linux/drivers/pci/hotplug/pciehp_hpc.c

1225 lines
30 KiB
C
Raw Normal View History

/*
* PCI Express PCI Hot Plug Driver
*
* Copyright (C) 1995,2001 Compaq Computer Corporation
* Copyright (C) 2001 Greg Kroah-Hartman (greg@kroah.com)
* Copyright (C) 2001 IBM Corp.
* Copyright (C) 2003-2004 Intel Corporation
*
* 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.
*
* Send feedback to <greg@kroah.com>,<kristen.c.accardi@intel.com>
*
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/signal.h>
#include <linux/jiffies.h>
#include <linux/timer.h>
#include <linux/pci.h>
#include <linux/interrupt.h>
#include <linux/time.h>
#include "../pci.h"
#include "pciehp.h"
static atomic_t pciehp_num_controllers = ATOMIC_INIT(0);
struct ctrl_reg {
u8 cap_id;
u8 nxt_ptr;
u16 cap_reg;
u32 dev_cap;
u16 dev_ctrl;
u16 dev_status;
u32 lnk_cap;
u16 lnk_ctrl;
u16 lnk_status;
u32 slot_cap;
u16 slot_ctrl;
u16 slot_status;
u16 root_ctrl;
u16 rsvp;
u32 root_status;
} __attribute__ ((packed));
/* offsets to the controller registers based on the above structure layout */
enum ctrl_offsets {
PCIECAPID = offsetof(struct ctrl_reg, cap_id),
NXTCAPPTR = offsetof(struct ctrl_reg, nxt_ptr),
CAPREG = offsetof(struct ctrl_reg, cap_reg),
DEVCAP = offsetof(struct ctrl_reg, dev_cap),
DEVCTRL = offsetof(struct ctrl_reg, dev_ctrl),
DEVSTATUS = offsetof(struct ctrl_reg, dev_status),
LNKCAP = offsetof(struct ctrl_reg, lnk_cap),
LNKCTRL = offsetof(struct ctrl_reg, lnk_ctrl),
LNKSTATUS = offsetof(struct ctrl_reg, lnk_status),
SLOTCAP = offsetof(struct ctrl_reg, slot_cap),
SLOTCTRL = offsetof(struct ctrl_reg, slot_ctrl),
SLOTSTATUS = offsetof(struct ctrl_reg, slot_status),
ROOTCTRL = offsetof(struct ctrl_reg, root_ctrl),
ROOTSTATUS = offsetof(struct ctrl_reg, root_status),
};
static inline int pciehp_readw(struct controller *ctrl, int reg, u16 *value)
{
struct pci_dev *dev = ctrl->pci_dev;
return pci_read_config_word(dev, ctrl->cap_base + reg, value);
}
static inline int pciehp_readl(struct controller *ctrl, int reg, u32 *value)
{
struct pci_dev *dev = ctrl->pci_dev;
return pci_read_config_dword(dev, ctrl->cap_base + reg, value);
}
static inline int pciehp_writew(struct controller *ctrl, int reg, u16 value)
{
struct pci_dev *dev = ctrl->pci_dev;
return pci_write_config_word(dev, ctrl->cap_base + reg, value);
}
static inline int pciehp_writel(struct controller *ctrl, int reg, u32 value)
{
struct pci_dev *dev = ctrl->pci_dev;
return pci_write_config_dword(dev, ctrl->cap_base + reg, value);
}
/* Field definitions in PCI Express Capabilities Register */
#define CAP_VER 0x000F
#define DEV_PORT_TYPE 0x00F0
#define SLOT_IMPL 0x0100
#define MSG_NUM 0x3E00
/* Device or Port Type */
#define NAT_ENDPT 0x00
#define LEG_ENDPT 0x01
#define ROOT_PORT 0x04
#define UP_STREAM 0x05
#define DN_STREAM 0x06
#define PCIE_PCI_BRDG 0x07
#define PCI_PCIE_BRDG 0x10
/* Field definitions in Device Capabilities Register */
#define DATTN_BUTTN_PRSN 0x1000
#define DATTN_LED_PRSN 0x2000
#define DPWR_LED_PRSN 0x4000
/* Field definitions in Link Capabilities Register */
#define MAX_LNK_SPEED 0x000F
#define MAX_LNK_WIDTH 0x03F0
/* Link Width Encoding */
#define LNK_X1 0x01
#define LNK_X2 0x02
#define LNK_X4 0x04
#define LNK_X8 0x08
#define LNK_X12 0x0C
#define LNK_X16 0x10
#define LNK_X32 0x20
/*Field definitions of Link Status Register */
#define LNK_SPEED 0x000F
#define NEG_LINK_WD 0x03F0
#define LNK_TRN_ERR 0x0400
#define LNK_TRN 0x0800
#define SLOT_CLK_CONF 0x1000
/* Field definitions in Slot Capabilities Register */
#define ATTN_BUTTN_PRSN 0x00000001
#define PWR_CTRL_PRSN 0x00000002
#define MRL_SENS_PRSN 0x00000004
#define ATTN_LED_PRSN 0x00000008
#define PWR_LED_PRSN 0x00000010
#define HP_SUPR_RM_SUP 0x00000020
#define HP_CAP 0x00000040
#define SLOT_PWR_VALUE 0x000003F8
#define SLOT_PWR_LIMIT 0x00000C00
#define PSN 0xFFF80000 /* PSN: Physical Slot Number */
/* Field definitions in Slot Control Register */
#define ATTN_BUTTN_ENABLE 0x0001
#define PWR_FAULT_DETECT_ENABLE 0x0002
#define MRL_DETECT_ENABLE 0x0004
#define PRSN_DETECT_ENABLE 0x0008
#define CMD_CMPL_INTR_ENABLE 0x0010
#define HP_INTR_ENABLE 0x0020
#define ATTN_LED_CTRL 0x00C0
#define PWR_LED_CTRL 0x0300
#define PWR_CTRL 0x0400
#define EMI_CTRL 0x0800
/* Attention indicator and Power indicator states */
#define LED_ON 0x01
#define LED_BLINK 0x10
#define LED_OFF 0x11
/* Power Control Command */
#define POWER_ON 0
#define POWER_OFF 0x0400
/* EMI Status defines */
#define EMI_DISENGAGED 0
#define EMI_ENGAGED 1
/* Field definitions in Slot Status Register */
#define ATTN_BUTTN_PRESSED 0x0001
#define PWR_FAULT_DETECTED 0x0002
