2005-04-17 02:20:36 +04:00
/*
* acpi_osl . c - OS - dependent functions ( $ Revision : 83 $ )
*
* Copyright ( C ) 2000 Andrew Henroid
* Copyright ( C ) 2001 , 2002 Andy Grover < andrew . grover @ intel . com >
* Copyright ( C ) 2001 , 2002 Paul Diefenbaugh < paul . s . diefenbaugh @ intel . com >
*
* ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
*
* 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 . 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 . , 59 Temple Place , Suite 330 , Boston , MA 02111 - 1307 USA
*
* ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
*
*/
# include <linux/module.h>
# include <linux/kernel.h>
# include <linux/slab.h>
# include <linux/mm.h>
# include <linux/pci.h>
# include <linux/interrupt.h>
# include <linux/kmod.h>
# include <linux/delay.h>
# include <linux/workqueue.h>
# include <linux/nmi.h>
2007-02-02 19:48:19 +03:00
# include <linux/acpi.h>
2005-04-17 02:20:36 +04:00
# include <acpi/acpi.h>
# include <asm/io.h>
# include <acpi/acpi_bus.h>
# include <acpi/processor.h>
# include <asm/uaccess.h>
# include <linux/efi.h>
# define _COMPONENT ACPI_OS_SERVICES
2007-02-13 06:42:12 +03:00
ACPI_MODULE_NAME ( " osl " ) ;
2005-04-17 02:20:36 +04:00
# define PREFIX "ACPI: "
2005-08-05 08:44:28 +04:00
struct acpi_os_dpc {
acpi_osd_exec_callback function ;
void * context ;
2006-11-22 17:55:48 +03:00
struct work_struct work ;
2005-04-17 02:20:36 +04:00
} ;
# ifdef CONFIG_ACPI_CUSTOM_DSDT
# include CONFIG_ACPI_CUSTOM_DSDT_FILE
# endif
# ifdef ENABLE_DEBUGGER
# include <linux/kdb.h>
/* stuff for debugger support */
int acpi_in_debugger ;
EXPORT_SYMBOL ( acpi_in_debugger ) ;
extern char line_buf [ 80 ] ;
2005-08-05 08:44:28 +04:00
# endif /*ENABLE_DEBUGGER */
2005-04-17 02:20:36 +04:00
static unsigned int acpi_irq_irq ;
static acpi_osd_handler acpi_irq_handler ;
static void * acpi_irq_context ;
static struct workqueue_struct * kacpid_wq ;
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
static struct workqueue_struct * kacpi_notify_wq ;
2005-04-17 02:20:36 +04:00
2007-05-30 02:43:33 +04:00
# define OSI_STRING_LENGTH_MAX 64 /* arbitrary */
static char osi_additional_string [ OSI_STRING_LENGTH_MAX ] ;
2007-01-19 02:42:55 +03:00
static void __init acpi_request_region ( struct acpi_generic_address * addr ,
unsigned int length , char * desc )
{
struct resource * res ;
if ( ! addr - > address | | ! length )
return ;
2007-02-03 09:38:16 +03:00
if ( addr - > space_id = = ACPI_ADR_SPACE_SYSTEM_IO )
2007-01-19 02:42:55 +03:00
res = request_region ( addr - > address , length , desc ) ;
2007-02-03 09:38:16 +03:00
else if ( addr - > space_id = = ACPI_ADR_SPACE_SYSTEM_MEMORY )
2007-01-19 02:42:55 +03:00
res = request_mem_region ( addr - > address , length , desc ) ;
}
static int __init acpi_reserve_resources ( void )
{
2007-02-03 09:38:16 +03:00
acpi_request_region ( & acpi_gbl_FADT . xpm1a_event_block , acpi_gbl_FADT . pm1_event_length ,
2007-01-19 02:42:55 +03:00
" ACPI PM1a_EVT_BLK " ) ;
2007-02-03 09:38:16 +03:00
acpi_request_region ( & acpi_gbl_FADT . xpm1b_event_block , acpi_gbl_FADT . pm1_event_length ,
2007-01-19 02:42:55 +03:00
" ACPI PM1b_EVT_BLK " ) ;
2007-02-03 09:38:16 +03:00
acpi_request_region ( & acpi_gbl_FADT . xpm1a_control_block , acpi_gbl_FADT . pm1_control_length ,
2007-01-19 02:42:55 +03:00
" ACPI PM1a_CNT_BLK " ) ;
2007-02-03 09:38:16 +03:00
acpi_request_region ( & acpi_gbl_FADT . xpm1b_control_block , acpi_gbl_FADT . pm1_control_length ,
2007-01-19 02:42:55 +03:00
" ACPI PM1b_CNT_BLK " ) ;
2007-02-03 09:38:16 +03:00
if ( acpi_gbl_FADT . pm_timer_length = = 4 )
acpi_request_region ( & acpi_gbl_FADT . xpm_timer_block , 4 , " ACPI PM_TMR " ) ;
2007-01-19 02:42:55 +03:00
2007-02-03 09:38:16 +03:00
acpi_request_region ( & acpi_gbl_FADT . xpm2_control_block , acpi_gbl_FADT . pm2_control_length ,
2007-01-19 02:42:55 +03:00
" ACPI PM2_CNT_BLK " ) ;
/* Length of GPE blocks must be a non-negative multiple of 2 */
2007-02-03 09:38:16 +03:00
if ( ! ( acpi_gbl_FADT . gpe0_block_length & 0x1 ) )
acpi_request_region ( & acpi_gbl_FADT . xgpe0_block ,
acpi_gbl_FADT . gpe0_block_length , " ACPI GPE0_BLK " ) ;
2007-01-19 02:42:55 +03:00
2007-02-03 09:38:16 +03:00
if ( ! ( acpi_gbl_FADT . gpe1_block_length & 0x1 ) )
acpi_request_region ( & acpi_gbl_FADT . xgpe1_block ,
acpi_gbl_FADT . gpe1_block_length , " ACPI GPE1_BLK " ) ;
2007-01-19 02:42:55 +03:00
return 0 ;
}
device_initcall ( acpi_reserve_resources ) ;
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_initialize ( void )
2005-04-17 02:20:36 +04:00
{
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_initialize1 ( void )
2005-04-17 02:20:36 +04:00
{
/*
* Initialize PCI configuration space access , as we ' ll need to access
* it while walking the namespace ( bus 0 and root bridges w / _BBNs ) .
