Bug 13338 - W83627DHG-A is not recognized
Summary: W83627DHG-A is not recognized
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Hardware Monitoring (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Jean Delvare
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-05-18 13:29 UTC by Frederic MASSOT
Modified: 2009-06-17 09:21 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.29
Subsystem:
Regression: No
Bisected commit-id:


Attachments
Dmesg (25.17 KB, text/plain)
2009-05-18 13:29 UTC, Frederic MASSOT
Details
Kernel config (41.94 KB, text/plain)
2009-05-18 13:29 UTC, Frederic MASSOT
Details
sensors-detect output (3.99 KB, text/plain)
2009-05-18 13:30 UTC, Frederic MASSOT
Details
Boot with w83627ehf.force_id=0xa020 (25.35 KB, text/plain)
2009-05-19 08:18 UTC, Frederic MASSOT
Details
w83627ehf: Add W83627DHG-P support (3.90 KB, patch)
2009-05-19 20:38 UTC, Jean Delvare
Details | Diff

Description Frederic MASSOT 2009-05-18 13:29:15 UTC
Created attachment 21407 [details]
Dmesg

Hi,

I have a motherboard Asus P5KPL-AM with a Winbond W83627DHG-A hardware monitor and it is not recognized by the module w83627ehf.

In attachment, dmesg, kernel config and sensors-detect output.

Regards.
Comment 1 Frederic MASSOT 2009-05-18 13:29:48 UTC
Created attachment 21408 [details]
Kernel config
Comment 2 Frederic MASSOT 2009-05-18 13:30:22 UTC
Created attachment 21409 [details]
sensors-detect output
Comment 3 Jean Delvare 2009-05-18 13:59:00 UTC
Please double-check and report the top marking of the chip (128-pin chip most probably in one corner of the board.)

Your chip has a device ID which is very different from what we know for the W83627DHG. I'll ask my Winbond/Nuvoton contact about it.
Comment 4 Frederic MASSOT 2009-05-18 15:28:16 UTC
Yes, it is a chip that is in a corner of the motherboard. The chip has pin on all four sides.

On the chip there are three lines:

WINBOND
W83627DHG-A
909G9C2908B003-ZZFA
Comment 5 Jean Delvare 2009-05-19 07:31:00 UTC
I received information from Nuvoton. The chip is branded as W83627DHG-P, this is a new incarnation of the W83627DHG. As far as I can see (but I did not read the full datasheet yet) the hardware monitoring block is essentially the same, except for an extra fan speed control mode.

So you might want to try booting with parameter "w83627ehf.force_id=0xa020", so that the w83627ehf treats your chip as a W83627DHG. It should work fine. Please report how it goes.
Comment 6 Frederic MASSOT 2009-05-19 08:16:24 UTC
OK, it works with boot parameter "w83627ehf.force_id=0xa020".

In dmesg there :

w83627ehf: Found W83627DHG chip at 0x290
ACPI: I/O resource w83627ehf [0x295-0x296] conflicts with ACPI region HWRE [0x290-0x299]
ACPI: Device needs an ACPI driver


And in syslog :

May 19 10:11:43 ithaqua sensord: sensord started
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: +12V: +12.36 V (min = +8.24 V, max = +1.48 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: AVCC: +3.36 V (min = +3.06 V, max = +0.88 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: 3VCC: +3.36 V (min = +2.98 V, max = +3.18 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: in4: +1.13 V (min = +1.77 V, max = +1.83 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: in5: +1.61 V (min = +1.24 V, max = +1.38 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: +5V: +3.07 V (min = +3.48 V, max = +4.56 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: VSB: +3.39 V (min = +2.80 V, max = +2.54 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: VBAT: +3.26 V (min = +1.73 V, max = +1.49 V) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: Case Fan: 1785 RPM (min = 1906 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: CPU Fan: 1117 RPM (min = 1757 RPM, div = 8) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: Aux Fan: 0 RPM (min = 73 RPM, div = 128) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: fan4: 0 RPM (min = 15340 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: fan5: 0 RPM (min = 168750 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord: Sensor alarm: Chip w83627dhg-isa-0290: Sys Temp: 33.0 C (limit = -59.0 C, hysteresis = -67.0 C) [ALARM]
May 19 10:11:43 ithaqua sensord: Chip: coretemp-isa-0000
May 19 10:11:43 ithaqua sensord: Adapter: ISA adapter
May 19 10:11:43 ithaqua sensord:   Core 0: 46.0 C
May 19 10:11:43 ithaqua sensord: Chip: coretemp-isa-0001
May 19 10:11:43 ithaqua sensord: Adapter: ISA adapter
May 19 10:11:43 ithaqua sensord:   Core 1: 44.0 C
May 19 10:11:43 ithaqua sensord: Chip: w83627dhg-isa-0290
May 19 10:11:43 ithaqua sensord: Adapter: ISA adapter
May 19 10:11:43 ithaqua sensord:   VCore: +1.22 V (min = +0.00 V, max = +1.74 V)
May 19 10:11:43 ithaqua sensord:   +12V: +12.36 V (min = +8.24 V, max = +1.48 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   AVCC: +3.36 V (min = +3.06 V, max = +0.88 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   3VCC: +3.36 V (min = +2.98 V, max = +3.18 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   in4: +1.13 V (min = +1.77 V, max = +1.83 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   in5: +1.61 V (min = +1.24 V, max = +1.38 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   +5V: +3.07 V (min = +3.48 V, max = +4.56 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   VSB: +3.39 V (min = +2.80 V, max = +2.54 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   VBAT: +3.26 V (min = +1.73 V, max = +1.49 V) [ALARM]
May 19 10:11:43 ithaqua sensord:   Case Fan: 1785 RPM (min = 1906 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord:   CPU Fan: 1117 RPM (min = 1757 RPM, div = 8) [ALARM]
May 19 10:11:43 ithaqua sensord:   Aux Fan: 0 RPM (min = 73 RPM, div = 128) [ALARM]
May 19 10:11:43 ithaqua sensord:   fan4: 0 RPM (min = 15340 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord:   fan5: 0 RPM (min = 168750 RPM, div = 4) [ALARM]
May 19 10:11:43 ithaqua sensord:   Sys Temp: 33.0 C (limit = -59.0 C, hysteresis = -67.0 C) [ALARM]
May 19 10:11:43 ithaqua sensord:   CPU Temp: 36.0 C (limit = 80.0 C, hysteresis = 75.0 C)
May 19 10:11:43 ithaqua sensord:   AUX Temp: 33.0 C (limit = 80.0 C, hysteresis = 75.0 C)
May 19 10:11:43 ithaqua sensord:   cpu0_vid: +0.000 V
Comment 7 Frederic MASSOT 2009-05-19 08:18:13 UTC
Created attachment 21425 [details]
Boot with w83627ehf.force_id=0xa020
Comment 8 Madhu 2009-05-19 08:23:49 UTC
(In reply to comment #5)
> I received information from Nuvoton. The chip is branded as W83627DHG-P, this
> is a new incarnation of the W83627DHG. As far as I can see (but I did not
> read
> the full datasheet yet) the hardware monitoring block is essentially the
> same,
> except for an extra fan speed control mode.
> 
> So you might want to try booting with parameter "w83627ehf.force_id=0xa020",
> so
> that the w83627ehf treats your chip as a W83627DHG. It should work fine.
> Please
> report how it goes.