#define MRL_SENS_CHANGED 0x0004
#define PRSN_DETECT_CHANGED 0x0008
#define CMD_COMPLETED 0x0010
#define MRL_STATE 0x0020
#define PRSN_STATE 0x0040
#define EMI_STATE 0x0080
#define EMI_STATUS_BIT 7
static irqreturn_t pcie_isr(int irq, void *dev_id);
static void start_int_poll_timer(struct controller *ctrl, int sec);
/* This is the interrupt polling timeout function. */
static void int_poll_timeout(unsigned long data)
{
struct controller *ctrl = (struct controller *)data;
/* Poll for interrupt events. regs == NULL => polling */
pcie_isr(0, ctrl);
init_timer(&ctrl->poll_timer);
if (!pciehp_poll_time)
pciehp_poll_time = 2; /* default polling interval is 2 sec */
start_int_poll_timer(ctrl, pciehp_poll_time);
}
/* This function starts the interrupt polling timer. */
static void start_int_poll_timer(struct controller *ctrl, int sec)
{
/* Clamp to sane value */
if ((sec <= 0) || (sec > 60))
sec = 2;
ctrl->poll_timer.function = &int_poll_timeout;
ctrl->poll_timer.data = (unsigned long)ctrl;
ctrl->poll_timer.expires = jiffies + sec * HZ;
add_timer(&ctrl->poll_timer);
}
static inline int pciehp_request_irq(struct controller *ctrl)
{
int retval, irq = ctrl->pcie->irq;
/* Install interrupt polling timer. Start with 10 sec delay */
if (pciehp_poll_mode) {
init_timer(&ctrl->poll_timer);
start_int_poll_timer(ctrl, 10);
return 0;
}
/* Installs the interrupt handler */
retval = request_irq(irq, pcie_isr, IRQF_SHARED, MY_NAME, ctrl);
if (retval)
ctrl_err(ctrl, "Cannot get irq %d for the hotplug controller\n",
irq);
return retval;
}
static inline void pciehp_free_irq(struct controller *ctrl)
{
if (pciehp_poll_mode)
del_timer_sync(&ctrl->poll_timer);
else
free_irq(ctrl->pcie->irq, ctrl);
}
static int pcie_poll_cmd(struct controller *ctrl)
{
u16 slot_status;
int timeout = 1000;
if (!pciehp_readw(ctrl, SLOTSTATUS, &slot_status)) {
if (slot_status & CMD_COMPLETED) {
pciehp_writew(ctrl, SLOTSTATUS, CMD_COMPLETED);
return 1;
}
}
while (timeout > 0) {
msleep(10);
timeout -= 10;
if (!pciehp_readw(ctrl, SLOTSTATUS, &slot_status)) {
if (slot_status & CMD_COMPLETED) {
pciehp_writew(ctrl, SLOTSTATUS, CMD_COMPLETED);
return 1;
}
}
}
return 0; /* timeout */
}
static void pcie_wait_cmd(struct controller *ctrl, int poll)
{
unsigned int msecs = pciehp_poll_mode ? 2500 : 1000;
unsigned long timeout = msecs_to_jiffies(msecs);
int rc;
if (poll)
rc = pcie_poll_cmd(ctrl);
else
rc = wait_event_timeout(ctrl->queue, !ctrl->cmd_busy, timeout);
if (!rc)
ctrl_dbg(ctrl, "Command not completed in 1000 msec\n");
}
/**
* pcie_write_cmd - Issue controller command
* @ctrl: controller to which the command is issued
* @cmd: command value written to slot control register
* @mask: bitmask of slot control register to be modified
*/
static int pcie_write_cmd(struct controller *ctrl, u16 cmd, u16 mask)
{
int retval = 0;
u16 slot_status;
u16 slot_ctrl;
mutex_lock(&ctrl->ctrl_lock);
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTSTATUS register\n",
__func__);
goto out;
}
pciehp: fix slow probing Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in http://bugzilla.kernel.org/show_bug.cgi?id=10751. The command completed bit in Slot Status register applies only to commands issued to control the attention indicator, power indicator, power controller, or electromechanical interlock. However, writes to other parts of the Slot Control register would end up writing to the control fields. Hence, any write to Slot Control register is considered as a command. However, if the controller doesn't support any of attention indicator, power indicator, power controller and electromechanical interlock, command completed bit would not set in writing to Slot Control register. In this case, we should not wait for command completed bit set, otherwise all commands would be considered not completed in timeout seconds (1 sec.). The cause of the problem is pciehp driver didn't take this situation into account. This patch changes pciehp to take it into account. This patch also add the check for "No Command Completed Support" bit in Slot Capability register. If it is set, we should not wait for command completed bit set as well. This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that pciehp did not wait for command completed properly (pciehp just ignored the command completion event). Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:04:30 +02:00
if (slot_status & CMD_COMPLETED) {
if (!ctrl->no_cmd_complete) {
/*
* After 1 sec and CMD_COMPLETED still not set, just
* proceed forward to issue the next command according
* to spec. Just print out the error message.
