struct pci_dn * handle_eeh_events (struct eeh_event *event)
{
struct device_node *frozen_dn;
struct pci_dn *frozen_pdn;
struct pci_bus *frozen_bus;
int rc = 0;
enum pci_ers_result result = PCI_ERS_RESULT_NONE;
const char *location, *pci_str, *drv_str;
frozen_dn = find_device_pe(event->dn);
if (!frozen_dn) {
location = of_get_property(event->dn, "ibm,loc-code", NULL);
location = location ? location : "unknown";
printk(KERN_ERR "EEH: Error: Cannot find partition endpoint "
"for location=%s pci addr=%s\n",
location, pci_name(event->dev));
return NULL;
}
frozen_bus = pcibios_find_pci_bus(frozen_dn);
location = of_get_property(frozen_dn, "ibm,loc-code", NULL);
location = location ? location : "unknown";
/* There are two different styles for coming up with the PE.
* In the old style, it was the highest EEH-capable device
* which was always an EADS pci bridge. In the new style,
* there might not be any EADS bridges, and even when there are,
* the firmware marks them as "EEH incapable". So another
* two-step is needed to find the pci bus.. */
if (!frozen_bus)
frozen_bus = pcibios_find_pci_bus (frozen_dn->parent);
if (!frozen_bus) {
printk(KERN_ERR "EEH: Cannot find PCI bus "
"for location=%s dn=%s\n",
location, frozen_dn->full_name);
return NULL;
}
frozen_pdn = PCI_DN(frozen_dn);
frozen_pdn->eeh_freeze_count++;
if (frozen_pdn->pcidev) {
pci_str = pci_name (frozen_pdn->pcidev);
drv_str = pcid_name (frozen_pdn->pcidev);
} else {
pci_str = pci_name (event->dev);
drv_str = pcid_name (event->dev);
}
if (frozen_pdn->eeh_freeze_count > EEH_MAX_ALLOWED_FREEZES)
goto excess_failures;
printk(KERN_WARNING
"EEH: This PCI device has failed %d times in the last hour:\n",
frozen_pdn->eeh_freeze_count);
printk(KERN_WARNING
"EEH: location=%s driver=%s pci addr=%s\n",
location, drv_str, pci_str);
/* Walk the various device drivers attached to this slot through
* a reset sequence, giving each an opportunity to do what it needs
* to accomplish the reset. Each child gets a report of the
* status ... if any child can't handle the reset, then the entire
* slot is dlpar removed and added.
*/
pci_walk_bus(frozen_bus, eeh_report_error, &result);
/* Get the current PCI slot state. This can take a long time,
* sometimes over 3 seconds for certain systems. */
rc = eeh_wait_for_slot_status (frozen_pdn, MAX_WAIT_FOR_RECOVERY*1000);
if (rc < 0) {
printk(KERN_WARNING "EEH: Permanent failure\n");
goto hard_fail;
}
/* Since rtas may enable MMIO when posting the error log,
* don't post the error log until after all dev drivers
* have been informed.
*/
eeh_slot_error_detail(frozen_pdn, EEH_LOG_TEMP_FAILURE);
/* If all device drivers were EEH-unaware, then shut
* down all of the device drivers, and hope they
* go down willingly, without panicing the system.
*/
if (result == PCI_ERS_RESULT_NONE) {
rc = eeh_reset_device(frozen_pdn, frozen_bus);
if (rc) {
printk(KERN_WARNING "EEH: Unable to reset, rc=%d\n", rc);
goto hard_fail;
}
}
/* If all devices reported they can proceed, then re-enable MMIO */
if (result == PCI_ERS_RESULT_CAN_RECOVER) {
rc = rtas_pci_enable(frozen_pdn, EEH_THAW_MMIO);
if (rc < 0)
goto hard_fail;
if (rc) {
result = PCI_ERS_RESULT_NEED_RESET;
} else {
result = PCI_ERS_RESULT_NONE;
pci_walk_bus(frozen_bus, eeh_report_mmio_enabled, &result);
}
}
/* If all devices reported they can proceed, then re-enable DMA */
if (result == PCI_ERS_RESULT_CAN_RECOVER) {
rc = rtas_pci_enable(frozen_pdn, EEH_THAW_DMA);
if (rc < 0)
goto hard_fail;
if (rc)
result = PCI_ERS_RESULT_NEED_RESET;
else
result = PCI_ERS_RESULT_RECOVERED;
}
/* If any device has a hard failure, then shut off everything. */
if (result == PCI_ERS_RESULT_DISCONNECT) {
printk(KERN_WARNING "EEH: Device driver gave up\n");
goto hard_fail;
}
/* If any device called out for a reset, then reset the slot */
if (result == PCI_ERS_RESULT_NEED_RESET) {
rc = eeh_reset_device(frozen_pdn, NULL);
if (rc) {
printk(KERN_WARNING "EEH: Cannot reset, rc=%d\n", rc);
goto hard_fail;
}
result = PCI_ERS_RESULT_NONE;
pci_walk_bus(frozen_bus, eeh_report_reset, &result);
}
/* All devices should claim they have recovered by now. */
if ((result != PCI_ERS_RESULT_RECOVERED) &&
(result != PCI_ERS_RESULT_NONE)) {
printk(KERN_WARNING "EEH: Not recovered\n");
goto hard_fail;
}
/* Tell all device drivers that they can resume operations */
pci_walk_bus(frozen_bus, eeh_report_resume, NULL);
return frozen_pdn;
excess_failures:
/*
* About 90% of all real-life EEH failures in the field
* are due to poorly seated PCI cards. Only 10% or so are
* due to actual, failed cards.
*/
printk(KERN_ERR
"EEH: PCI device at location=%s driver=%s pci addr=%s \n"
"has failed %d times in the last hour "
"and has been permanently disabled. \n"
"Please try reseating this device or replacing it.\n",
location, drv_str, pci_str, frozen_pdn->eeh_freeze_count);
goto perm_error;
hard_fail:
printk(KERN_ERR
"EEH: Unable to recover from failure of PCI device "
"at location=%s driver=%s pci addr=%s \n"
"Please try reseating this device or replacing it.\n",
location, drv_str, pci_str);
perm_error:
eeh_slot_error_detail(frozen_pdn, EEH_LOG_PERM_FAILURE);
/* Notify all devices that they're about to go down. */
pci_walk_bus(frozen_bus, eeh_report_failure, NULL);
/* Shut down the device drivers for good. */
pcibios_remove_pci_devices(frozen_bus);
return NULL;
}
struct pci_dn * handle_eeh_events (struct eeh_event *event)
{
struct device_node *frozen_dn;
struct pci_dn *frozen_pdn;
struct pci_bus *frozen_bus;
int rc = 0;
enum pci_ers_result result = PCI_ERS_RESULT_NONE;
const char *location, *pci_str, *drv_str;
frozen_dn = find_device_pe(event->dn);
frozen_bus = pcibios_find_pci_bus(frozen_dn);
if (!frozen_dn) {
location = (char *) get_property(event->dn, "ibm,loc-code", NULL);
location = location ? location : "unknown";
printk(KERN_ERR "EEH: Error: Cannot find partition endpoint "
"for location=%s pci addr=%s\n",
location, pci_name(event->dev));
return NULL;
}
location = (char *) get_property(frozen_dn, "ibm,loc-code", NULL);
location = location ? location : "unknown";
/* There are two different styles for coming up with the PE.
* In the old style, it was the highest EEH-capable device
* which was always an EADS pci bridge. In the new style,
* there might not be any EADS bridges, and even when there are,
* the firmware marks them as "EEH incapable". So another
* two-step is needed to find the pci bus.. */
if (!frozen_bus)
frozen_bus = pcibios_find_pci_bus (frozen_dn->parent);
if (!frozen_bus) {
printk(KERN_ERR "EEH: Cannot find PCI bus "
"for location=%s dn=%s\n",
location, frozen_dn->full_name);
return NULL;
}
#if 0
/* We may get "permanent failure" messages on empty slots.