*/
if ( ! raw_pci_ops ) {
2005-08-05 08:44:28 +04:00
printk ( KERN_ERR PREFIX
" Access to PCI configuration space unavailable \n " ) ;
2005-04-17 02:20:36 +04:00
return AE_NULL_ENTRY ;
}
kacpid_wq = create_singlethread_workqueue ( " kacpid " ) ;
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
kacpi_notify_wq = create_singlethread_workqueue ( " kacpi_notify " ) ;
2005-04-17 02:20:36 +04:00
BUG_ON ( ! kacpid_wq ) ;
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
BUG_ON ( ! kacpi_notify_wq ) ;
2005-04-17 02:20:36 +04:00
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_terminate ( void )
2005-04-17 02:20:36 +04:00
{
if ( acpi_irq_handler ) {
acpi_os_remove_interrupt_handler ( acpi_irq_irq ,
acpi_irq_handler ) ;
}
destroy_workqueue ( kacpid_wq ) ;
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
destroy_workqueue ( kacpi_notify_wq ) ;
2005-04-17 02:20:36 +04:00
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
void acpi_os_printf ( const char * fmt , . . . )
2005-04-17 02:20:36 +04:00
{
va_list args ;
va_start ( args , fmt ) ;
acpi_os_vprintf ( fmt , args ) ;
va_end ( args ) ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_printf ) ;
2005-08-05 08:44:28 +04:00
void acpi_os_vprintf ( const char * fmt , va_list args )
2005-04-17 02:20:36 +04:00
{
static char buffer [ 512 ] ;
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
vsprintf ( buffer , fmt , args ) ;
# ifdef ENABLE_DEBUGGER
if ( acpi_in_debugger ) {
kdb_printf ( " %s " , buffer ) ;
} else {
printk ( " %s " , buffer ) ;
}
# else
printk ( " %s " , buffer ) ;
# endif
}
2007-02-02 19:48:19 +03:00
acpi_physical_address __init acpi_os_get_root_pointer ( void )
2005-04-17 02:20:36 +04:00
{
if ( efi_enabled ) {
2006-03-26 13:37:08 +04:00
if ( efi . acpi20 ! = EFI_INVALID_TABLE_ADDR )
2007-02-02 19:48:19 +03:00
return efi . acpi20 ;
2006-03-26 13:37:08 +04:00
else if ( efi . acpi ! = EFI_INVALID_TABLE_ADDR )
2007-02-02 19:48:19 +03:00
return efi . acpi ;
2005-04-17 02:20:36 +04:00
else {
2005-08-05 08:44:28 +04:00
printk ( KERN_ERR PREFIX
" System description tables not found \n " ) ;
2007-02-02 19:48:19 +03:00
return 0 ;
2005-04-17 02:20:36 +04:00
}
2007-02-02 19:48:19 +03:00
} else
return acpi_find_rsdp ( ) ;
2005-04-17 02:20:36 +04:00
}
2007-02-02 19:48:19 +03:00
void __iomem * acpi_os_map_memory ( acpi_physical_address phys , acpi_size size )
2005-04-17 02:20:36 +04:00
{
2006-03-26 13:37:10 +04:00
if ( phys > ULONG_MAX ) {
printk ( KERN_ERR PREFIX " Cannot map memory that high \n " ) ;
2007-02-14 03:11:36 +03:00
return NULL ;
2005-04-17 02:20:36 +04:00
}
2007-02-02 19:48:19 +03:00
if ( acpi_gbl_permanent_mmap )
/*
* ioremap checks to ensure this is in reserved space
*/
return ioremap ( ( unsigned long ) phys , size ) ;
else
return __acpi_map_table ( ( unsigned long ) phys , size ) ;
2005-04-17 02:20:36 +04:00
}
2006-01-08 12:03:15 +03:00
EXPORT_SYMBOL_GPL ( acpi_os_map_memory ) ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
void acpi_os_unmap_memory ( void __iomem * virt , acpi_size size )
2005-04-17 02:20:36 +04:00
{
2007-02-02 19:48:19 +03:00
if ( acpi_gbl_permanent_mmap ) {
iounmap ( virt ) ;
}
2005-04-17 02:20:36 +04:00
}
2006-01-08 12:03:15 +03:00
EXPORT_SYMBOL_GPL ( acpi_os_unmap_memory ) ;
2005-04-17 02:20:36 +04:00
# ifdef ACPI_FUTURE_USAGE
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_get_physical_address ( void * virt , acpi_physical_address * phys )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
if ( ! phys | | ! virt )
2005-04-17 02:20:36 +04:00
return AE_BAD_PARAMETER ;
* phys = virt_to_phys ( virt ) ;
return AE_OK ;
}
# endif
# define ACPI_MAX_OVERRIDE_LEN 100
static char acpi_os_name [ ACPI_MAX_OVERRIDE_LEN ] ;
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_predefined_override ( const struct acpi_predefined_names * init_val ,
acpi_string * new_val )
2005-04-17 02:20:36 +04:00
{
if ( ! init_val | | ! new_val )
return AE_BAD_PARAMETER ;
* new_val = NULL ;
2005-08-05 08:44:28 +04:00
if ( ! memcmp ( init_val - > name , " _OS_ " , 4 ) & & strlen ( acpi_os_name ) ) {
2005-04-17 02:20:36 +04:00
printk ( KERN_INFO PREFIX " Overriding _OS definition to '%s' \n " ,
2005-08-05 08:44:28 +04:00
acpi_os_name ) ;
2005-04-17 02:20:36 +04:00
* new_val = acpi_os_name ;
}
return AE_OK ;
}
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_table_override ( struct acpi_table_header * existing_table ,
struct acpi_table_header * * new_table )
2005-04-17 02:20:36 +04:00
{
if ( ! existing_table | | ! new_table )
return AE_BAD_PARAMETER ;
# ifdef CONFIG_ACPI_CUSTOM_DSDT
if ( strncmp ( existing_table - > signature , " DSDT " , 4 ) = = 0 )
2005-08-05 08:44:28 +04:00
* new_table = ( struct acpi_table_header * ) AmlCode ;
2005-04-17 02:20:36 +04:00
else
* new_table = NULL ;
# else
* new_table = NULL ;
# endif
return AE_OK ;
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 17:55:46 +04:00
static irqreturn_t acpi_irq ( int irq , void * dev_id )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
return ( * acpi_irq_handler ) ( acpi_irq_context ) ? IRQ_HANDLED : IRQ_NONE ;
2005-04-17 02:20:36 +04:00
}
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_install_interrupt_handler ( u32 gsi , acpi_osd_handler handler ,
void * context )
2005-04-17 02:20:36 +04:00
{
unsigned int irq ;
/*
* Ignore the GSI from the core , and use the value in our copy of the
* FADT . It may not be the same if an interrupt source override exists
* for the SCI .