*/
ctrl_dbg(ctrl,
"%s: CMD_COMPLETED not clear after 1 sec.\n",
__func__);
pciehp: fix slow probing Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in http://bugzilla.kernel.org/show_bug.cgi?id=10751. The command completed bit in Slot Status register applies only to commands issued to control the attention indicator, power indicator, power controller, or electromechanical interlock. However, writes to other parts of the Slot Control register would end up writing to the control fields. Hence, any write to Slot Control register is considered as a command. However, if the controller doesn't support any of attention indicator, power indicator, power controller and electromechanical interlock, command completed bit would not set in writing to Slot Control register. In this case, we should not wait for command completed bit set, otherwise all commands would be considered not completed in timeout seconds (1 sec.). The cause of the problem is pciehp driver didn't take this situation into account. This patch changes pciehp to take it into account. This patch also add the check for "No Command Completed Support" bit in Slot Capability register. If it is set, we should not wait for command completed bit set as well. This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that pciehp did not wait for command completed properly (pciehp just ignored the command completion event). Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:04:30 +02:00
} else if (!NO_CMD_CMPL(ctrl)) {
/*
* This controller semms to notify of command completed
* event even though it supports none of power
* controller, attention led, power led and EMI.
*/
ctrl_dbg(ctrl, "%s: Unexpected CMD_COMPLETED. Need to "
"wait for command completed event.\n",
__func__);
pciehp: fix slow probing Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in http://bugzilla.kernel.org/show_bug.cgi?id=10751. The command completed bit in Slot Status register applies only to commands issued to control the attention indicator, power indicator, power controller, or electromechanical interlock. However, writes to other parts of the Slot Control register would end up writing to the control fields. Hence, any write to Slot Control register is considered as a command. However, if the controller doesn't support any of attention indicator, power indicator, power controller and electromechanical interlock, command completed bit would not set in writing to Slot Control register. In this case, we should not wait for command completed bit set, otherwise all commands would be considered not completed in timeout seconds (1 sec.). The cause of the problem is pciehp driver didn't take this situation into account. This patch changes pciehp to take it into account. This patch also add the check for "No Command Completed Support" bit in Slot Capability register. If it is set, we should not wait for command completed bit set as well. This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that pciehp did not wait for command completed properly (pciehp just ignored the command completion event). Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:04:30 +02:00
ctrl->no_cmd_complete = 0;
} else {
ctrl_dbg(ctrl, "%s: Unexpected CMD_COMPLETED. Maybe "
"the controller is broken.\n", __func__);
pciehp: fix slow probing Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in http://bugzilla.kernel.org/show_bug.cgi?id=10751. The command completed bit in Slot Status register applies only to commands issued to control the attention indicator, power indicator, power controller, or electromechanical interlock. However, writes to other parts of the Slot Control register would end up writing to the control fields. Hence, any write to Slot Control register is considered as a command. However, if the controller doesn't support any of attention indicator, power indicator, power controller and electromechanical interlock, command completed bit would not set in writing to Slot Control register. In this case, we should not wait for command completed bit set, otherwise all commands would be considered not completed in timeout seconds (1 sec.). The cause of the problem is pciehp driver didn't take this situation into account. This patch changes pciehp to take it into account. This patch also add the check for "No Command Completed Support" bit in Slot Capability register. If it is set, we should not wait for command completed bit set as well. This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that pciehp did not wait for command completed properly (pciehp just ignored the command completion event). Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:04:30 +02:00
}
}
retval = pciehp_readw(ctrl, SLOTCTRL, &slot_ctrl);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTCTRL register\n", __func__);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
goto out;
}
slot_ctrl &= ~mask;
slot_ctrl |= (cmd & mask);
ctrl->cmd_busy = 1;
smp_mb();
retval = pciehp_writew(ctrl, SLOTCTRL, slot_ctrl);
if (retval)
ctrl_err(ctrl, "%s: Cannot write to SLOTCTRL register\n",
__func__);
/*
* Wait for command completion.
*/
if (!retval && !ctrl->no_cmd_complete) {
int poll = 0;
/*
* if hotplug interrupt is not enabled or command
* completed interrupt is not enabled, we need to poll
* command completed event.
