diff options
author | Bob Moore <robert.moore@intel.com> | 2006-06-23 17:04:00 -0400 |
---|---|---|
committer | Len Brown <len.brown@intel.com> | 2006-06-28 03:11:38 -0400 |
commit | 967440e3be1af06ad4dc7bb18d2e3c16130fe067 (patch) | |
tree | c9bbf70475333f4f169838ed88233f8410010677 /drivers/acpi/hardware | |
parent | 95b38b3f453c16de0f8cddcde3e71050bbfb37b9 (diff) | |
download | op-kernel-dev-967440e3be1af06ad4dc7bb18d2e3c16130fe067.zip op-kernel-dev-967440e3be1af06ad4dc7bb18d2e3c16130fe067.tar.gz |
ACPI: ACPICA 20060623
Implemented a new acpi_spinlock type for the OSL lock
interfaces. This allows the type to be customized to
the host OS for improved efficiency (since a spinlock is
usually a very small object.)
Implemented support for "ignored" bits in the ACPI
registers. According to the ACPI specification, these
bits should be preserved when writing the registers via
a read/modify/write cycle. There are 3 bits preserved
in this manner: PM1_CONTROL[0] (SCI_EN), PM1_CONTROL[9],
and PM1_STATUS[11].
http://bugzilla.kernel.org/show_bug.cgi?id=3691
Implemented the initial deployment of new OSL mutex
interfaces. Since some host operating systems have
separate mutex and semaphore objects, this feature was
requested. The base code now uses mutexes (and the new
mutex interfaces) wherever a binary semaphore was used
previously. However, for the current release, the mutex
interfaces are defined as macros to map them to the
existing semaphore interfaces.
Fixed several problems with the support for the control
method SyncLevel parameter. The SyncLevel now works
according to the ACPI specification and in concert with the
Mutex SyncLevel parameter, since the current SyncLevel is
a property of the executing thread. Mutual exclusion for
control methods is now implemented with a mutex instead
of a semaphore.
Fixed three instances of the use of the C shift operator
in the bitfield support code (exfldio.c) to avoid the use
of a shift value larger than the target data width. The
behavior of C compilers is undefined in this case and can
cause unpredictable results, and therefore the case must
be detected and avoided. (Fiodor Suietov)
Added an info message whenever an SSDT or OEM table
is loaded dynamically via the Load() or LoadTable()
ASL operators. This should improve debugging capability
since it will show exactly what tables have been loaded
(beyond the tables present in the RSDT/XSDT.)
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Diffstat (limited to 'drivers/acpi/hardware')
-rw-r--r-- | drivers/acpi/hardware/hwregs.c | 77 |
1 files changed, 62 insertions, 15 deletions
diff --git a/drivers/acpi/hardware/hwregs.c b/drivers/acpi/hardware/hwregs.c index ae142de..3143f36 100644 --- a/drivers/acpi/hardware/hwregs.c +++ b/drivers/acpi/hardware/hwregs.c @@ -172,9 +172,9 @@ acpi_get_sleep_type_data(u8 sleep_state, u8 * sleep_type_a, u8 * sleep_type_b) } /* - * The package must have at least two elements. NOTE (March 2005): This + * The package must have at least two elements. NOTE (March 2005): This * goes against the current ACPI spec which defines this object as a - * package with one encoded DWORD element. However, existing practice + * package with one encoded DWORD element. However, existing practice * by BIOS vendors seems to be to have 2 or more elements, at least * one per sleep type (A/B). */ @@ -255,7 +255,7 @@ struct acpi_bit_register_info *acpi_hw_get_bit_register_info(u32 register_id) * return_value - Value that was read from the register * Flags - Lock the hardware or not * - * RETURN: Status and the value read from specified Register. Value + * RETURN: Status and the value read from specified Register. Value * returned is normalized to bit0 (is shifted all the way right) * * DESCRIPTION: ACPI bit_register read function. @@ -361,8 +361,8 @@ acpi_status acpi_set_register(u32 register_id, u32 value, u32 flags) case ACPI_REGISTER_PM1_STATUS: /* - * Status Registers are different from the rest. Clear by - * writing 1, and writing 0 has no effect. So, the only relevant + * Status Registers are