[lm-sensors] Driver for SMSC SCH5627 needed

[lm-sensors] Driver for SMSC SCH5627 needed

[Rainer Koenig]

Hi,

I have a board that has a SMSC SCH5627 chip on it. sensors detect report
"unknow chip SMSC 0xc601".
Where can I get a driver for this chip?

Best regards
Rainer

-- 
Dipl.-Inf. (FH) Rainer Koenig
Project Manager Linux Clients
Dept. TSP WPS R&D SW OSE

Fujitsu Technology Solutions
Bürgermeister-Ullrich-Str. 100
86199 Augsburg
Germany

Telephone: +49-821-804-3321
Telefax:   +49-821-804-2131
Mail:      mailto:Rainer.Koenig@ts.fujitsu.com 

Internet         ts.fujtsu.com
Company Details  ts.fujitsu.com/imprint.html


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

Hi Rainer,


> Hi,
>
> I have a board that has a SMSC SCH5627 chip on it. sensors detect report
> "unknow chip SMSC 0xc601".
> Where can I get a driver for this chip?

First we need a datasheet. Then we can figure out if we already have a
driver that supports a similar device and could be easily extended. No
datasheet usually means no driver.

What board model/vendor is this? Dell machine?

...Juerg



> Best regards
> Rainer
>

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Rainer Koenig]

Hi Juerg,

Juerg Haefliger schrieb:
>> I have a board that has a SMSC SCH5627 chip on it. sensors detect report
>> "unknow chip SMSC 0xc601".
>> Where can I get a driver for this chip?
>>     
>
> First we need a datasheet. Then we can figure out if we already have a
> driver that supports a similar device and could be easily extended. No
> datasheet usually means no driver.
>   
I perfectly understand that. So we used our contacts to SMSC to ask them
for a chip documentation that is not under NDA. Didn't get an answer yet,
but still hoping.
> What board model/vendor is this? Dell machine?
>   
The board is a D2984 board used in the ESPRIMO E5645 from Fujitsu
Technology
Solutions. Data sheet is here:
http://docs.ts.fujitsu.com/dl.aspx?idO300628-7547-44b6-ac82-172ebae232c1

Best regards
Rainer

-- 
Dipl.-Inf. (FH) Rainer Koenig
Project Manager Linux Clients
Dept. TSP WPS R&D SW OSE

Fujitsu Technology Solutions
Bürgermeister-Ullrich-Str. 100
86199 Augsburg
Germany

Telephone: +49-821-804-3321
Telefax:   +49-821-804-2131
Mail:      mailto:Rainer.Koenig@ts.fujitsu.com 

Internet         ts.fujtsu.com
Company Details  ts.fujitsu.com/imprint.html


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

> Hi Juerg,

Hi Rainer,

>> First we need a datasheet. Then we can figure out if we already have a
>> driver that supports a similar device and could be easily extended. No
>> datasheet usually means no driver.
>>
> I perfectly understand that. So we used our contacts to SMSC to ask them
> for a chip documentation that is not under NDA. Didn't get an answer yet,
> but still hoping.

I sent an email to my SMSC contact. We'll see what happens.

...Juerg

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

Hi Rainer,


>> Hi Juerg,
>
> Hi Rainer,
>
>>> First we need a datasheet. Then we can figure out if we already have a
>>> driver that supports a similar device and could be easily extended. No
>>> datasheet usually means no driver.
>>>
>> I perfectly understand that. So we used our contacts to SMSC to ask them
>> for a chip documentation that is not under NDA. Didn't get an answer yet,
>> but still hoping.
>
> I sent an email to my SMSC contact. We'll see what happens.

I received the datasheet from SMSC but under NDA, so I can't share it.
I'll check it against the currently supported chips to see if we have
something compatible or fairly close. Give me a few days though...

...Juerg

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

Hi Rainer,


> I received the datasheet from SMSC but under NDA, so I can't share it.
> I'll check it against the currently supported chips to see if we have
> something compatible or fairly close. Give me a few days though...

Took a quick glance at the datasheet. This is a totally different
beast that is not even remotely compatible to anything we support thus
far AFAIK. It contains an embedded microcontroller (!) which handles
all the hwmon related tasks. The datasheet that I received doesn't
contain any information about the microcontroller other than a simple
block diagram. So at the moment it's absolutely a no go and even with
all the technical data we would have to start from scratch with a new
driver.