*/
if (!(slot_ctrl & HP_INTR_ENABLE) ||
!(slot_ctrl & CMD_CMPL_INTR_ENABLE))
poll = 1;
pcie_wait_cmd(ctrl, poll);
}
out:
mutex_unlock(&ctrl->ctrl_lock);
return retval;
}
static int hpc_check_lnk_status(struct controller *ctrl)
{
u16 lnk_status;
int retval = 0;
retval = pciehp_readw(ctrl, LNKSTATUS, &lnk_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read LNKSTATUS register\n",
__func__);
return retval;
}
ctrl_dbg(ctrl, "%s: lnk_status = %x\n", __func__, lnk_status);
if ( (lnk_status & LNK_TRN) || (lnk_status & LNK_TRN_ERR) ||
!(lnk_status & NEG_LINK_WD)) {
ctrl_err(ctrl, "%s : Link Training Error occurs \n", __func__);
retval = -1;
return retval;
}
return retval;
}
static int hpc_get_attention_status(struct slot *slot, u8 *status)
{
struct controller *ctrl = slot->ctrl;
u16 slot_ctrl;
u8 atten_led_state;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTCTRL, &slot_ctrl);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTCTRL register\n", __func__);
return retval;
}
ctrl_dbg(ctrl, "%s: SLOTCTRL %x, value read %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_ctrl);
atten_led_state = (slot_ctrl & ATTN_LED_CTRL) >> 6;
switch (atten_led_state) {
case 0:
*status = 0xFF; /* Reserved */
break;
case 1:
*status = 1; /* On */
break;
case 2:
*status = 2; /* Blink */
break;
case 3:
*status = 0; /* Off */
break;
default:
*status = 0xFF;
break;
}
return 0;
}
static int hpc_get_power_status(struct slot *slot, u8 *status)
{
struct controller *ctrl = slot->ctrl;
u16 slot_ctrl;
u8 pwr_state;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTCTRL, &slot_ctrl);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTCTRL register\n", __func__);
return retval;
}
ctrl_dbg(ctrl, "%s: SLOTCTRL %x value read %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_ctrl);
pwr_state = (slot_ctrl & PWR_CTRL) >> 10;
switch (pwr_state) {
case 0:
*status = 1;
break;
case 1:
*status = 0;
break;
default:
*status = 0xFF;
break;
}
return retval;
}
static int hpc_get_latch_status(struct slot *slot, u8 *status)
{
struct controller *ctrl = slot->ctrl;
u16 slot_status;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTSTATUS register\n",
__func__);
return retval;
}
*status = (((slot_status & MRL_STATE) >> 5) == 0) ? 0 : 1;
return 0;
}
static int hpc_get_adapter_status(struct slot *slot, u8 *status)
{
struct controller *ctrl = slot->ctrl;
u16 slot_status;
u8 card_state;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTSTATUS register\n",
__func__);
return retval;
}
card_state = (u8)((slot_status & PRSN_STATE) >> 6);
*status = (card_state == 1) ? 1 : 0;
return 0;
}
static int hpc_query_power_fault(struct slot *slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_status;
u8 pwr_fault;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot check for power fault\n", __func__);
return retval;
}
pwr_fault = (u8)((slot_status & PWR_FAULT_DETECTED) >> 1);
return pwr_fault;
}
static int hpc_get_emi_status(struct slot *slot, u8 *status)
{
struct controller *ctrl = slot->ctrl;
u16 slot_status;
int retval = 0;
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s : Cannot check EMI status\n", __func__);
return retval;
}
*status = (slot_status & EMI_STATE) >> EMI_STATUS_BIT;
return retval;
}
static int hpc_toggle_emi(struct slot *slot)
{
u16 slot_cmd;
u16 cmd_mask;
int rc;
slot_cmd = EMI_CTRL;
cmd_mask = EMI_CTRL;
rc = pcie_write_cmd(slot->ctrl, slot_cmd, cmd_mask);
slot->last_emi_toggle = get_seconds();
return rc;
}
static int hpc_set_attention_status(struct slot *slot, u8 value)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
int rc;
cmd_mask = ATTN_LED_CTRL;
switch (value) {
case 0 : /* turn off */
slot_cmd = 0x00C0;
break;
case 1: /* turn on */
slot_cmd = 0x0040;
break;
case 2: /* turn blink */
slot_cmd = 0x0080;
break;
default:
return -1;
}
rc = pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
return rc;
}
static void hpc_set_green_led_on(struct slot *slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
slot_cmd = 0x0100;
cmd_mask = PWR_LED_CTRL;
pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
}
static void hpc_set_green_led_off(struct slot *slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
slot_cmd = 0x0300;
cmd_mask = PWR_LED_CTRL;
pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
}
static void hpc_set_green_led_blink(struct slot *slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
slot_cmd = 0x0200;
cmd_mask = PWR_LED_CTRL;
pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
}
static int hpc_power_on_slot(struct slot * slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
u16 slot_status;
int retval = 0;
ctrl_dbg(ctrl, "%s: slot->hp_slot %x\n", __func__, slot->hp_slot);
/* Clear sticky power-fault bit from previous power failures */
retval = pciehp_readw(ctrl, SLOTSTATUS, &slot_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read SLOTSTATUS register\n",
__func__);
return retval;
}
slot_status &= PWR_FAULT_DETECTED;
if (slot_status) {
retval = pciehp_writew(ctrl, SLOTSTATUS, slot_status);
if (retval) {
ctrl_err(ctrl,
"%s: Cannot write to SLOTSTATUS register\n",
__func__);
return retval;
}
}
slot_cmd = POWER_ON;
cmd_mask = PWR_CTRL;
/* Enable detection that we turned off at slot power-off time */
if (!pciehp_poll_mode) {
slot_cmd |= (PWR_FAULT_DETECT_ENABLE | MRL_DETECT_ENABLE |
PRSN_DETECT_ENABLE);
cmd_mask |= (PWR_FAULT_DETECT_ENABLE | MRL_DETECT_ENABLE |
PRSN_DETECT_ENABLE);
}
retval = pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
if (retval) {
ctrl_err(ctrl, "%s: Write %x command failed!\n",
__func__, slot_cmd);
return -1;
}
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
return retval;
}
static inline int pcie_mask_bad_dllp(struct controller *ctrl)
{
struct pci_dev *dev = ctrl->pci_dev;
int pos;
u32 reg;
pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
if (!pos)
return 0;
pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, &reg);
if (reg & PCI_ERR_COR_BAD_DLLP)
return 0;
reg |= PCI_ERR_COR_BAD_DLLP;
pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg);
return 1;
}
static inline void pcie_unmask_bad_dllp(struct controller *ctrl)
{
struct pci_dev *dev = ctrl->pci_dev;
u32 reg;
int pos;
pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
if (!pos)
return;
pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, &reg);
if (!(reg & PCI_ERR_COR_BAD_DLLP))
return;
reg &= ~PCI_ERR_COR_BAD_DLLP;
pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg);
}
static int hpc_power_off_slot(struct slot * slot)
{
struct controller *ctrl = slot->ctrl;
u16 slot_cmd;
u16 cmd_mask;
int retval = 0;
int changed;
ctrl_dbg(ctrl, "%s: slot->hp_slot %x\n", __func__, slot->hp_slot);
/*
* Set Bad DLLP Mask bit in Correctable Error Mask
* Register. This is the workaround against Bad DLLP error
* that sometimes happens during turning power off the slot
* which conforms to PCI Express 1.0a spec.
