iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

ACPI fixes for v4.7

Browse Source 
- Fix a recent regression in the ACPI EC driver introduced by a fix
    of another problem that uncovered a latent code ordering issue in
    the driver (Lv Zheng).
 
  - Revert a recent ACPICA commit that attempted to address a lock
    ordering issue introduced by a previous fix, but caused Dell
    Precision 5510 to fail to boot, revert that previous fix too and
    finally revert the commit that caused the original problem (a
    deadlock in the ACPICA code) to happen (Rafael Wysocki).
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.22 (GNU/Linux)
 
 iQIcBAABCAAGBQJXhVOqAAoJEILEb/54YlRxWXsP/2W3HYWabMnCWVobJYyxP0TO
 mCmyZZmPpGqzPi9HV9VLTcKPZrhrLzFwk3LnDkxWWH6aEfDz5gTjrME4DNJdhTsu
 SqNXf77bDdCYX6lgbdvB/ZsZkStV8b5k8gldXaBJHs7MNi33+/xepqXgoUcawB0n
 H/M5udwiDMQyh8m78pkOksuicD+phO+QfdpxTW/qvmFfu7IlqeFlTCepYUyEXu5l
 XwNoaptrWfsEDprg04z4jeB+xta92eUQ+BUzqpQcGAfmXYVUm7XPjhE0wepqgtta
 sfxkdtM66uNZw/r3XbErVPwM/8Fl7Rywdi8qIR5KaQol+icgM62e+WWL1+XMiIiw
 IZSk7XYF7QQVgwMSzPXhBd6HOzYgAgYH5UITtWlQKyi15C1R2ULSW91g92xcjSFP
 Zyb6sN2/VWbN30WVsaZJ1GEGTFUOSE/iE2/R3rnRdgmb5nh4b5d83HmAXf90uBSX
 2U008aqW1mQoRQrVupk/j4Lf6YKZjAGlwb0pmfguI6YROzgCcuaopTEI69bUpgfj
 4LfbNi/b6tOtNrCAsJSvS8A+vYmRuqk5HpJNuECefJbJgSy2xCBWBFT9Il50o/6V
 zl6IOE6WS/Rz77l0tAkzsiQhq4vYWTbQsrtQp7op98R1XHioLABp4fA00kkQpzbw
 SQcHzGFA7sMoWOcuKDbi
 =hkH2
 -----END PGP SIGNATURE-----

Merge tag 'acpi-urgent-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull ACPI fixes from Rafael Wysocki:
 "One ACPI EC driver regression fix (code ordering) and three reverts of
  ACPICA commits, one that introduced a problem and two unsuccessful
  attempted fixes on top of it.

  Specifics:

   - Fix a recent regression in the ACPI EC driver introduced by a fix
     of another problem that uncovered a latent code ordering issue in
     the driver (Lv Zheng).

   - Revert a recent ACPICA commit that attempted to address a lock
     ordering issue introduced by a previous fix, but caused Dell
     Precision 5510 to fail to boot, revert that previous fix too and
     finally revert the commit that caused the original problem (a
     deadlock in the ACPICA code) to happen (Rafael Wysocki)"

* tag 'acpi-urgent-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  Revert "ACPI 2.0 / AML: Improve module level execution by moving the If/Else/While execution to per-table basis"
  Revert "ACPICA: Namespace: Fix deadlock triggered by MLC support in dynamic table loading"
  Revert "ACPICA: Namespace: Fix namespace/interpreter lock ordering"
  ACPI / EC: Fix code ordering issue in ec_remove_handlers()
This commit is contained in:
Linus Torvalds 2016-07-13 06:37:03 +09:00
commit f1b5e4fac1
4 changed files with 15 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
drivers/acpi/acpica/exconfig.c
View File

@ -108,9 +108,7 @@ acpi_ex_add_table(u32 table_index,
/* Add the table to the namespace */
acpi_ex_exit_interpreter();
status = acpi_ns_load_table(table_index, parent_node);
acpi_ex_enter_interpreter();
if (ACPI_FAILURE(status)) {
acpi_ut_remove_reference(obj_desc);
*ddb_handle = NULL;

7
drivers/acpi/acpica/nsload.c
View File

@ -46,7 +46,6 @@
#include "acnamesp.h"
#include "acdispat.h"
#include "actables.h"
#include "acinterp.h"
#define _COMPONENT ACPI_NAMESPACE
ACPI_MODULE_NAME("nsload")
@ -79,8 +78,6 @@ acpi_ns_load_table(u32 table_index, struct acpi_namespace_node *node)
ACPI_FUNCTION_TRACE(ns_load_table);
acpi_ex_enter_interpreter();
/*
* Parse the table and load the namespace with all named
* objects found within. Control methods are NOT parsed
@ -92,7 +89,7 @@ acpi_ns_load_table(u32 table_index, struct acpi_namespace_node *node)
*/
status = acpi_ut_acquire_mutex(ACPI_MTX_NAMESPACE);
if (ACPI_FAILURE(status)) {
goto unlock_interp;
return_ACPI_STATUS(status);
}
/* If table already loaded into namespace, just return */
@ -133,8 +130,6 @@ acpi_ns_load_table(u32 table_index, struct acpi_namespace_node *node)
unlock:
(void)acpi_ut_release_mutex(ACPI_MTX_NAMESPACE);
unlock_interp:
(void)acpi_ex_exit_interpreter();
if (ACPI_FAILURE(status)) {
return_ACPI_STATUS(status);

15
drivers/acpi/ec.c
View File

@ -1331,8 +1331,6 @@ static int ec_install_handlers(struct acpi_ec *ec)
static void ec_remove_handlers(struct acpi_ec *ec)
{
acpi_ec_stop(ec, false);
if (test_bit(EC_FLAGS_EC_HANDLER_INSTALLED, &ec->flags)) {
if (ACPI_FAILURE(acpi_remove_address_space_handler(ec->handle,
ACPI_ADR_SPACE_EC, &acpi_ec_space_handler)))
@ -1340,6 +1338,19 @@ static void ec_remove_handlers(struct acpi_ec *ec)
clear_bit(EC_FLAGS_EC_HANDLER_INSTALLED, &ec->flags);
}
/*
* Stops handling the EC transactions after removing the operation
* region handler. This is required because _REG(DISCONNECT)
* invoked during the removal can result in new EC transactions.
*
* Flushes the EC requests and thus disables the GPE before
* removing the GPE handler. This is required by the current ACPICA
* GPE core. ACPICA GPE core will automatically disable a GPE when
* it is indicated but there is no way to handle it. So the drivers
* must disable the GPEs prior to removing the GPE handlers.
*/
acpi_ec_stop(ec, false);
if (test_bit(EC_FLAGS_GPE_HANDLER_INSTALLED, &ec->flags)) {
if (ACPI_FAILURE(acpi_remove_gpe_handler(NULL, ec->gpe,
&acpi_ec_gpe_handler)))

2
include/acpi/acpixf.h
View File

@ -192,7 +192,7 @@ ACPI_INIT_GLOBAL(u8, acpi_gbl_do_not_use_xsdt, FALSE);
/*
* Optionally support group module level code.
*/
ACPI_INIT_GLOBAL(u8, acpi_gbl_group_module_level_code, FALSE);
ACPI_INIT_GLOBAL(u8, acpi_gbl_group_module_level_code, TRUE);
/*
* Optionally use 32-bit FADT addresses if and when there is a conflict