...Juerg




> ...Juerg
>

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Jean Delvare]

On Thu, 21 Oct 2010 17:47:20 +0200, Juerg Haefliger wrote:
> Hi Rainer,
> 
> 
> > I received the datasheet from SMSC but under NDA, so I can't share it.
> > I'll check it against the currently supported chips to see if we have
> > something compatible or fairly close. Give me a few days though...
> 
> Took a quick glance at the datasheet. This is a totally different
> beast that is not even remotely compatible to anything we support thus
> far AFAIK. It contains an embedded microcontroller (!) which handles
> all the hwmon related tasks. The datasheet that I received doesn't
> contain any information about the microcontroller other than a simple
> block diagram. So at the moment it's absolutely a no go and even with
> all the technical data we would have to start from scratch with a new
> driver.

Thanks Juerg for checking this. Can you please update sensors-detect
with the device name and features, and also create an entry in
wiki/Devices? So that other user with that chip know the current
status, even if it's bad.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

> Thanks Juerg for checking this. Can you please update sensors-detect
> with the device name and features, and also create an entry in
> wiki/Devices? So that other user with that chip know the current
> status, even if it's bad.


Will do.

...Juerg


> --
> Jean Delvare
>

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

Hi Jean

>> Took a quick glance at the datasheet. This is a totally different
>> beast that is not even remotely compatible to anything we support thus
>> far AFAIK. It contains an embedded microcontroller (!) which handles
>> all the hwmon related tasks. The datasheet that I received doesn't
>> contain any information about the microcontroller other than a simple
>> block diagram. So at the moment it's absolutely a no go and even with
>> all the technical data we would have to start from scratch with a new
>> driver.
>
> Thanks Juerg for checking this. Can you please update sensors-detect
> with the device name and features, and also create an entry in
> wiki/Devices? So that other user with that chip know the current
> status, even if it's bad.

I added detection to sensors-detect but I'm not too happy with the outcome:
Trying family `SMSC'...                                     Yes
Found `SMSC SCH5627 Desktop Embedded Controller'
   (but not activated)

The chip shouldn't be probed at all since it's non-standard. Users
might get confused by the 'not activated' message and will ask how to
activate it. Is there a way to bypass the probing and just print a
message that there is no driver for this chip?

...Juerg




> --
> Jean Delvare
>

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Juerg Haefliger]

Hi Jean,

While I have your attention :-)

> Hi Jean
>
>>> Took a quick glance at the datasheet. This is a totally different
>>> beast that is not even remotely compatible to anything we support thus
>>> far AFAIK. It contains an embedded microcontroller (!) which handles
>>> all the hwmon related tasks. The datasheet that I received doesn't
>>> contain any information about the microcontroller other than a simple
>>> block diagram. So at the moment it's absolutely a no go and even with
>>> all the technical data we would have to start from scratch with a new
>>> driver.
>>
>> Thanks Juerg for checking this. Can you please update sensors-detect
>> with the device name and features, and also create an entry in
>> wiki/Devices? So that other user with that chip know the current
>> status, even if it's bad.
>
> I added detection to sensors-detect but I'm not too happy with the outcome:
> Trying family `SMSC'...                                     Yes
> Found `SMSC SCH5627 Desktop Embedded Controller'
>   (but not activated)
>
> The chip shouldn't be probed at all since it's non-standard. Users
> might get confused by the 'not activated' message and will ask how to
> activate it. Is there a way to bypass the probing and just print a
> message that there is no driver for this chip?

Any suggestions on how to accomplish this?

Thanks
...Juerg

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Jean Delvare]

Hi Juerg,

Sorry for the late answer.

On Tue, 9 Nov 2010 10:35:51 +0100, Juerg Haefliger wrote:
> Hi Jean
> 
> >> Took a quick glance at the datasheet. This is a totally different
> >> beast that is not even remotely compatible to anything we support thus
> >> far AFAIK. It contains an embedded microcontroller (!) which handles
> >> all the hwmon related tasks. The datasheet that I received doesn't
> >> contain any information about the microcontroller other than a simple
> >> block diagram. So at the moment it's absolutely a no go and even with
> >> all the technical data we would have to start from scratch with a new
> >> driver.
> >
> > Thanks Juerg for checking this. Can you please update sensors-detect
> > with the device name and features, and also create an entry in
> > wiki/Devices? So that other user with that chip know the current
> > status, even if it's bad.
> 
> I added detection to sensors-detect but I'm not too happy with the outcome:
> Trying family `SMSC'...                                     Yes
> Found `SMSC SCH5627 Desktop Embedded Controller'
>    (but not activated)
> 
> The chip shouldn't be probed at all since it's non-standard. Users
> might get confused by the 'not activated' message and will ask how to
> activate it. Is there a way to bypass the probing and just print a
> message that there is no driver for this chip?

Hans submitted a candidate driver for this device meanwhile, so it
would make sense to add support to sensors-detect too.

I am surprised by your report above. Looking at Hans's driver code, the
only non-standard thing about the SCH5627 is that the base I/O register
is at 0x66 instead of 0x60, with swapped bytes. The activation bit is
still 0x30[0], while your message above suggests that the bit in
question wasn't enabled on your system (although it's a little
difficult to comment, given that you never shared your sensors-detect
patch with us; maybe you selected the wrong logical device?)