*/
changed = pcie_mask_bad_dllp(ctrl);
slot_cmd = POWER_OFF;
cmd_mask = PWR_CTRL;
/*
* If we get MRL or presence detect interrupts now, the isr
* will notice the sticky power-fault bit too and issue power
* indicator change commands. This will lead to an endless loop
* of command completions, since the power-fault bit remains on
* till the slot is powered on again.
*/
if (!pciehp_poll_mode) {
slot_cmd &= ~(PWR_FAULT_DETECT_ENABLE | MRL_DETECT_ENABLE |
PRSN_DETECT_ENABLE);
cmd_mask |= (PWR_FAULT_DETECT_ENABLE | MRL_DETECT_ENABLE |
PRSN_DETECT_ENABLE);
}
retval = pcie_write_cmd(ctrl, slot_cmd, cmd_mask);
if (retval) {
ctrl_err(ctrl, "%s: Write command failed!\n", __func__);
retval = -1;
goto out;
}
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n",
__func__, ctrl->cap_base + SLOTCTRL, slot_cmd);
out:
if (changed)
pcie_unmask_bad_dllp(ctrl);
return retval;
}
static irqreturn_t pcie_isr(int irq, void *dev_id)
{
struct controller *ctrl = (struct controller *)dev_id;
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
u16 detected, intr_loc;
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
struct slot *p_slot;
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/*
* In order to guarantee that all interrupt events are
* serviced, we need to re-inspect Slot Status register after
* clearing what is presumed to be the last pending interrupt.
*/
intr_loc = 0;
do {
if (pciehp_readw(ctrl, SLOTSTATUS, &detected)) {
ctrl_err(ctrl, "%s: Cannot read SLOTSTATUS\n",
__func__);
return IRQ_NONE;
}
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
detected &= (ATTN_BUTTN_PRESSED | PWR_FAULT_DETECTED |
MRL_SENS_CHANGED | PRSN_DETECT_CHANGED |
CMD_COMPLETED);
intr_loc |= detected;
if (!intr_loc)
return IRQ_NONE;
if (detected && pciehp_writew(ctrl, SLOTSTATUS, detected)) {
ctrl_err(ctrl, "%s: Cannot write to SLOTSTATUS\n",
__func__);
return IRQ_NONE;
}
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
} while (detected);
ctrl_dbg(ctrl, "%s: intr_loc %x\n", __func__, intr_loc);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/* Check Command Complete Interrupt Pending */
if (intr_loc & CMD_COMPLETED) {
ctrl->cmd_busy = 0;
smp_mb();
wake_up(&ctrl->queue);
}
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
if (!(intr_loc & ~CMD_COMPLETED))
return IRQ_HANDLED;
p_slot = pciehp_find_slot(ctrl, ctrl->slot_device_offset);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/* Check MRL Sensor Changed */
if (intr_loc & MRL_SENS_CHANGED)
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
pciehp_handle_switch_change(p_slot);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/* Check Attention Button Pressed */
if (intr_loc & ATTN_BUTTN_PRESSED)
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
pciehp_handle_attention_button(p_slot);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/* Check Presence Detect Changed */
if (intr_loc & PRSN_DETECT_CHANGED)
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
pciehp_handle_presence_change(p_slot);
pciehp: Fix interrupt event handlig Current pciehp implementation disables and re-enables hotplug interrupts in its interrupt handler. This operation might be intend to guarantee that interrupts for the events newly occured during previous events are being handled will be successfully generated. But current implementaion has the following prolems. - Current interrupt service routin clears status changes without waiting command completion. Because of this, events might not be cleared properly. - Current interrupt service routine clears status changes caused by disabling or enabling hotplug interrupts itself. This will lose new events that occurs during previous interrupts are being handled. - Current implementation doesn't have any serialization mechanism between the code to wait for command completion and the interrupt handler that clears the command completion events caused by itself. There is clearly race conditions between them, and it may cause the problem that waiting for command completion doesn't work for example. To fix those problems, this patch stops disabling/re-enabling hotplug interrupts in interrupt service routine. Instead of this, this patch re-inspects Slot Status register after clearing what is presumed to be the last bending interrupt in order to guarantee that all interrupt events are serviced. Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-04-25 23:38:57 +02:00
/* Check Power Fault Detected */
if (intr_loc & PWR_FAULT_DETECTED)
pciehp: fix NULL dereference in interrupt handler Fix the following NULL dereference problem reported from Pierre Ossman and Ingo Molnar. pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0 pciehp: pciehp_find_slot: slot (device=0x0) not found BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 PGD 0 Oops: 0000 [1] CPU 0 Modules linked in: Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170 RIP: 0010:[<ffffffff80494a8b>] [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP: 0000:ffff81003f83fbb0 EFLAGS: 00010046 RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046 RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103 R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30 R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926 FS: 0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000) Stack: 0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008 ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002 0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30 Call Trace: [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc [<ffffffff80260831>] request_irq+0x106/0x12f [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc [<ffffffff804933a3>] pciehp_probe+0x64/0x541 [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76 [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0 [<ffffffff8054b108>] __driver_attach+0x7c/0x7e [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8054b288>] driver_register+0x5f/0x13e [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49 [<ffffffff80c09d52>] pcied_init+0x15/0x8b [<ffffffff80bf3938>] kernel_init+0x75/0x243 [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a [<ffffffff8020c258>] child_rip+0xa/0x12 [<ffffffff8020bcec>] ? restore_args+0x0/0x30 [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243 [<ffffffff8020c24e>] ? child_rip+0x0/0x12 Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48 RIP [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113 RSP <ffff81003f83fbb0> CR2: 0000000000000070 Kernel panic - not syncing: Fatal exception The situation under which it occurs is hw and timing related: it appears to happen on a system that has PCI hotplug hardware but with no active hotplug cards, and another interrupt in the same (shared) IRQ line arrives too early, before the hotplug-slot entry has been set up - as triggered by CONFIG_DEBUG_SHIRQ=y: This patch contains the following two fixes. (1) Clear all events bits in Slot Status register to prevent the pciehp driver from detecting the spurious events that would have been occur before pciehp loading. (2) Add check whether slot initialization had been already done. This is short term fix. We need more structural fixes to install interrupt handler after slot initialization is done. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:03:16 +02:00
pciehp_handle_power_fault(p_slot);
return IRQ_HANDLED;
}
static int hpc_get_max_lnk_speed(struct slot *slot, enum pci_bus_speed *value)
{
struct controller *ctrl = slot->ctrl;
enum pcie_link_speed lnk_speed;
u32 lnk_cap;
int retval = 0;
retval = pciehp_readl(ctrl, LNKCAP, &lnk_cap);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read LNKCAP register\n", __func__);
return retval;
}
switch (lnk_cap & 0x000F) {
case 1:
lnk_speed = PCIE_2PT5GB;
break;
default:
lnk_speed = PCIE_LNK_SPEED_UNKNOWN;
break;
}
*value = lnk_speed;
ctrl_dbg(ctrl, "Max link speed = %d\n", lnk_speed);
return retval;
}
static int hpc_get_max_lnk_width(struct slot *slot,
enum pcie_link_width *value)
{
struct controller *ctrl = slot->ctrl;
enum pcie_link_width lnk_wdth;
u32 lnk_cap;
int retval = 0;
retval = pciehp_readl(ctrl, LNKCAP, &lnk_cap);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read LNKCAP register\n", __func__);
return retval;
}
switch ((lnk_cap & 0x03F0) >> 4){
case 0:
lnk_wdth = PCIE_LNK_WIDTH_RESRV;
break;
case 1:
lnk_wdth = PCIE_LNK_X1;
break;
case 2:
lnk_wdth = PCIE_LNK_X2;
break;
case 4:
lnk_wdth = PCIE_LNK_X4;
break;
case 8:
lnk_wdth = PCIE_LNK_X8;
break;
case 12:
lnk_wdth = PCIE_LNK_X12;
break;
case 16:
lnk_wdth = PCIE_LNK_X16;
break;
case 32:
lnk_wdth = PCIE_LNK_X32;
break;
default:
lnk_wdth = PCIE_LNK_WIDTH_UNKNOWN;
break;
}
*value = lnk_wdth;
ctrl_dbg(ctrl, "Max link width = %d\n", lnk_wdth);
return retval;
}
static int hpc_get_cur_lnk_speed(struct slot *slot, enum pci_bus_speed *value)
{
struct controller *ctrl = slot->ctrl;
enum pcie_link_speed lnk_speed = PCI_SPEED_UNKNOWN;
int retval = 0;
u16 lnk_status;
retval = pciehp_readw(ctrl, LNKSTATUS, &lnk_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read LNKSTATUS register\n",
__func__);
return retval;
}
switch (lnk_status & 0x0F) {
case 1:
lnk_speed = PCIE_2PT5GB;
break;
default:
lnk_speed = PCIE_LNK_SPEED_UNKNOWN;
break;
}
*value = lnk_speed;
ctrl_dbg(ctrl, "Current link speed = %d\n", lnk_speed);
return retval;
}
static int hpc_get_cur_lnk_width(struct slot *slot,
enum pcie_link_width *value)
{
struct controller *ctrl = slot->ctrl;
enum pcie_link_width lnk_wdth = PCIE_LNK_WIDTH_UNKNOWN;
int retval = 0;
u16 lnk_status;
retval = pciehp_readw(ctrl, LNKSTATUS, &lnk_status);
if (retval) {
ctrl_err(ctrl, "%s: Cannot read LNKSTATUS register\n",
__func__);
return retval;
}
switch ((lnk_status & 0x03F0) >> 4){
case 0:
lnk_wdth = PCIE_LNK_WIDTH_RESRV;
break;
case 1:
lnk_wdth = PCIE_LNK_X1;
break;
case 2:
lnk_wdth = PCIE_LNK_X2;
break;
case 4:
lnk_wdth = PCIE_LNK_X4;
break;
case 8:
lnk_wdth = PCIE_LNK_X8;
break;
case 12:
lnk_wdth = PCIE_LNK_X12;
break;
case 16:
lnk_wdth = PCIE_LNK_X16;
break;
case 32:
lnk_wdth = PCIE_LNK_X32;
break;
default:
lnk_wdth = PCIE_LNK_WIDTH_UNKNOWN;
break;
}
*value = lnk_wdth;
ctrl_dbg(ctrl, "Current link width = %d\n", lnk_wdth);
return retval;
}
static void pcie_release_ctrl(struct controller *ctrl);
static struct hpc_ops pciehp_hpc_ops = {
.power_on_slot = hpc_power_on_slot,
.power_off_slot = hpc_power_off_slot,
.set_attention_status = hpc_set_attention_status,
.get_power_status = hpc_get_power_status,
.get_attention_status = hpc_get_attention_status,
.get_latch_status = hpc_get_latch_status,
.get_adapter_status = hpc_get_adapter_status,
.get_emi_status = hpc_get_emi_status,
.toggle_emi = hpc_toggle_emi,
.get_max_bus_speed = hpc_get_max_lnk_speed,
.get_cur_bus_speed = hpc_get_cur_lnk_speed,
.get_max_lnk_width = hpc_get_max_lnk_width,
.get_cur_lnk_width = hpc_get_cur_lnk_width,
.query_power_fault = hpc_query_power_fault,
.green_led_on = hpc_set_green_led_on,
.green_led_off = hpc_set_green_led_off,
.green_led_blink = hpc_set_green_led_blink,
.release_ctlr = pcie_release_ctrl,
.check_lnk_status = hpc_check_lnk_status,
};
int pcie_enable_notification(struct controller *ctrl)
{
u16 cmd, mask;
cmd = PRSN_DETECT_ENABLE;
if (ATTN_BUTTN(ctrl))
cmd |= ATTN_BUTTN_ENABLE;
if (POWER_CTRL(ctrl))
cmd |= PWR_FAULT_DETECT_ENABLE;
if (MRL_SENS(ctrl))
cmd |= MRL_DETECT_ENABLE;
if (!pciehp_poll_mode)
cmd |= HP_INTR_ENABLE | CMD_CMPL_INTR_ENABLE;
mask = PRSN_DETECT_ENABLE | ATTN_BUTTN_ENABLE | MRL_DETECT_ENABLE |
PWR_FAULT_DETECT_ENABLE | HP_INTR_ENABLE | CMD_CMPL_INTR_ENABLE;
if (pcie_write_cmd(ctrl, cmd, mask)) {
ctrl_err(ctrl, "%s: Cannot enable software notification\n",
__func__);
return -1;
}
return 0;
}
static void pcie_disable_notification(struct controller *ctrl)
{
u16 mask;
mask = PRSN_DETECT_ENABLE | ATTN_BUTTN_ENABLE | MRL_DETECT_ENABLE |
PWR_FAULT_DETECT_ENABLE | HP_INTR_ENABLE | CMD_CMPL_INTR_ENABLE;
if (pcie_write_cmd(ctrl, 0, mask))
ctrl_warn(ctrl, "%s: Cannot disable software notification\n",
__func__);
}
static int pcie_init_notification(struct controller *ctrl)
{
if (pciehp_request_irq(ctrl))
return -1;
if (pcie_enable_notification(ctrl)) {
pciehp_free_irq(ctrl);
return -1;
}
return 0;
}
static void pcie_shutdown_notification(struct controller *ctrl)
{
pcie_disable_notification(ctrl);
pciehp_free_irq(ctrl);
}
static int pcie_init_slot(struct controller *ctrl)
{
struct slot *slot;
slot = kzalloc(sizeof(*slot), GFP_KERNEL);
if (!slot)
return -ENOMEM;
slot->hp_slot = 0;
slot->ctrl = ctrl;
slot->bus = ctrl->pci_dev->subordinate->number;
slot->device = ctrl->slot_device_offset + slot->hp_slot;
slot->hpc_ops = ctrl->hpc_ops;
slot->number = ctrl->first_slot;
mutex_init(&slot->lock);
INIT_DELAYED_WORK(&slot->work, pciehp_queue_pushbutton_work);
list_add(&slot->slot_list, &ctrl->slot_list);
return 0;
}
static void pcie_cleanup_slot(struct controller *ctrl)
{
struct slot *slot;
slot = list_first_entry(&ctrl->slot_list, struct slot, slot_list);
list_del(&slot->slot_list);
cancel_delayed_work(&slot->work);
flush_scheduled_work();
flush_workqueue(pciehp_wq);
kfree(slot);
}
static inline void dbg_ctrl(struct controller *ctrl)
{
int i;
u16 reg16;
struct pci_dev *pdev = ctrl->pci_dev;
if (!pciehp_debug)
return;
ctrl_info(ctrl, "Hotplug Controller:\n");
ctrl_info(ctrl, " Seg/Bus/Dev/Func/IRQ : %s IRQ %d\n",
pci_name(pdev), pdev->irq);
ctrl_info(ctrl, " Vendor ID : 0x%04x\n", pdev->vendor);
ctrl_info(ctrl, " Device ID : 0x%04x\n", pdev->device);
ctrl_info(ctrl, " Subsystem ID : 0x%04x\n",
pdev->subsystem_device);
ctrl_info(ctrl, " Subsystem Vendor ID : 0x%04x\n",
pdev->subsystem_vendor);
ctrl_info(ctrl, " PCIe Cap offset : 0x%02x\n", ctrl->cap_base);
for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
if (!pci_resource_len(pdev, i))
continue;
ctrl_info(ctrl, " PCI resource [%d] : 0x%llx@0x%llx\n",
i, (unsigned long long)pci_resource_len(pdev, i),
(unsigned long long)pci_resource_start(pdev, i));
}
ctrl_info(ctrl, "Slot Capabilities : 0x%08x\n", ctrl->slot_cap);
