Linux ACPI
 help / color / mirror / Atom feed
From: Lv Zheng <lv.zheng@intel.com>
To: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
	Len Brown <len.brown@intel.com>
Cc: Lv Zheng <lv.zheng@intel.com>, Lv Zheng <zetalog@gmail.com>,
	linux-kernel@vger.kernel.org, linux-acpi@vger.kernel.org,
	Bob Moore <robert.moore@intel.com>
Subject: [PATCH 38/42] ACPICA: Events: Enhance acpi_ev_execute_reg_method() to ensure no _REG evaluations can happen during OS early boot stages
Date: Tue, 29 Dec 2015 14:03:08 +0800	[thread overview]
Message-ID: <ff7df012fbc0628131b14aa7368dc428ca6dbdbe.1451367389.git.lv.zheng@intel.com> (raw)
In-Reply-To: <cover.1451367388.git.lv.zheng@intel.com>

ACPICA commit 31178590dde82368fdb0f6b0e466b6c0add96c57

We can ensure no early _REG evaluations by ensuring the following rules in
acpi_ev_execute_reg_method():
1. If an address space handler is installed during early stage,
   _REG(CONNECT) evaluations are blocked. This is achieved using
   acpi_gbl_reg_methods_enabled which is renamed from
   acpi_gbl_reg_methods_executed.
2. If _REG(CONNECT) has never been evalauted for the region object,
   _REG(DISCONNECT) evaluations are blocked. This is achieved by a new
   region object flag: AOPOBJ_REG_CONNECTED.
Note that, after applying this patch, we can ensure _REG(DISCONNECT) is
always paired to _REG(CONNECT). Lv Zheng

Link: https://github.com/acpica/acpica/commit/31178590
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
---
 drivers/acpi/acpica/acglobal.h |    2 +-
 drivers/acpi/acpica/acobject.h |    7 ++++---
 drivers/acpi/acpica/evregion.c |   25 ++++++++++++++++++++++---
 drivers/acpi/acpica/evxfregn.c |   33 ---------------------------------
 drivers/acpi/acpica/excreate.c |    3 ++-
 drivers/acpi/acpica/utinit.c   |    1 -
 drivers/acpi/acpica/utxfinit.c |    1 +
 7 files changed, 30 insertions(+), 42 deletions(-)

diff --git a/drivers/acpi/acpica/acglobal.h b/drivers/acpi/acpica/acglobal.h
index ef0abf4..73462ca 100644
--- a/drivers/acpi/acpica/acglobal.h
+++ b/drivers/acpi/acpica/acglobal.h
@@ -165,7 +165,7 @@ ACPI_GLOBAL(u8, acpi_gbl_next_owner_id_offset);
 
 /* Initialization sequencing */
 
-ACPI_GLOBAL(u8, acpi_gbl_reg_methods_executed);
+ACPI_INIT_GLOBAL(u8, acpi_gbl_reg_methods_enabled, FALSE);
 
 /* Misc */
 
diff --git a/drivers/acpi/acpica/acobject.h b/drivers/acpi/acpica/acobject.h
index 0bd02c4..2b154cf 100644
--- a/drivers/acpi/acpica/acobject.h
+++ b/drivers/acpi/acpica/acobject.h
@@ -93,9 +93,10 @@
 #define AOPOBJ_AML_CONSTANT         0x01	/* Integer is an AML constant */
 #define AOPOBJ_STATIC_POINTER       0x02	/* Data is part of an ACPI table, don't delete */
 #define AOPOBJ_DATA_VALID           0x04	/* Object is initialized and data is valid */
-#define AOPOBJ_OBJECT_INITIALIZED   0x08	/* Region is initialized, _REG was run */
-#define AOPOBJ_SETUP_COMPLETE       0x10	/* Region setup is complete */
-#define AOPOBJ_INVALID              0x20	/* Host OS won't allow a Region address */
+#define AOPOBJ_OBJECT_INITIALIZED   0x08	/* Region is initialized */
+#define AOPOBJ_REG_CONNECTED        0x10	/* _REG was run */
+#define AOPOBJ_SETUP_COMPLETE       0x20	/* Region setup is complete */
+#define AOPOBJ_INVALID              0x40	/* Host OS won't allow a Region address */
 