*/
2007-02-02 19:48:22 +03:00
gsi = acpi_gbl_FADT . sci_interrupt ;
2005-04-17 02:20:36 +04:00
if ( acpi_gsi_to_irq ( gsi , & irq ) < 0 ) {
printk ( KERN_ERR PREFIX " SCI (ACPI GSI %d) not registered \n " ,
gsi ) ;
return AE_OK ;
}
acpi_irq_handler = handler ;
acpi_irq_context = context ;
2006-07-02 06:29:38 +04:00
if ( request_irq ( irq , acpi_irq , IRQF_SHARED , " acpi " , acpi_irq ) ) {
2005-04-17 02:20:36 +04:00
printk ( KERN_ERR PREFIX " SCI (IRQ%d) allocation failed \n " , irq ) ;
return AE_NOT_ACQUIRED ;
}
acpi_irq_irq = irq ;
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_remove_interrupt_handler ( u32 irq , acpi_osd_handler handler )
2005-04-17 02:20:36 +04:00
{
if ( irq ) {
free_irq ( irq , acpi_irq ) ;
acpi_irq_handler = NULL ;
acpi_irq_irq = 0 ;
}
return AE_OK ;
}
/*
* Running in interpreter thread context , safe to sleep
*/
2005-08-05 08:44:28 +04:00
void acpi_os_sleep ( acpi_integer ms )
2005-04-17 02:20:36 +04:00
{
2005-11-07 12:01:14 +03:00
schedule_timeout_interruptible ( msecs_to_jiffies ( ms ) ) ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_sleep ) ;
2005-08-05 08:44:28 +04:00
void acpi_os_stall ( u32 us )
2005-04-17 02:20:36 +04:00
{
while ( us ) {
u32 delay = 1000 ;
if ( delay > us )
delay = us ;
udelay ( delay ) ;
touch_nmi_watchdog ( ) ;
us - = delay ;
}
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_stall ) ;
/*
* Support ACPI 3.0 AML Timer operand
* Returns 64 - bit free - running , monotonically increasing timer
* with 100 ns granularity
*/
2005-08-05 08:44:28 +04:00
u64 acpi_os_get_timer ( void )
2005-04-17 02:20:36 +04:00
{
static u64 t ;
# ifdef CONFIG_HPET
/* TBD: use HPET if available */
# endif
# ifdef CONFIG_X86_PM_TIMER
/* TBD: default to PM timer if HPET was not available */
# endif
if ( ! t )
printk ( KERN_ERR PREFIX " acpi_os_get_timer() TBD \n " ) ;
return + + t ;
}
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_read_port ( acpi_io_address port , u32 * value , u32 width )
2005-04-17 02:20:36 +04:00
{
u32 dummy ;
if ( ! value )
value = & dummy ;
2005-08-05 08:44:28 +04:00
switch ( width ) {
2005-04-17 02:20:36 +04:00
case 8 :
2005-08-05 08:44:28 +04:00
* ( u8 * ) value = inb ( port ) ;
2005-04-17 02:20:36 +04:00
break ;
case 16 :
2005-08-05 08:44:28 +04:00
* ( u16 * ) value = inw ( port ) ;
2005-04-17 02:20:36 +04:00
break ;
case 32 :
2005-08-05 08:44:28 +04:00
* ( u32 * ) value = inl ( port ) ;
2005-04-17 02:20:36 +04:00
break ;
default :
BUG ( ) ;
}
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_read_port ) ;
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_write_port ( acpi_io_address port , u32 value , u32 width )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
switch ( width ) {
2005-04-17 02:20:36 +04:00
case 8 :
outb ( value , port ) ;
break ;
case 16 :
outw ( value , port ) ;
break ;
case 32 :
outl ( value , port ) ;
break ;
default :
BUG ( ) ;
}
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_write_port ) ;
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_read_memory ( acpi_physical_address phys_addr , u32 * value , u32 width )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
u32 dummy ;
void __iomem * virt_addr ;
2005-04-17 02:20:36 +04:00
2006-03-26 13:37:10 +04:00
virt_addr = ioremap ( phys_addr , width ) ;
2005-04-17 02:20:36 +04:00
if ( ! value )
value = & dummy ;
switch ( width ) {
case 8 :
2005-08-05 08:44:28 +04:00
* ( u8 * ) value = readb ( virt_addr ) ;
2005-04-17 02:20:36 +04:00
break ;
case 16 :
2005-08-05 08:44:28 +04:00
* ( u16 * ) value = readw ( virt_addr ) ;
2005-04-17 02:20:36 +04:00
break ;
case 32 :
2005-08-05 08:44:28 +04:00
* ( u32 * ) value = readl ( virt_addr ) ;
2005-04-17 02:20:36 +04:00
break ;
default :
BUG ( ) ;
}
2006-03-26 13:37:10 +04:00
iounmap ( virt_addr ) ;
2005-04-17 02:20:36 +04:00
return AE_OK ;
}
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_write_memory ( acpi_physical_address phys_addr , u32 value , u32 width )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
void __iomem * virt_addr ;
2005-04-17 02:20:36 +04:00
2006-03-26 13:37:10 +04:00
virt_addr = ioremap ( phys_addr , width ) ;
2005-04-17 02:20:36 +04:00
switch ( width ) {
case 8 :
writeb ( value , virt_addr ) ;
break ;
case 16 :
writew ( value , virt_addr ) ;
break ;
case 32 :
writel ( value , virt_addr ) ;
break ;
default :
BUG ( ) ;
}
2006-03-26 13:37:10 +04:00
iounmap ( virt_addr ) ;
2005-04-17 02:20:36 +04:00
return AE_OK ;
}
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_read_pci_configuration ( struct acpi_pci_id * pci_id , u32 reg ,
void * value , u32 width )
2005-04-17 02:20:36 +04:00
{
int result , size ;
if ( ! value )
return AE_BAD_PARAMETER ;
switch ( width ) {
case 8 :
size = 1 ;
break ;
case 16 :
size = 2 ;
break ;
case 32 :
size = 4 ;
break ;
default :
return AE_ERROR ;
}
BUG_ON ( ! raw_pci_ops ) ;
result = raw_pci_ops - > read ( pci_id - > segment , pci_id - > bus ,
2005-08-05 08:44:28 +04:00
PCI_DEVFN ( pci_id - > device , pci_id - > function ) ,
reg , size , value ) ;
2005-04-17 02:20:36 +04:00
return ( result ? AE_ERROR : AE_OK ) ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_read_pci_configuration ) ;
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_write_pci_configuration ( struct acpi_pci_id * pci_id , u32 reg ,
acpi_integer value , u32 width )
2005-04-17 02:20:36 +04:00
{
int result , size ;
switch ( width ) {
case 8 :
size = 1 ;
break ;
case 16 :
size = 2 ;
break ;
case 32 :
size = 4 ;
break ;
default :
return AE_ERROR ;
}
BUG_ON ( ! raw_pci_ops ) ;
result = raw_pci_ops - > write ( pci_id - > segment , pci_id - > bus ,
2005-08-05 08:44:28 +04:00
PCI_DEVFN ( pci_id - > device , pci_id - > function ) ,
reg , size , value ) ;
2005-04-17 02:20:36 +04:00
return ( result ? AE_ERROR : AE_OK ) ;
}
/* TODO: Change code to take advantage of driver model more */
2005-08-05 08:44:28 +04:00
static void acpi_os_derive_pci_id_2 ( acpi_handle rhandle , /* upper bound */
acpi_handle chandle , /* current node */
struct acpi_pci_id * * id ,
int * is_bridge , u8 * bus_number )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
acpi_handle handle ;
struct acpi_pci_id * pci_id = * id ;
acpi_status status ;
unsigned long temp ;
acpi_object_type type ;
u8 tu8 ;
2005-04-17 02:20:36 +04:00
acpi_get_parent ( chandle , & handle ) ;
if ( handle ! = rhandle ) {
2005-08-05 08:44:28 +04:00
acpi_os_derive_pci_id_2 ( rhandle , handle , & pci_id , is_bridge ,
bus_number ) ;
2005-04-17 02:20:36 +04:00
status = acpi_get_type ( handle , & type ) ;
2005-08-05 08:44:28 +04:00
if ( ( ACPI_FAILURE ( status ) ) | | ( type ! = ACPI_TYPE_DEVICE ) )
2005-04-17 02:20:36 +04:00
return ;
2005-08-05 08:44:28 +04:00
status =
acpi_evaluate_integer ( handle , METHOD_NAME__ADR , NULL ,
& temp ) ;
2005-04-17 02:20:36 +04:00
if ( ACPI_SUCCESS ( status ) ) {
2005-08-05 08:44:28 +04:00
pci_id - > device = ACPI_HIWORD ( ACPI_LODWORD ( temp ) ) ;
pci_id - > function = ACPI_LOWORD ( ACPI_LODWORD ( temp ) ) ;
