[WATCHDOG] hpwdt: Add NMI sourcing

Add NMI sourcing functionality (Can only be active if nmi_watchdog is inactive). Signed-off-by: Thomas Mingarelli <thomas.mingarelli@hp.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>

[WATCHDOG] hpwdt: Add NMI sourcing
Add NMI sourcing functionality (Can only be active if nmi_watchdog is inactive). Signed-off-by: Thomas Mingarelli <thomas.mingarelli@hp.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Thomas Mingarelli · Wim Van Sebroeck
1 parent 55e8ddecec
Showing 2 changed files with 128 additions and 15 deletions Side-by-side Diff
Documentation/watchdog/hpwdt.txt
drivers/watchdog/hpwdt.c
+Last reviewed: 06/02/2009
+
+                     HP iLO2 NMI Watchdog Driver
+              NMI sourcing for iLO2 based ProLiant Servers
+                     Documentation and Driver by
+              Thomas Mingarelli <thomas.mingarelli@hp.com>
+
+ The HP iLO2 NMI Watchdog driver is a kernel module that provides basic
+ watchdog functionality and the added benefit of NMI sourcing. Both the
+ watchdog functionality and the NMI sourcing capability need to be enabled
+ by the user. Remember that the two modes are not dependant on one another.
+ A user can have the NMI sourcing without the watchdog timer and vice-versa.
+
+ Watchdog functionality is enabled like any other common watchdog driver. That
+ is, an application needs to be started that kicks off the watchdog timer. A
+ basic application exists in the Documentation/watchdog/src directory called
+ watchdog-test.c. Simply compile the C file and kick it off. If the system
+ gets into a bad state and hangs, the HP ProLiant iLO 2 timer register will
+ not be updated in a timely fashion and a hardware system reset (also known as
+ an Automatic Server Recovery (ASR)) event will occur.
+
+ The hpwdt driver also has three (3) module parameters. They are the following:
+
+ soft_margin - allows the user to set the watchdog timer value
+ allow_kdump - allows the user to save off a kernel dump image after an NMI
+ nowayout    - basic watchdog parameter that does not allow the timer to
+               be restarted or an impending ASR to be escaped.
+
+ NOTE: More information about watchdog drivers in general, including the ioctl
+       interface to /dev/watchdog can be found in
+       Documentation/watchdog/watchdog-api.txt and Documentation/IPMI.txt.
+
+ The NMI sourcing capability is disabled when the driver discovers that the
+ nmi_watchdog is turned on (nmi_watchdog = 1). This is due to the inability to
+ distinguish between "NMI Watchdog Ticks" and "HW generated NMI events" in the
+ Linux kernel. What this means is that the hpwdt nmi handler code is called
+ each time the NMI signal fires off. This could amount to several thousands of
+ NMIs in a matter of seconds. If a user sees the Linux kernel's "dazed and
+ confused" message in the logs or if the system gets into a hung state, then
+ the user should reboot with nmi_watchdog=0.
+
+ 1. If the kernel has not been booted with nmi_watchdog turned off then
+    edit /boot/grub/menu.lst and place the nmi_watchdog=0 at the end of the
+    currently booting kernel line.
+ 2. reboot the sever
+
+ Now, the hpwdt can successfully receive and source the NMI and provide a log
+ message that details the reason for the NMI (as determined by the HP BIOS).
+
+ Below is a list of NMIs the HP BIOS understands along with the associated
+ code (reason):
+
+	No source found                00h
+
+	Uncorrectable Memory Error     01h
+
+	ASR NMI                        1Bh
+
+	PCI Parity Error               20h
+
+	NMI Button Press               27h
+
+	SB_BUS_NMI                     28h
+
+	ILO Doorbell NMI               29h
+
+	ILO IOP NMI                    2Ah
+
+	ILO Watchdog NMI               2Bh
+
+	Proc Throt NMI                 2Ch
+
+	Front Side Bus NMI             2Dh
+
+	PCI Express Error              2Fh
+
+	DMA controller NMI             30h
+
+	Hypertransport/CSI Error       31h
+
+
+
+ -- Tom Mingarelli
+    (thomas.mingarelli@hp.com)
@@ -19,6 +19,7 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/irq.h>
+#include <linux/nmi.h>
 #include <linux/kernel.h>
 #include <linux/miscdevice.h>
 #include <linux/mm.h>
@@ -47,7 +48,7 @@
 #define PCI_BIOS32_PARAGRAPH_LEN	16
 #define PCI_ROM_BASE1			0x000F0000
 #define ROM_SIZE			0x10000
-#define HPWDT_VERSION			"1.01"
+#define HPWDT_VERSION			"1.1.1"
  