 /******************************************************************************
  *
diff --git a/drivers/acpi/acpica/evregion.c b/drivers/acpi/acpica/evregion.c
index 432b810..cf6e878 100644
--- a/drivers/acpi/acpica/evregion.c
+++ b/drivers/acpi/acpica/evregion.c
@@ -104,8 +104,6 @@ acpi_status acpi_ev_initialize_op_regions(void)
 		}
 	}
 
-	acpi_gbl_reg_methods_executed = TRUE;
-
 	(void)acpi_ut_release_mutex(ACPI_MTX_NAMESPACE);
 	return_ACPI_STATUS(status);
 }
@@ -601,7 +599,18 @@ acpi_ev_execute_reg_method(union acpi_operand_object *region_obj, u32 function)
 		return_ACPI_STATUS(AE_NOT_EXIST);
 	}
 
-	if (region_obj2->extra.method_REG == NULL) {
+	if (region_obj2->extra.method_REG == NULL ||
+	    region_obj->region.handler == NULL ||
+	    !acpi_gbl_reg_methods_enabled) {
+		return_ACPI_STATUS(AE_OK);
+	}
+
+	/* _REG(DISCONNECT) should be paired with _REG(CONNECT) */
+
+	if ((function == ACPI_REG_CONNECT &&
+	     region_obj->common.flags & AOPOBJ_REG_CONNECTED) ||
+	    (function == ACPI_REG_DISCONNECT &&
+	     !(region_obj->common.flags & AOPOBJ_REG_CONNECTED))) {
 		return_ACPI_STATUS(AE_OK);
 	}
 
@@ -650,6 +659,16 @@ acpi_ev_execute_reg_method(union acpi_operand_object *region_obj, u32 function)
 	status = acpi_ns_evaluate(info);
 	acpi_ut_remove_reference(args[1]);
 
+	if (ACPI_FAILURE(status)) {
+		goto cleanup2;
+	}
+
+	if (function == ACPI_REG_CONNECT) {
+		region_obj->common.flags |= AOPOBJ_REG_CONNECTED;
+	} else {
+		region_obj->common.flags &= ~AOPOBJ_REG_CONNECTED;
+	}
+
 cleanup2:
 	acpi_ut_remove_reference(args[0]);
 
diff --git a/drivers/acpi/acpica/evxfregn.c b/drivers/acpi/acpica/evxfregn.c
index 3c90508..29f9f39 100644
--- a/drivers/acpi/acpica/evxfregn.c
+++ b/drivers/acpi/acpica/evxfregn.c
@@ -112,39 +112,6 @@ acpi_install_address_space_handler(acpi_handle device,
 		goto unlock_and_exit;
 	}
 
-	/*
-	 * For the default space_IDs, (the IDs for which there are default region handlers
-	 * installed) Only execute the _REG methods if the global initialization _REG
-	 * methods have already been run (via acpi_initialize_objects). In other words,
-	 * we will defer the execution of the _REG methods for these space_IDs until
-	 * execution of acpi_initialize_objects. This is done because we need the handlers
-	 * for the default spaces (mem/io/pci/table) to be installed before we can run
-	 * any control methods (or _REG methods). There is known BIOS code that depends
-	 * on this.
-	 *
-	 * For all other space_IDs, we can safely execute the _REG methods immediately.
-	 * This means that for IDs like embedded_controller, this function should be called
-	 * only after acpi_enable_subsystem has been called.
-	 */
-	switch (space_id) {
-	case ACPI_ADR_SPACE_SYSTEM_MEMORY:
-	case ACPI_ADR_SPACE_SYSTEM_IO:
-	case ACPI_ADR_SPACE_PCI_CONFIG:
-	case ACPI_ADR_SPACE_DATA_TABLE:
-
-		if (!acpi_gbl_reg_methods_executed) {
-
-			/* We will defer execution of the _REG methods for this space */
-
-			goto unlock_and_exit;
-		}
-		break;
-
-	default:
-
-		break;
-	}
-
 	/* Run all _REG methods for this address space */
 
 	status = acpi_ev_execute_reg_methods(node, space_id);
diff --git a/drivers/acpi/acpica/excreate.c b/drivers/acpi/acpica/excreate.c
index 6a1396e..46be5a2 100644
--- a/drivers/acpi/acpica/excreate.c
+++ b/drivers/acpi/acpica/excreate.c
@@ -345,7 +345,8 @@ acpi_ex_create_region(u8 * aml_start,
 	obj_desc->region.node = node;
 	obj_desc->region.handler = NULL;
 	obj_desc->common.flags &=
-	    ~(AOPOBJ_SETUP_COMPLETE | AOPOBJ_OBJECT_INITIALIZED);
+	    ~(AOPOBJ_SETUP_COMPLETE | AOPOBJ_REG_CONNECTED |
+	      AOPOBJ_OBJECT_INITIALIZED);
 
 	/* Install the new region object in the parent Node */
 
diff --git a/drivers/acpi/acpica/utinit.c b/drivers/acpi/acpica/utinit.c
index ccd0745..fd82a12 100644
--- a/drivers/acpi/acpica/utinit.c
+++ b/drivers/acpi/acpica/utinit.c
@@ -206,7 +206,6 @@ acpi_status acpi_ut_init_globals(void)
 	acpi_gbl_next_owner_id_offset = 0;
 	acpi_gbl_debugger_configuration = DEBUGGER_THREADING;
 	acpi_gbl_osi_mutex = NULL;
-	acpi_gbl_reg_methods_executed = FALSE;
 	acpi_gbl_max_loop_iterations = 0xFFFF;
 
 	/* Hardware oriented */
diff --git a/drivers/acpi/acpica/utxfinit.c b/drivers/acpi/acpica/utxfinit.c
index 7d8eb60..1c7ed49 100644
--- a/drivers/acpi/acpica/utxfinit.c
+++ b/drivers/acpi/acpica/utxfinit.c
@@ -267,6 +267,7 @@ acpi_status __init acpi_initialize_objects(u32 flags)
 	 * initialized, even if they contain executable AML (see the call to
 	 * acpi_ns_initialize_objects below).
 	 */
+	acpi_gbl_reg_methods_enabled = TRUE;
 	if (!(flags & ACPI_NO_ADDRESS_SPACE_INIT)) {
 		ACPI_DEBUG_PRINT((ACPI_DB_EXEC,
 				  "[Init] Executing _REG OpRegion methods\n"));
-- 
1.7.10

