/arch/x86_64/kernel/irq.c	2004-09-12 22:32:26.000000000 -0700
+++ linux-irq/arch/x86_64/kernel/irq.c	2004-09-04 11:48:23.000000000 -0700
@@ -835,6 +835,8 @@
 	return len;
 }
 
+int no_irq_affinity;
+
 static int irq_affinity_write_proc (struct file *file,
 					const char __user *buffer,
 					unsigned long count, void *data)
@@ -842,7 +844,7 @@
 	int irq = (long) data, full_count = count, err;
 	cpumask_t tmp, new_value;
 
-	if (!irq_desc[irq].handler->set_affinity)
+	if (!irq_desc[irq].handler->set_affinity || no_irq_affinity)
 		return -EIO;
 
 	err = cpumask_parse(buffer, count, new_value);
diff -Nru linux-2.6.9-rc2/drivers/pci/quirks.c linux-irq/drivers/pci/quirks.c
--- linux-2.6.9-rc2/drivers/pci/quirks.c	2004-09-12 22:31:27.000000000 -0700
+++ linux-irq/drivers/pci/quirks.c	2004-09-04 12:33:54.373316312 -0700
@@ -814,6 +814,64 @@
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_82801DB_12,	asus_hides_smbus_lpc );
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_82801EB_0,	asus_hides_smbus_lpc );
 
+#ifdef CONFIG_X86_IO_APIC
+#include <asm/hw_irq.h>
+#ifdef CONFIG_IRQBALANCE
+extern int irqbalance_disable(char *str);
+#endif
+extern int no_irq_affinity;
+extern int noirqdebug_setup(char *str);
+
+
+void __devinit quirk_intel_irqbalance(struct pci_dev *dev)
+{
+	u8 config, rev;
+	u32 word;
+	extern struct pci_raw_ops *raw_pci_ops;
+#ifdef CONFIG_IRQBALANCE
+	extern int irqbalance_disabled;
+
+	/* we might have come through check_hw_irqbalance()
+	 * and enabled the workaround
+	 */
+	if (irqbalance_disabled == 1 && no_irq_affinity)
+		return;
+#endif
+
+	pci_read_config_byte(dev, PCI_CLASS_REVISION, &rev);
+	if (rev > 0x9)
+		return;
+
+	printk(KERN_INFO "Intel E7520/7320/7525 detected.");
+
+	/* enable access to config space*/
+	pci_read_config_byte(dev, 0xf4, &config);
+	config |= 0x2;
+	pci_write_config_byte(dev, 0xf4, config);
+
+	/* read xTPR register */
+	raw_pci_ops->read(0, 0, 0x40, 0x4c, 2, &word);
+
+	if (!(word & (1 << 13))) {
+		printk(KERN_INFO "Disabling irq balancing and affinity\n");
+#ifdef __i386__
+#ifdef CONFIG_IRQBALANCE
+		irqbalance_disable("");
+#endif
+		noirqdebug_setup("");
+#endif
+		no_irq_affinity = 1;
+	}
+
+	config &= ~0x2;
+	/* disable access to config space*/
+	pci_write_config_byte(dev, 0xf4, config);
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_SMCH,	quirk_intel_irqbalance);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_E7320_MCH,	quirk_intel_irqbalance);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_E7525_MCH,	quirk_intel_irqbalance);
+#endif
+
 /*
  * SiS 96x south bridge: BIOS typically hides SMBus device...
  */
diff -Nru linux-2.6.9-rc2/include/linux/pci_ids.h linux-irq/include/linux/pci_ids.h
--- linux-2.6.9-rc2/include/linux/pci_ids.h	2004-09-12 22:33:54.000000000 -0700
+++ linux-irq/include/linux/pci_ids.h	2004-09-04 11:48:26.000000000 -0700
@@ -2204,6 +2204,8 @@
 #define PCI_DEVICE_ID_INTEL_82855GM_HB	0x3580
 #define PCI_DEVICE_ID_INTEL_82855GM_IG	0x3582
 #define PCI_DEVICE_ID_INTEL_SMCH	0x3590
+#define PCI_DEVICE_ID_INTEL_E7320_MCH	0x3592
+#define PCI_DEVICE_ID_INTEL_E7525_MCH	0x359e
 #define PCI_DEVICE_ID_INTEL_80310	0x530d
 #define PCI_DEVICE_ID_INTEL_82371SB_0	0x7000
 #define PCI_DEVICE_ID_INTEL_82371SB_1	0x7010

  reply	other threads:[~2004-09-24 22:21 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-09-24  6:34 Suresh Siddha
2004-09-24 15:05 ` Zwane Mwaikambo
2004-09-24 20:24   ` Jeff Garzik
2004-09-24 22:20     ` Suresh Siddha [this message]
2004-09-25  9:15       ` Zwane Mwaikambo
2004-09-27 18:03         ` Suresh Siddha
2004-09-28 21:11           ` Suresh Siddha
2004-09-29 15:16             ` Zwane Mwaikambo
2004-09-30 22:51             ` Andrew Morton
2004-10-01  1:15               ` Suresh Siddha
2004-09-29 13:36           ` Zwane Mwaikambo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20040924152026.A25742@unix-os.sc.intel.com \
    --to=suresh.b.siddha@intel.com \
    --cc=asit.k.mallick@intel.com \
    --cc=jgarzik@pobox.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=zwane@fsmlabs.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome