linux/drivers/acpi/events/evmisc.c
Alexey Starikovskiy 5f7748cf91 Execute AML Notify() requests on stack.
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.

The notify handler in the ACPI thermal driver is a C-routine,
which may invoke the ACPI interpreter again to get access
to some ACPI variables such as temperature.  (acpi_evaluate_xxx)
On these HP machines such an evaluation changes state of an ASL variable
and lets the loop above break.

In the current ACPI implementation, Notify requests are being deferred
to the same kacpid workqueue on which the above GPE handler with
infinite loop is executing. Thus we have a deadlock -- loop will
continue to spin, sending notify events, and at the same time
preventing these notify events from being run on a workqueue. All
notify events are deferred, thus we see explosion in memory consumption.

Also as GPE handling is blocked, machines overheat because ACPI-based
fan control is stalled.  Eventually by external poll of the same
acpi_evaluate, kacpid is released and all the queued notify events are
free to run, thus 100% CPU utilization by kacpid for several seconds
or more.

To prevent this failure,  Linux must not send notify events to the
kacpid workqueue -- either executing them immediately or putting them
on some other thread.

The first attempt to create a new thread was done by Peter Wainwright
He created a bunch of threads, which were stealing work from a kacpid
workqueue.
This patch appeared in 2.6.15-based kernel shipped with Ubuntu 6.06 LTS.

Second attempt was done by Alexey Starikovskiy, who created a new thread
for each Notify event. This worked OK on HP nx machines,
but broke Linus' Compaq n620c, by producing threads with a speed what
they stopped the machine completely.
Thus this patch was reverted from 2.6.18-rc2.

Alexey re-made the patch to create second workqueue just for notify events,
thus hopping it will not break Linus' machine. Patch was tested on the
same HP nx machines in #5534 and #7122, but this broke Linus' machine
also and was reverted from 2.6.19-rc with much fanfair.

The 4th patch inserted schedule_timeout(1) into deferred
execution of kacpid, if we had any notify requests pending, but Linus
decided that it was too complex (involved either changes to workqueue
to see if it's empty or atomic inc/dec).  Then a 5th attempt did a
yield() to every GPE execution.

Finally, this 6th generation patch simply executes the notify handler
on the stack.  Previous attempts to do this simple solution failed
because of issues in AML mutex re-entrancy which are now fixed
by the previous patch in this series.