different from the rest. Clear by + * writing 1, and writing 0 has no effect. So, the only relevant * information is the single bit we're interested in, all others should * be written as 0 so they will be left unchanged. */ @@ -467,14 +467,13 @@ ACPI_EXPORT_SYMBOL(acpi_set_register) * * FUNCTION: acpi_hw_register_read * - * PARAMETERS: use_lock - Mutex hw access - * register_id - register_iD + Offset + * PARAMETERS: use_lock - Lock hardware? True/False + * register_id - ACPI Register ID * return_value - Where the register value is returned * * RETURN: Status and the value read. * - * DESCRIPTION: Acpi register read function. Registers are read at the - * given offset. + * DESCRIPTION: Read from the specified ACPI register * ******************************************************************************/ acpi_status @@ -580,14 +579,26 @@ acpi_hw_register_read(u8 use_lock, u32 register_id, u32 * return_value) * * FUNCTION: acpi_hw_register_write * - * PARAMETERS: use_lock - Mutex hw access - * register_id - register_iD + Offset + * PARAMETERS: use_lock - Lock hardware? True/False + * register_id - ACPI Register ID * Value - The value to write * * RETURN: Status * - * DESCRIPTION: Acpi register Write function. Registers are written at the - * given offset. + * DESCRIPTION: Write to the specified ACPI register + * + * NOTE: In accordance with the ACPI specification, this function automatically + * preserves the value of the following bits, meaning that these bits cannot be + * changed via this interface: + * + * PM1_CONTROL[0] = SCI_EN + * PM1_CONTROL[9] + * PM1_STATUS[11] + * + * ACPI References: + * 1) Hardware Ignored Bits: When software writes to a register with ignored + * bit fields, it preserves the ignored bit fields + * 2) SCI_EN: OSPM always preserves this bit position * ******************************************************************************/ @@ -595,6 +606,7 @@ acpi_status acpi_hw_register_write(u8 use_lock, u32 register_id, u32 value) { acpi_status status; acpi_cpu_flags lock_flags = 0; + u32 read_value; ACPI_FUNCTION_TRACE(hw_register_write); @@ -605,6 +617,22 @@ acpi_status acpi_hw_register_write(u8 use_lock, u32 register_id, u32 value) switch (register_id) { case ACPI_REGISTER_PM1_STATUS: /* 16-bit access */ + /* Perform a read first to preserve certain bits (per ACPI spec) */ + + status = acpi_hw_register_read(ACPI_MTX_DO_NOT_LOCK, + ACPI_REGISTER_PM1_STATUS, + &read_value); + if (ACPI_FAILURE(status)) { + goto unlock_and_exit; + } + + /* Insert the bits to be preserved */ + + ACPI_INSERT_BITS(value, ACPI_PM1_STATUS_PRESERVED_BITS, + read_value); + + /* Now we can write the data */ + status = acpi_hw_low_level_write(16, value, &acpi_gbl_FADT->xpm1a_evt_blk); @@ -635,6 +663,25 @@ acpi_status acpi_hw_register_write(u8 use_lock, u32 register_id, u32 value) case ACPI_REGISTER_PM1_CONTROL: /* 16-bit access */ + /* + * Perform a read first to preserve certain bits (per ACPI spec) + * + * Note: This includes SCI_EN, we never want to change this bit + */ + status = acpi_hw_register_read(ACPI_MTX_DO_NOT_LOCK, + ACPI_REGISTER_PM1_CONTROL, + &read_value); + if (ACPI_FAILURE(status)) { + goto unlock_and_exit; + } + + /* Insert the bits to be preserved */ + + ACPI_INSERT_BITS(value, ACPI_PM1_CONTROL_PRESERVED_BITS, + read_value); + + /* Now we can write the data */ + status = acpi_hw_low_level_write(16, value, &acpi_gbl_FADT->xpm1a_cnt_blk); @@ -726,7 +773,7 @@ acpi_hw_low_level_read(u32 width, u32 * value, struct acpi_generic_address *reg) return (AE_OK); } - /* Get a local copy of the address. Handles possible alignment issues */ + /* Get a local copy of the address. Handles possible alignment issues */ ACPI_MOVE_64_TO_64(&address, ®->address); if (!address) { @@ -798,7 +845,7 @@ acpi_hw_low_level_write(u32 width, u32 value, struct acpi_generic_address * reg) return (AE_OK); } - /* Get a local copy of the address. Handles possible alignment issues */ + /* Get a local copy of the address. Handles possible alignment issues */ ACPI_MOVE_64_TO_64(&address, ®->address); if (!address) { |