  parent reply	other threads:[~2015-12-29  6:03 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-29  5:52 [PATCH 00/42] ACPICA: 20151218 Release Lv Zheng
2015-12-29  5:52 ` [PATCH 01/42] ACPICA: Linuxize: reduce divergences for 20151218 release Lv Zheng
2015-12-29  5:52 ` [PATCH 02/42] ACPICA: acpi_get_sleep_type_data: Reduce warnings Lv Zheng
2015-12-29  5:52 ` [PATCH 03/42] ACPICA: Namespace: Fix wrong error log Lv Zheng
2015-12-29  5:53 ` [PATCH 04/42] ACPICA: Debugger: reduce old external path format Lv Zheng
2015-12-29  5:53 ` [PATCH 05/42] ACPICA: Fix SyncLevel support interaction with method auto-serialization Lv Zheng
2015-12-29  5:54 ` [PATCH 06/42] ACPICA: Add "const" to some functions that return fixed strings Lv Zheng
2015-12-29  5:54 ` [PATCH 07/42] ACPICA: exmutex: General cleanup, restructured some code Lv Zheng
2015-12-29  5:54 ` [PATCH 08/42] ACPICA: Core: Major update for code formatting, no functional changes Lv Zheng
2015-12-29  5:54 ` [PATCH 09/42] ACPICA: Split interpreter tracing functions to a new file Lv Zheng
2015-12-29  5:54 ` [PATCH 10/42] ACPICA: acpiexec: Add support for AML files containing multiple tables Lv Zheng
2015-12-29  5:54 ` [PATCH 11/42] ACPICA: Disassembler/tools: Support for multiple ACPI tables in one file Lv Zheng
2015-12-29  5:55 ` [PATCH 12/42] ACPICA: iasl/acpiexec: Update input file handling and verification Lv Zheng
2015-12-29  5:55 ` [PATCH 13/42] ACPICA: Debugger: Remove some unecessary NULL checks Lv Zheng
2015-12-29  5:55 ` [PATCH 14/42] ACPICA: Revert "acpi_get_object_info: Add support for ACPI 5.0 _SUB method." Lv Zheng
2015-12-29  5:56 ` [PATCH 15/42] ACPICA: Add comment explaining _SUB removal Lv Zheng
2015-12-29  5:56 ` [PATCH 16/42] ACPICA: acpiexec/acpinames: Update for error checking macros Lv Zheng
2015-12-29  5:56 ` [PATCH 17/42] ACPICA: Concatenate operator: Add extensions to support all ACPI objects Lv Zheng
2015-12-29  5:57 ` [PATCH 18/42] ACPICA: Debug Object: Cleanup output Lv Zheng
2015-12-29  5:57 ` [PATCH 19/42] ACPICA: Debug object: Fix output for a NULL object Lv Zheng
2015-12-29  5:57 ` [PATCH 20/42] ACPICA: Update for output of the Debug Object Lv Zheng
2015-12-29  5:57 ` [PATCH 21/42] ACPICA: Namespace: Add scope information to the simple object repair mechanism Lv Zheng
2015-12-29  5:58 ` [PATCH 22/42] ACPICA: Namespace: Add String -> ObjectReference conversion support Lv Zheng
2015-12-29  5:58 ` [PATCH 23/42] ACPICA: getopt: Comment update, no functional change Lv Zheng
2015-12-29  5:59 ` [PATCH 24/42] ACPICA: Tools: Add spacing and missing options in acpibin tool Lv Zheng
2015-12-29  5:59 ` [PATCH 25/42] ACPICA: Add new exception code, AE_IO_ERROR Lv Zheng
2015-12-29  6:00 ` [PATCH 26/42] ACPICA: iasl/Disassembler: Support ASL ElseIf operator Lv Zheng
2015-12-29  6:00 ` [PATCH 27/42] ACPICA: Parser: Add constants for internal namepath function Lv Zheng
2015-12-29  6:00 ` [PATCH 28/42] ACPICA: Parser: Fix for SuperName method invocation Lv Zheng
2015-12-29  6:00 ` [PATCH 29/42] ACPICA: Update parameter type for ObjectType operator Lv Zheng
2015-12-29  6:00 ` [PATCH 30/42] ACPICA: Update internal #defines for ObjectType operator. No functional change Lv Zheng
2015-12-29  6:01 ` [PATCH 31/42] ACPICA: Update for CondRefOf and RefOf operators Lv Zheng
2015-12-29  6:01 ` [PATCH 32/42] ACPICA: Cleanup code related to the per-table module level improvement Lv Zheng
2015-12-29  6:02 ` [PATCH 33/42] ACPICA: Events: Deploys acpi_ev_find_region_handler() Lv Zheng
2015-12-29  6:02 ` [PATCH 34/42] ACPICA: Events: Uses common_notify for address space handlers Lv Zheng
2015-12-29  6:02 ` [PATCH 35/42] ACPICA: Utilities: Reorder initialization code Lv Zheng
2015-12-29  6:02 ` [PATCH 36/42] ACPICA: Events: Fix an issue that region object is re-attached to another scope when it is already attached Lv Zheng
2015-12-29  6:02 ` [PATCH 37/42] ACPICA: Events: Split acpi_ev_associate_reg_method() from region initialization code Lv Zheng
2015-12-29  6:03 ` Lv Zheng [this message]
2015-12-29  6:03 ` [PATCH 39/42] ACPICA: Events: Introduce ACPI_REG_DISCONNECT invocation to acpi_ev_execute_reg_methods() Lv Zheng
2015-12-29  6:04 ` [PATCH 40/42] ACPICA: Add "root node" case to the ACPI name repair code Lv Zheng
2015-12-29  6:04 ` [PATCH 41/42] ACPICA: Add per-table execution of module-level code Lv Zheng
2015-12-29  6:04 ` [PATCH 42/42] ACPICA: Update version to 20151218 Lv Zheng
2016-01-01  3:04 ` [PATCH 00/42] ACPICA: 20151218 Release Rafael J. Wysocki
2016-01-01  3:24   ` Rafael J. Wysocki
2016-01-04  2:29   ` Zheng, Lv
2016-01-03  0:49 ` Rafael J. Wysocki
2016-01-03  6:45 ` David Lang
2016-01-04  2:27   ` Zheng, Lv
2016-01-04 13:16     ` Rafael J. Wysocki

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ff7df012fbc0628131b14aa7368dc428ca6dbdbe.1451367389.git.lv.zheng@intel.com \
    --to=lv.zheng@intel.com \
    --cc=len.brown@intel.com \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rafael.j.wysocki@intel.com \
    --cc=robert.moore@intel.com \
    --cc=zetalog@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox