Use 'init=0' to skip initialization, this should fix
those pesky PC's that crash when you load w83781d.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1209 7894878c-1315-0410-8ee3-d5d059ff63e0
Fail and printk if still busy.
Fail and printk if completion timeout at end of transaction.
Reduce timeout from 5 sec to 1 sec.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1190 7894878c-1315-0410-8ee3-d5d059ff63e0
'supported' table. i2c adapter name and io region will
always be piix4 now - too hard to do 5 different names.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1188 7894878c-1315-0410-8ee3-d5d059ff63e0
adding a rather large delay after resetting the device
via register 0x40. There is obviously a state machine
in the chip that takes a while to run after you reset it.
Figured out the delay time by trial and error.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1168 7894878c-1315-0410-8ee3-d5d059ff63e0
so that each temp. sensor for each chip can be printed with
the right number of digits. I tried to set each chip to
the right precision but there may still be some tweaking to do.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1161 7894878c-1315-0410-8ee3-d5d059ff63e0
when encountering a long list of unsupported chips.
All these chips have the same 1039/0008 ID as the 5595 but are
incompatible.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1150 7894878c-1315-0410-8ee3-d5d059ff63e0
This was broken by 2.6.0 because
2.6.0 didn't reset the fan clock setting at startup.
Therefore the AS_FAN_* macros were no longer needed to specially
correct the fan speed. So remove the special handling and the
AS_FAN_* macros and things should work now.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1135 7894878c-1315-0410-8ee3-d5d059ff63e0
_sysctl_real
_register_entry
_proc_real
_detect
_deregister_entry
_chips_data
which were the global symbols renamed in i2c-proc.[ch] in the
i2c package.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1101 7894878c-1315-0410-8ee3-d5d059ff63e0
which ensures i2c Bus is tristated correctly. It isn't clear from
the datasheet if the 586's GPIO pins are open drain or not,
but if they aren't, then the driver was driving the pins high
rather than letting them get pulled high as the spec requires.
This is corrected by toggling the direction rather than the
data register to control the bus.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1093 7894878c-1315-0410-8ee3-d5d059ff63e0
Also automatically enable ACPI if it was disabled.
Forcing address doesn't work on my test system, but hey,
at least I tried to test it...
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1077 7894878c-1315-0410-8ee3-d5d059ff63e0
to force_addr. It is now consistent with the bus PCI adapters
that support base address setting (piix4, i801, viapro).
Other benefit is that via686a_find can now return -ENODEV,
and module will fail to load when base address is not set and
parameter is not used.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1073 7894878c-1315-0410-8ee3-d5d059ff63e0
The temperature registers are in different places for the two
versions. Both versions unfortunately have the same PCI ID.
Apparently the temperature conversion formulas are different
for the two versions. Neither is yet implemented in the driver.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1070 7894878c-1315-0410-8ee3-d5d059ff63e0
in a different place than the LM78, which this driver was copied from,
so temp was completely broken.
Temp shares a pin and registers with in4; add in4 support, recognize
chip configuration and act accordingly.
git-svn-id: http://lm-sensors.org/svn/lm-sensors/trunk@1067 7894878c-1315-0410-8ee3-d5d059ff63e0
Mark D. Studebaker