 struct bios32_service_dir {
 	u32 signature;
@@ -119,6 +120,7 @@
 static char expect_release;
 static unsigned long hpwdt_is_open;
 static unsigned int allow_kdump;
+static int hpwdt_nmi_sourcing;
  
 static void __iomem *pci_mem_addr;		/* the PCI-memory address */
 static unsigned long __iomem *hpwdt_timer_reg;
  
@@ -468,21 +470,22 @@
 	if (ulReason != DIE_NMI && ulReason != DIE_NMI_IPI)
 		return NOTIFY_OK;
  
-	spin_lock_irqsave(&rom_lock, rom_pl);
-	if (!die_nmi_called)
-		asminline_call(&cmn_regs, cru_rom_addr);
-	die_nmi_called = 1;
-	spin_unlock_irqrestore(&rom_lock, rom_pl);
-	if (cmn_regs.u1.ral == 0) {
-		printk(KERN_WARNING "hpwdt: An NMI occurred, "
-			"but unable to determine source.\n");
-	} else {
-		if (allow_kdump)
-			hpwdt_stop();
-		panic("An NMI occurred, please see the Integrated "
-			"Management Log for details.\n");
+	if (hpwdt_nmi_sourcing) {
+		spin_lock_irqsave(&rom_lock, rom_pl);
+		if (!die_nmi_called)
+			asminline_call(&cmn_regs, cru_rom_addr);
+		die_nmi_called = 1;
+		spin_unlock_irqrestore(&rom_lock, rom_pl);
+		if (cmn_regs.u1.ral == 0) {
+			printk(KERN_WARNING "hpwdt: An NMI occurred, "
+				"but unable to determine source.\n");
+		} else {
+			if (allow_kdump)
+				hpwdt_stop();
+			panic("An NMI occurred, please see the Integrated "
+				"Management Log for details.\n");
+		}
 	}
-
 	return NOTIFY_OK;
 }
  
  
@@ -627,10 +630,36 @@
  *	Init & Exit
  */
  
+#ifdef ARCH_HAS_NMI_WATCHDOG
+static void __devinit hpwdt_check_nmi_sourcing(struct pci_dev *dev)
+{
+	/*
+	 * If nmi_watchdog is turned off then we can turn on
+	 * our nmi sourcing capability.
+	 */
+	if (!nmi_watchdog_active())
+		hpwdt_nmi_sourcing = 1;
+	else
+		dev_warn(&dev->dev, "NMI sourcing is disabled. To enable this "
+			"functionality you must reboot with nmi_watchdog=0.\n");
+}
+#else
+static void __devinit hpwdt_check_nmi_sourcing(struct pci_dev *dev)
+{
+	dev_warn(&dev->dev, "NMI sourcing is disabled. "
+		"Your kernel does not support a NMI Watchdog.\n");
+}
+#endif
+
 static int __devinit hpwdt_init_one(struct pci_dev *dev,
 					const struct pci_device_id *ent)
 {
 	int retval;
+
+	/*
+	 * Check if we can do NMI sourcing or not
+	 */
+	hpwdt_check_nmi_sourcing(dev);
  
 	/*
 	 * First let's find out if we are on an iLO2 server. We will
	1	+Last reviewed: 06/02/2009
	2	+
	3	+ HP iLO2 NMI Watchdog Driver
	4	+ NMI sourcing for iLO2 based ProLiant Servers
	5	+ Documentation and Driver by
	6	+ Thomas Mingarelli <thomas.mingarelli@hp.com>
	7	+
	8	+ The HP iLO2 NMI Watchdog driver is a kernel module that provides basic
	9	+ watchdog functionality and the added benefit of NMI sourcing. Both the
	10	+ watchdog functionality and the NMI sourcing capability need to be enabled
	11	+ by the user. Remember that the two modes are not dependant on one another.
	12	+ A user can have the NMI sourcing without the watchdog timer and vice-versa.
	13	+
	14	+ Watchdog functionality is enabled like any other common watchdog driver. That
	15	+ is, an application needs to be started that kicks off the watchdog timer. A
	16	+ basic application exists in the Documentation/watchdog/src directory called
	17	+ watchdog-test.c. Simply compile the C file and kick it off. If the system
	18	+ gets into a bad state and hangs, the HP ProLiant iLO 2 timer register will
	19	+ not be updated in a timely fashion and a hardware system reset (also known as
	20	+ an Automatic Server Recovery (ASR)) event will occur.
	21	+
	22	+ The hpwdt driver also has three (3) module parameters. They are the following:
	23	+
	24	+ soft_margin - allows the user to set the watchdog timer value
	25	+ allow_kdump - allows the user to save off a kernel dump image after an NMI
	26	+ nowayout - basic watchdog parameter that does not allow the timer to
	27	+ be restarted or an impending ASR to be escaped.
	28	+
	29	+ NOTE: More information about watchdog drivers in general, including the ioctl
	30	+ interface to /dev/watchdog can be found in
	31	+ Documentation/watchdog/watchdog-api.txt and Documentation/IPMI.txt.
	32	+
	33	+ The NMI sourcing capability is disabled when the driver discovers that the
	34	+ nmi_watchdog is turned on (nmi_watchdog = 1). This is due to the inability to
	35	+ distinguish between "NMI Watchdog Ticks" and "HW generated NMI events" in the
	36	+ Linux kernel. What this means is that the hpwdt nmi handler code is called
	37	+ each time the NMI signal fires off. This could amount to several thousands of
	38	+ NMIs in a matter of seconds. If a user sees the Linux kernel's "dazed and
	39	+ confused" message in the logs or if the system gets into a hung state, then
	40	+ the user should reboot with nmi_watchdog=0.
	41	+
	42	+ 1. If the kernel has not been booted with nmi_watchdog turned off then
	43	+ edit /boot/grub/menu.lst and place the nmi_watchdog=0 at the end of the
	44	+ currently booting kernel line.
	45	+ 2. reboot the sever
	46	+
	47	+ Now, the hpwdt can successfully receive and source the NMI and provide a log
	48	+ message that details the reason for the NMI (as determined by the HP BIOS).
	49	+
	50	+ Below is a list of NMIs the HP BIOS understands along with the associated
	51	+ code (reason):
	52	+
	53	+ No source found 00h
	54	+
	55	+ Uncorrectable Memory Error 01h
	56	+
	57	+ ASR NMI 1Bh
	58	+
	59	+ PCI Parity Error 20h
	60	+
	61	+ NMI Button Press 27h
	62	+
	63	+ SB_BUS_NMI 28h
	64	+
	65	+ ILO Doorbell NMI 29h
	66	+
	67	+ ILO IOP NMI 2Ah
	68	+
	69	+ ILO Watchdog NMI 2Bh
	70	+
	71	+ Proc Throt NMI 2Ch
	72	+
	73	+ Front Side Bus NMI 2Dh
	74	+