2005-04-17 02:20:36 +04:00
if ( * is_bridge )
pci_id - > bus = * bus_number ;
/* any nicer way to get bus number of bridge ? */
2005-08-05 08:44:28 +04:00
status =
acpi_os_read_pci_configuration ( pci_id , 0x0e , & tu8 ,
8 ) ;
if ( ACPI_SUCCESS ( status )
& & ( ( tu8 & 0x7f ) = = 1 | | ( tu8 & 0x7f ) = = 2 ) ) {
status =
acpi_os_read_pci_configuration ( pci_id , 0x18 ,
& tu8 , 8 ) ;
2005-04-17 02:20:36 +04:00
if ( ! ACPI_SUCCESS ( status ) ) {
/* Certainly broken... FIX ME */
return ;
}
* is_bridge = 1 ;
pci_id - > bus = tu8 ;
2005-08-05 08:44:28 +04:00
status =
acpi_os_read_pci_configuration ( pci_id , 0x19 ,
& tu8 , 8 ) ;
2005-04-17 02:20:36 +04:00
if ( ACPI_SUCCESS ( status ) ) {
* bus_number = tu8 ;
}
} else
* is_bridge = 0 ;
}
}
}
2005-08-05 08:44:28 +04:00
void acpi_os_derive_pci_id ( acpi_handle rhandle , /* upper bound */
acpi_handle chandle , /* current node */
struct acpi_pci_id * * id )
2005-04-17 02:20:36 +04:00
{
int is_bridge = 1 ;
u8 bus_number = ( * id ) - > bus ;
acpi_os_derive_pci_id_2 ( rhandle , chandle , id , & is_bridge , & bus_number ) ;
}
2006-11-22 17:55:48 +03:00
static void acpi_os_execute_deferred ( struct work_struct * work )
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
{
struct acpi_os_dpc * dpc = container_of ( work , struct acpi_os_dpc , work ) ;
if ( ! dpc ) {
printk ( KERN_ERR PREFIX " Invalid (NULL) context \n " ) ;
return ;
}
dpc - > function ( dpc - > context ) ;
kfree ( dpc ) ;
/* Yield cpu to notify thread */
cond_resched ( ) ;
return ;
}
static void acpi_os_execute_notify ( struct work_struct * work )
2005-04-17 02:20:36 +04:00
{
2006-11-22 17:55:48 +03:00
struct acpi_os_dpc * dpc = container_of ( work , struct acpi_os_dpc , work ) ;
2005-04-17 02:20:36 +04:00
if ( ! dpc ) {
2006-06-27 07:41:38 +04:00
printk ( KERN_ERR PREFIX " Invalid (NULL) context \n " ) ;
2006-06-27 08:41:40 +04:00
return ;
2005-04-17 02:20:36 +04:00
}
dpc - > function ( dpc - > context ) ;
kfree ( dpc ) ;
2006-06-27 08:41:40 +04:00
return ;
2005-04-17 02:20:36 +04:00
}
2006-05-05 11:23:00 +04:00
/*******************************************************************************
*
* FUNCTION : acpi_os_execute
*
* PARAMETERS : Type - Type of the callback
* Function - Function to be executed
* Context - Function parameters
*
* RETURN : Status
*
* DESCRIPTION : Depending on type , either queues function for deferred execution or
* immediately executes function on a separate thread .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
acpi_status acpi_os_execute ( acpi_execute_type type ,
2005-08-05 08:44:28 +04:00
acpi_osd_exec_callback function , void * context )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
acpi_status status = AE_OK ;
struct acpi_os_dpc * dpc ;
2006-07-13 06:46:42 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_EXEC ,
" Scheduling function [%p(%p)] for deferred execution. \n " ,
function , context ) ) ;
2005-04-17 02:20:36 +04:00
if ( ! function )
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
return AE_BAD_PARAMETER ;
2006-07-13 06:46:42 +04:00
2005-04-17 02:20:36 +04:00
/*
* Allocate / initialize DPC structure . Note that this memory will be
2006-11-22 17:55:48 +03:00
* freed by the callee . The kernel handles the work_struct list in a
2005-04-17 02:20:36 +04:00
* way that allows us to also free its memory inside the callee .
* Because we may want to schedule several tasks with different
* parameters we can ' t use the approach some kernel code uses of
2006-11-22 17:55:48 +03:00
* having a static work_struct .
2005-04-17 02:20:36 +04:00
*/
2006-07-13 06:46:42 +04:00
2006-11-22 17:55:48 +03:00
dpc = kmalloc ( sizeof ( struct acpi_os_dpc ) , GFP_ATOMIC ) ;
2005-04-17 02:20:36 +04:00
if ( ! dpc )
2006-11-18 06:31:09 +03:00
return_ACPI_STATUS ( AE_NO_MEMORY ) ;
2005-04-17 02:20:36 +04:00
dpc - > function = function ;
dpc - > context = context ;
2006-11-18 06:31:09 +03:00
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
if ( type = = OSL_NOTIFY_HANDLER ) {
INIT_WORK ( & dpc - > work , acpi_os_execute_notify ) ;
if ( ! queue_work ( kacpi_notify_wq , & dpc - > work ) ) {
status = AE_ERROR ;
kfree ( dpc ) ;
}
} else {
INIT_WORK ( & dpc - > work , acpi_os_execute_deferred ) ;
if ( ! queue_work ( kacpid_wq , & dpc - > work ) ) {
ACPI_DEBUG_PRINT ( ( ACPI_DB_ERROR ,
2006-11-18 06:31:09 +03:00
" Call to queue_work() failed. \n " ) ) ;
ACPI: created a dedicated workqueue for notify() execution
HP nx6125/nx6325/... machines have a _GPE handler with an infinite
loop sending Notify() events to different ACPI subsystems.
Notify handler in ACPI driver is a C-routine, which may call ACPI
interpreter again to get access to some ACPI variables
(acpi_evaluate_xxx).
On these HP machines such an evaluation changes state of some 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 increase in memory consumption
noticed by author of the thread. Also as GPE handling is bloked,
machines overheat. 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 all these horrors it's needed to not put notify events to
kacpid workqueue by either executing them immediately or putting them
on some other thread. It's dangerous to execute notify events in
place, as it will put several ACPI interpreter stacks on top of each
other (at least 4 in case of nx6125), thus causing kernel stack
overflow.
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 kernel shipped with Ubuntu 6.06 LTS.
Second attempt was done by me, I 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 18-rc2 as I remember.
I 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 I did not received reply
from Linus on a test patch sent to him.
Patch went to 19-rc and was rejected with much fanfare again.
There was 4th patch, which 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).
Now you see last variant which adds yield() to every GPE execution.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
http://bugzilla.kernel.org/show_bug.cgi?id=8385
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2007-05-10 07:31:03 +04:00
status = AE_ERROR ;
kfree ( dpc ) ;
}
2005-04-17 02:20:36 +04:00
}
2006-11-18 06:31:09 +03:00
return_ACPI_STATUS ( status ) ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
2006-05-05 11:23:00 +04:00
EXPORT_SYMBOL ( acpi_os_execute ) ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
void acpi_os_wait_events_complete ( void * context )
2005-04-17 02:20:36 +04:00
{
flush_workqueue ( kacpid_wq ) ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_wait_events_complete ) ;
/*
* Allocate the memory for a spinlock and initialize it .
*/
2006-06-24 01:04:00 +04:00
acpi_status acpi_os_create_lock ( acpi_spinlock * handle )
2005-04-17 02:20:36 +04:00
{
2006-06-24 01:04:00 +04:00
spin_lock_init ( * handle ) ;
2005-04-17 02:20:36 +04:00
2006-06-27 08:41:40 +04:00
return AE_OK ;
2005-04-17 02:20:36 +04:00
}
/*
* Deallocate the memory for a spinlock .
*/
2006-06-24 01:04:00 +04:00
void acpi_os_delete_lock ( acpi_spinlock handle )
2005-04-17 02:20:36 +04:00
{
2006-06-27 08:41:40 +04:00
return ;
2005-04-17 02:20:36 +04:00
}
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_create_semaphore ( u32 max_units , u32 initial_units , acpi_handle * handle )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
struct semaphore * sem = NULL ;
2005-04-17 02:20:36 +04:00
sem = acpi_os_allocate ( sizeof ( struct semaphore ) ) ;
if ( ! sem )
2006-06-27 08:41:40 +04:00
return AE_NO_MEMORY ;
2005-04-17 02:20:36 +04:00
memset ( sem , 0 , sizeof ( struct semaphore ) ) ;
sema_init ( sem , initial_units ) ;
2005-08-05 08:44:28 +04:00
* handle = ( acpi_handle * ) sem ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX , " Creating semaphore[%p|%d]. \n " ,
* handle , initial_units ) ) ;
2005-04-17 02:20:36 +04:00
2006-06-27 08:41:40 +04:00
return AE_OK ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
EXPORT_SYMBOL ( acpi_os_create_semaphore ) ;
2005-04-17 02:20:36 +04:00
/*
* TODO : A better way to delete semaphores ? Linux doesn ' t have a
* ' delete_semaphore ( ) ' function - - may result in an invalid
* pointer dereference for non - synchronized consumers . Should
* we at least check for blocked threads and signal / cancel them ?
*/
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_delete_semaphore ( acpi_handle handle )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
struct semaphore * sem = ( struct semaphore * ) handle ;
2005-04-17 02:20:36 +04:00
if ( ! sem )
2006-06-27 08:41:40 +04:00
return AE_BAD_PARAMETER ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX , " Deleting semaphore[%p]. \n " , handle ) ) ;
2005-04-17 02:20:36 +04:00
2006-06-30 11:19:10 +04:00
kfree ( sem ) ;
2005-08-05 08:44:28 +04:00
sem = NULL ;
2005-04-17 02:20:36 +04:00
2006-06-27 08:41:40 +04:00
return AE_OK ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
EXPORT_SYMBOL ( acpi_os_delete_semaphore ) ;
2005-04-17 02:20:36 +04:00
/*
* TODO : The kernel doesn ' t have a ' down_timeout ' function - - had to
* improvise . The process is to sleep for one scheduler quantum
* until the semaphore becomes available . Downside is that this
* may result in starvation for timeout - based waits when there ' s
* lots of semaphore activity .
*
* TODO : Support for units > 1 ?
*/
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_wait_semaphore ( acpi_handle handle , u32 units , u16 timeout )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
acpi_status status = AE_OK ;
struct semaphore * sem = ( struct semaphore * ) handle ;
int ret = 0 ;
2005-04-17 02:20:36 +04:00
if ( ! sem | | ( units < 1 ) )
2006-06-27 08:41:40 +04:00
return AE_BAD_PARAMETER ;
2005-04-17 02:20:36 +04:00
if ( units > 1 )
2006-06-27 08:41:40 +04:00
return AE_SUPPORT ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX , " Waiting for semaphore[%p|%d|%d] \n " ,
handle , units , timeout ) ) ;
2005-04-17 02:20:36 +04:00
2006-08-17 03:16:58 +04:00
/*
* This can be called during resume with interrupts off .
* Like boot - time , we should be single threaded and will
* always get the lock if we try - - timeout or not .
* If this doesn ' t succeed , then we will oops courtesy of
* might_sleep ( ) in down ( ) .
*/
if ( ! down_trylock ( sem ) )
return AE_OK ;
2005-08-05 08:44:28 +04:00
switch ( timeout ) {
2005-04-17 02:20:36 +04:00
/*
* No Wait :
* - - - - - - - -
* A zero timeout value indicates that we shouldn ' t wait - just
* acquire the semaphore if available otherwise return AE_TIME
* ( a . k . a . ' would block ' ) .
*/
2005-08-05 08:44:28 +04:00
case 0 :
if ( down_trylock ( sem ) )
2005-04-17 02:20:36 +04:00
status = AE_TIME ;
break ;
/*
* Wait Indefinitely :
* - - - - - - - - - - - - - - - - - -
*/
2005-08-05 08:44:28 +04:00
case ACPI_WAIT_FOREVER :
2005-04-17 02:20:36 +04:00
down ( sem ) ;
break ;
/*
* Wait w / Timeout :
* - - - - - - - - - - - - - - - -
*/
2005-08-05 08:44:28 +04:00
default :
2005-04-17 02:20:36 +04:00
// TODO: A better timeout algorithm?
{
int i = 0 ;
2005-08-05 08:44:28 +04:00
static const int quantum_ms = 1000 / HZ ;
2005-04-17 02:20:36 +04:00
ret = down_trylock ( sem ) ;
2005-12-31 09:45:00 +03:00
for ( i = timeout ; ( i > 0 & & ret ! = 0 ) ; i - = quantum_ms ) {
2005-11-07 12:01:14 +03:00
schedule_timeout_interruptible ( 1 ) ;
2005-04-17 02:20:36 +04:00
ret = down_trylock ( sem ) ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
if ( ret ! = 0 )
status = AE_TIME ;
}
break ;
}
if ( ACPI_FAILURE ( status ) ) {
2006-04-27 13:25:00 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX ,
2006-06-27 07:58:43 +04:00
" Failed to acquire semaphore[%p|%d|%d], %s " ,
2005-08-05 08:44:28 +04:00
handle , units , timeout ,
acpi_format_exception ( status ) ) ) ;
} else {
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX ,
2006-06-27 07:58:43 +04:00
" Acquired semaphore[%p|%d|%d] " , handle ,
2005-08-05 08:44:28 +04:00
units , timeout ) ) ;
2005-04-17 02:20:36 +04:00
}
2006-06-27 08:41:40 +04:00
return status ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
EXPORT_SYMBOL ( acpi_os_wait_semaphore ) ;
2005-04-17 02:20:36 +04:00
/*
* TODO : Support for units > 1 ?
*/
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_signal_semaphore ( acpi_handle handle , u32 units )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
struct semaphore * sem = ( struct semaphore * ) handle ;
2005-04-17 02:20:36 +04:00
if ( ! sem | | ( units < 1 ) )
2006-06-27 08:41:40 +04:00
return AE_BAD_PARAMETER ;
2005-04-17 02:20:36 +04:00
if ( units > 1 )
2006-06-27 08:41:40 +04:00
return AE_SUPPORT ;
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
ACPI_DEBUG_PRINT ( ( ACPI_DB_MUTEX , " Signaling semaphore[%p|%d] \n " , handle ,
units ) ) ;
2005-04-17 02:20:36 +04:00
up ( sem ) ;
2006-06-27 08:41:40 +04:00
return AE_OK ;
2005-04-17 02:20:36 +04:00
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_signal_semaphore ) ;
# ifdef ACPI_FUTURE_USAGE
2005-08-05 08:44:28 +04:00
u32 acpi_os_get_line ( char * buffer )
2005-04-17 02:20:36 +04:00
{
# ifdef ENABLE_DEBUGGER
if ( acpi_in_debugger ) {
u32 chars ;
kdb_read ( buffer , sizeof ( line_buf ) ) ;
/* remove the CR kdb includes */
chars = strlen ( buffer ) - 1 ;
buffer [ chars ] = ' \0 ' ;
}
# endif
return 0 ;
}
2005-08-05 08:44:28 +04:00
# endif /* ACPI_FUTURE_USAGE */
2005-04-17 02:20:36 +04:00
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_signal ( u32 function , void * info )
2005-04-17 02:20:36 +04:00
{
2005-08-05 08:44:28 +04:00
switch ( function ) {
2005-04-17 02:20:36 +04:00
case ACPI_SIGNAL_FATAL :
printk ( KERN_ERR PREFIX " Fatal opcode executed \n " ) ;
break ;
case ACPI_SIGNAL_BREAKPOINT :
/*
* AML Breakpoint
* ACPI spec . says to treat it as a NOP unless
* you are debugging . So if / when we integrate
* AML debugger into the kernel debugger its
* hook will go here . But until then it is
* not useful to print anything on breakpoints .
*/
break ;
default :
break ;
}
return AE_OK ;
}
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
EXPORT_SYMBOL ( acpi_os_signal ) ;
2005-08-05 08:44:28 +04:00
static int __init acpi_os_name_setup ( char * str )
2005-04-17 02:20:36 +04:00
{
char * p = acpi_os_name ;
2005-08-05 08:44:28 +04:00
int count = ACPI_MAX_OVERRIDE_LEN - 1 ;
2005-04-17 02:20:36 +04:00
if ( ! str | | ! * str )
return 0 ;
for ( ; count - - & & str & & * str ; str + + ) {
if ( isalnum ( * str ) | | * str = = ' ' | | * str = = ' : ' )
* p + + = * str ;
else if ( * str = = ' \' ' | | * str = = ' " ' )
continue ;
else
break ;
}
* p = 0 ;
return 1 ;
2005-08-05 08:44:28 +04:00
2005-04-17 02:20:36 +04:00
}
__setup ( " acpi_os_name= " , acpi_os_name_setup ) ;
/*
2007-05-30 02:43:33 +04:00
* Modify the list of " OS Interfaces " reported to BIOS via _OSI
*
2005-04-17 02:20:36 +04:00
* empty string disables _OSI
2007-05-30 02:43:33 +04:00
* string starting with ' ! ' disables that string
* otherwise string is added to list , augmenting built - in strings
2005-04-17 02:20:36 +04:00
*/
2005-08-05 08:44:28 +04:00
static int __init acpi_osi_setup ( char * str )
2005-04-17 02:20:36 +04:00
{
if ( str = = NULL | | * str = = ' \0 ' ) {
printk ( KERN_INFO PREFIX " _OSI method disabled \n " ) ;
acpi_gbl_create_osi_method = FALSE ;
2007-05-30 02:43:33 +04:00
} else if ( * str = = ' ! ' ) {
if ( acpi_osi_invalidate ( + + str ) = = AE_OK )
printk ( KERN_INFO PREFIX " Deleted _OSI(%s) \n " , str ) ;
} else if ( * osi_additional_string = = ' \0 ' ) {
strncpy ( osi_additional_string , str , OSI_STRING_LENGTH_MAX ) ;
printk ( KERN_INFO PREFIX " Added _OSI(%s) \n " , str ) ;
2005-04-17 02:20:36 +04:00
}
return 1 ;
}
__setup ( " acpi_osi= " , acpi_osi_setup ) ;
/* enable serialization to combat AE_ALREADY_EXISTS errors */
2005-08-05 08:44:28 +04:00
static int __init acpi_serialize_setup ( char * str )
2005-04-17 02:20:36 +04:00
{
printk ( KERN_INFO PREFIX " serialize enabled \n " ) ;
acpi_gbl_all_methods_serialized = TRUE ;
return 1 ;
}
__setup ( " acpi_serialize " , acpi_serialize_setup ) ;
/*
* Wake and Run - Time GPES are expected to be separate .
* We disable wake - GPEs at run - time to prevent spurious
* interrupts .
*
* However , if a system exists that shares Wake and
* Run - time events on the same GPE this flag is available
* to tell Linux to keep the wake - time GPEs enabled at run - time .
*/
2005-08-05 08:44:28 +04:00
static int __init acpi_wake_gpes_always_on_setup ( char * str )
2005-04-17 02:20:36 +04:00
{
printk ( KERN_INFO PREFIX " wake GPEs not disabled \n " ) ;
acpi_gbl_leave_wake_gpes_disabled = FALSE ;
return 1 ;
}
__setup ( " acpi_wake_gpes_always_on " , acpi_wake_gpes_always_on_setup ) ;
/*
* max_cstate is defined in the base kernel so modules can
* change it w / o depending on the state of the processor module .
*/
unsigned int max_cstate = ACPI_PROCESSOR_MAX_POWER ;
EXPORT_SYMBOL ( max_cstate ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
/*
* Acquire a spinlock .
*
* handle is a pointer to the spinlock_t .
*/
2006-06-24 01:04:00 +04:00
acpi_cpu_flags acpi_os_acquire_lock ( acpi_spinlock lockp )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2006-01-28 00:43:00 +03:00
acpi_cpu_flags flags ;
2006-06-24 01:04:00 +04:00
spin_lock_irqsave ( lockp , flags ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
return flags ;
}
/*
* Release a spinlock . See above .
*/
2006-06-24 01:04:00 +04:00
void acpi_os_release_lock ( acpi_spinlock lockp , acpi_cpu_flags flags )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2006-06-24 01:04:00 +04:00
spin_unlock_irqrestore ( lockp , flags ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
}
# ifndef ACPI_USE_LOCAL_CACHE
/*******************************************************************************
*
* FUNCTION : acpi_os_create_cache
*
ACPI: ACPICA 20060421
Removed a device initialization optimization introduced in
20051216 where the _STA method was not run unless an _INI
was also present for the same device. This optimization
could cause problems because it could allow _INI methods
to be run within a not-present device subtree (If a
not-present device had no _INI, _STA would not be run,
the not-present status would not be discovered, and the
children of the device would be incorrectly traversed.)
Implemented a new _STA optimization where namespace
subtrees that do not contain _INI are identified and
ignored during device initialization. Selectively running
_STA can significantly improve boot time on large machines
(with assistance from Len Brown.)
Implemented support for the device initialization case
where the returned _STA flags indicate a device not-present
but functioning. In this case, _INI is not run, but the
device children are examined for presence, as per the
ACPI specification.
Implemented an additional change to the IndexField support
in order to conform to MS behavior. The value written to
the Index Register is not simply a byte offset, it is a
byte offset in units of the access width of the parent
Index Field. (Fiodor Suietov)
Defined and deployed a new OSL interface,
acpi_os_validate_address(). This interface is called during
the creation of all AML operation regions, and allows
the host OS to exert control over what addresses it will
allow the AML code to access. Operation Regions whose
addresses are disallowed will cause a runtime exception
when they are actually accessed (will not affect or abort
table loading.)
Defined and deployed a new OSL interface,
acpi_os_validate_interface(). This interface allows the host OS
to match the various "optional" interface/behavior strings
for the _OSI predefined control method as appropriate
(with assistance from Bjorn Helgaas.)
Restructured and corrected various problems in the
exception handling code paths within DsCallControlMethod
and DsTerminateControlMethod in dsmethod (with assistance
from Takayoshi Kochi.)
