Hello Yuan, Please can you help (and new datasheet would be nice too :) ? I'm afraid that I cant move any further without your help. Thanks Regards Rudolf Daniel Nilsson wrote: > Hi Rudolf, > > On Wed, Nov 30, 2005 at 11:11:00PM +0100, Rudolf Marek wrote: > >>Never give up, never surrender :) Lets focus on SMI. This should disable >>the SMI from this chip. So we can at least eliminate this source. >>OVT# is also unused so this should point if you have HW or SW (bus driver) >>problem. > > > That's the spirit :-) > > >>Please can you try with this sequence, which is same except of new first >>line. >> >>Quoting Jean: >> >>You can use the i2cset program to write to the W83792D chip directly. You >>can try using the following commands in your script, without loading the >>w83792d driver. Let us know if it hangs of not. >> >>i2cset -y 0 0x2c 0x40 0x1 # DISABLE SMI >>i2cset -y 0 0x2c 0x2b 0xff # in0_max >>... >>This should completly disable the SMI. Lets see how it works :) > > > Ok, I've tried this and a few other things. First of all, I can't shut > off SMI which seems really odd. > > The first time a ran this sequence I got this result and the machine > did not hang: > > oden:~# ./set_safe_direct_wo_smi.sh > + i2cset -y 0 0x2c 0x40 0x1 > No size specified (using byte-data access) > Warning - data mismatch - wrote 0x01, read back 0x06 > + i2cset -y 0 0x2c 0x2b 0xff > No size specified (using byte-data access) > Value 0xff written, readback matched > + i2cset -y 0 0x2c 0x2c 0x00 > [all the remaining writes succeeded] > > The second time I got this result and the machine did still not hang: > > oden:~# ./set_safe_direct_wo_smi.sh > + i2cset -y 0 0x2c 0x40 0x1 > No size specified (using byte-data access) > Warning - data mismatch - wrote 0x01, read back 0x03 > + i2cset -y 0 0x2c 0x2b 0xff > No size specified (using byte-data access) > Value 0xff written, readback matched > + i2cset -y 0 0x2c 0x2c 0x00 > [all the remaining writes succeeded] > > Finally the third time the machine just got this far and then hung: > > oden:~# ./set_safe_direct_wo_smi.sh > + i2cset -y 0 0x2c 0x40 0x1 > No size specified (using byte-data access) > > > So according to the datasheet writing 0x01 to this register should > shut off EN_SMI, but that doesn't seem to work since the readback > still shows 0x03 (should be 0x01, right?). I've been trying all sorts > of things to disable this EN_SMI bit but I can't seem to do it. > > For example, writing 0x80 should re-init the whole chip: > > oden:~# i2cset -y 0 0x2c 0x40 0x80 > No size specified (using byte-data access) > Warning - data mismatch - wrote 0x80, read back 0x03 > > But that still has the EN_SMI bit set??? It shouldn't according to my > datasheet at least, that bit should be reset to '0'... > > I've also played around in the BIOS for a while, there seems to be > some king of "MiniBMC" functionality in the BIOS that is logging > overtemp events someplace. That log was filled with events since BIOS > enabled all fan monitors by default. I've shut off all those monitors > and even the whole MiniBMC funtionality but I still get the same hangs > when writing the sensor limits. I honestly don't know if there is some > other device on this bus that is communicating with the w82792d chip > to read these levels as well that might be interfering here. Since > this MiniBMC thing might be of interest I actually wrote Gigabyte > support about this whole issue, I'm keeping my fingers crossed but I > doubt that the question will reach the appropriate people at Gigabyte. > > One more thing I observed; > > oden:~# i2cdump -y 0 0x2c > No size specified (using byte-data access) > 0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef > 00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > 10: 00 00 64 64 bf 00 00 63 00 ff 00 00 f5 00 8f 80 ..dd?..c....?.?? > 20: d3 d4 b9 b9 b8 b8 cd 15 ff ff ff ff 00 96 64 d7 ????????.....?d? > 30: af d9 b1 d9 b1 d7 af ff 00 25 00 f0 f0 f0 2f 39 ???????..%.???/9 > 40: 03 00 00 00 00 ff ff 11 2c 13 40 c3 51 ff 80 5c ?......?,?@?Q.?\ > 50: ff ff ff ff ff ff ff ff 7a 60 ff 11 11 c1 05 7f ........z`.????? > 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ................ > 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ................ > 80: 01 8f 01 8f 00 00 00 00 11 11 ff ff 3c 3c 0a 0a ????....??..<<?? > 90: 00 00 00 01 8f ff 00 00 11 ff 3c 00 00 01 01 ff ...??...?.<..??. > a0: 01 01 01 8f 8f 8f 8f ff 3e e2 00 e0 ff ff 00 00 ???????.>?.?.... > b0: d3 cb ff ff ff 00 e2 b9 ff ff ff f0 f0 f0 ff ff ??....??...???.. > c0: f2 80 02 39 00 3c 00 ff 2c 00 00 5d 00 60 00 ff ???9.<..,..].`.. > d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ................ > e0: 8b 8b 8b 28 3c 50 28 3c 50 28 3c 50 ff ff ff ff ???(<P(<P(<P.... > f0: ff ff 80 ff 00 01 00 00 ff 02 ff 00 00 00 00 ff ..?..?...?...... > > Look at index 49, it says "13". According to the datasheet I have a > revision C chip should have "12". Is this a newer chip then what the > datasheet covers? In this case there might be some confusion on > disabling SMI... > > Yuan, could you offer some help and guidance as to why the EN_SMI bit > can't be turned off and what this revision "13" chip is? My > motherboard is a Gigabyte GA-4MXSV by the way. > > Regards