http://bugzilla.kernel.org/show_bug.cgi?id=5534

Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-02-15 16:13:51 -05:00

605 lines
17 KiB
C

/******************************************************************************
*
* Module Name: evmisc - Miscellaneous event manager support functions
*
*****************************************************************************/
/*
* Copyright (C) 2000 - 2007, R. Byron Moore
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions, and the following disclaimer,
* without modification.
* 2. Redistributions in binary form must reproduce at minimum a disclaimer
* substantially similar to the "NO WARRANTY" disclaimer below
* ("Disclaimer") and any redistribution must be conditioned upon
* including a substantially similar Disclaimer requirement for further
* binary redistribution.
* 3. Neither the names of the above-listed copyright holders nor the names
* of any contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* Alternatively, this software may be distributed under the terms of the
* GNU General Public License ("GPL") version 2 as published by the Free
* Software Foundation.
*
* NO WARRANTY
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
* IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGES.
*/
#include <acpi/acpi.h>
#include <acpi/acevents.h>
#include <acpi/acnamesp.h>
#include <acpi/acinterp.h>
#define _COMPONENT ACPI_EVENTS
ACPI_MODULE_NAME("evmisc")
/* Names for Notify() values, used for debug output */
#ifdef ACPI_DEBUG_OUTPUT
static const char *acpi_notify_value_names[] = {
"Bus Check",
"Device Check",
"Device Wake",
"Eject Request",
"Device Check Light",
"Frequency Mismatch",
"Bus Mode Mismatch",
"Power Fault"
};
#endif
/* Pointer to FACS needed for the Global Lock */
static struct acpi_table_facs *facs = NULL;
/* Local prototypes */
static void ACPI_SYSTEM_XFACE acpi_ev_notify_dispatch(void *context);
static u32 acpi_ev_global_lock_handler(void *context);
static acpi_status acpi_ev_remove_global_lock_handler(void);
/*******************************************************************************
*
* FUNCTION: acpi_ev_is_notify_object
*
* PARAMETERS: Node - Node to check
*
* RETURN: TRUE if notifies allowed on this object
*
* DESCRIPTION: Check type of node for a object that supports notifies.
*
* TBD: This could be replaced by a flag bit in the node.
*
******************************************************************************/
u8 acpi_ev_is_notify_object(struct acpi_namespace_node *node)
{
switch (node->type) {
case ACPI_TYPE_DEVICE:
case ACPI_TYPE_PROCESSOR:
case ACPI_TYPE_POWER:
case ACPI_TYPE_THERMAL:
/*
* These are the ONLY objects that can receive ACPI notifications
*/
return (TRUE);
default:
return (FALSE);
}
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_queue_notify_request
*
* PARAMETERS: Node - NS node for the notified object
* notify_value - Value from the Notify() request
*
* RETURN: Status
*
* DESCRIPTION: Dispatch a device notification event to a previously
* installed handler.
*
******************************************************************************/
acpi_status
acpi_ev_queue_notify_request(struct acpi_namespace_node * node,
u32 notify_value)
{
union acpi_operand_object *obj_desc;
union acpi_operand_object *handler_obj = NULL;
union acpi_generic_state *notify_info;
acpi_status status = AE_OK;
ACPI_FUNCTION_NAME(ev_queue_notify_request);
/*
* For value 3 (Ejection Request), some device method may need to be run.
* For value 2 (Device Wake) if _PRW exists, the _PS0 method may need
* to be run.
* For value 0x80 (Status Change) on the power button or sleep button,
* initiate soft-off or sleep operation?
*/
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"Dispatching Notify(%X) on node %p\n", notify_value,
node));
if (notify_value <= 7) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Notify value: %s\n",
acpi_notify_value_names[notify_value]));
} else {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"Notify value: 0x%2.2X **Device Specific**\n",
notify_value));
}
/* Get the notify object attached to the NS Node */
obj_desc = acpi_ns_get_attached_object(node);
if (obj_desc) {
/* We have the notify object, Get the right handler */
switch (node->type) {
case ACPI_TYPE_DEVICE:
case ACPI_TYPE_THERMAL:
case ACPI_TYPE_PROCESSOR:
case ACPI_TYPE_POWER:
if (notify_value <= ACPI_MAX_SYS_NOTIFY) {
handler_obj =
obj_desc->common_notify.system_notify;
} else {
handler_obj =
obj_desc->common_notify.device_notify;
}
break;
default:
/* All other types are not supported */
return (AE_TYPE);
}
}
/* If there is any handler to run, schedule the dispatcher */
if ((acpi_gbl_system_notify.handler
&& (notify_value <= ACPI_MAX_SYS_NOTIFY))
|| (acpi_gbl_device_notify.handler
&& (notify_value > ACPI_MAX_SYS_NOTIFY)) || handler_obj) {
notify_info = acpi_ut_create_generic_state();
if (!notify_info) {
return (AE_NO_MEMORY);
}
notify_info->common.descriptor_type =
ACPI_DESC_TYPE_STATE_NOTIFY;
notify_info->notify.node = node;
