[LinuxBIOS] m57sli-s4 and i2c

[Ward Vandewege]

Hi all,

Thanks to Stepan's perseverence (thanks!), the m57sli-s4 code is now
sufficiently merged into the v2 tree; the following is based on v2 checkout
2611.

I've reported before that there still seems to be something missing in the
i2c initialization for the gigabyte m57sli-s4 board. The symptom I was seeing
before was a very slow X start (30-50s delay) which can be fixed by adding

Option   "NoDDC2"

to the "device" section of the X config file.

I tried playing with lm-sensors today, and sure enough, I can only read out a
couple of sensors (the ones controlled by the k8temp module). There is
another set of sensors on the SuperIO, which is detected by the
lmsensors-detect script as

Found `ITE IT8716F Super IO Sensors'
    (but not activated)

The sensors-detect script is a perl script part of the lm-sensors package;
you can also download the latest version here

http://www.lm-sensors.org/browser/lm-sensors/trunk/prog/detect/sensors-detec...

Under LinuxBIOS, the it87 module can not be loaded:

FATAL: Error inserting it87
(/lib/modules/2.6.20.1/kernel/drivers/hwmon/it87.ko): No such device

Under the proprietary bios, the sensors-detect script detects the IT8716F
just fine. The it87 module can be loaded, which allows reading out of various
system temperatures and fan speeds, as well as manual fan speed control.

Any clues as to what is going on here, and how to fix this?

Thanks,
Ward.

[Peter Stuge]

On Wed, Apr 18, 2007 at 05:40:26PM -0400, Ward Vandewege wrote:
> Found `ITE IT8716F Super IO Sensors'
>     (but not activated)

--8<--
  # Switch to the sensor logical device
  outb($addrreg, $superio{logdevreg});
  outb($datareg, $chip->{logdev});

# Check the activation register
  outb($addrreg, $superio{actreg});
  $val = inb($datareg);
  if (!($val & $superio{actmask})) {
    print "\n    (but not activated)\n";
    return;
  }
-->8--

I guess LB doesn't activate the sensors part of the superio?
> Under LinuxBIOS, the it87 module can not be loaded:
>
> FATAL: Error inserting it87
> (/lib/modules/2.6.20.1/kernel/drivers/hwmon/it87.ko): No such device

Is there anything informative in dmesg after this error?

//Peter

[yhlu]

Ward,

The easy way You could compare the superio (ite8716) reg info between
legacy BIOS and LinuxBIOS. Then you could update superio.c for ite8716
to set related registers, like some HWM for winbond 83627...
I wonder some conf for fan etc may need some calibration.

YH

[Ward Vandewege]

Hi YH,

On Wed, Apr 18, 2007 at 08:47:23PM -0700, yhlu wrote:
> The easy way You could compare the superio (ite8716) reg info between
> legacy BIOS and LinuxBIOS.

Right. I've actually succeeded to activate the EC for the it8716f superio,
and to enable fan control, by working from the datasheet and the it87 driver
in the Linux kernel.

But I'd like to compare the registers to see the temperature/voltage sensor
calibration that is done in the proprietary BIOS.

How would I read those registers from userspace? I've looked around with
Google, but I can't seem to find an answer. I presume lspci is out, since the
superio hangs off the isa bus.

Thanks,
Ward.

[ST]

Hi Ward

Thanks for your updated build description on the wiki. I think there is a list 
of build instructions on the list, where your description is not linked?
http://www.linuxbios.org/Documentation#How-To.27s

Also the description of the filo payload still postulates that filo doesn't 
load sata disks. Which it does according to your build instructions.

I know i am a little biased... :-) but i think my proposed solution has some 
advantages over desoldering the flashrom. First desoldering with a heat gun 
poses the risk that you simple blow some of the tiny resistors away and 
second it is pretty easy to switch between the two flashroms which is quite 
neat in comparing register settings...
> How would I read those registers from userspace? I've looked around with
> Google, but I can't seem to find an answer. I presume lspci is out, since
> the superio hangs off the isa bus.

I think thats inb for byte access and inw for word access (16 bit). You need 
to be root to do this though. The addresses can be taken right from the 
datasheet.

Best regards
ST

[Ward Vandewege]

Hi ST,

On Fri, Apr 27, 2007 at 10:32:43PM +0200, ST wrote:
> Thanks for your updated build description on the wiki. I think there is a list 
> of build instructions on the list, where your description is not linked?
> http://www.linuxbios.org/Documentation#How-To.27s

Right, I've added it.
> Also the description of the filo payload still postulates that filo doesn't 
> load sata disks. Which it does according to your build instructions.