Modified the Linux source converter to ignore quoted string
literals while converting identifiers from mixed to lower
case. This will correct problems with the disassembler
and other areas where such strings must not be modified.
The ACPI_FUNCTION_* macros no longer require quotes around
the function name. This allows the Linux source converter
to convert the names, now that the converter ignores
quoted strings.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2006-04-22 01:15:00 +04:00
* PARAMETERS : name - Ascii name for the cache
* size - Size of each cached object
* depth - Maximum depth of the cache ( in objects ) < ignored >
* cache - Where the new cache object is returned
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
*
ACPI: ACPICA 20060421
Removed a device initialization optimization introduced in
20051216 where the _STA method was not run unless an _INI
was also present for the same device. This optimization
could cause problems because it could allow _INI methods
to be run within a not-present device subtree (If a
not-present device had no _INI, _STA would not be run,
the not-present status would not be discovered, and the
children of the device would be incorrectly traversed.)
Implemented a new _STA optimization where namespace
subtrees that do not contain _INI are identified and
ignored during device initialization. Selectively running
_STA can significantly improve boot time on large machines
(with assistance from Len Brown.)
Implemented support for the device initialization case
where the returned _STA flags indicate a device not-present
but functioning. In this case, _INI is not run, but the
device children are examined for presence, as per the
ACPI specification.
Implemented an additional change to the IndexField support
in order to conform to MS behavior. The value written to
the Index Register is not simply a byte offset, it is a
byte offset in units of the access width of the parent
Index Field. (Fiodor Suietov)
Defined and deployed a new OSL interface,
acpi_os_validate_address(). This interface is called during
the creation of all AML operation regions, and allows
the host OS to exert control over what addresses it will
allow the AML code to access. Operation Regions whose
addresses are disallowed will cause a runtime exception
when they are actually accessed (will not affect or abort
table loading.)
Defined and deployed a new OSL interface,
acpi_os_validate_interface(). This interface allows the host OS
to match the various "optional" interface/behavior strings
for the _OSI predefined control method as appropriate
(with assistance from Bjorn Helgaas.)
Restructured and corrected various problems in the
exception handling code paths within DsCallControlMethod
and DsTerminateControlMethod in dsmethod (with assistance
from Takayoshi Kochi.)
Modified the Linux source converter to ignore quoted string
literals while converting identifiers from mixed to lower
case. This will correct problems with the disassembler
and other areas where such strings must not be modified.
The ACPI_FUNCTION_* macros no longer require quotes around
the function name. This allows the Linux source converter
to convert the names, now that the converter ignores
quoted strings.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2006-04-22 01:15:00 +04:00
* RETURN : status
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
*
* DESCRIPTION : Create a cache object
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
acpi_status
2005-08-05 08:44:28 +04:00
acpi_os_create_cache ( char * name , u16 size , u16 depth , acpi_cache_t * * cache )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2005-08-05 08:44:28 +04:00
* cache = kmem_cache_create ( name , size , 0 , 0 , NULL , NULL ) ;
2006-12-19 23:56:13 +03:00
if ( * cache = = NULL )
ACPI: ACPICA 20060421
Removed a device initialization optimization introduced in
20051216 where the _STA method was not run unless an _INI
was also present for the same device. This optimization
could cause problems because it could allow _INI methods
to be run within a not-present device subtree (If a
not-present device had no _INI, _STA would not be run,
the not-present status would not be discovered, and the
children of the device would be incorrectly traversed.)
Implemented a new _STA optimization where namespace
subtrees that do not contain _INI are identified and
ignored during device initialization. Selectively running
_STA can significantly improve boot time on large machines
(with assistance from Len Brown.)
Implemented support for the device initialization case
where the returned _STA flags indicate a device not-present
but functioning. In this case, _INI is not run, but the
device children are examined for presence, as per the
ACPI specification.
Implemented an additional change to the IndexField support
in order to conform to MS behavior. The value written to
the Index Register is not simply a byte offset, it is a
byte offset in units of the access width of the parent
Index Field. (Fiodor Suietov)
Defined and deployed a new OSL interface,
acpi_os_validate_address(). This interface is called during
the creation of all AML operation regions, and allows
the host OS to exert control over what addresses it will
allow the AML code to access. Operation Regions whose
addresses are disallowed will cause a runtime exception
when they are actually accessed (will not affect or abort
table loading.)
Defined and deployed a new OSL interface,
acpi_os_validate_interface(). This interface allows the host OS
to match the various "optional" interface/behavior strings
for the _OSI predefined control method as appropriate
(with assistance from Bjorn Helgaas.)
Restructured and corrected various problems in the
exception handling code paths within DsCallControlMethod
and DsTerminateControlMethod in dsmethod (with assistance
from Takayoshi Kochi.)
Modified the Linux source converter to ignore quoted string
literals while converting identifiers from mixed to lower
case. This will correct problems with the disassembler
and other areas where such strings must not be modified.
The ACPI_FUNCTION_* macros no longer require quotes around
the function name. This allows the Linux source converter
to convert the names, now that the converter ignores
quoted strings.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2006-04-22 01:15:00 +04:00
return AE_ERROR ;
else
return AE_OK ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
}
/*******************************************************************************
*
* FUNCTION : acpi_os_purge_cache
*
* PARAMETERS : Cache - Handle to cache object
*
* RETURN : Status
*
* DESCRIPTION : Free all objects within the requested cache .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_purge_cache ( acpi_cache_t * cache )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2006-10-01 02:28:50 +04:00
kmem_cache_shrink ( cache ) ;
2005-08-05 08:44:28 +04:00
return ( AE_OK ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
}
/*******************************************************************************
*
* FUNCTION : acpi_os_delete_cache
*
* PARAMETERS : Cache - Handle to cache object
*
* RETURN : Status
*
* DESCRIPTION : Free all objects within the requested cache and delete the
* cache object .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_delete_cache ( acpi_cache_t * cache )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2006-09-27 12:49:40 +04:00
kmem_cache_destroy ( cache ) ;
2005-08-05 08:44:28 +04:00
return ( AE_OK ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
}
/*******************************************************************************
*
* FUNCTION : acpi_os_release_object
*
* PARAMETERS : Cache - Handle to cache object
* Object - The object to be released
*
* RETURN : None
*
* DESCRIPTION : Release an object to the specified cache . If cache is full ,
* the object is deleted .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2005-08-05 08:44:28 +04:00
acpi_status acpi_os_release_object ( acpi_cache_t * cache , void * object )
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
{
2005-08-05 08:44:28 +04:00
kmem_cache_free ( cache , object ) ;
return ( AE_OK ) ;
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
}
ACPI: ACPICA 20060421
Removed a device initialization optimization introduced in
20051216 where the _STA method was not run unless an _INI
was also present for the same device. This optimization
could cause problems because it could allow _INI methods
to be run within a not-present device subtree (If a
not-present device had no _INI, _STA would not be run,
the not-present status would not be discovered, and the
children of the device would be incorrectly traversed.)
Implemented a new _STA optimization where namespace
subtrees that do not contain _INI are identified and
ignored during device initialization. Selectively running
_STA can significantly improve boot time on large machines
(with assistance from Len Brown.)
Implemented support for the device initialization case
where the returned _STA flags indicate a device not-present
but functioning. In this case, _INI is not run, but the
device children are examined for presence, as per the
ACPI specification.