notify_info->notify.value = (u16) notify_value;
notify_info->notify.handler_obj = handler_obj;
acpi_ex_relinquish_interpreter();
acpi_ev_notify_dispatch(notify_info);
acpi_ex_reacquire_interpreter();
}
if (!handler_obj) {
/*
* There is no per-device notify handler for this device.
* This may or may not be a problem.
*/
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"No notify handler for Notify(%4.4s, %X) node %p\n",
acpi_ut_get_node_name(node), notify_value,
node));
}
return (status);
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_notify_dispatch
*
* PARAMETERS: Context - To be passed to the notify handler
*
* RETURN: None.
*
* DESCRIPTION: Dispatch a device notification event to a previously
* installed handler.
*
******************************************************************************/
static void ACPI_SYSTEM_XFACE acpi_ev_notify_dispatch(void *context)
{
union acpi_generic_state *notify_info =
(union acpi_generic_state *)context;
acpi_notify_handler global_handler = NULL;
void *global_context = NULL;
union acpi_operand_object *handler_obj;
ACPI_FUNCTION_ENTRY();
/*
* We will invoke a global notify handler if installed.
* This is done _before_ we invoke the per-device handler attached
* to the device.
*/
if (notify_info->notify.value <= ACPI_MAX_SYS_NOTIFY) {
/* Global system notification handler */
if (acpi_gbl_system_notify.handler) {
global_handler = acpi_gbl_system_notify.handler;
global_context = acpi_gbl_system_notify.context;
}
} else {
/* Global driver notification handler */
if (acpi_gbl_device_notify.handler) {
global_handler = acpi_gbl_device_notify.handler;
global_context = acpi_gbl_device_notify.context;
}
}
/* Invoke the system handler first, if present */
if (global_handler) {
global_handler(notify_info->notify.node,
notify_info->notify.value, global_context);
}
/* Now invoke the per-device handler, if present */
handler_obj = notify_info->notify.handler_obj;
if (handler_obj) {
handler_obj->notify.handler(notify_info->notify.node,
notify_info->notify.value,
handler_obj->notify.context);
}
/* All done with the info object */
acpi_ut_delete_generic_state(notify_info);
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_global_lock_handler
*
* PARAMETERS: Context - From thread interface, not used
*
* RETURN: ACPI_INTERRUPT_HANDLED
*
* DESCRIPTION: Invoked directly from the SCI handler when a global lock
* release interrupt occurs. Attempt to acquire the global lock,
* if successful, signal the thread waiting for the lock.
*
* NOTE: Assumes that the semaphore can be signaled from interrupt level. If
* this is not possible for some reason, a separate thread will have to be
* scheduled to do this.
*
******************************************************************************/
static u32 acpi_ev_global_lock_handler(void *context)
{
u8 acquired = FALSE;
/*
* Attempt to get the lock.
*
* If we don't get it now, it will be marked pending and we will
* take another interrupt when it becomes free.
*/
ACPI_ACQUIRE_GLOBAL_LOCK(facs, acquired);
if (acquired) {
/* Got the lock, now wake all threads waiting for it */
acpi_gbl_global_lock_acquired = TRUE;
/* Send a unit to the semaphore */
if (ACPI_FAILURE(acpi_os_signal_semaphore(
acpi_gbl_global_lock_semaphore, 1))) {
ACPI_ERROR((AE_INFO,
"Could not signal Global Lock semaphore"));
}
}
return (ACPI_INTERRUPT_HANDLED);
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_init_global_lock_handler
*
* PARAMETERS: None
*
* RETURN: Status
*
* DESCRIPTION: Install a handler for the global lock release event
*
******************************************************************************/
acpi_status acpi_ev_init_global_lock_handler(void)
{
acpi_status status;
ACPI_FUNCTION_TRACE(ev_init_global_lock_handler);
status =
acpi_get_table_by_index(ACPI_TABLE_INDEX_FACS,
(struct acpi_table_header **)&facs);
if (ACPI_FAILURE(status)) {
return_ACPI_STATUS(status);
}
acpi_gbl_global_lock_present = TRUE;
status = acpi_install_fixed_event_handler(ACPI_EVENT_GLOBAL,
acpi_ev_global_lock_handler,
NULL);
/*
* If the global lock does not exist on this platform, the attempt
* to enable GBL_STATUS will fail (the GBL_ENABLE bit will not stick)
* Map to AE_OK, but mark global lock as not present.
* Any attempt to actually use the global lock will be flagged
* with an error.
*/
if (status == AE_NO_HARDWARE_RESPONSE) {
ACPI_ERROR((AE_INFO,
"No response from Global Lock hardware, disabling lock"));
acpi_gbl_global_lock_present = FALSE;
status = AE_OK;
}
return_ACPI_STATUS(status);
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_remove_global_lock_handler
*
* PARAMETERS: None
*
* RETURN: Status
*
* DESCRIPTION: Remove the handler for the Global Lock
*
******************************************************************************/
static acpi_status acpi_ev_remove_global_lock_handler(void)
{
acpi_status status;
ACPI_FUNCTION_TRACE(ev_remove_global_lock_handler);
acpi_gbl_global_lock_present = FALSE;
status = acpi_remove_fixed_event_handler(ACPI_EVENT_GLOBAL,
