Commit ac325acd authored by Linas Vepstas's avatar Linas Vepstas Committed by Paul Mackerras
Browse files

[PATCH] powerpc/pseries: clear PCI failure counter if no new failures 



The current PCI error recovery system keeps track of the number of PCI card
resets, and refuses to bring a card back up if this number is too large.
The goal of doing this was to avoid an infinite loop of resets if a card is
obviously dead.  However, if the failures are rare, but the machine has a
high uptime, this mechanism might still be triggered; this is too harsh.

This patch will avoids this problem by decrementing the fail count after an
hour.  Thus, as long as a pci card BSOD's less than 6 times an hour, it
will continue to be reset indefinitely.  If it's failure rate is greater
than that, it will be taken off-line permanently.

This patch is larger than it might otherwise be because it changes
indentation by removing a pointless while-loop.  The while loop is not
needed, as the handler is invoked once fo each event (by schedule_work());
the loop is leftover cruft from an earlier implementation.

Signed-off-by: default avatarLinas Vepstas <linas@austin.ibm.com>
Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
Signed-off-by: default avatarPaul Mackerras <paulus@samba.org>
parent 4bd174fe

arch/powerpc/platforms/pseries/eeh_driver.c

+7 −6
Original line number Diff line number Diff line
@@ -23,9 +23,8 @@ 
 *

 */

#include <linux/delay.h>

#include <linux/irq.h>

#include <linux/interrupt.h>

#include <linux/notifier.h>

#include <linux/irq.h>

#include <linux/pci.h>

#include <asm/eeh.h>

#include <asm/eeh_event.h>
@@ -250,7 +249,7 @@ static int eeh_reset_device (struct pci_dn *pe_dn, struct pci_bus *bus) 
 */

#define MAX_WAIT_FOR_RECOVERY 15



void handle_eeh_events (struct eeh_event *event)

struct pci_dn * handle_eeh_events (struct eeh_event *event)

{

	struct device_node *frozen_dn;

	struct pci_dn *frozen_pdn;
@@ -265,7 +264,7 @@ void handle_eeh_events (struct eeh_event *event) 
	if (!frozen_dn) {

		printk(KERN_ERR "EEH: Error: Cannot find partition endpoint for %s\n",

		        pci_name(event->dev));

		return;

		return NULL;

	}



	/* There are two different styles for coming up with the PE.
@@ -280,7 +279,7 @@ void handle_eeh_events (struct eeh_event *event) 
	if (!frozen_bus) {

		printk(KERN_ERR "EEH: Cannot find PCI bus for %s\n",

		        frozen_dn->full_name);

		return;

		return NULL;

	}



#if 0
@@ -355,7 +354,7 @@ void handle_eeh_events (struct eeh_event *event) 
	/* Tell all device drivers that they can resume operations */

	pci_walk_bus(frozen_bus, eeh_report_resume, NULL);



	return;

	return frozen_pdn;

	

excess_failures:

	/*
@@ -384,6 +383,8 @@ perm_error: 


	/* Shut down the device drivers for good. */

	pcibios_remove_pci_devices(frozen_bus);



	return NULL;

}



/* ---------- end of file ---------- */

arch/powerpc/platforms/pseries/eeh_event.c

+28 −22
Original line number Diff line number Diff line
@@ -18,6 +18,7 @@ 
 * Copyright (c) 2005 Linas Vepstas <linas@linas.org>

 */



#include <linux/delay.h>

#include <linux/list.h>

#include <linux/mutex.h>

#include <linux/pci.h>
@@ -56,10 +57,9 @@ static int eeh_event_handler(void * dummy) 
{

	unsigned long flags;

	struct eeh_event	*event;

	struct pci_dn *pdn;



	daemonize ("eehd");



	while (1) {

	set_current_state(TASK_INTERRUPTIBLE);



	spin_lock_irqsave(&eeh_eventlist_lock, flags);
@@ -73,7 +73,7 @@ static int eeh_event_handler(void * dummy) 
	spin_unlock_irqrestore(&eeh_eventlist_lock, flags);



	if (event == NULL)

			break;

		return 0;



	/* Serialize processing of EEH events */

	mutex_lock(&eeh_event_mutex);
@@ -82,12 +82,18 @@ static int eeh_event_handler(void * dummy) 
	printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",

	       pci_name(event->dev));



		handle_eeh_events(event);

	pdn = handle_eeh_events(event);



	eeh_clear_slot(event->dn, EEH_MODE_RECOVERING);

	pci_dev_put(event->dev);

	kfree(event);

	mutex_unlock(&eeh_event_mutex);



	/* If there are no new errors after an hour, clear the counter. */

	if (pdn && pdn->eeh_freeze_count>0) {

		msleep_interruptible (3600*1000);

		if (pdn->eeh_freeze_count>0)

			pdn->eeh_freeze_count--;

	}



	return 0;

include/asm-powerpc/eeh_event.h

+5 −5
Original line number Diff line number Diff line
@@ -18,8 +18,8 @@ 
 * Copyright (c) 2005 Linas Vepstas <linas@linas.org>

 */



#ifndef ASM_PPC64_EEH_EVENT_H

#define ASM_PPC64_EEH_EVENT_H

#ifndef ASM_POWERPC_EEH_EVENT_H

#define ASM_POWERPC_EEH_EVENT_H

#ifdef __KERNEL__



/** EEH event -- structure holding pci controller data that describes
@@ -39,7 +39,7 @@ struct eeh_event { 
 * @dev pci device

 *

 * This routine builds a PCI error event which will be delivered

 * to all listeners on the peh_notifier_chain.

 * to all listeners on the eeh_notifier_chain.

 *

 * This routine can be called within an interrupt context;

 * the actual event will be delivered in a normal context
@@ -51,7 +51,7 @@ int eeh_send_failure_event (struct device_node *dn, 
                            int time_unavail);



/* Main recovery function */

void handle_eeh_events (struct eeh_event *);

struct pci_dn * handle_eeh_events (struct eeh_event *);



#endif /* __KERNEL__ */

#endif /* ASM_PPC64_EEH_EVENT_H */
 
#endif /* ASM_POWERPC_EEH_EVENT_H */

CRA Git | Maintained and supported by SUSTech CRA and CCSE