ctrl_info(ctrl, " Physical Slot Number : %d\n", ctrl->first_slot);
ctrl_info(ctrl, " Attention Button : %3s\n",
ATTN_BUTTN(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " Power Controller : %3s\n",
POWER_CTRL(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " MRL Sensor : %3s\n",
MRL_SENS(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " Attention Indicator : %3s\n",
ATTN_LED(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " Power Indicator : %3s\n",
PWR_LED(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " Hot-Plug Surprise : %3s\n",
HP_SUPR_RM(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " EMI Present : %3s\n",
EMI(ctrl) ? "yes" : "no");
ctrl_info(ctrl, " Command Completed : %3s\n",
NO_CMD_CMPL(ctrl) ? "no" : "yes");
pciehp_readw(ctrl, SLOTSTATUS, &reg16);
ctrl_info(ctrl, "Slot Status : 0x%04x\n", reg16);
pciehp_readw(ctrl, SLOTCTRL, &reg16);
ctrl_info(ctrl, "Slot Control : 0x%04x\n", reg16);
}
struct controller *pcie_init(struct pcie_device *dev)
{
struct controller *ctrl;
u32 slot_cap;
struct pci_dev *pdev = dev->port;
ctrl = kzalloc(sizeof(*ctrl), GFP_KERNEL);
if (!ctrl) {
dev_err(&dev->device, "%s : out of memory\n", __func__);
goto abort;
}
INIT_LIST_HEAD(&ctrl->slot_list);
ctrl->pcie = dev;
ctrl->pci_dev = pdev;
ctrl->cap_base = pci_find_capability(pdev, PCI_CAP_ID_EXP);
if (!ctrl->cap_base) {
ctrl_err(ctrl, "%s: Cannot find PCI Express capability\n",
__func__);
goto abort;
}
if (pciehp_readl(ctrl, SLOTCAP, &slot_cap)) {
ctrl_err(ctrl, "%s: Cannot read SLOTCAP register\n", __func__);
goto abort;
}
ctrl->slot_cap = slot_cap;
ctrl->first_slot = slot_cap >> 19;
ctrl->slot_device_offset = 0;
ctrl->num_slots = 1;
ctrl->hpc_ops = &pciehp_hpc_ops;
mutex_init(&ctrl->crit_sect);
mutex_init(&ctrl->ctrl_lock);
init_waitqueue_head(&ctrl->queue);
dbg_ctrl(ctrl);
pciehp: fix slow probing Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in http://bugzilla.kernel.org/show_bug.cgi?id=10751. The command completed bit in Slot Status register applies only to commands issued to control the attention indicator, power indicator, power controller, or electromechanical interlock. However, writes to other parts of the Slot Control register would end up writing to the control fields. Hence, any write to Slot Control register is considered as a command. However, if the controller doesn't support any of attention indicator, power indicator, power controller and electromechanical interlock, command completed bit would not set in writing to Slot Control register. In this case, we should not wait for command completed bit set, otherwise all commands would be considered not completed in timeout seconds (1 sec.). The cause of the problem is pciehp driver didn't take this situation into account. This patch changes pciehp to take it into account. This patch also add the check for "No Command Completed Support" bit in Slot Capability register. If it is set, we should not wait for command completed bit set as well. This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that pciehp did not wait for command completed properly (pciehp just ignored the command completion event). Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com> Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2008-05-27 12:04:30 +02:00
/*
* Controller doesn't notify of command completion if the "No
* Command Completed Support" bit is set in Slot Capability
* register or the controller supports none of power
* controller, attention led, power led and EMI.
*/
if (NO_CMD_CMPL(ctrl) ||
!(POWER_CTRL(ctrl) | ATTN_LED(ctrl) | PWR_LED(ctrl) | EMI(ctrl)))
ctrl->no_cmd_complete = 1;
/* Clear all remaining event bits in Slot Status register */
if (pciehp_writew(ctrl, SLOTSTATUS, 0x1f))
goto abort_ctrl;
/* Disable sotfware notification */
pcie_disable_notification(ctrl);
/*
* If this is the first controller to be initialized,
* initialize the pciehp work queue
*/
if (atomic_add_return(1, &pciehp_num_controllers) == 1) {
pciehp_wq = create_singlethread_workqueue("pciehpd");
if (!pciehp_wq)
goto abort_ctrl;
}
ctrl_info(ctrl, "HPC vendor_id %x device_id %x ss_vid %x ss_did %x\n",
pdev->vendor, pdev->device, pdev->subsystem_vendor,
pdev->subsystem_device);
if (pcie_init_slot(ctrl))
goto abort_ctrl;
if (pcie_init_notification(ctrl))
goto abort_slot;
return ctrl;
abort_slot:
pcie_cleanup_slot(ctrl);
abort_ctrl:
kfree(ctrl);
abort:
return NULL;
}
void pcie_release_ctrl(struct controller *ctrl)
{
pcie_shutdown_notification(ctrl);
pcie_cleanup_slot(ctrl);
/*
* If this is the last controller to be released, destroy the
* pciehp work queue
*/
if (atomic_dec_and_test(&pciehp_num_controllers))
destroy_workqueue(pciehp_wq);
kfree(ctrl);
}