Anyway, I agree that the non-standard location of the base I/O register
isn't supported by sensors-detect yet. We already have some code to
handle non-standard SMSC chips, but the SCH5627 is non-standard in a
different way. I'll try to come up with something, stay tuned.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Jean Delvare]

[-- Attachment #1: Type: text/plain, Size: 657 bytes --]

On Sat, 12 Mar 2011 12:19:09 +0100, Jean Delvare wrote:
> Anyway, I agree that the non-standard location of the base I/O register
> isn't supported by sensors-detect yet. We already have some code to
> handle non-standard SMSC chips, but the SCH5627 is non-standard in a
> different way. I'll try to come up with something, stay tuned.

Can anyone with access to a SCH5627 chip test the sensors-detect patch
in attachment? Alternatively a pre-patched sensors-detect can be
downloaded at:
  http://khali.linux-fr.org/devel/misc/sensors-detect

It should properly detect the SCH5627, but I can't test it as I don't
have such a chip.

Thanks,
-- 
Jean Delvare

[-- Attachment #2: smsc-sch5627-detection.patch --]
[-- Type: text/x-patch, Size: 4590 bytes --]

Add detection of the SMSC SCH5627. Some code is needed because the
base I/O registers live at non-standard addresses.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
---
 prog/detect/sensors-detect |   60 ++++++++++++++++++++++++++++++++++----------
 1 file changed, 47 insertions(+), 13 deletions(-)

--- lm-sensors.orig/prog/detect/sensors-detect	2011-03-04 15:59:20.000000000 +0100
+++ lm-sensors/prog/detect/sensors-detect	2011-03-12 14:05:38.000000000 +0100
@@ -1402,6 +1402,7 @@ use vars qw(@i2c_adapter_names);
 #	 monitoring registers can only be accessed via the SMBus
 #  devid: The device ID we have to match (base device)
 #  devid_mask (optional): Bitmask to apply before checking the device ID
+#  regs (optional): Register definitions, where they differ from the standard.
 #  logdev: The logical device containing the sensors
 #  check (optional): A function to refine the detection. Will be passed
 #      the index and data ports as parameters. Must return 1 for a matching
@@ -1761,6 +1762,16 @@ use constant FEAT_SMBUS	=> (1 << 7);
 		# No datasheet
 		driver => "not-a-sensor",
 		devid => 0x83,
+	}, {
+		name => "SMSC SCH5627 Super IO",
+		driver => "sch5627",
+		devid => 0xc6,
+		regs => {
+			basereg_lsb => 0x66,
+			basereg_msb => 0x67,
+		},
+		logdev => 0x0c,
+		features => FEAT_IN | FEAT_FAN | FEAT_TEMP,
 	}
 );
 
@@ -2236,6 +2247,20 @@ sub any_list_match
 	return 0;
 }
 
+# $_[0]: Reference to base hash
+# $_[1]: Reference to overlay hash
+# Result: Overlayed hash
+sub overlay_hash
+{
+	my ($base, $overlay) = @_;
+	my %result = %{$base};
+
+	foreach my $key (keys %{$overlay}) {
+		$result{$key} = $overlay->{$key};
+	}
+	return %result;
+}
+
 ###################
 # I/O PORT ACCESS #
 ###################
@@ -3544,7 +3569,7 @@ sub scan_isa_bus
 	$| = 0;
 }
 
-use vars qw(%superio);
+use vars qw(%standard_superio);
 
 # The following are taken from the PNP ISA spec (so it's supposed
 # to be common to all Super I/O chips):
@@ -3552,13 +3577,14 @@ use vars qw(%superio);
 #  logdevreg: The logical device register
 #  actreg: The activation register within the logical device
 #  actmask: The activation bit in the activation register
-#  basereg: The I/O base register within the logical device
-%superio = (
+#  basereg_*: The I/O base registers within the logical device
+%standard_superio = (
 	devidreg => 0x20,
 	logdevreg => 0x07,
 	actreg => 0x30,
 	actmask => 0x01,
-	basereg => 0x60,
+	basereg_msb => 0x60,
+	basereg_lsb => 0x61,
 );
 
 sub exit_superio
@@ -3580,18 +3606,18 @@ sub guess_superio_ld
 	my ($oldldn, $ldn, $addr);
 
 	# Save logical device number
-	outb($addrreg, $superio{logdevreg});
+	outb($addrreg, $standard_superio{logdevreg});
 	$oldldn = inb($datareg);
 
 	for ($ldn = 0; $ldn < 16; $ldn++) {
 		# Select logical device
-		outb($addrreg, $superio{logdevreg});
+		outb($addrreg, $standard_superio{logdevreg});
 		outb($datareg, $ldn);
 
