2005-11-04 07:52:49 +07:00
|
|
|
/*
|
|
|
|
* PCI Error Recovery Driver for RPA-compliant PPC64 platform.
|
2007-05-24 00:28:01 +07:00
|
|
|
* Copyright IBM Corp. 2004 2005
|
|
|
|
* Copyright Linas Vepstas <linas@linas.org> 2004, 2005
|
2005-11-04 07:52:49 +07:00
|
|
|
*
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or (at
|
|
|
|
* your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful, but
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
|
|
|
|
* NON INFRINGEMENT. See the GNU General Public License for more
|
|
|
|
* details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|
|
|
*
|
2007-05-24 00:28:01 +07:00
|
|
|
* Send comments and feedback to Linas Vepstas <linas@austin.ibm.com>
|
2005-11-04 07:52:49 +07:00
|
|
|
*/
|
|
|
|
#include <linux/delay.h>
|
|
|
|
#include <linux/interrupt.h>
|
2006-04-19 11:05:21 +07:00
|
|
|
#include <linux/irq.h>
|
2012-09-17 11:34:27 +07:00
|
|
|
#include <linux/module.h>
|
2005-11-04 07:52:49 +07:00
|
|
|
#include <linux/pci.h>
|
|
|
|
#include <asm/eeh.h>
|
|
|
|
#include <asm/eeh_event.h>
|
|
|
|
#include <asm/ppc-pci.h>
|
|
|
|
#include <asm/pci-bridge.h>
|
|
|
|
#include <asm/prom.h>
|
|
|
|
#include <asm/rtas.h>
|
|
|
|
|
2012-02-28 03:04:02 +07:00
|
|
|
/**
|
|
|
|
* eeh_pcid_name - Retrieve name of PCI device driver
|
|
|
|
* @pdev: PCI device
|
|
|
|
*
|
|
|
|
* This routine is used to retrieve the name of PCI device driver
|
|
|
|
* if that's valid.
|
|
|
|
*/
|
2012-02-28 03:04:08 +07:00
|
|
|
static inline const char *eeh_pcid_name(struct pci_dev *pdev)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2006-02-27 21:52:59 +07:00
|
|
|
if (pdev && pdev->dev.driver)
|
2005-11-04 07:52:49 +07:00
|
|
|
return pdev->dev.driver->name;
|
|
|
|
return "";
|
|
|
|
}
|
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
/**
|
|
|
|
* eeh_pcid_get - Get the PCI device driver
|
|
|
|
* @pdev: PCI device
|
|
|
|
*
|
|
|
|
* The function is used to retrieve the PCI device driver for
|
|
|
|
* the indicated PCI device. Besides, we will increase the reference
|
|
|
|
* of the PCI device driver to prevent that being unloaded on
|
|
|
|
* the fly. Otherwise, kernel crash would be seen.
|
|
|
|
*/
|
|
|
|
static inline struct pci_driver *eeh_pcid_get(struct pci_dev *pdev)
|
|
|
|
{
|
|
|
|
if (!pdev || !pdev->driver)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
if (!try_module_get(pdev->driver->driver.owner))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
return pdev->driver;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* eeh_pcid_put - Dereference on the PCI device driver
|
|
|
|
* @pdev: PCI device
|
|
|
|
*
|
|
|
|
* The function is called to do dereference on the PCI device
|
|
|
|
* driver of the indicated PCI device.
|
|
|
|
*/
|
|
|
|
static inline void eeh_pcid_put(struct pci_dev *pdev)
|
|
|
|
{
|
|
|
|
if (!pdev || !pdev->driver)
|
|
|
|
return;
|
|
|
|
|
|
|
|
module_put(pdev->driver->driver.owner);
|
|
|
|
}
|
|
|
|
|
2008-07-31 13:54:20 +07:00
|
|
|
#if 0
|
2008-06-10 06:26:10 +07:00
|
|
|
static void print_device_node_tree(struct pci_dn *pdn, int dent)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
|
|
|
int i;
|
2008-06-10 06:26:10 +07:00
|
|
|
struct device_node *pc;
|
|
|
|
|
|
|
|
if (!pdn)
|
|
|
|
return;
|
|
|
|
for (i = 0; i < dent; i++)
|
2005-11-04 07:52:49 +07:00
|
|
|
printk(" ");
|
|
|
|
printk("dn=%s mode=%x \tcfg_addr=%x pe_addr=%x \tfull=%s\n",
|
|
|
|
pdn->node->name, pdn->eeh_mode, pdn->eeh_config_addr,
|
|
|
|
pdn->eeh_pe_config_addr, pdn->node->full_name);
|
|
|
|
dent += 3;
|
2008-06-10 06:26:10 +07:00
|
|
|
pc = pdn->node->child;
|
2005-11-04 07:52:49 +07:00
|
|
|
while (pc) {
|
|
|
|
print_device_node_tree(PCI_DN(pc), dent);
|
|
|
|
pc = pc->sibling;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2009-02-10 18:12:21 +07:00
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_disable_irq - Disable interrupt for the recovering device
|
|
|
|
* @dev: PCI device
|
|
|
|
*
|
|
|
|
* This routine must be called when reporting temporary or permanent
|
|
|
|
* error to the particular PCI device to disable interrupt of that
|
|
|
|
* device. If the device has enabled MSI or MSI-X interrupt, we needn't
|
|
|
|
* do real work because EEH should freeze DMA transfers for those PCI
|
|
|
|
* devices encountering EEH errors, which includes MSI or MSI-X.