It most certainly does. I've fixed the FILO page.
> I know i am a little biased... :-) but i think my proposed solution has some 
> advantages over desoldering the flashrom. First desoldering with a heat gun 
> poses the risk that you simple blow some of the tiny resistors away and 
> second it is pretty easy to switch between the two flashroms which is quite 
> neat in comparing register settings...

Yes, absolutely. Your solution is better.
> > How would I read those registers from userspace? I've looked around with
> > Google, but I can't seem to find an answer. I presume lspci is out, since
> > the superio hangs off the isa bus.
>
> I think thats inb for byte access and inw for word access (16 bit). You need 
> to be root to do this though. The addresses can be taken right from the 
> datasheet.

Right - I thought I could do this by dumping /dev/ports, but I can only see 4
(the superio EC controller lives at 0x290, with a control register at 0x295
and a data register at 0x296):

00000200  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff
*
00000290  ff ff ff ff 32 0a 32 32  ff ff ff ff ff ff ff ff
000002a0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff
*

So then I thought I needed to do what was described in the datasheet: write
the index that I want to read out to the control register, and read the data
from the data register:

#!/usr/bin/perl

sub io_read {
    my $offset = shift;
    open(TMP,'/dev/port');
    seek(TMP,0x295,0);
    write(TMP,$offset,1);
    seek(TMP,0x296, 0);
    read(TMP,$tmp,1);
    close(TMP);
    printf("0x%02x\n",$tmp);
    return $tmp;
  }

io_read(0x14);

I'm obviously doing something wrong because I get 0x0 back, for any index,
when running under LinuxBIOS, and 0x2 when running under the proprietary
BIOS.

Thanks,
Ward.

[Uwe Hermann]

On Fri, Apr 27, 2007 at 10:32:43PM +0200, ST wrote:
> I know i am a little biased... :-) but i think my proposed solution has some 
> advantages over desoldering the flashrom. First desoldering with a heat gun 
> poses the risk that you simple blow some of the tiny resistors away and 
> second it is pretty easy to switch between the two flashroms which is quite 
> neat in comparing register settings...

Btw, do you want to include your instructions page into the wiki and
maintain it there? And/or create a copy of the instructions in the wiki?

Uwe.

[Stefan Reinauer]

* Ward Vandewege  [070427 22:09]:
> How would I read those registers from userspace? I've looked around with
> Google, but I can't seem to find an answer. I presume lspci is out, since the
> superio hangs off the isa bus.

check LinuxBIOSv2/util/probe_superio

[Ward Vandewege]

On Thu, Apr 19, 2007 at 05:41:49AM +0200, Peter Stuge wrote:
> I guess LB doesn't activate the sensors part of the superio?

That was my guess too.
> > Under LinuxBIOS, the it87 module can not be loaded:
> >
> > FATAL: Error inserting it87
> > (/lib/modules/2.6.20.1/kernel/drivers/hwmon/it87.ko): No such device
>
> Is there anything informative in dmesg after this error?

[20987.920000] i2c_adapter i2c-9191: ISA main adapter registered
[20987.936000] it87: Device not activated, skipping

Looks like that might be the problem.

Thanks,
Ward.

[Ward Vandewege]

On Wed, Apr 18, 2007 at 11:48:42PM -0400, Ward Vandewege wrote:
> On Thu, Apr 19, 2007 at 05:41:49AM +0200, Peter Stuge wrote:
> > I guess LB doesn't activate the sensors part of the superio?
>
> That was my guess too.

I've managed to activate the 'environmental controller' (EC) on the gigabyte
m57sli-s4's superio.

This now works:

======================================================================
# perl sensors-detect
Some Super I/O chips may also contain sensors. We have to write to
standard I/O ports to probe them. This is usually safe.
Do you want to scan for Super I/O sensors? (YES/no): 
Probing for Super-I/O at 0x2e/0x2f
Trying family `National Semiconductor'...                   No
Trying family `SMSC'...                                     No
Trying family `VIA/Winbond/Fintek'...                       No
Trying family `ITE'...                                      Yes
Found `ITE IT8716F Super IO Sensors'                        Success!
    (address 0x290, driver `it87')
======================================================================

And I get output with the 'sensors' command:

======================================================================
# sensors
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp:
             +39°C
Core1 Temp:
             +40°C

it8716-isa-0290
Adapter: ISA adapter
VCore:     +1.30 V  (min =  +0.00 V, max =  +4.08 V)
VDDR:      +1.94 V  (min =  +0.00 V, max =  +4.08 V)
+3.3V:     +3.26 V  (min =  +0.00 V, max =  +4.08 V)
+5V:       +4.87 V  (min =  +0.00 V, max =  +6.85 V)
+12V:     +12.42 V  (min =  +0.00 V, max = +16.32 V)
in5:       +0.54 V  (min =  +0.00 V, max =  +4.08 V)
in6:       +1.07 V  (min =  +0.00 V, max =  +4.08 V)
5VSB:      +4.87 V  (min =  +0.00 V, max =  +6.85 V)
VBat:      +3.09 V
fan1:     1618 RPM  (min =    0 RPM)
fan2:     1333 RPM  (min =    0 RPM)
fan3:        0 RPM  (min =    0 RPM)
temp1:       +25°C  (low  =    -1°C, high =  +127°C)   sensor = thermistor
temp2:      +128°C  (low  =    -1°C, high =  +127°C)   sensor = diode
temp3:       +81°C  (low  =    -1°C, high =  +127°C)   sensor = thermistor
vid:      +1.300 V
======================================================================

Most of these values seem ok or at least plausable - fan3 is not connected so
that's correct; the value for temp2 and possibly temp3 is not correct.

This is what sensors tells me under the proprietary BIOS:
======================================================================
# sensors
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp:
             +26°C
Core1 Temp:
             +29°C

it8716-isa-0290
Adapter: ISA adapter
VCore:     +1.07 V  (min =  +0.00 V, max =  +4.08 V)
VDDR:      +1.94 V  (min =  +0.00 V, max =  +4.08 V)
+3.3V:     +3.26 V  (min =  +0.00 V, max =  +4.08 V)
+5V:       +4.84 V  (min =  +0.00 V, max =  +6.85 V)
+12V:     +12.42 V  (min =  +0.00 V, max = +16.32 V)
in5:       +0.90 V  (min =  +0.00 V, max =  +4.08 V)
in6:       +0.66 V  (min =  +0.00 V, max =  +4.08 V)
5VSB:      +4.84 V  (min =  +0.00 V, max =  +6.85 V)
VBat:      +3.10 V
fan1:     1622 RPM  (min =    0 RPM)
fan2:     1336 RPM  (min =    0 RPM)
fan3:        0 RPM  (min =    0 RPM)
temp1:       +25°C  (low  =  +127°C, high =  +127°C)   sensor = thermistor
temp2:       -21°C  (low  =  +127°C, high =  +127°C)   sensor = thermistor
temp3:       +31°C  (low  =  +127°C, high =  +127°C)   sensor = diode
vid:      +1.100 V
======================================================================

Some of these are in agreement with what I get under LinuxBIOS - but the
core0/core1 temperatures are more than 10 degrees off (that's in k8temp);
Vcore is different; in5 and in6 are different; temp2 is *also* wrong (funny
:), and temp3 is different; vid is different.

I guess I'm not quite there yet; this is what's logged when loading the
it8716 module:

[   14.860000] it87: Found IT8716F chip at 0x290, revision 0
[   14.864000] it87: in3 is VCC (+5V)
[   14.868000] it87: in7 is VCCH (+5V Stand-By)
[   14.872000] i2c_adapter i2c-9191: Driver it87-isa registered
[   14.872000] i2c_adapter i2c-9191: client [it8716] registered with bus id
9191-0290
[   14.872000] it87-isa 9191-0290: Detected broken BIOS defaults, disabling
PWM interface

Funnily enough, if I load the it87 module with the following parameter

modprobe it87 fix_pwm_polarity=1

(which is a fix for some broken proprietary BIOS'es out there ;) it seems to
be happier:

[  797.956000] it87: Found IT8716F chip at 0x290, revision 0
[  797.960000] it87: in3 is VCC (+5V)
[  797.964000] it87: in7 is VCCH (+5V Stand-By)
[  797.968000] i2c_adapter i2c-9191: Driver it87-isa registered
[  797.968000] i2c_adapter i2c-9191: client [it8716] registered with bus id
9191-0290
[  797.968000] it87-isa 9191-0290: Reconfiguring PWM to active high polarity

And then I can set the fan speed :)

YH mentioned some initialization might be required for some of these
readouts, so I guess I'll go back to the datasheet to see if I can figure
something out.

In any case, patch attached to get at least this far.