* These are false alarms. Empty slots have no child dn. */
if ((event->state == pci_channel_io_perm_failure) && (frozen_device == NULL))
return;
#endif
frozen_pdn = PCI_DN(frozen_dn);
frozen_pdn->eeh_freeze_count++;
if (frozen_pdn->pcidev) {
pci_str = pci_name (frozen_pdn->pcidev);
drv_str = pcid_name (frozen_pdn->pcidev);
} else {
pci_str = pci_name (event->dev);
drv_str = pcid_name (event->dev);
}
if (frozen_pdn->eeh_freeze_count > EEH_MAX_ALLOWED_FREEZES)
goto excess_failures;
/* If the reset state is a '5' and the time to reset is 0 (infinity)
* or is more then 15 seconds, then mark this as a permanent failure.
*/
if ((event->state == pci_channel_io_perm_failure) &&
((event->time_unavail <= 0) ||
(event->time_unavail > MAX_WAIT_FOR_RECOVERY*1000)))
goto hard_fail;
eeh_slot_error_detail(frozen_pdn, 1 /* Temporary Error */);
printk(KERN_WARNING
"EEH: This PCI device has failed %d times since last reboot: "
"location=%s driver=%s pci addr=%s\n",
frozen_pdn->eeh_freeze_count, location, drv_str, pci_str);
/* Walk the various device drivers attached to this slot through
* a reset sequence, giving each an opportunity to do what it needs
* to accomplish the reset. Each child gets a report of the
* status ... if any child can't handle the reset, then the entire
* slot is dlpar removed and added.
*/
pci_walk_bus(frozen_bus, eeh_report_error, &result);
/* If all device drivers were EEH-unaware, then shut
* down all of the device drivers, and hope they
* go down willingly, without panicing the system.
*/
if (result == PCI_ERS_RESULT_NONE) {
rc = eeh_reset_device(frozen_pdn, frozen_bus);
if (rc)
goto hard_fail;
}
/* If any device called out for a reset, then reset the slot */
if (result == PCI_ERS_RESULT_NEED_RESET) {
rc = eeh_reset_device(frozen_pdn, NULL);
if (rc)
goto hard_fail;
pci_walk_bus(frozen_bus, eeh_report_reset, NULL);
}
/* If all devices reported they can proceed, the re-enable PIO */
if (result == PCI_ERS_RESULT_CAN_RECOVER) {
/* XXX Not supported; we brute-force reset the device */
rc = eeh_reset_device(frozen_pdn, NULL);
if (rc)
goto hard_fail;
pci_walk_bus(frozen_bus, eeh_report_reset, NULL);
}
/* Tell all device drivers that they can resume operations */
pci_walk_bus(frozen_bus, eeh_report_resume, NULL);
return frozen_pdn;
excess_failures:
/*
* About 90% of all real-life EEH failures in the field
* are due to poorly seated PCI cards. Only 10% or so are
* due to actual, failed cards.
*/
printk(KERN_ERR
"EEH: PCI device at location=%s driver=%s pci addr=%s \n"
"has failed %d times and has been permanently disabled. \n"
"Please try reseating this device or replacing it.\n",
location, drv_str, pci_str, frozen_pdn->eeh_freeze_count);
goto perm_error;
hard_fail:
printk(KERN_ERR
"EEH: Unable to recover from failure of PCI device "
"at location=%s driver=%s pci addr=%s \n"
"Please try reseating this device or replacing it.\n",
location, drv_str, pci_str);
perm_error:
eeh_slot_error_detail(frozen_pdn, 2 /* Permanent Error */);
/* Notify all devices that they're about to go down. */
pci_walk_bus(frozen_bus, eeh_report_failure, NULL);
/* Shut down the device drivers for good. */
pcibios_remove_pci_devices(frozen_bus);
return NULL;
}