Implemented an additional change to the IndexField support
in order to conform to MS behavior. The value written to
the Index Register is not simply a byte offset, it is a
byte offset in units of the access width of the parent
Index Field. (Fiodor Suietov)
Defined and deployed a new OSL interface,
acpi_os_validate_address(). This interface is called during
the creation of all AML operation regions, and allows
the host OS to exert control over what addresses it will
allow the AML code to access. Operation Regions whose
addresses are disallowed will cause a runtime exception
when they are actually accessed (will not affect or abort
table loading.)
Defined and deployed a new OSL interface,
acpi_os_validate_interface(). This interface allows the host OS
to match the various "optional" interface/behavior strings
for the _OSI predefined control method as appropriate
(with assistance from Bjorn Helgaas.)
Restructured and corrected various problems in the
exception handling code paths within DsCallControlMethod
and DsTerminateControlMethod in dsmethod (with assistance
from Takayoshi Kochi.)
Modified the Linux source converter to ignore quoted string
literals while converting identifiers from mixed to lower
case. This will correct problems with the disassembler
and other areas where such strings must not be modified.
The ACPI_FUNCTION_* macros no longer require quotes around
the function name. This allows the Linux source converter
to convert the names, now that the converter ignores
quoted strings.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2006-04-22 01:15:00 +04:00
/******************************************************************************
*
* FUNCTION : acpi_os_validate_interface
*
* PARAMETERS : interface - Requested interface to be validated
*
* RETURN : AE_OK if interface is supported , AE_SUPPORT otherwise
*
* DESCRIPTION : Match an interface string to the interfaces supported by the
* host . Strings originate from an AML call to the _OSI method .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
acpi_status
acpi_os_validate_interface ( char * interface )
{
2007-05-30 02:43:33 +04:00
if ( ! strncmp ( osi_additional_string , interface , OSI_STRING_LENGTH_MAX ) )
return AE_OK ;
return AE_SUPPORT ;
ACPI: ACPICA 20060421
Removed a device initialization optimization introduced in
20051216 where the _STA method was not run unless an _INI
was also present for the same device. This optimization
could cause problems because it could allow _INI methods
to be run within a not-present device subtree (If a
not-present device had no _INI, _STA would not be run,
the not-present status would not be discovered, and the
children of the device would be incorrectly traversed.)
Implemented a new _STA optimization where namespace
subtrees that do not contain _INI are identified and
ignored during device initialization. Selectively running
_STA can significantly improve boot time on large machines
(with assistance from Len Brown.)
Implemented support for the device initialization case
where the returned _STA flags indicate a device not-present
but functioning. In this case, _INI is not run, but the
device children are examined for presence, as per the
ACPI specification.
Implemented an additional change to the IndexField support
in order to conform to MS behavior. The value written to
the Index Register is not simply a byte offset, it is a
byte offset in units of the access width of the parent
Index Field. (Fiodor Suietov)
Defined and deployed a new OSL interface,
acpi_os_validate_address(). This interface is called during
the creation of all AML operation regions, and allows
the host OS to exert control over what addresses it will
allow the AML code to access. Operation Regions whose
addresses are disallowed will cause a runtime exception
when they are actually accessed (will not affect or abort
table loading.)
Defined and deployed a new OSL interface,
acpi_os_validate_interface(). This interface allows the host OS
to match the various "optional" interface/behavior strings
for the _OSI predefined control method as appropriate
(with assistance from Bjorn Helgaas.)
Restructured and corrected various problems in the
exception handling code paths within DsCallControlMethod
and DsTerminateControlMethod in dsmethod (with assistance
from Takayoshi Kochi.)
Modified the Linux source converter to ignore quoted string
literals while converting identifiers from mixed to lower
case. This will correct problems with the disassembler
and other areas where such strings must not be modified.
The ACPI_FUNCTION_* macros no longer require quotes around
the function name. This allows the Linux source converter
to convert the names, now that the converter ignores
quoted strings.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
2006-04-22 01:15:00 +04:00
}
/******************************************************************************
*
* FUNCTION : acpi_os_validate_address
*
* PARAMETERS : space_id - ACPI space ID
* address - Physical address
* length - Address length
*
* RETURN : AE_OK if address / length is valid for the space_id . Otherwise ,
* should return AE_AML_ILLEGAL_ADDRESS .
*
* DESCRIPTION : Validate a system address via the host OS . Used to validate
* the addresses accessed by AML operation regions .
*
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
acpi_status
acpi_os_validate_address (
u8 space_id ,
acpi_physical_address address ,
acpi_size length )
{
return AE_OK ;
}
ACPICA 20050617-0624 from Bob Moore <robert.moore@intel.com>
ACPICA 20050617:
Moved the object cache operations into the OS interface
layer (OSL) to allow the host OS to handle these operations
if desired (for example, the Linux OSL will invoke the
slab allocator). This support is optional; the compile
time define ACPI_USE_LOCAL_CACHE may be used to utilize
the original cache code in the ACPI CA core. The new OSL
interfaces are shown below. See utalloc.c for an example
implementation, and acpiosxf.h for the exact interface
definitions. Thanks to Alexey Starikovskiy.
acpi_os_create_cache
acpi_os_delete_cache
acpi_os_purge_cache
acpi_os_acquire_object
acpi_os_release_object
Modified the interfaces to acpi_os_acquire_lock and
acpi_os_release_lock to return and restore a flags
parameter. This fits better with many OS lock models.
Note: the current execution state (interrupt handler
or not) is no longer passed to these interfaces. If
necessary, the OSL must determine this state by itself, a
simple and fast operation. Thanks to Alexey Starikovskiy.
Fixed a problem in the ACPI table handling where a valid
XSDT was assumed present if the revision of the RSDP
was 2 or greater. According to the ACPI specification,
the XSDT is optional in all cases, and the table manager
therefore now checks for both an RSDP >=2 and a valid
XSDT pointer. Otherwise, the RSDT pointer is used.
Some ACPI 2.0 compliant BIOSs contain only the RSDT.
Fixed an interpreter problem with the Mid() operator in the
case of an input string where the resulting output string
is of zero length. It now correctly returns a valid,
null terminated string object instead of a string object
with a null pointer.
Fixed a problem with the control method argument handling
to allow a store to an Arg object that already contains an
object of type Device. The Device object is now correctly
overwritten. Previously, an error was returned.
ACPICA 20050624:
Modified the new OSL cache interfaces to use ACPI_CACHE_T
as the type for the host-defined cache object. This allows
the OSL implementation to define and type this object in
any manner desired, simplifying the OSL implementation.
For example, ACPI_CACHE_T is defined as kmem_cache_t for
Linux, and should be defined in the OS-specific header
file for other operating systems as required.
Changed the interface to AcpiOsAcquireObject to directly
return the requested object as the function return (instead
of ACPI_STATUS.) This change was made for performance
reasons, since this is the purpose of the interface in the
first place. acpi_os_acquire_object is now similar to the
acpi_os_allocate interface. Thanks to Alexey Starikovskiy.
Modified the initialization sequence in
acpi_initialize_subsystem to call the OSL interface
acpi_osl_initialize first, before any local initialization.
This change was required because the global initialization
now calls OSL interfaces.
Restructured the code base to split some files because
of size and/or because the code logically belonged in a
separate file. New files are listed below.
utilities/utcache.c /* Local cache interfaces */
utilities/utmutex.c /* Local mutex support */
utilities/utstate.c /* State object support */
parser/psloop.c /* Main AML parse loop */
Signed-off-by: Len Brown <len.brown@intel.com>
2005-06-24 08:00:00 +04:00
# endif