Thanks,
Ward.

[LinuxBIOS] [PATCH] Re: m57sli-s4 and i2c

[Ward Vandewege]

Hi all,

So, here's a patch to enable fan control on the it8716f superio, as used in
the gigabyte m57sli-s4 board. Patch attached.

This allows fan control by loading the it87 module (or compiling it87 support
in your kernel), for instance via the /sys/ interface:

cat /sys/bus/i2c/devices/9191-0290/pwm1
  128 -> fan running at half speed
  echo 255 > /sys/bus/i2c/devices/9191-0290/pwm1
  -> fan goes to full speed

Be careful; there are no safeguards against turning the fans *off*...

This works great for me, but I'm not 100% sure about the pnp_dev_info
structure values, see my other e-mail.

Thanks,
Ward.

[LinuxBIOS] [PATCH] Re: m57sli-s4 and i2c

[Uwe Hermann]

On Fri, Apr 27, 2007 at 05:08:50PM -0400, Ward Vandewege wrote:
> Add fan control support.
>
> Signed-off-by: Ward Vandewege

Just for book-keeping: this was committed in r2620.

Uwe.

[LinuxBIOS] [PATCH] Re: m57sli-s4 and i2c

[ST]

Hi
> This allows fan control by loading the it87 module (or compiling it87
> support in your kernel), for instance via the /sys/ interface:
>
> cat /sys/bus/i2c/devices/9191-0290/pwm1
>   128 -> fan running at half speed
>   echo 255 > /sys/bus/i2c/devices/9191-0290/pwm1
>   -> fan goes to full speed
>
> Be careful; there are no safeguards against turning the fans *off*...

Mh, i just tried it and doesn't work for me :-(. Probably i am to tired and 
did a stupid mistake. But well i still can't load the it87 module. But my 
tree has the changes in superio.c. Hints are welcome :-). Probably im am 
missing a kernel driver or so?
Ward i have also the problem that i don't see any classic PCI cards on my 
machine (with linuxbios, works with stock bios), can you confirm this? I see 
all the nvidia stuff and the video card though.

I have written a small c program which reads out the values of the ec 
controller. It works for me under stock bios but gives "ff" on all reads 
under linuxbios :-(. Ward it would be nice if you could try it on your 
machine... and yes i was a little bit off with my statement can be take right 
of the datasheet. The program needs to be run as root or gives an access 
error.

Attached is also the output of the program on my machine under i386 and amd64. 
Keep in mind that some of the registers are fan and temperature input so they 
differ.

Concerning the question of Uwe, yes if you like i can maintain it on the wiki 
but i need an account?

Good night
ST

[LinuxBIOS] [PATCH] Re: m57sli-s4 and i2c

[Uwe Hermann]

On Sat, Apr 28, 2007 at 01:43:56AM +0200, ST wrote:
> Concerning the question of Uwe, yes if you like i can maintain it on the wiki 
> but i need an account?

Sure, no problem. Stefan or Ron can surely create one, see
http://linuxbios.org/index.php?title=Special:Userlogin

Uwe.

[LinuxBIOS] [PATCH] Re: m57sli-s4 and i2c

[Ward Vandewege]

Hi ST,

On Sat, Apr 28, 2007 at 01:43:56AM +0200, ST wrote:
> Mh, i just tried it and doesn't work for me :-(. Probably i am to tired and 
> did a stupid mistake. But well i still can't load the it87 module. But my 
> tree has the changes in superio.c. Hints are welcome :-). Probably im am 
> missing a kernel driver or so?

I suspect you don't have the other patch that was submitted under r2620 - the
changes to Config.lb for the m57sli-s4 ?
> Ward i have also the problem that i don't see any classic PCI cards on my 
> machine (with linuxbios, works with stock bios), can you confirm this? I see 
> all the nvidia stuff and the video card though.

I'll look into that next week, when I'm back at the office.
> I have written a small c program which reads out the values of the ec 
> controller. It works for me under stock bios but gives "ff" on all reads 
> under linuxbios :-(. Ward it would be nice if you could try it on your 
> machine... and yes i was a little bit off with my statement can be take right 
> of the datasheet. The program needs to be run as root or gives an access 
> error.

Sure, I had figured that much; my perl script still didn't work. I'll give
your program a go and get back to you next week.
> Attached is also the output of the program on my machine under i386 and amd64. 
> Keep in mind that some of the registers are fan and temperature input so they 
> differ.

Very helpful, thanks!

Thanks,
Ward.