	75	+ PCI Express Error 2Fh
	76	+
	77	+ DMA controller NMI 30h
	78	+
	79	+ Hypertransport/CSI Error 31h
	80	+
	81	+
	82	+
	83	+ -- Tom Mingarelli
	84	+ (thomas.mingarelli@hp.com)
...	...	@@ -19,6 +19,7 @@
19	19	#include <linux/interrupt.h>
20	20	#include <linux/io.h>
21	21	#include <linux/irq.h>
	22	+#include <linux/nmi.h>
22	23	#include <linux/kernel.h>
23	24	#include <linux/miscdevice.h>
24	25	#include <linux/mm.h>
...	...	@@ -47,7 +48,7 @@
47	48	#define PCI_BIOS32_PARAGRAPH_LEN 16
48	49	#define PCI_ROM_BASE1 0x000F0000
49	50	#define ROM_SIZE 0x10000
50		-#define HPWDT_VERSION "1.01"
	51	+#define HPWDT_VERSION "1.1.1"
51	52
52	53	struct bios32_service_dir {
53	54	u32 signature;
...	...	@@ -119,6 +120,7 @@
119	120	static char expect_release;
120	121	static unsigned long hpwdt_is_open;
121	122	static unsigned int allow_kdump;
	123	+static int hpwdt_nmi_sourcing;
122	124
123	125	static void __iomem pci_mem_addr; / the PCI-memory address */
124	126	static unsigned long __iomem *hpwdt_timer_reg;
125	127
...	...	@@ -468,21 +470,22 @@
468	470	if (ulReason != DIE_NMI && ulReason != DIE_NMI_IPI)
469	471	return NOTIFY_OK;
470	472
471		- spin_lock_irqsave(&rom_lock, rom_pl);
472		- if (!die_nmi_called)
473		- asminline_call(&cmn_regs, cru_rom_addr);
474		- die_nmi_called = 1;
475		- spin_unlock_irqrestore(&rom_lock, rom_pl);
476		- if (cmn_regs.u1.ral == 0) {
477		- printk(KERN_WARNING "hpwdt: An NMI occurred, "
478		- "but unable to determine source.\n");
479		- } else {
480		- if (allow_kdump)
481		- hpwdt_stop();
482		- panic("An NMI occurred, please see the Integrated "
483		- "Management Log for details.\n");
	473	+ if (hpwdt_nmi_sourcing) {
	474	+ spin_lock_irqsave(&rom_lock, rom_pl);
	475	+ if (!die_nmi_called)
	476	+ asminline_call(&cmn_regs, cru_rom_addr);
	477	+ die_nmi_called = 1;
	478	+ spin_unlock_irqrestore(&rom_lock, rom_pl);
	479	+ if (cmn_regs.u1.ral == 0) {
	480	+ printk(KERN_WARNING "hpwdt: An NMI occurred, "
	481	+ "but unable to determine source.\n");
	482	+ } else {
	483	+ if (allow_kdump)
	484	+ hpwdt_stop();
	485	+ panic("An NMI occurred, please see the Integrated "
	486	+ "Management Log for details.\n");
	487	+ }
484	488	}
485		-
486	489	return NOTIFY_OK;
487	490	}
488	491
489	492
...	...	@@ -627,10 +630,36 @@
627	630	* Init & Exit
628	631	*/
629	632
	633	+#ifdef ARCH_HAS_NMI_WATCHDOG
	634	+static void __devinit hpwdt_check_nmi_sourcing(struct pci_dev *dev)
	635	+{
	636	+ /*
	637	+ * If nmi_watchdog is turned off then we can turn on
	638	+ * our nmi sourcing capability.
	639	+ */
	640	+ if (!nmi_watchdog_active())
	641	+ hpwdt_nmi_sourcing = 1;
	642	+ else
	643	+ dev_warn(&dev->dev, "NMI sourcing is disabled. To enable this "
	644	+ "functionality you must reboot with nmi_watchdog=0.\n");
	645	+}
	646	+#else
	647	+static void __devinit hpwdt_check_nmi_sourcing(struct pci_dev *dev)
	648	+{
	649	+ dev_warn(&dev->dev, "NMI sourcing is disabled. "
	650	+ "Your kernel does not support a NMI Watchdog.\n");
	651	+}
	652	+#endif
	653	+
630	654	static int __devinit hpwdt_init_one(struct pci_dev *dev,
631	655	const struct pci_device_id *ent)
632	656	{
633	657	int retval;
	658	+
	659	+ /*
	660	+ * Check if we can do NMI sourcing or not
	661	+ */
	662	+ hpwdt_check_nmi_sourcing(dev);
634	663
635	664	/*
636	665	* First let's find out if we are on an iLO2 server. We will