|
2009-02-10 18:12:21 +07:00
|
|
|
*/
|
|
|
|
static void eeh_disable_irq(struct pci_dev *dev)
|
|
|
|
{
|
2012-02-28 03:04:08 +07:00
|
|
|
struct eeh_dev *edev = pci_dev_to_eeh_dev(dev);
|
2009-02-10 18:12:21 +07:00
|
|
|
|
|
|
|
/* Don't disable MSI and MSI-X interrupts. They are
|
|
|
|
* effectively disabled by the DMA Stopped state
|
|
|
|
* when an EEH error occurs.
|
2012-02-28 03:04:02 +07:00
|
|
|
*/
|
2009-02-10 18:12:21 +07:00
|
|
|
if (dev->msi_enabled || dev->msix_enabled)
|
|
|
|
return;
|
|
|
|
|
2009-10-14 02:44:47 +07:00
|
|
|
if (!irq_has_action(dev->irq))
|
2009-02-10 18:12:21 +07:00
|
|
|
return;
|
|
|
|
|
2012-09-08 05:44:20 +07:00
|
|
|
edev->mode |= EEH_DEV_IRQ_DISABLED;
|
2009-02-10 18:12:21 +07:00
|
|
|
disable_irq_nosync(dev->irq);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_enable_irq - Enable interrupt for the recovering device
|
|
|
|
* @dev: PCI device
|
|
|
|
*
|
|
|
|
* This routine must be called to enable interrupt while failed
|
|
|
|
* device could be resumed.
|
2009-02-10 18:12:21 +07:00
|
|
|
*/
|
|
|
|
static void eeh_enable_irq(struct pci_dev *dev)
|
|
|
|
{
|
2012-02-28 03:04:08 +07:00
|
|
|
struct eeh_dev *edev = pci_dev_to_eeh_dev(dev);
|
2009-02-10 18:12:21 +07:00
|
|
|
|
2012-09-08 05:44:20 +07:00
|
|
|
if ((edev->mode) & EEH_DEV_IRQ_DISABLED) {
|
|
|
|
edev->mode &= ~EEH_DEV_IRQ_DISABLED;
|
2009-02-10 18:12:21 +07:00
|
|
|
enable_irq(dev->irq);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-09-16 06:56:35 +07:00
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_report_error - Report pci error to each device driver
|
2012-09-08 05:44:19 +07:00
|
|
|
* @data: eeh device
|
2012-02-28 03:04:02 +07:00
|
|
|
* @userdata: return value
|
2006-09-16 06:56:35 +07:00
|
|
|
*
|
|
|
|
* Report an EEH error to each device driver, collect up and
|
|
|
|
* merge the device driver responses. Cumulative response
|
|
|
|
* passed back in "userdata".
|
2005-11-04 07:52:49 +07:00
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static void *eeh_report_error(void *data, void *userdata)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2012-09-08 05:44:19 +07:00
|
|
|
struct eeh_dev *edev = (struct eeh_dev *)data;
|
|
|
|
struct pci_dev *dev = eeh_dev_to_pci_dev(edev);
|
2005-11-29 13:17:02 +07:00
|
|
|
enum pci_ers_result rc, *res = userdata;
|
2012-09-17 11:34:27 +07:00
|
|
|
struct pci_driver *driver;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-09-08 05:44:19 +07:00
|
|
|
/* We might not have the associated PCI device,
|
|
|
|
* then we should continue for next one.