Thanks Jean, I will also try this as I have the same W83627DHG-P on my DG41TY
motherboard.
Comment 9 Jean Delvare 2009-05-19 08:26:04 UTC
(In reply to comment #6)
> OK, it works with boot parameter "w83627ehf.force_id=0xa020".
> 
> In dmesg there :
> 
> w83627ehf: Found W83627DHG chip at 0x290
> ACPI: I/O resource w83627ehf [0x295-0x296] conflicts with ACPI region HWRE
> [0x290-0x299]
> ACPI: Device needs an ACPI driver

This means that ACPI wants you off the hardware monitoring chip, as it must be using it already. Bad luck. Starting with kernel 2.6.30 you will no longer be able to use the device by default (and you already shouldn't use it), even if we add support for it to the w83627ehf driver.

In kernel 2.6.30 you might have some luck with the new asus_atk0110 driver though.
Comment 10 Frederic MASSOT 2009-05-19 09:03:15 UTC
If I understand you correctly, the hardware monitoring chip must be accessed by ACPI and not directly?
Comment 11 Jean Delvare 2009-05-19 09:07:22 UTC
On your board, yes, exactly.
Comment 12 Madhu 2009-05-19 11:05:01 UTC
Jean, just want some clarification here, 

1. If the motherboard supports the acpi to the sensor chip, we should not use the sensor driver.

2. If that doesnt, we can use the sensor driver.


Also from the dmesg log attached in this bug, I see the acpi supports the CPU cores as cooling device 1 and 2, in that case, 'coretemp' sensor driver will not conflict with acpi? Is that because 'coretemp' uses MSR (machine specific register) not the IO resource to read the thermal values?
Comment 13 Jean Delvare 2009-05-19 11:14:00 UTC
Madhu, you are correct about ACPI vs. native driver for the W83627DHG chip.

As for your second question, on this system ACPI knows of the CPU cores as _cooling_ devices, not thermal sensors. This simply means that ACPI knows how to scale down the frequency of the CPU cores, and/or throttle them down, and/or put them in sleep states. This is all about passive cooling, not temperature measurement. So this is totally different from what the coretemp driver does.

That being said, I indeed believe that MSR-based temperature readings are immune to problems caused by concurrent I/O access, and as such the coretemp driver should always be usable regardless of what ACPI does.
Comment 14 Madhu 2009-05-19 11:42:56 UTC
thanks for clarifying that, Jean.
Comment 15 Jean Delvare 2009-05-19 20:38:32 UTC
Created attachment 21437 [details]
w83627ehf: Add W83627DHG-P support

Add support for the new incarnation of the Winbond/Nuvoton W83627DHG chip known as W83627DHG-P. It is basically the same as the original W83627DHG with an additional automatic can speed control mode (not supported by the driver yet.)
Comment 16 Frederic MASSOT 2009-05-25 07:53:13 UTC
Hi,

Will you push the patch to the 2.6.30-rc8 or for the next merge window?

Regards.
Comment 17 Jean Delvare 2009-05-25 13:35:10 UTC
Next merge window, it's too late for 2.6.30.

Note You need to log in before you can comment on or make changes to this bug.