acpi_ev_global_lock_handler);
return_ACPI_STATUS(status);
}
/******************************************************************************
*
* FUNCTION: acpi_ev_acquire_global_lock
*
* PARAMETERS: Timeout - Max time to wait for the lock, in millisec.
*
* RETURN: Status
*
* DESCRIPTION: Attempt to gain ownership of the Global Lock.
*
* MUTEX: Interpreter must be locked
*
* Note: The original implementation allowed multiple threads to "acquire" the
* Global Lock, and the OS would hold the lock until the last thread had
* released it. However, this could potentially starve the BIOS out of the
* lock, especially in the case where there is a tight handshake between the
* Embedded Controller driver and the BIOS. Therefore, this implementation
* allows only one thread to acquire the HW Global Lock at a time, and makes
* the global lock appear as a standard mutex on the OS side.
*
*****************************************************************************/
acpi_status acpi_ev_acquire_global_lock(u16 timeout)
{
acpi_status status = AE_OK;
u8 acquired = FALSE;
ACPI_FUNCTION_TRACE(ev_acquire_global_lock);
/*
* Only one thread can acquire the GL at a time, the global_lock_mutex
* enforces this. This interface releases the interpreter if we must wait.
*/
status = acpi_ex_system_wait_mutex(acpi_gbl_global_lock_mutex, timeout);
if (ACPI_FAILURE(status)) {
return_ACPI_STATUS(status);
}
/*
* Make sure that a global lock actually exists. If not, just treat
* the lock as a standard mutex.
*/
if (!acpi_gbl_global_lock_present) {
acpi_gbl_global_lock_acquired = TRUE;
return_ACPI_STATUS(AE_OK);
}
/* Attempt to acquire the actual hardware lock */
ACPI_ACQUIRE_GLOBAL_LOCK(facs, acquired);
if (acquired) {
/* We got the lock */
ACPI_DEBUG_PRINT((ACPI_DB_EXEC,
"Acquired hardware Global Lock\n"));
acpi_gbl_global_lock_acquired = TRUE;
return_ACPI_STATUS(AE_OK);
}
/*
* Did not get the lock. The pending bit was set above, and we must now
* wait until we get the global lock released interrupt.
*/
ACPI_DEBUG_PRINT((ACPI_DB_EXEC, "Waiting for hardware Global Lock\n"));
/*
* Wait for handshake with the global lock interrupt handler.
* This interface releases the interpreter if we must wait.
*/
status = acpi_ex_system_wait_semaphore(acpi_gbl_global_lock_semaphore,
ACPI_WAIT_FOREVER);
return_ACPI_STATUS(status);
}
/*******************************************************************************
*
* FUNCTION: acpi_ev_release_global_lock
*
* PARAMETERS: None
*
* RETURN: Status
*
* DESCRIPTION: Releases ownership of the Global Lock.
*
******************************************************************************/
acpi_status acpi_ev_release_global_lock(void)
{
u8 pending = FALSE;
acpi_status status = AE_OK;
ACPI_FUNCTION_TRACE(ev_release_global_lock);
/* Lock must be already acquired */
if (!acpi_gbl_global_lock_acquired) {
ACPI_WARNING((AE_INFO,
"Cannot release the ACPI Global Lock, it has not been acquired"));
return_ACPI_STATUS(AE_NOT_ACQUIRED);
}
if (acpi_gbl_global_lock_present) {
/* Allow any thread to release the lock */
ACPI_RELEASE_GLOBAL_LOCK(facs, pending);
/*
* If the pending bit was set, we must write GBL_RLS to the control
* register
*/
if (pending) {
status =
acpi_set_register(ACPI_BITREG_GLOBAL_LOCK_RELEASE,
1);
}
ACPI_DEBUG_PRINT((ACPI_DB_EXEC,
"Released hardware Global Lock\n"));
}
acpi_gbl_global_lock_acquired = FALSE;
/* Release the local GL mutex */
acpi_os_release_mutex(acpi_gbl_global_lock_mutex);
return_ACPI_STATUS(status);
}
/******************************************************************************
*
* FUNCTION: acpi_ev_terminate
*
* PARAMETERS: none
*
* RETURN: none
*
* DESCRIPTION: Disable events and free memory allocated for table storage.
*
******************************************************************************/
void acpi_ev_terminate(void)
{
acpi_native_uint i;
acpi_status status;
ACPI_FUNCTION_TRACE(ev_terminate);
if (acpi_gbl_events_initialized) {
/*
* Disable all event-related functionality.
* In all cases, on error, print a message but obviously we don't abort.
*/
/* Disable all fixed events */
for (i = 0; i < ACPI_NUM_FIXED_EVENTS; i++) {
status = acpi_disable_event((u32) i, 0);
if (ACPI_FAILURE(status)) {
ACPI_ERROR((AE_INFO,
"Could not disable fixed event %d",
(u32) i));
}
}
/* Disable all GPEs in all GPE blocks */
status = acpi_ev_walk_gpe_list(acpi_hw_disable_gpe_block);
/* Remove SCI handler */
status = acpi_ev_remove_sci_handler();
if (ACPI_FAILURE(status)) {
ACPI_ERROR((AE_INFO, "Could not remove SCI handler"));
}
status = acpi_ev_remove_global_lock_handler();
if (ACPI_FAILURE(status)) {
ACPI_ERROR((AE_INFO,
"Could not remove Global Lock handler"));
}
}
/* Deallocate all handler objects installed within GPE info structs */
status = acpi_ev_walk_gpe_list(acpi_ev_delete_gpe_handlers);
/* Return to original mode if necessary */
if (acpi_gbl_original_mode == ACPI_SYS_MODE_LEGACY) {
status = acpi_disable();
if (ACPI_FAILURE(status)) {
ACPI_WARNING((AE_INFO, "AcpiDisable failed"));
}
}
return_VOID;
}