diff options
author | Alexey Starikovskiy <alexey.y.starikovskiy@linux.intel.com> | 2007-02-15 21:12:23 (GMT) |
---|---|---|
committer | Len Brown <len.brown@intel.com> | 2007-02-15 21:13:16 (GMT) |
commit | c0d127b56937c3e72c2b1819161d2f6718eee877 (patch) | |
tree | 40ac0fd3a628685ce25d11d1b00360344279ec5b /drivers/acpi/dispatcher | |
parent | 724339d76d9407cd1a8ad32a9c1fdf64840cc51b (diff) | |
download | linux-fsl-qoriq-c0d127b56937c3e72c2b1819161d2f6718eee877.tar.xz |
ACPICA: fix AML mutex re-entrancy
ACPI AML supports "serialized" methods which are protected
by an implicit mutex. The mutex is re-entrant for that AML thread
to allow recursion.
However, Linux implements notify() by creating a new AML thread.
So for systems where notify() re-enters a serialized method,
deadlock results.
The fix is to use the Linux thread_id as the key to allowing
re-entrancy, not the AML thread pointer.
http://bugzilla.kernel.org/show_bug.cgi?id=5534
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Diffstat (limited to 'drivers/acpi/dispatcher')
-rw-r--r-- | drivers/acpi/dispatcher/dsmethod.c | 12 |
1 files changed, 5 insertions, 7 deletions
diff --git a/drivers/acpi/dispatcher/dsmethod.c b/drivers/acpi/dispatcher/dsmethod.c index 1cbe619..1683e5c 100644 --- a/drivers/acpi/dispatcher/dsmethod.c +++ b/drivers/acpi/dispatcher/dsmethod.c @@ -231,10 +231,8 @@ acpi_ds_begin_method_execution(struct acpi_namespace_node *method_node, * Obtain the method mutex if necessary. Do not acquire mutex for a * recursive call. */ - if (!walk_state || - !obj_desc->method.mutex->mutex.owner_thread || - (walk_state->thread != - obj_desc->method.mutex->mutex.owner_thread)) { + if (acpi_os_get_thread_id() != + obj_desc->method.mutex->mutex.owner_thread_id) { /* * Acquire the method mutex. This releases the interpreter if we * block (and reacquires it before it returns) @@ -248,14 +246,14 @@ acpi_ds_begin_method_execution(struct acpi_namespace_node *method_node, } /* Update the mutex and walk info and save the original sync_level */ + obj_desc->method.mutex->mutex.owner_thread_id = + acpi_os_get_thread_id(); if (walk_state) { obj_desc->method.mutex->mutex. original_sync_level = walk_state->thread->current_sync_level; - obj_desc->method.mutex->mutex.owner_thread = - walk_state->thread; walk_state->thread->current_sync_level = obj_desc->method.sync_level; } else { @@ -569,7 +567,7 @@ acpi_ds_terminate_control_method(union acpi_operand_object *method_desc, acpi_os_release_mutex(method_desc->method.mutex->mutex. os_mutex); - method_desc->method.mutex->mutex.owner_thread = NULL; + method_desc->method.mutex->mutex.owner_thread_id = ACPI_MUTEX_NOT_ACQUIRED; } } |