 		# Read base I/O address
-		outb($addrreg, $superio{basereg});
+		outb($addrreg, $standard_superio{basereg_msb});
 		$addr = inb($datareg) << 8;
-		outb($addrreg, $superio{basereg} + 1);
+		outb($addrreg, $standard_superio{basereg_lsb});
 		$addr |= inb($datareg);
 		next unless ($addr & 0xfff8) == $typical_addr;
 
@@ -3601,7 +3627,7 @@ sub guess_superio_ld
 	}
 
 	# Be nice, restore original logical device
-	outb($addrreg, $superio{logdevreg});
+	outb($addrreg, $standard_superio{logdevreg});
 	outb($datareg, $oldldn);
 }
 
@@ -3610,6 +3636,14 @@ sub probe_superio
 {
 	my ($addrreg, $datareg, $chip) = @_;
 	my ($val, $addr);
+	my %superio;
+
+	# Use chip-specific registers if provided
+	if (exists $chip->{regs}) {
+		%superio = overlay_hash(\%standard_superio, $chip->{regs});
+	} else {
+		%superio = %standard_superio;
+	}
 
 	if (exists $chip->{check}) {
 		return 0 unless $chip->{check}($addrreg, $datareg);
@@ -3636,9 +3670,9 @@ sub probe_superio
 	outb($datareg, $chip->{logdev});
 
 	# Get the IO base address
-	outb($addrreg, $superio{basereg});
+	outb($addrreg, $superio{basereg_msb});
 	$addr = inb($datareg);
-	outb($addrreg, $superio{basereg} + 1);
+	outb($addrreg, $superio{basereg_lsb});
 	$addr = ($addr << 8) | inb($datareg);
 
 	# Check the activation register and base address
@@ -3721,9 +3755,9 @@ sub scan_superio
 		}
 
 		# did it work?
-		outb($addrreg, $superio{devidreg});
+		outb($addrreg, $standard_superio{devidreg});
 		$val = inb($datareg);
-		outb($addrreg, $superio{devidreg} + 1);
+		outb($addrreg, $standard_superio{devidreg} + 1);
 		$val = ($val << 8) | inb($datareg);
 		if ($val == 0x0000 || $val == 0xffff) {
 			print "No\n";

[-- Attachment #3: Type: text/plain, Size: 153 bytes --]

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Hans de Goede]

Hi,

On 03/12/2011 02:21 PM, Jean Delvare wrote:
> On Sat, 12 Mar 2011 12:19:09 +0100, Jean Delvare wrote:
>> Anyway, I agree that the non-standard location of the base I/O register
>> isn't supported by sensors-detect yet. We already have some code to
>> handle non-standard SMSC chips, but the SCH5627 is non-standard in a
>> different way. I'll try to come up with something, stay tuned.
>
> Can anyone with access to a SCH5627 chip test the sensors-detect patch
> in attachment?

Thanks for doing this! I had this on my todo. Likewise I've extending
sensors detect for the new Fintek models supported by the f71882fg driver,
as well as updating the devices wiki page on my todo.

Works like a charm:

Trying family `SMSC'...                                     Yes
Found `SMSC SCH5627 Super IO'                               Success!
     (address 0x580, driver `sch5627')

(and yes that is the correct io address).

Regards,

Hans

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] Driver for SMSC SCH5627 needed

[Jean Delvare]

On Sun, 13 Mar 2011 13:55:02 +0100, Hans de Goede wrote:
> On 03/12/2011 02:21 PM, Jean Delvare wrote:
> > On Sat, 12 Mar 2011 12:19:09 +0100, Jean Delvare wrote:
> >> Anyway, I agree that the non-standard location of the base I/O register
> >> isn't supported by sensors-detect yet. We already have some code to
> >> handle non-standard SMSC chips, but the SCH5627 is non-standard in a
> >> different way. I'll try to come up with something, stay tuned.
> >
> > Can anyone with access to a SCH5627 chip test the sensors-detect patch
> > in attachment?
> 
> Thanks for doing this! I had this on my todo. Likewise I've extending
> sensors detect for the new Fintek models supported by the f71882fg driver,
> as well as updating the devices wiki page on my todo.

I didn't mean to step on your toes, simply I was under the impression
that nobody but me was willing to touch sensors-detect's core code.

Thanks for keeping sensors-detect and wiki/Devices up-to-date, this is
very appreciated.

> Works like a charm:
> 
> Trying family `SMSC'...                                     Yes
> Found `SMSC SCH5627 Super IO'                               Success!
>      (address 0x580, driver `sch5627')
> 
> (and yes that is the correct io address).

Thanks for testing! I've just committed the patch to SVN.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors