> So I will just include what I know and what works perfect for all mbm > user so far: > > - before all detections I do: IOByte($4E,$00,IOWrite); so set any > bank there might be to 00 We don't do that. Maybe we should consider doing so. I guess this makes identification easier. Are there any side effects? On the other hand, our policy is that our detection script should never write to the chips. > AS99127F_1 & AS99127F_2 & ASB100: > - scan smbus between $29 and $2F We use the full original range ($20 to $2F). What makes you think the range for the as99127f/asb100 is different? Shouldn't it be $28 to $2F (3-bit address selection, 8 addresses)? Oh, BTW, I don't think we ever saw these chips at any other address than $2d. Did you? > - if register $58 holds $31 then we have an asus (either ASB100 > or AS99127F > - if (IOByte($4F,$0,IORead)=$94) or (IOByte($4F,$0,IORead)=$06) > then ASB100 > - if (IOByte($4F,$0,IORead)=$12) or (IOByte($4F,$0,IORead)=$C3) > then AS99127F_1 That's what we have. > - if (IOByte($4F,$0,IORead)=$5C) or (IOByte($4F,$0,IORead)=$A3) > then AS99127F_2 We did not have this one, I'll add it. > Mozart2: > - scan smbus on $77 > - if register $58 holds $56 or $10 then we are talking Mozart2 > - of the mozart there are 3 types: > - //Asus ASM58 Mozart-2 $58=$56 $4E=$94 $4F=$36 > - //Asus AS2K129R Mozart-2 $58=$56 $4E=$94 $4F=$06 > - //Asus XXXXXXXX Mozart-2 $58=$10 $4E=$5C $4F=$A3 > - you can handle all 3 the exact same way!!!!! We have absolutely no support for this one, but we could consider adding at least detection. It's great that the three of them are handled the same way. > so that is detection, next temp sensors: > > ASB100: > - no diode selection option like the winbonds have > - sensor 1: offset $27 of the sensor chip > - sensor 2 & 3 are the 2 LM75's on the smbus > - sensor 4: offset $17 of the sensor chip > - remark: I noticed that on intel boards sensor 2 is used for the cpu > and 4 is ignored/stuck, on amd boards sensor 4 is the cpu and sensor 2 > is either ignored or a socket temp We don't have temp4 in the driver yet. Could be added. > AS99127F_1 & AS99127F_2: > - no diode selection option like the winbonds have > - they both are the same temperature wise > - sensor 1: offset $27 of the sensor chip > - sensor 2 & 3 are the 2 LM75's on the smbus > > Mozart2: > - no diode selection option like the winbonds have > - sensor 1: offset $27 of the sensor chip > - sensor 2: offset $13 of the sensor chip So you mean the Mozart2 has no LM75s? > and finally the voltages, this is where there is a diffrence between > the 2 AS99127F versions: > > ASB100: > V0V1=$20*0.016 > V0V2=$21*0.016 > V0V3=$22*0.016 > V0V4=$23*0.016*1.68 > V0V5=$24*0.016*3.8 > V0V6=($25*(-0.016))*3.97 > V0V7=($26*(-0.016))*1.666 > > AS99127F_1: > V0=AS99127F Standard 1 > V0V1=$20*0.016 > V0V2=$21*0.016 > V0V3=$22*0.016 > V0V4=$23*0.016*1.68 > V0V5=$24*0.016*3.8 > V0V6=($25*(-0.016))*3.97 > V0V7=($26*(-0.016))*1.503 > > AS99127F_2: > V1=AS99127F Standard 2 > V1V1=$20*0.016 > V1V2=$21*0.016 > V1V3=$22*0.016 > V1V4=$23*0.016*1.68 > V1V5=$24*0.016*3.8 > V1V6=(($25*0.016)-3.6)*5.14+3.6 > V1V7=(($26*0.016)-3.6)*3.14+3.6 Great. As long as the registers are the same, we don't care about the formulae. They can be added to our config file, so this require no driver code change :) > Mozart2: > V0=Mozart2 Standard > V0V1=$20*0.016 > V0V2=255 > V0V3=$22*0.016 > V0V4=$23*0.016*1.68 > V0V5=$24*0.016*4 > V0V6=255 > V0V7=255 > > I have send a lot of info in the last month to lm_sensor users, so > it's not always clear to me who is asking this for lm_sensor or for > himselves. This time it was for the team. I don't know about the other times. I'm really astonished that our users ask you about that. These details are only interesting to us as the driver writers. Users shouldn't really matter about them. > so if possible maybe post this somewhere on your site/forum/anything > so I can direct requests from lm_sensor users to it. I think of including your detailed expanations into our w83781d driver doc, which is available online at: http://www2.lm-sensors.nu/~lm78/cvs/lm_sensors2/doc/chips/w83781d Do you agree? I would eventually re-edit a bit to make it more like a doc. > The Asus boards are the most popular so I can say for 99.999% that > the above is correct, there "might" be a diode selection possible but > never used by Asus so far. and believe me, if it's wrong people know > my e-mail adres very well ;) Precisely, that is where the current thread is supposed to be about. I don't know if you support w83781d's (and clones such as the asus chips we are talking about right now) PWM capabilities. We do. These registers are at 0x5A and 0x5B (bank 0). I tried using PWM on my board, it didn't work but I noticed that on my as99127f (version 1) the register at 0x5B seem to be a temperature type selector. Bits 1, 3 and 5 (starting with LSB=0) could be temp type selector for temp1, temp2 and temp3 respectively. I don't know how useful the information can be, but that's something I wanted to share with you. > please let me know if this was usefull and if you posted it so I have > the url :) This was very, *very* useful. Thanks *a lot* for your help and time. -- Jean Delvare http://www.ensicaen.ismra.fr/~delvare/
