The announced changes to the LM78 voltages, and an example config file.

Note: GL518SM driver need this change, too; this is not yet done.


git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@106 7894878c-1315-0410-8ee3-d5d059ff63e0

INSTALL

@@ -30,6 +30,9 @@ this package. Some interesting variables:
   The prefix which is added to most directory locations below.
 * MODDIR (/lib/modules/extra/misc)
   The directory where 'make install' installs your modules.
+* ETCDIR (/etc)
+  The directory where 'make install' installs the example configuration file
+  if none is found there.
 * LIBDIR ($(PREFIX)/lib)
   The directory where 'make install' installs the shared and static libraries.
 * BINDIR ($(PREFIX)/bin)
@@ -78,14 +81,21 @@ If you have installed the modules, as described within doc/modules, you can
 do 'modprobe isa && modprobe piix4 && modprobe lm78 && modprobe i2c-proc' 
 (the last is optional) from any directory to reach the same situation.
 
-This package also contains a library of functions which can be used to
+Do not panic if some (or all) of the values in the /proc/sys/dev/sensors/*/*
+files do not seem to correspond to anything in earlier versions. Starting
+with lm_sensors version 2.1.0, these files will reflect the real chip
+measurements, without any scaling or adjusting. This is most obvious for
+the voltage files. You can specify the conversion factors in a configuration
+file.  Use an application linked to libsensors to read these real values.
+
+This package contains a library of functions which can be used to
 access sensor data in a more programmer-friendly way. It will be built as
 both a shared and a static library, and installed in your LIB directory
 (normally /usr/local/lib). It needs a configuration file, which is usually
 called 'sensors.conf', and can be placed in /etc, /usr/lib/sensors,
 /usr/local/lib/sensors, /usr/lib, or /usr/local/lib. [NOTE: No real
-standard config file is written yet. A very simple example can be found in 
-the prog/sensors directory.]
+config file documentation is written yet. A relatively well-commented
+example can be found in the etc directory.]
 
 Where version 1 of this package had the human-readable /proc/sensors file,
 this version includes a user-space program which gathers all data

Makefile

@@ -56,6 +56,10 @@ PREFIX := /usr/local
 # This is the directory into which the modules will be installed.
 MODDIR := /lib/modules/extra/misc
 
+# This is the directory where sensors.conf will be installed, if no other
+# configuration file is found
+ETCDIR := /etc
+
 # You should not need to change this. It is the directory into which the
 # library files (both static and shared) will be installed.
 LIBDIR := $(PREFIX)/lib
@@ -95,7 +99,7 @@ LIBINCLUDEDIR := $(INCLUDEDIR)/sensors
 # to do this. 
 
 # The subdirectories we need to build things in 
-SRCDIRS := src lib prog/sensors
+SRCDIRS := src lib prog/sensors etc
 ifeq ($(I2C),1)
 SRCDIRS += i2c i2c/detect i2c/drivers i2c/eeprom
 endif

README

@@ -53,7 +53,7 @@ support both ways of accessing them.
 
 Because the SMBus is just a special case of the generalized I2C bus, we
 can simulate the SMBus protocol on plain I2C busses. These busses are
-sometimes used in other parts of your computer. If a supported chips is
+sometimes used in other parts of your computer. If a supported chip is
 attached to one of these additional busses, they can be used too.
 
 Please read INSTALL before trying to compile and install these modules.

README.directories

@@ -7,12 +7,14 @@ The directories within this package:
   (mod-????????.tar.gz).
 * doc
   Documentation about the modules within this package
+* etc
+  Configuration files
 * i2c
   Simon Vogl's i2c module, including any necessary patches, excluding any
   files which are removed by 'make clean'.
-* src
-  Our own module code.
 * lib
   The user-space sensors support library code
 * prog
   Several supporting programs, some of them use the library, some do not.
+* src
+  Our own module code.

etc/Module.mk

@@ -0,0 +1,36 @@
+#  Module.mk - Makefile for a Linux module for reading sensor data.
+#  Copyright (c) 1998  Frodo Looijaard <frodol@dds.nl>
+#
+#  This program is free software; you can redistribute it and/or modify
+#  it under the terms of the GNU General Public License as published by
+#  the Free Software Foundation; either version 2 of the License, or
+#  (at your option) any later version.
+#
+#  This program is distributed in the hope that it will be useful,
+#  but WITHOUT ANY WARRANTY; without even the implied warranty of
+#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#  GNU General Public License for more details.
+#
+#  You should have received a copy of the GNU General Public License
+#  along with this program; if not, write to the Free Software
+#  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+# Note that MODULE_DIR (the directory in which this file resides) is a
+# 'simply expanded variable'. That means that its value is substituted
+# verbatim in the rules, until it is redefined. 
+MODULE_DIR := etc
+
+ETCTARGET := $(MODULE_DIR)/sensors.conf.eg
+ETCINSTALL := $(ETCDIR)/sensors.conf
+
+
+# No all rule
+
+install-etc:
+	$(MKDIR) $(ETCDIR)
+	if [ ! -e $(ETCINSTALL) ] ; then \
+	  $(INSTALL) -o  root -g root -m 644 $(ETCTARGET) $(ETCINSTALL); \
+	fi
+install :: install-etc
+
+# No clean rule

etc/sensors.conf.eg

@@ -0,0 +1,120 @@
+# Hardware sensors configuration file
+
+# This configuration file will be used by all applcations linked to 
+# libsensors.
+
+# Hash marks introduce comments, which continue until the end of a line
+#
+# Identifiers consisting of only digits and letters can be used
+# unquoted; other identifiers must be quoted. Escape characters within
+# quotes operate like those in C.
+
+# A 'chip' line specifies what the following 'label', 'compute', 'set' and 
+# (not yet implemented) 'valid' lines refer to. In this case, until the
+# next 'chip' line, everything refers to all lm78, lm78-j and lm79
+# chips. Other examples are *-isa-* for everything on the ISA bus, and
+# lm78-j-i2c-*-4e for all lm78-j chips on address 0x4e of any I2C bus.
+#
+# If more chip statements match a specific chip, they are all considered.
+# Later lines overrule earlier lines, so if you set the in0 label for
+# lm78-* to "This", and later on the in0 label for lm78-isa-* to "That",
+# "That" is used for LM78 chips on the ISA bus, and "This" for LM78
+# chips on a non-ISA bus.
+
+chip "lm78-*" "lm78-j-*" "lm79-*"
+
+# A label line describes what a certain feature stands for on your
+# mainboard. Programs can retrieve these names and display them.
+# If no label is specified for a certain feature, the default name
+# (ie. 'fan1' for fan1) is used.
+
+# If you specify a label for in1, this label is also used for in1_min and
+# in1_max, unless they have their own labels declared. There are several
+# of these logical groups.
+
+# These are as advised in the LM78 and LM79 data sheets, and used on almost
+# any # mainboard we have seen.
+
+  label in0 "VCore 1"
+  label in1 "VCore 2"
+  label in2 "+3.3V"
+  label in3 "+5V"
+  label in4 "+12V"
+  label in5 "-12V"
+  label in6 "-5V"
+  
+# A compute line describes how to scale a certain feature. There are
+# two expressions in it: the first describes how the /proc value must
+# be translated to a user value, the second how a user value must be
+# translated to a /proc value. 'val' is the value to operate on.
+
+# Like for the label statement, there are logical groups here. They are
+# sometimes a bit different, though. For example, fan1_div is in the
+# logical label group of fan1 (it gets the same label if none is declared
+# for it), but it is not in the compute group of fan1 (as it uses a
+# completely different system of values).
+
+# For positive voltages (in0..in4), two resistors are used, with the following
+# formula (R1,R2: resistor values, Vs: read voltage, Vin: pin voltage)
+#   R1 = R2 * (Vs/Vin - 1)
+# For negative voltages (in5, in6) two resistors are used, with the following
+# formula (Rin,Rf: resistor values, Vs: read voltage, Vin: pin voltage)
+#   Rin = (Vs * Rf) / Vin
+
+# Here are the official LM78 and LM79 data sheet values.
+#       Vs     R1,Rin   R2,Rf    Vin
+# in3   +5.0      6.8    10     +2.98
+# in4  +12.0     30      10     +3.00
+# in5  -12.0    240      60     +3.00
+# in6   -5.0    100      60     +3.00
+
+# These would lead to these declarations:
+# compute in3 ((6.8/10)+1)*val ,  val/((6.8/10)+1)
+# compute in4 ((30/10)+1)*val  ,  val/((30/10)+1)
+# compute in5 -(240/60)*val    ,  -val/(240/60)
+# compute in6 -(100/60)*val    ,  -val/(100/60)
+
+# On almost any mainboard we have seen, the Winbond compute values lead to
+# much better results, though. 
+
+#       Vs     R1,Rin   R2,Rf    Vin
+# in4  +12.0     28      10     +3.00
+# in5  -12.0    210      60.4   +3.00
+# in6   -5.0     90.9    60.4   +3.00
+
+# These leads to these declarations:
+  compute in3 ((6.8/10)+1)*val ,  val/((6.8/10)+1)
+  compute in4 ((28/10)+1)*val  ,  val/((28/10)+1)
+  compute in5 -(210/60.4)*val  ,  -val/(210/60.4)
+  compute in6 -(90.9/60.4)*val ,  -val/(90.9/60.4)
+
+# Set statements set things like limits. Complete expressions can be
+# used. Not everything can sensibly be set: setting 'in0', for example,
+# is impossible! These settings # are put through the compute translations;
+# so if we specify '12.8' # for in6, '3.2' will actually be written!
+
+# Though set statements are correctly parsed at this moment, nothing will
+# be done with them yet by libsensors
+
+# Invalid statements tell certain features are not wanted. User programs can
+# still read them if they really want, though; this is just an advisory
+# marking. 'in0' would also invalidate 'in0_max' and 'in0_min'.
+# The invalid keyword is not yet recognized by the parser.
+
+
+# There is one other feature: the 'bus' statement. An example is below.
+
+bus "i2c-0" "SMBus PIIX4 adapter at e800" "Non-I2C SMBus adapter"
+
+# If we refer from now on to 'i2c-0' in 'chip' lines, this will run-time
+# be matched to this bus. So even if the PIIX4 is called 'i2c-5' at that
+# moment, because five other adapters were detected first, 'i2c-0' in
+# the config file would always only match this physical bus. In the above
+# config file, this feature is not needed; but the next lines would
+# only affect the LM75 chips on the PIIX4 adapter:
+
+chip "lm75-i2c-0-*"
+
+# You should really use the output of /proc/bus/chips to generate bus lines,
+# because one mistyped characted will inhibit the match. Wildcards are not
+# yet supported; spaces at the end are ignored, though.

kernel/chips/lm78.c

@@ -65,11 +65,8 @@
 
 
 /* Conversions. Rounding is only done on the TO_REG variants. */
-static int lm78_in_conv[7] = {10000, 10000, 10000, 16892, 38000, 
-                              -34768, -15050 };
-#define IN_TO_REG(val,nr) (((((val) * 100000 / lm78_in_conv[nr]) + 8) / 16) \
-                           & 0xff)
-#define IN_FROM_REG(val,nr) (((val) *  16 * lm78_in_conv[nr]) / 100000)
+#define IN_TO_REG(val,nr)  (((val) * 10 + 8)/16)
+#define IN_FROM_REG(val,nr) (((val) *  16) / 10)
 
 #define FAN_TO_REG(val) ((val)==0?255:((1350000+(val))/((val)*2)) & 0xff)
 #define FAN_FROM_REG(val) ((val)==0?-1:(val)==255?0:1350000/((val)*2))
@@ -84,14 +81,15 @@ static int lm78_in_conv[7] = {10000, 10000, 10000, 16892, 38000,
 #define DIV_FROM_REG(val) (1 << (val))
 #define DIV_TO_REG(val) ((val)==8?3:(val)==4?2:(val)==1?0:1)
 
-/* Initial limits */
+/* Initial limits. To keep them sane, we use the 'standard' translation as
+   specified in the LM78 sheet. Use the config file to set better limits. */
 #define LM78_INIT_IN_0 (vid==350?280:vid)
 #define LM78_INIT_IN_1 (vid==350?280:vid)
 #define LM78_INIT_IN_2 330
-#define LM78_INIT_IN_3 500
-#define LM78_INIT_IN_4 1200
-#define LM78_INIT_IN_5 -1200
-#define LM78_INIT_IN_6 -500
+#define LM78_INIT_IN_3 (((500)   * 100)/168)
+#define LM78_INIT_IN_4 (((1200)  * 10)/38)
+#define LM78_INIT_IN_5 (((-1200) * -604)/2100)
+#define LM78_INIT_IN_6 (((-500)  * -604)/909)
 
 #define LM78_INIT_IN_PERCENTAGE 10
 

src/lm78.c

@@ -65,11 +65,8 @@
 
 
 /* Conversions. Rounding is only done on the TO_REG variants. */
-static int lm78_in_conv[7] = {10000, 10000, 10000, 16892, 38000, 
-                              -34768, -15050 };
-#define IN_TO_REG(val,nr) (((((val) * 100000 / lm78_in_conv[nr]) + 8) / 16) \
-                           & 0xff)
-#define IN_FROM_REG(val,nr) (((val) *  16 * lm78_in_conv[nr]) / 100000)
+#define IN_TO_REG(val,nr)  (((val) * 10 + 8)/16)
+#define IN_FROM_REG(val,nr) (((val) *  16) / 10)
 
 #define FAN_TO_REG(val) ((val)==0?255:((1350000+(val))/((val)*2)) & 0xff)
 #define FAN_FROM_REG(val) ((val)==0?-1:(val)==255?0:1350000/((val)*2))
@@ -84,14 +81,15 @@ static int lm78_in_conv[7] = {10000, 10000, 10000, 16892, 38000,
 #define DIV_FROM_REG(val) (1 << (val))
 #define DIV_TO_REG(val) ((val)==8?3:(val)==4?2:(val)==1?0:1)
 
-/* Initial limits */
+/* Initial limits. To keep them sane, we use the 'standard' translation as
+   specified in the LM78 sheet. Use the config file to set better limits. */
 #define LM78_INIT_IN_0 (vid==350?280:vid)
 #define LM78_INIT_IN_1 (vid==350?280:vid)
 #define LM78_INIT_IN_2 330
-#define LM78_INIT_IN_3 500
-#define LM78_INIT_IN_4 1200
-#define LM78_INIT_IN_5 -1200
-#define LM78_INIT_IN_6 -500
+#define LM78_INIT_IN_3 (((500)   * 100)/168)
+#define LM78_INIT_IN_4 (((1200)  * 10)/38)
+#define LM78_INIT_IN_5 (((-1200) * -604)/2100)
+#define LM78_INIT_IN_6 (((-500)  * -604)/909)
 
 #define LM78_INIT_IN_PERCENTAGE 10