aboutsummaryrefslogtreecommitdiffstats
path: root/Documentation/i2c
diff options
context:
space:
mode:
authorJean Delvare <khali@linux-fr.org>2007-02-13 16:09:00 -0500
committerJean Delvare <khali@arrakis.delvare>2007-02-13 16:09:00 -0500
commit099ab118b6b194ad43865f62776a0d36d4b1c7d2 (patch)
treefbf07ef85555a784c84d4e80ea617a9244c744dd /Documentation/i2c
parentfcdd96ecaf04fb4f229ac1a64fe77fda890dffd5 (diff)
i2c-i801: Document the SMBus unhiding quirk
This is a frequently asked question so it deserves a paragraph in the driver documentation. Signed-off-by: Jean Delvare <khali@linux-fr.org>
Diffstat (limited to 'Documentation/i2c')
-rw-r--r--Documentation/i2c/busses/i2c-i80160
1 files changed, 55 insertions, 5 deletions
diff --git a/Documentation/i2c/busses/i2c-i801 b/Documentation/i2c/busses/i2c-i801
index 3db69a086c41..c34f0db78a30 100644
--- a/Documentation/i2c/busses/i2c-i801
+++ b/Documentation/i2c/busses/i2c-i801
@@ -48,14 +48,9 @@ following:
48The SMBus controller is function 3 in device 1f. Class 0c05 is SMBus Serial 48The SMBus controller is function 3 in device 1f. Class 0c05 is SMBus Serial
49Controller. 49Controller.
50 50
51If you do NOT see the 24x3 device at function 3, and you can't figure out
52any way in the BIOS to enable it,
53
54The ICH chips are quite similar to Intel's PIIX4 chip, at least in the 51The ICH chips are quite similar to Intel's PIIX4 chip, at least in the
55SMBus controller. 52SMBus controller.
56 53
57See the file i2c-piix4 for some additional information.
58
59 54
60Process Call Support 55Process Call Support
61-------------------- 56--------------------
@@ -74,6 +69,61 @@ SMBus 2.0 Support
74 69
75The 82801DB (ICH4) and later chips support several SMBus 2.0 features. 70The 82801DB (ICH4) and later chips support several SMBus 2.0 features.
76 71
72
73Hidden ICH SMBus
74----------------
75
76If your system has an Intel ICH south bridge, but you do NOT see the
77SMBus device at 00:1f.3 in lspci, and you can't figure out any way in the
78BIOS to enable it, it means it has been hidden by the BIOS code. Asus is
79well known for first doing this on their P4B motherboard, and many other
80boards after that. Some vendor machines are affected as well.
81
82The first thing to try is the "i2c_ec" ACPI driver. It could be that the
83SMBus was hidden on purpose because it'll be driven by ACPI. If the
84i2c_ec driver works for you, just forget about the i2c-i801 driver and
85don't try to unhide the ICH SMBus. Even if i2c_ec doesn't work, you
86better make sure that the SMBus isn't used by the ACPI code. Try loading
87the "fan" and "thermal" drivers, and check in /proc/acpi/fan and
88/proc/acpi/thermal_zone. If you find anything there, it's likely that
89the ACPI is accessing the SMBus and it's safer not to unhide it. Only
90once you are certain that ACPI isn't using the SMBus, you can attempt
91to unhide it.
92
93In order to unhide the SMBus, we need to change the value of a PCI
94register before the kernel enumerates the PCI devices. This is done in
95drivers/pci/quirks.c, where all affected boards must be listed (see
96function asus_hides_smbus_hostbridge.) If the SMBus device is missing,
97and you think there's something interesting on the SMBus (e.g. a
98hardware monitoring chip), you need to add your board to the list.
99
100The motherboard is identified using the subvendor and subdevice IDs of the
101host bridge PCI device. Get yours with "lspci -n -v -s 00:00.0":
102
10300:00.0 Class 0600: 8086:2570 (rev 02)
104 Subsystem: 1043:80f2
105 Flags: bus master, fast devsel, latency 0
106 Memory at fc000000 (32-bit, prefetchable) [size=32M]
107 Capabilities: [e4] #09 [2106]
108 Capabilities: [a0] AGP version 3.0
109
110Here the host bridge ID is 2570 (82865G/PE/P), the subvendor ID is 1043
111(Asus) and the subdevice ID is 80f2 (P4P800-X). You can find the symbolic
112names for the bridge ID and the subvendor ID in include/linux/pci_ids.h,
113and then add a case for your subdevice ID at the right place in
114drivers/pci/quirks.c. Then please give it very good testing, to make sure
115that the unhidden SMBus doesn't conflict with e.g. ACPI.
116
117If it works, proves useful (i.e. there are usable chips on the SMBus)
118and seems safe, please submit a patch for inclusion into the kernel.
119
120Note: There's a useful script in lm_sensors 2.10.2 and later, named
121unhide_ICH_SMBus (in prog/hotplug), which uses the fakephp driver to
122temporarily unhide the SMBus without having to patch and recompile your
123kernel. It's very convenient if you just want to check if there's
124anything interesting on your hidden ICH SMBus.
125
126
77********************** 127**********************
78The lm_sensors project gratefully acknowledges the support of Texas 128The lm_sensors project gratefully acknowledges the support of Texas
79Instruments in the initial development of this driver. 129Instruments in the initial development of this driver.