|
|
|
|
*/
|
|
|
|
if (!dev) return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
dev->error_state = pci_channel_io_frozen;
|
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
driver = eeh_pcid_get(dev);
|
|
|
|
if (!driver) return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2009-02-10 18:12:21 +07:00
|
|
|
eeh_disable_irq(dev);
|
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
if (!driver->err_handler ||
|
2012-09-17 11:34:27 +07:00
|
|
|
!driver->err_handler->error_detected) {
|
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2012-09-17 11:34:27 +07:00
|
|
|
}
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-02-28 03:04:02 +07:00
|
|
|
rc = driver->err_handler->error_detected(dev, pci_channel_io_frozen);
|
2007-11-03 03:27:50 +07:00
|
|
|
|
|
|
|
/* A driver that needs a reset trumps all others */
|
|
|
|
if (rc == PCI_ERS_RESULT_NEED_RESET) *res = rc;
|
2005-11-29 13:17:02 +07:00
|
|
|
if (*res == PCI_ERS_RESULT_NONE) *res = rc;
|
2009-06-16 12:34:38 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2006-09-16 06:58:59 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_report_mmio_enabled - Tell drivers that MMIO has been enabled
|
2012-09-08 05:44:19 +07:00
|
|
|
* @data: eeh device
|
2012-02-28 03:04:02 +07:00
|
|
|
* @userdata: return value
|
2006-09-16 06:58:59 +07:00
|
|
|
*
|
2007-11-03 03:25:55 +07:00
|
|
|
* Tells each device driver that IO ports, MMIO and config space I/O
|
|
|
|
* are now enabled. Collects up and merges the device driver responses.
|
|
|
|
* Cumulative response passed back in "userdata".
|
2006-09-16 06:58:59 +07:00
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static void *eeh_report_mmio_enabled(void *data, void *userdata)
|
2006-09-16 06:58:59 +07:00
|
|
|
{
|
2012-09-08 05:44:19 +07:00
|
|
|
struct eeh_dev *edev = (struct eeh_dev *)data;
|
|
|
|
struct pci_dev *dev = eeh_dev_to_pci_dev(edev);
|
2006-09-16 06:58:59 +07:00
|
|
|
enum pci_ers_result rc, *res = userdata;
|
2012-09-08 05:44:19 +07:00
|
|
|
struct pci_driver *driver;
|
2006-09-16 06:58:59 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
driver = eeh_pcid_get(dev);
|
|
|
|
if (!driver) return NULL;
|
2012-09-08 05:44:19 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
if (!driver->err_handler ||
|
|
|
|
!driver->err_handler->mmio_enabled) {
|
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2012-09-17 11:34:27 +07:00
|
|
|
}
|
2006-09-16 06:58:59 +07:00
|
|
|
|
2012-02-28 03:04:02 +07:00
|
|
|
rc = driver->err_handler->mmio_enabled(dev);
|
2007-11-03 03:27:50 +07:00
|
|
|
|
|
|
|
/* A driver that needs a reset trumps all others */
|
|
|
|
if (rc == PCI_ERS_RESULT_NEED_RESET) *res = rc;
|
2006-09-16 06:58:59 +07:00
|
|
|
if (*res == PCI_ERS_RESULT_NONE) *res = rc;
|
2009-06-16 12:34:38 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:56:35 +07:00
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_report_reset - Tell device that slot has been reset
|
2012-09-08 05:44:19 +07:00
|
|
|
* @data: eeh device
|
2012-02-28 03:04:02 +07:00
|
|
|
* @userdata: return value
|
|
|
|
*
|
|
|
|
* This routine must be called while EEH tries to reset particular
|
|
|
|
* PCI device so that the associated PCI device driver could take
|
|
|
|
* some actions, usually to save data the driver needs so that the
|
|
|
|
* driver can work again while the device is recovered.
|
2005-11-04 07:52:49 +07:00
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static void *eeh_report_reset(void *data, void *userdata)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2012-09-08 05:44:19 +07:00
|
|
|
struct eeh_dev *edev = (struct eeh_dev *)data;
|
|
|
|
struct pci_dev *dev = eeh_dev_to_pci_dev(edev);
|
2006-09-16 06:58:59 +07:00
|
|
|
enum pci_ers_result rc, *res = userdata;
|
2012-09-08 05:44:19 +07:00
|
|
|
struct pci_driver *driver;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
if (!dev) return NULL;
|
powerpc/pseries: Set error_state to pci_channel_io_normal in eeh_report_reset()
While adding native EEH support to Emulex and Qlogic drivers, it was
discovered that dev->error_state was set to pci_io_channel_normal too
late in the recovery process. These drivers rely on error_state to
determine if they can access the device in their slot_reset callback,
thus error_state needs to be set to pci_io_channel_normal in
eeh_report_reset(). Below is a detailed explanation (courtesy of Richard
Lary) as to why this is necessary.
Background:
PCI MMIO or DMA accesses to a frozen slot generate additional EEH
errors. If the number of additional EEH errors exceeds EEH_MAX_FAILS the
adapter will be shutdown. To avoid triggering excessive EEH errors and
an undesirable adapter shutdown, some drivers use the
pci_channel_offline(dev) wrapper function to return a Boolean value
based on the value of pci_dev->error_state to determine if PCI MMIO or
DMA accesses are safe. If the wrapper returns TRUE, drivers must not
make PCI MMIO or DMA access to their hardware.
The pci_dev structure member error_state reflects one of three values,
1) pci_channel_io_normal, 2) pci_channel_io_frozen, 3)
pci_channel_io_perm_failure. Function pci_channel_offline(dev) returns
TRUE if error_state is pci_channel_io_frozen or pci_channel_io_perm_failure.
The EEH driver sets pci_dev->error_state to pci_channel_io_frozen at the
point where the PCI slot is frozen. Currently, the EEH driver restores
dev->error_state to pci_channel_io_normal in eeh_report_resume() before
calling the driver's resume callback. However, when the EEH driver calls
the driver's slot_reset callback() from eeh_report_reset(), it
incorrectly indicates the error state is still pci_channel_io_frozen.
Waiting until eeh_report_resume() to restore dev->error_state to
pci_channel_io_normal is too late for Emulex and QLogic FC drivers and
any other drivers which are designed to use common code paths in these
two cases: i) those called after the driver's slot_reset callback() and
ii) those called after the PCI slot is frozen but before the driver's
slot_reset callback is called. Case i) all driver paths executed to
reinitialize the hardware after a reset and case ii) all code paths
executed by driver kernel threads that run asynchronous to the main
driver thread, such as interrupt handlers and worker threads to process
driver work queues.
Emulex and QLogic FC drivers are designed with common code paths which
require that pci_channel_offline(dev) reflect the true state of the
hardware. The state transitions that the hardware takes from Normal
Operations to Slot Frozen to Reset to Normal Operations are documented
in the Power Architecture™ Platform Requirements+ (PAPR+) in Table 75.
PE State Control.
PAPR defines the following 3 states:
0 -- Not reset, Not EEH stopped, MMIO load/store allowed, DMA allowed
(Normal Operations)
1 -- Reset, Not EEH stopped, MMIO load/store disabled, DMA disabled
2 -- Not reset, EEH stopped, MMIO load/store disabled, DMA disabled
(Slot Frozen)
An EEH error places the slot in state 2 (Frozen) and the adapter driver
is notified that an EEH error was detected. If the adapter driver
returns PCI_ERS_RESULT_NEED_RESET, the EEH driver calls
eeh_reset_device() to place the slot into state 1 (Reset) and
eeh_reset_device completes by placing the slot into State 0 (Normal
Operations). Upon return from eeh_reset_device(), the EEH driver calls
eeh_report_reset, which then calls the adapter's slot_reset callback. At
the time the adapter's slot_reset callback is called, the true state of
the hardware is Normal Operations and should be accurately reflected by
setting dev->error_state to pci_channel_io_normal.
The current implementation of EEH driver does not do so and requires
this change to correct this deficiency.
Signed-off-by: Mike Mason <mmlnx@us.ibm.com>
Acked-by: Linas Vepstas <linasvepstas@gmail.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
2009-04-10 15:57:03 +07:00
|
|
|
dev->error_state = pci_channel_io_normal;
|
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
driver = eeh_pcid_get(dev);
|
|
|
|
if (!driver) return NULL;
|
|
|
|
|
2009-02-10 18:12:21 +07:00
|
|
|
eeh_enable_irq(dev);
|
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
if (!driver->err_handler ||
|
2012-09-17 11:34:27 +07:00
|
|
|
!driver->err_handler->slot_reset) {
|
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2012-09-17 11:34:27 +07:00
|
|
|
}
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
rc = driver->err_handler->slot_reset(dev);
|
2007-03-20 02:55:51 +07:00
|
|
|
if ((*res == PCI_ERS_RESULT_NONE) ||
|
|
|
|
(*res == PCI_ERS_RESULT_RECOVERED)) *res = rc;
|
2006-09-16 06:58:59 +07:00
|
|
|
if (*res == PCI_ERS_RESULT_DISCONNECT &&
|
|
|
|
rc == PCI_ERS_RESULT_NEED_RESET) *res = rc;
|
2009-06-16 12:34:38 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:56:35 +07:00
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_report_resume - Tell device to resume normal operations
|
2012-09-08 05:44:19 +07:00
|
|
|
* @data: eeh device
|
2012-02-28 03:04:02 +07:00
|
|
|
* @userdata: return value
|
|
|
|
*
|
|
|
|
* This routine must be called to notify the device driver that it
|
|
|
|
* could resume so that the device driver can do some initialization
|
|
|
|
* to make the recovered device work again.
|
2006-09-16 06:56:35 +07:00
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static void *eeh_report_resume(void *data, void *userdata)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2012-09-08 05:44:19 +07:00
|
|
|
struct eeh_dev *edev = (struct eeh_dev *)data;
|
|
|
|
struct pci_dev *dev = eeh_dev_to_pci_dev(edev);
|
|
|
|
struct pci_driver *driver;
|
|
|
|
|
|
|
|
if (!dev) return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
dev->error_state = pci_channel_io_normal;
|
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
driver = eeh_pcid_get(dev);
|
|
|
|
if (!driver) return NULL;
|
2006-12-07 01:32:20 +07:00
|
|
|
|
2009-02-10 18:12:21 +07:00
|
|
|
eeh_enable_irq(dev);
|
|
|
|
|
2006-12-07 01:32:20 +07:00
|
|
|
if (!driver->err_handler ||
|
2012-09-17 11:34:27 +07:00
|
|
|
!driver->err_handler->resume) {
|
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2012-09-17 11:34:27 +07:00
|
|
|
}
|
2005-11-04 07:52:49 +07:00
|
|
|
|
|
|
|
driver->err_handler->resume(dev);
|
2009-06-16 12:34:38 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:56:35 +07:00
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_report_failure - Tell device driver that device is dead.
|
2012-09-08 05:44:19 +07:00
|
|
|
* @data: eeh device
|
2012-02-28 03:04:02 +07:00
|
|
|
* @userdata: return value
|
2006-09-16 06:56:35 +07:00
|
|
|
*
|
|
|
|
* This informs the device driver that the device is permanently
|
|
|
|
* dead, and that no further recovery attempts will be made on it.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static void *eeh_report_failure(void *data, void *userdata)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2012-09-08 05:44:19 +07:00
|
|
|
struct eeh_dev *edev = (struct eeh_dev *)data;
|
|
|
|
struct pci_dev *dev = eeh_dev_to_pci_dev(edev);
|
|
|
|
struct pci_driver *driver;
|
|
|
|
|
|
|
|
if (!dev) return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
dev->error_state = pci_channel_io_perm_failure;
|
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
driver = eeh_pcid_get(dev);
|
|
|
|
if (!driver) return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2009-02-10 18:12:21 +07:00
|
|
|
eeh_disable_irq(dev);
|
|
|
|
|
|
|
|
if (!driver->err_handler ||
|
2012-09-17 11:34:27 +07:00
|
|
|
!driver->err_handler->error_detected) {
|
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2012-09-17 11:34:27 +07:00
|
|
|
}
|
2009-02-10 18:12:21 +07:00
|
|
|
|
2005-11-04 07:52:49 +07:00
|
|
|
driver->err_handler->error_detected(dev, pci_channel_io_perm_failure);
|
2009-06-16 12:34:38 +07:00
|
|
|
|
2012-09-17 11:34:27 +07:00
|
|
|
eeh_pcid_put(dev);
|
2012-09-08 05:44:19 +07:00
|
|
|
return NULL;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2012-02-28 03:04:02 +07:00
|
|
|
* eeh_reset_device - Perform actual reset of a pci slot
|
2012-09-08 05:44:19 +07:00
|
|
|
* @pe: EEH PE
|
2012-02-28 03:04:02 +07:00
|
|
|
* @bus: PCI bus corresponding to the isolcated slot
|
2005-11-04 07:52:49 +07:00
|
|
|
*
|
2012-02-28 03:04:02 +07:00
|
|
|
* This routine must be called to do reset on the indicated PE.
|
|
|
|
* During the reset, udev might be invoked because those affected
|
|
|
|
* PCI devices will be removed and then added.
|
2005-11-04 07:52:49 +07:00
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
static int eeh_reset_device(struct eeh_pe *pe, struct pci_bus *bus)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
2006-04-29 05:39:38 +07:00
|
|
|
int cnt, rc;
|
|
|
|
|
|
|
|
/* pcibios will clear the counter; save the value */
|
2012-09-08 05:44:19 +07:00
|
|
|
cnt = pe->freeze_count;
|
2006-04-29 05:39:38 +07:00
|
|
|
|
2012-09-12 02:16:17 +07:00
|
|
|
/*
|
|
|
|
* We don't remove the corresponding PE instances because
|
|
|
|
* we need the information afterwords. The attached EEH
|
|
|
|
* devices are expected to be attached soon when calling
|
|
|
|
* into pcibios_add_pci_devices().
|
|
|
|
*/
|
2005-11-04 07:52:49 +07:00
|
|
|
if (bus)
|
2012-09-12 02:16:17 +07:00
|
|
|
__pcibios_remove_pci_devices(bus, 0);
|
2005-11-04 07:52:49 +07:00
|
|
|
|
|
|
|
/* Reset the pci controller. (Asserts RST#; resets config space).
|
2005-11-04 07:54:54 +07:00
|
|
|
* Reconfigure bridges and devices. Don't try to bring the system
|
2012-02-28 03:04:02 +07:00
|
|
|
* up if the reset failed for some reason.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_reset_pe(pe);
|
2005-11-04 07:54:54 +07:00
|
|
|
if (rc)
|
|
|
|
return rc;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-09-08 05:44:19 +07:00
|
|
|
/* Restore PE */
|
|
|
|
eeh_ops->configure_bridge(pe);
|
|
|
|
eeh_pe_restore_bars(pe);
|
2005-11-04 07:52:49 +07:00
|
|
|
|
|
|
|
/* Give the system 5 seconds to finish running the user-space
|
|
|
|
* hotplug shutdown scripts, e.g. ifdown for ethernet. Yes,
|
|
|
|
* this is a hack, but if we don't do this, and try to bring
|
|
|
|
* the device up before the scripts have taken it down,
|
|
|
|
* potentially weird things happen.
|
|
|
|
*/
|
|
|
|
if (bus) {
|
2012-02-28 03:04:02 +07:00
|
|
|
ssleep(5);
|
2005-11-04 07:52:49 +07:00
|
|
|
pcibios_add_pci_devices(bus);
|
|
|
|
}
|
2012-09-08 05:44:19 +07:00
|
|
|
pe->freeze_count = cnt;
|
2005-11-04 07:54:54 +07:00
|
|
|
|
|
|
|
return 0;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* The longest amount of time to wait for a pci device
|
|
|
|
* to come back on line, in seconds.
|
|
|
|
*/
|
2007-03-20 02:59:59 +07:00
|
|
|
#define MAX_WAIT_FOR_RECOVERY 150
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-02-28 03:04:02 +07:00
|
|
|
/**
|
|
|
|
* eeh_handle_event - Reset a PCI device after hard lockup.
|
2012-09-08 05:44:19 +07:00
|
|
|
* @pe: EEH PE
|
2012-02-28 03:04:02 +07:00
|
|
|
*
|
|
|
|
* While PHB detects address or data parity errors on particular PCI
|
|
|
|
* slot, the associated PE will be frozen. Besides, DMA's occurring
|
|
|
|
* to wild addresses (which usually happen due to bugs in device
|
|
|
|
* drivers or in PCI adapter firmware) can cause EEH error. #SERR,
|
|
|
|
* #PERR or other misc PCI-related errors also can trigger EEH errors.
|
|
|
|
*
|
|
|
|
* Recovery process consists of unplugging the device driver (which
|
|
|
|
* generated hotplug events to userspace), then issuing a PCI #RST to
|
|
|
|
* the device, then reconfiguring the PCI config space for all bridges
|
|
|
|
* & devices under this slot, and then finally restarting the device
|
|
|
|
* drivers (which cause a second set of hotplug events to go out to
|
|
|
|
* userspace).
