[lm-sensors] HowTo Asus P5Q, W83667HG vs ATK0110 - fancontrol

Jean Delvare khali at linux-fr.org
Thu Jun 25 10:31:25 CEST 2009


Hallo Konstantin,

On Wed, 24 Jun 2009 16:23:50 +0200, Konstantin Kletschke wrote:
> Okay, I even moved the w83627ehf.ko and asus_atk0110.ko away and
> rebooted.
> 
> Am 2009-06-24 14:53 +0200 schrieb Jean Delvare:
> 
> > isaset 0x295 0x296 0x4e 0 # set bank to 0
> > isadump 0x295 0x296
> 
>  >isaset 0x295 0x296 0x4e 0
>  >isadump 0x295 0x296
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> 00: 04 ff 04 3d 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 10: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff c5
> 20: 80 dc cf cd d3 ff c3 2c ff 99 ff da 00 06 71 10
> 30: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> 40: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> 50: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> 60: 04 3d 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> 70: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2c ff
> 80: 04 ff 04 3d 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 90: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff c5
> a0: 80 dc cf cd d3 ff c3 2c ff 99 ff da 00 06 71 10
> b0: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> c0: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> d0: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> e0: 04 3d 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> f0: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2c ff
> 
> > isaset 0x295 0x296 0x01 0xaa
> > isadump 0x295 0x296
> 
>  >isaset 0x295 0x296 0x01 0xaa
> Data mismatch, wrote 0xaa, read 0xff back.
>  >isadump 0x295 0x296
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> 00: 04 ff 04 3d 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 10: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff cf
> 20: 80 dc cf ce d3 ff c3 2c ff 99 ff da 00 06 71 10
> 30: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> 40: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> 50: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> 60: 04 3d 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> 70: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2c ff
> 80: 04 ff 04 3d 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 90: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff cf
> a0: 80 dc cf ce d3 ff c3 2c ff 99 ff da 00 06 71 10
> b0: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> c0: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> d0: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> e0: 04 3d 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> f0: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2c ff
> 
> 0x01 seems to be readonly.
> 
> > isaset 0x295 0x296 0x03 0xbb
> > isadump 0x295 0x296
> 
>  >isaset 0x295 0x296 0x03 0xbb
> Data mismatch, wrote 0xbb, read 0x37 back.
>  >isadump 0x295 0x296
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> 00: 04 bb 04 37 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 10: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff d1
> 20: 80 dc cf ce d3 ff c3 2b ff 9d ff da 00 06 71 10
> 30: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> 40: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> 50: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> 60: 04 37 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> 70: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2b ff
> 80: 04 bb 04 37 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 90: 04 99 30 00 00 01 01 3c 43 17 00 00 32 ff ff d1
> a0: 80 dc cf ce d3 ff c3 2b ff 9d ff da 00 06 71 10
> b0: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> c0: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> d0: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> e0: 04 37 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> f0: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2b ff
> 
> 0x03 writes to 0x01...
> 
> > isaset 0x295 0x296 0x11 0xcc
> > isadump 0x295 0x296
> 
>  >isaset 0x295 0x296 0x11 0xcc
>  >isadump 0x295 0x296
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> 00: 04 bb 04 37 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 10: 04 cc 30 00 00 01 01 3c 43 17 00 00 32 ff ff d3
> 20: 80 dc cf ce d3 ff c3 2b ff 9d ff da 00 06 71 10
> 30: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> 40: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> 50: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> 60: 04 37 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> 70: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2b ff
> 80: 04 bb 04 37 00 00 46 20 01 9b 01 01 3c 3c 04 04
> 90: 04 cc 30 00 00 01 01 3c 43 17 00 00 32 ff ff d3
> a0: 80 dc cf ce d3 ff c3 2b ff 9d ff da 00 06 71 10
> b0: f7 94 c8 45 50 b2 8a 13 c1 90 85 01 27 01 01 ff
> c0: 03 00 00 ff ff 00 3f 35 2d 21 20 40 10 95 00 a3
> d0: ff ff 00 ff ff ff 00 80 c1 6f ff ff 19 45 04 05
> e0: 04 37 50 00 01 01 3c ff 0a ff 01 ff ff ff ff 28
> f0: 46 46 46 46 46 37 ff ff ff ff ff 4b 22 00 2b ff
> 
> 0x11 indeed writes to 0x11.
> 
> This reflects somehow how echo into the /sys filesystem reacts.

Indeed. So this is not a bug in the driver, but rather the hardware
misbehaving. I'll ask out contact at Nuvoton about this.

> > If this works fine then that would be a bug in the driver. If it
> > doesn't, that would be a hardware bug.
> 
> What I wonder, is there any success report pwm controlling cpu fan from
> userspace with this W83667HG-A chip?

Support for the W83667HG is still relatively new. Officially this chip
is only supported since kernel 2.6.30 which was released just 2 weeks
ago. So we didn't have much feedback yet.

> Meanwhile I read many reports not achieving this. There are two other
> fan header (3-PIN non PWM) and many people connect their fans there and
> doing DC regulation, even on windows platform (the other pin on some
> p5q-xxx boards are able to do DC regulating there, I don't know if mine
> is able to do yet).

Could as well be that the chip has a silicon bug...

-- 
Jean Delvare




More information about the lm-sensors mailing list