[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: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

2 files changed, 35 insertions, 28 deletions

diff --git a/arch/powerpc/platforms/pseries/eeh_driver.c b/arch/powerpc/platforms/pseries/eeh_driver.c
index 1fba695e32e8..2a9eb2630730 100644
--- a/arch/powerpc/platforms/pseries/eeh_driver.c
+++ b/arch/powerpc/platforms/pseries/eeh_driver.c
@@ -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 ---------- */
diff --git a/arch/powerpc/platforms/pseries/eeh_event.c b/arch/powerpc/platforms/pseries/eeh_event.c
index a1bda6f96fd1..a0b39640a00f 100644
--- a/arch/powerpc/platforms/pseries/eeh_event.c
+++ b/arch/powerpc/platforms/pseries/eeh_event.c
@@ -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>
@@ -56,38 +57,43 @@ static int eeh_event_handler(void * dummy)
 {
         unsigned long flags;
         struct eeh_event        *event;
+        struct pci_dn *pdn;
 
         daemonize ("eehd");
+        set_current_state(TASK_INTERRUPTIBLE);
 
-        while (1) {
-                set_current_state(TASK_INTERRUPTIBLE);
+        spin_lock_irqsave(&eeh_eventlist_lock, flags);
+        event = NULL;
 
-                spin_lock_irqsave(&eeh_eventlist_lock, flags);
-                event = NULL;
+        /* Unqueue the event, get ready to process. */
+        if (!list_empty(&eeh_eventlist)) {
+                event = list_entry(eeh_eventlist.next, struct eeh_event, list);
+                list_del(&event->list);
+        }
+        spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
 
-                /* Unqueue the event, get ready to process. */
-                if (!list_empty(&eeh_eventlist)) {
-                        event = list_entry(eeh_eventlist.next, struct eeh_event, list);
-                        list_del(&event->list);
-                }
-                spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
+        if (event == NULL)
+                return 0;
 
-                if (event == NULL)
-                        break;
+        /* Serialize processing of EEH events */
+        mutex_lock(&eeh_event_mutex);
+        eeh_mark_slot(event->dn, EEH_MODE_RECOVERING);
 
-                /* Serialize processing of EEH events */
-                mutex_lock(&eeh_event_mutex);
-                eeh_mark_slot(event->dn, EEH_MODE_RECOVERING);
+        printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
+               pci_name(event->dev));
 
-                printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
-                       pci_name(event->dev));
+        pdn = handle_eeh_events(event);
 
-                handle_eeh_events(event);
+        eeh_clear_slot(event->dn, EEH_MODE_RECOVERING);
+        pci_dev_put(event->dev);
+        kfree(event);
+        mutex_unlock(&eeh_event_mutex);
 
-                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;