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
void eeh_handle_event(struct eeh_pe *pe)
|
2005-11-04 07:52:49 +07:00
|
|
|
{
|
|
|
|
struct pci_bus *frozen_bus;
|
2005-11-04 07:54:54 +07:00
|
|
|
int rc = 0;
|
2005-11-29 13:17:02 +07:00
|
|
|
enum pci_ers_result result = PCI_ERS_RESULT_NONE;
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2012-09-08 05:44:19 +07:00
|
|
|
frozen_bus = eeh_pe_bus_get(pe);
|
2005-11-04 07:52:49 +07:00
|
|
|
if (!frozen_bus) {
|
2012-09-08 05:44:19 +07:00
|
|
|
pr_err("%s: Cannot find PCI bus for PHB#%d-PE#%x\n",
|
|
|
|
__func__, pe->phb->global_number, pe->addr);
|
|
|
|
return;
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2012-09-08 05:44:19 +07:00
|
|
|
pe->freeze_count++;
|
|
|
|
if (pe->freeze_count > EEH_MAX_ALLOWED_FREEZES)
|
2006-03-30 04:31:04 +07:00
|
|
|
goto excess_failures;
|
2012-09-08 05:44:19 +07:00
|
|
|
pr_warning("EEH: This PCI device has failed %d times in the last hour\n",
|
|
|
|
pe->freeze_count);
|
2005-11-04 07:52:49 +07:00
|
|
|
|
|
|
|
/* 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.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_pe_dev_traverse(pe, eeh_report_error, &result);
|
2005-11-04 07:52:49 +07:00
|
|
|
|
2007-11-16 01:58:36 +07:00
|
|
|
/* Get the current PCI slot state. This can take a long time,
|
2012-02-28 03:04:02 +07:00
|
|
|
* sometimes over 3 seconds for certain systems.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_ops->wait_state(pe, MAX_WAIT_FOR_RECOVERY*1000);
|
2012-02-28 03:03:57 +07:00
|
|
|
if (rc < 0 || rc == EEH_STATE_NOT_SUPPORT) {
|
2007-11-16 01:58:36 +07:00
|
|
|
printk(KERN_WARNING "EEH: Permanent failure\n");
|
|
|
|
goto hard_fail;
|
|
|
|
}
|
|
|
|
|
2007-05-09 06:33:29 +07:00
|
|
|
/* Since rtas may enable MMIO when posting the error log,
|
|
|
|
* don't post the error log until after all dev drivers
|
2007-05-09 23:38:11 +07:00
|
|
|
* have been informed.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_slot_error_detail(pe, EEH_LOG_TEMP);
|
2007-05-09 06:33:29 +07:00
|
|
|
|
2005-11-04 07:52:49 +07:00
|
|
|
/* 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.
|
|
|
|
*/
|
2005-11-29 13:17:02 +07:00
|
|
|
if (result == PCI_ERS_RESULT_NONE) {
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_reset_device(pe, frozen_bus);
|
2007-03-20 02:52:04 +07:00
|
|
|
if (rc) {
|
|
|
|
printk(KERN_WARNING "EEH: Unable to reset, rc=%d\n", rc);
|
2005-11-04 07:54:54 +07:00
|
|
|
goto hard_fail;
|
2007-03-20 02:52:04 +07:00
|
|
|
}
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
/* If all devices reported they can proceed, then re-enable MMIO */
|
|
|
|
if (result == PCI_ERS_RESULT_CAN_RECOVER) {
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_pci_enable(pe, EEH_OPT_THAW_MMIO);
|
2006-09-16 06:58:59 +07:00
|
|
|
|
2007-03-20 02:59:59 +07:00
|
|
|
if (rc < 0)
|
|
|
|
goto hard_fail;
|
2006-09-16 06:58:59 +07:00
|
|
|
if (rc) {
|
|
|
|
result = PCI_ERS_RESULT_NEED_RESET;
|
|
|
|
} else {
|
|
|
|
result = PCI_ERS_RESULT_NONE;
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_pe_dev_traverse(pe, eeh_report_mmio_enabled, &result);
|
2006-09-16 06:58:59 +07:00
|
|
|
}
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
/* If all devices reported they can proceed, then re-enable DMA */
|
2005-11-29 13:17:02 +07:00
|
|
|
if (result == PCI_ERS_RESULT_CAN_RECOVER) {
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_pci_enable(pe, EEH_OPT_THAW_DMA);
|
2006-09-16 06:58:59 +07:00
|
|
|
|
2007-03-20 02:59:59 +07:00
|
|
|
if (rc < 0)
|
|
|
|
goto hard_fail;
|
2006-09-16 06:58:59 +07:00
|
|
|
if (rc)
|
|
|
|
result = PCI_ERS_RESULT_NEED_RESET;
|
2006-12-07 01:32:20 +07:00
|
|
|
else
|
|
|
|
result = PCI_ERS_RESULT_RECOVERED;
|
2006-09-16 06:58:59 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* If any device has a hard failure, then shut off everything. */
|
2007-03-20 02:52:04 +07:00
|
|
|
if (result == PCI_ERS_RESULT_DISCONNECT) {
|
|
|
|
printk(KERN_WARNING "EEH: Device driver gave up\n");
|
2006-09-16 06:58:59 +07:00
|
|
|
goto hard_fail;
|
2007-03-20 02:52:04 +07:00
|
|
|
}
|
2006-09-16 06:58:59 +07:00
|
|
|
|
|
|
|
/* If any device called out for a reset, then reset the slot */
|
|
|
|
if (result == PCI_ERS_RESULT_NEED_RESET) {
|
2012-09-08 05:44:19 +07:00
|
|
|
rc = eeh_reset_device(pe, NULL);
|
2007-03-20 02:52:04 +07:00
|
|
|
if (rc) {
|
|
|
|
printk(KERN_WARNING "EEH: Cannot reset, rc=%d\n", rc);
|
2005-11-04 07:54:54 +07:00
|
|
|
goto hard_fail;
|
2007-03-20 02:52:04 +07:00
|
|
|
}
|
2006-09-16 06:58:59 +07:00
|
|
|
result = PCI_ERS_RESULT_NONE;
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_pe_dev_traverse(pe, eeh_report_reset, &result);
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|
2006-09-16 06:58:59 +07:00
|
|
|
/* All devices should claim they have recovered by now. */
|
2007-03-20 02:55:10 +07:00
|
|
|
if ((result != PCI_ERS_RESULT_RECOVERED) &&
|
|
|
|
(result != PCI_ERS_RESULT_NONE)) {
|
2007-03-20 02:52:04 +07:00
|
|
|
printk(KERN_WARNING "EEH: Not recovered\n");
|
2006-09-16 06:58:59 +07:00
|
|
|
goto hard_fail;
|
2007-03-20 02:52:04 +07:00
|
|
|
}
|
2006-09-16 06:58:59 +07:00
|
|
|
|
2005-11-04 07:52:49 +07:00
|
|
|
/* Tell all device drivers that they can resume operations */
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_pe_dev_traverse(pe, eeh_report_resume, NULL);
|
2005-11-04 07:54:54 +07:00
|
|
|
|
2012-09-08 05:44:19 +07:00
|
|
|
return;
|
2005-11-04 07:54:54 +07:00
|
|
|
|
2006-03-30 04:31:04 +07:00
|
|
|
excess_failures:
|
2005-11-04 07:54:54 +07:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
2012-09-08 05:44:19 +07:00
|
|
|
pr_err("EEH: PHB#%d-PE#%x has failed %d times in the\n"
|
|
|
|
"last hour and has been permanently disabled.\n"
|
|
|
|
"Please try reseating or replacing it.\n",
|
|
|
|
pe->phb->global_number, pe->addr,
|
|
|
|
pe->freeze_count);
|
2006-03-30 04:31:04 +07:00
|
|
|
goto perm_error;
|
|
|
|
|
|
|
|
hard_fail:
|
2012-09-08 05:44:19 +07:00
|
|
|
pr_err("EEH: Unable to recover from failure from PHB#%d-PE#%x.\n"
|
|
|
|
"Please try reseating or replacing it\n",
|
|
|
|
pe->phb->global_number, pe->addr);
|
2005-11-04 07:54:54 +07:00
|
|
|
|
2006-03-30 04:31:04 +07:00
|
|
|
perm_error:
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_slot_error_detail(pe, EEH_LOG_PERM);
|
2005-11-04 07:54:54 +07:00
|
|
|
|
|
|
|
/* Notify all devices that they're about to go down. */
|
2012-09-08 05:44:19 +07:00
|
|
|
eeh_pe_dev_traverse(pe, eeh_report_failure, NULL);
|
2005-11-04 07:54:54 +07:00
|
|
|
|
|
|
|
/* Shut down the device drivers for good. */
|
2012-09-08 05:44:19 +07:00
|
|
|
if (frozen_bus)
|
|
|
|
pcibios_remove_pci_devices(frozen_bus);
|
2005-11-04 07:52:49 +07:00
|
|
|
}
|
|
|
|
|