public inbox for linux-acpi@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] ACPI: make some IO ports off-limits to AML
@ 2009-02-04  3:59 Len Brown
  2009-02-04  6:15 ` Matthew Garrett
  0 siblings, 1 reply; 3+ messages in thread
From: Len Brown @ 2009-02-04  3:59 UTC (permalink / raw)
  To: linux-acpi; +Cc: Linux Kernel Mailing List

From: Len Brown <len.brown@intel.com>

ACPICA exports acpi_os_validate_address() so the OS
can prevent BIOS AML from accessing specified addresses.

Start using this interface to prevent AML from accessing
some well known IO addresses that the OS "owns".

Signed-off-by: Len Brown <len.brown@intel.com>
---
 drivers/acpi/osl.c |   50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 50 insertions(+), 0 deletions(-)

diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index 6729a49..4fb01b0 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -1317,6 +1317,54 @@ acpi_os_validate_interface (char *interface)
 	return AE_SUPPORT;
 }
 
+#ifdef	CONFIG_X86
+
+struct aml_port_desc {
+	uint	start;
+	uint	end;
+	char*   name;
+	char	warned;
+};
+
+static struct aml_port_desc aml_invalid_port_list[] = {
+	{0x20, 0x21, "PIC0", 0},
+	{0xA0, 0xA1, "PIC1", 0},
+	{0x4D0, 0x4D1, "ELCR", 0}
+};
+
+/*
+ * valid_aml_io_address()
+ *
+ * if valid, return true
+ * else invalid, warn once, return false
+ */
+static bool valid_aml_io_address(uint address, uint length)
+{
+	int i;
+	int entries = sizeof(aml_invalid_port_list) / sizeof(struct aml_port_desc);
+
+	for (i = 0; i < entries; ++i) {
+		if ((address >= aml_invalid_port_list[i].start &&
+			address <= aml_invalid_port_list[i].end) ||
+			(address + length >= aml_invalid_port_list[i].start &&
+			address  + length <= aml_invalid_port_list[i].end))
+		{
+			if (!aml_invalid_port_list[i].warned)
+			{
+				printk(KERN_ERR "ACPI: Denied BIOS AML access"
+					" to invalid port 0x%x+0x%x (%s)\n",
+					address, length,
+					aml_invalid_port_list[i].name);
+				aml_invalid_port_list[i].warned = 1;
+			}
+			return false;	/* invalid */
+		}
+	}
+	return true;	/* valid */
+}
+#else
+static inline bool valid_aml_io_address(uint address, uint length) { return true; }
+#endif
 /******************************************************************************
  *
  * FUNCTION:    acpi_os_validate_address
@@ -1346,6 +1394,8 @@ acpi_os_validate_address (
 
 	switch (space_id) {
 	case ACPI_ADR_SPACE_SYSTEM_IO:
+		if (!valid_aml_io_address(address, length))
+			return AE_AML_ILLEGAL_ADDRESS;
 	case ACPI_ADR_SPACE_SYSTEM_MEMORY:
 		/* Only interference checks against SystemIO and SytemMemory
 		   are needed */
-- 
1.6.1.2.319.gbd9e


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] ACPI: make some IO ports off-limits to AML
  2009-02-04  3:59 [PATCH] ACPI: make some IO ports off-limits to AML Len Brown
@ 2009-02-04  6:15 ` Matthew Garrett
  2009-03-28  1:30   ` Len Brown
  0 siblings, 1 reply; 3+ messages in thread
From: Matthew Garrett @ 2009-02-04  6:15 UTC (permalink / raw)
  To: Len Brown; +Cc: linux-acpi, Linux Kernel Mailing List

On Tue, Feb 03, 2009 at 10:59:43PM -0500, Len Brown wrote:

> Start using this interface to prevent AML from accessing
> some well known IO addresses that the OS "owns".

It might be nice to extend this to cover the other ports listed at 
http://www.microsoft.com/whdc/archive/BIOSAML.mspx if (and only if) the 
firmware requested _OSI("Windows 2001") or later.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] ACPI: make some IO ports off-limits to AML
  2009-02-04  6:15 ` Matthew Garrett
@ 2009-03-28  1:30   ` Len Brown
  0 siblings, 0 replies; 3+ messages in thread
From: Len Brown @ 2009-03-28  1:30 UTC (permalink / raw)
  To: Matthew Garrett; +Cc: linux-acpi, Linux Kernel Mailing List

On Wed, 4 Feb 2009, Matthew Garrett wrote:

> On Tue, Feb 03, 2009 at 10:59:43PM -0500, Len Brown wrote:
> 
> > Start using this interface to prevent AML from accessing
> > some well known IO addresses that the OS "owns".
> 
> It might be nice to extend this to cover the other ports listed at 
> http://www.microsoft.com/whdc/archive/BIOSAML.mspx if (and only if) the 
> firmware requested _OSI("Windows 2001") or later.

Right.  We reverted this initialization-time check for 2.6.29
b/c it had false positives of opregions that were never used.

For 2.6.30 we replaced it with a run-time check, and the check
follows the compatibility rules referenced above.

http://git.kernel.org/?p=linux/kernel/git/lenb/linux-acpi-2.6.git;a=commitdiff;h=7f0719039085cc40114abce84cf29fe57da226f4

cheers,
Len Brown, Intel Open Source Technology Center


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-03-28  1:31 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-02-04  3:59 [PATCH] ACPI: make some IO ports off-limits to AML Len Brown
2009-02-04  6:15 ` Matthew Garrett
2009-03-28  1:30   ` Len Brown

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox