On Fri, Nov 25, 2011 at 05:22:41AM -0500, Eddi De Pieri wrote: > Hi Karsten, > > about the complete power cycle issue: > > While testing I had same problem like you... I had to remove power > cord to get my server back... > > Now I tried a soft "reboot" and I get any issue.. > > So my patch seems to be fully working and I'd like to know how to > commit into mainstream. > 1) Don't top-post. Your e-mail may otherwise be ignored. 2) There is conflicting information if your patch is causing trouble or not. Please clarify if and when a reboot fails, and what you changed to solve the problem. 3) This is an I2C multiplexer. Provide a patch which introduces an I2C multiplexer driver. 4) Read and follow Documentation/hwmon/submitting-patches for submitting your patch. Thanks, Guenter > I think may be useful for other pc based on same chipset like ours. > > Regards, > > Eddi > > On Mon, Nov 21, 2011 at 12:03 AM, Karsten de Freese <karsten@xxxxxxxxxxx> wrote: > > hello Eddi, > > > > a less-than-welcome update: > > > > at least when the three modules i2c-piix4.ko, i2c-piix4-n36l.ko, w83795.ko > > are loaded, a restart (sudo shutdown -r now) of my N36L fails, the system > > does not come up until I do a complete power cycle. I'll have to look more > > closely at this, once I have physical access to the box again. > > > > regards - > > > > Karsten > > > > Am 20.11.2011 14:38, schrieb Karsten de Freese: > >> > >> hello Eddi, > >> > >> Success! > >> > >> Thanks for the detailed steps. This all works as described, at least as > >> far as the 'sensors' output is concerned. With the respective modules, I get > >> all the measurements you're listing! (See attached.) > >> > >> 'i2cdetect -l' will however not produce any output for whatsoever reason. > >> > >> The next thing I'd like to find out is how to load the modules > >> automatically in a clean / kernel-upgrade-proof way.. > >> > >> thanks and regards - > >> > >> Karsten > >> > >> > >> Am 18.11.2011 11:09, schrieb Eddi De Pieri: > >>> > >>> I suggest you: > >>> > >>> > >>> mkdir ~/piix4-n36l > >>> cp > >>> [yourlinuxtree]/usr/src/linux-source-2.6.32/drivers/i2c/busses/i2c-piix4.c > >>> ~/piix4-n36l/ > >>> cp ~/Downloads/Makefile ~/piix4-n36l/ > >>> cd ~/piix4-n36l/ > >>> > >>> patch -p1< proliant-microserver-n36l.diff > >>> make > >>> you will have your new module inside ~/piix4-n36l/ > >>> > >>> rmmod i2c-piix4 > >>> insmod ./i2c-piix4.ko > >>> insmod ./i2c-piix4-n36l.ko > >>> > >>> insmod w83795 (you should already build separately) > >>> > >>> sensors > >>> > >>> regards Eddi. > >>> > >>> I hope an answer from Jean... I'd like if this patch may be inserted > >>> in mainstream linux with the necessary adjustment.. > >>> > >>> On Thu, Nov 17, 2011 at 11:39 PM, Karsten de Freese<karsten@xxxxxxxxxxx> > >>> wrote: > >>>> > >>>> hello Eddi, > >>>> > >>>> I'm ready for two questions ;-) > >>>> > >>>> .diff / patch looks simple, and I also found i2c-amd756-s4882.c - > >>>> however: > >>>> - the diff file actually describes the patch to be on top of > >>>> i2c-piix4.c? > >>>> - if I'd compile the patched file, would that give me a standalone > >>>> driver? > >>>> > >>>> thanks for your help - > >>>> > >>>> Karsten > >>>> > >>>> On 17.11.2011 22:25, Karsten de Freese wrote: > >>>>> > >>>>> hi Eddi, > >>>>> > >>>>> thanks very much - I had almost forgotten about the topic, but stumbled > >>>>> across it again a few days ago, so that's very interesting. > >>>>> I'm not yet familiar with .diff patches, and may welcome a little help. > >>>>> However, let me try to find out on my own first.. > >>>>> > >>>>> thanks& regards - > >>>>> > >>>>> Karsten > >>>>> > >>>>> On 16.11.2011 17:18, Eddi De Pieri wrote: > >>>>>> > >>>>>> Finally I get W83795ADG working on my proliant microserver > >>>>>> > >>>>>> The patch is based on i2c-amd756-s4882.c and adapted for sb800 > >>>>>> southbus. > >>>>>> > >>>>>> Karsten, please try it too... > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -l > >>>>>> i2c-0 smbus SMBus piix4 adapter (SDA0) SMBus > >>>>>> adapter > >>>>>> i2c-1 smbus SMBus piix4 adapter (SDA2) SMBus > >>>>>> adapter > >>>>>> i2c-2 smbus SMBus piix4 adapter (SDA3) SMBus > >>>>>> adapter > >>>>>> i2c-3 smbus SMBus piix4 adapter (SDA4) SMBus > >>>>>> adapter > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# > >>>>>> > >>>>>> yes SDA1 is reserved... so i can't multiplex it > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# sensors > >>>>>> k10temp-pci-00c3 > >>>>>> Adapter: PCI adapter > >>>>>> temp1: +24.5°C (high = +70.0°C, crit = +100.0°C) > >>>>>> > >>>>>> w83795adg-i2c-1-2f > >>>>>> Adapter: SMBus piix4 adapter (SDA2) > >>>>>> in0: +1.02 V (min = +0.00 V, max = +2.05 V) > >>>>>> in1: +1.52 V (min = +0.00 V, max = +2.05 V) > >>>>>> in2: +1.10 V (min = +0.00 V, max = +2.05 V) > >>>>>> in3: +0.89 V (min = +0.00 V, max = +2.05 V) > >>>>>> in12: +3.35 V (min = +0.00 V, max = +6.14 V) > >>>>>> in13: +3.28 V (min = +0.00 V, max = +6.14 V) > >>>>>> fan1: 703 RPM (min = 329 RPM) > >>>>>> temp1: +23.0°C (high = +109.0°C, hyst = +109.0°C) > >>>>>> (crit = +109.0°C, hyst = +109.0°C) sensor = > >>>>>> thermal diode > >>>>>> temp2: +33.2°C (high = +105.0°C, hyst = +105.0°C) > >>>>>> (crit = +105.0°C, hyst = +105.0°C) sensor = > >>>>>> thermal diode > >>>>>> temp5: +14.0°C (high = +39.0°C, hyst = +39.0°C) > >>>>>> (crit = +44.0°C, hyst = +44.0°C) sensor = > >>>>>> thermistor > >>>>>> beep_enable:disabled > >>>>>> > >>>>>> jc42-i2c-0-18 > >>>>>> Adapter: SMBus piix4 adapter (SDA0) > >>>>>> temp1: +20.5°C (low = +0.0°C, high = +0.0°C) ALARM > >>>>>> (crit = +0.0°C, hyst = +0.0°C) ALARM > >>>>>> > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 0 > >>>>>> 0 1 2 3 4 5 6 7 8 9 a b c d e f > >>>>>> 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 10: -- -- -- -- -- -- -- -- UU -- -- -- -- -- -- -- > >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 50: 50 -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 70: -- -- -- -- -- -- -- -- > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 1 > >>>>>> 0 1 2 3 4 5 6 7 8 9 a b c d e f > >>>>>> 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- UU > >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 60: -- 61 -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 70: -- -- -- -- -- -- -- -- > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 2 > >>>>>> 0 1 2 3 4 5 6 7 8 9 a b c d e f > >>>>>> 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- 4c -- -- -- > >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 70: -- -- -- -- -- -- -- -- > >>>>>> > >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 3 > >>>>>> 0 1 2 3 4 5 6 7 8 9 a b c d e f > >>>>>> 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > >>>>>> 70: -- -- -- -- -- -- -- -- > >>>>>> > >>>>>> pay attention that the msleep seems to be really needed... > >>>>>> > >>>>>> Regards, Eddi > >>>>>> > >>>>>> On Tue, Aug 16, 2011 at 9:26 AM, Eddi De Pieri<eddi@xxxxxxxxxxx> > >>>>>> wrote: > >>>>>>> > >>>>>>> Hi Jean... > >>>>>>> > >>>>>>> The bios show fan rpm and temperature. Since my own proliant > >>>>>>> microserver > >>>>>>> lacks of ipmi card, I doubt the bios give me the results using ipmi > >>>>>>> call. > >>>>>>> > >>>>>>> The HP Proliant Microserver chipset (SB820M) have a multiplexed smbus > >>>>>>> like your. > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> http://support.amd.com/it/Embedded_TechDocs/47283_sb820m_ds_pub_2.00.pdf > >>>>>>> > >>>>>>> Here is technical document from amd... > >>>>>>> > >>>>>>> > >>>>>>> at p52/53 they say: > >>>>>>> > >>>>>>> (1) The SDA1 and SCL1 SMBus interface is dedicated for ASF devices > >>>>>>> only. > >>>>>>> It should not be used to > >>>>>>> connect to any other devices. > >>>>>>> (2) There are only two SMBus controllers. The SCL1/SDA1 pair is > >>>>>>> controlled by SMBus controller 1. > >>>>>>> > >>>>>>> > >>>>>>> SCL0/SDA0, SCL2/SDA2, SCL3/SDA3 and SCL4/SDA4 are multiplexed pins > >>>>>>> that > >>>>>>> are all controlled by > >>>>>>> SMBus controller 0, and only 1 pair of those pins can be active at > >>>>>>> any > >>>>>>> time. > >>>>>>> > >>>>>>> I can't understand how to enable 2, 3 and 4 muxed bus.. > >>>>>>> > >>>>>>> > >>>>>>> If you consolidate a solution for i2c-i801.c I hope we can be > >>>>>>> replicated > >>>>>>> for i2c-piix4.c too.. > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> Hi Karsten, > >>>>>>> > >>>>>>> On Fri, 22 Oct 2010 16:04:10 +0200, karsten at maxi-dsl.de wrote: > >>>>>>>> > >>>>>>>> hi Jean, > >>>>>>>>> > >>>>>>>>> The w83795 driver uses the same detection mechanism used in > >>>>>>>>> sensors-detect. So if sensors-detect doesn't see your chip, I am > >>>>>>>>> not > >>>>>>>>> surprised that the driver didn't either. > >>>>>>>>> The full output of sensors-detect is needed to diagnose your > >>>>>>>>> problem > >>>>>>>>> further. It could be that your SMBus controller isn't supported, or > >>>>>>>>> maybe you are supposed to access the chip through IPMI. > >>>>>>>> > >>>>>>>> IPMI is probably the key.. > >>>>>>>> From the sensors-detect output: > >>>>>>>> "Warning: the required module ipmisensors is not currently installed > >>>>>>>> on your system. If it is built into the kernel then it's OK. > >>>>>>>> Otherwise, checkhttp://www.lm-sensors.org/wiki/Devices for > >>>>>>>> driver availability" (full ouput attached) > >>>>>>>> > >>>>>>>> I did see the output before, but didn't really understand the > >>>>>>>> implications (IPMI seems to be an alternative way of accessing the > >>>>>>>> monitoring chips?) > >>>>>>> > >>>>>>> Yes. The strength of IPMI access is that it can be used remotely too, > >>>>>>> for example from a BMC plugged into your system. > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>>> The suggested ipmisensors driver however does not seem to be > >>>>>>>> available, > >>>>>>>> right? > >>>>>>>> (http://lm-sensors.org/wiki/Devices says "(2004-12-12) Port to > >>>>>>>> Linux > >>>>>>>> 2.6 in progress by Yani Ioannou") > >>>>>>> > >>>>>>> That project is essentially dead. At the time being, the best way to > >>>>>>> retrieve IPMI-based sensor values is using "ipmitool sensor". The > >>>>>>> drawback is that this isn't integrated with libsensors, so the > >>>>>>> usually > >>>>>>> > >>>>>>> > >>>>>>> monitoring applications won't work. > >>>>>>> > >>>>>>> If you have a BMC or are otherwise already using IPMI on your > >>>>>>> machine, > >>>>>>> just use ipmitool and you're done. > >>>>>>> > >>>>>>> If not, then it might be worth trying ipmitool once, and then try > >>>>>>> > >>>>>>> > >>>>>>> sensors-detect again. There is a small chance that, once ipmitool > >>>>>>> managed to access the monitoring device, sensors-detect will see it. > >>>>>>> My > >>>>>>> own mainboard which has a W83795ADG chip has a multiplexed SMBus, and > >>>>>>> I > >>>>>>> can only see the W83795ADG in sensors-detect when the multiplexer is > >>>>>>> > >>>>>>> > >>>>>>> set properly. Otherwise I would see the memory modules, which _do_ > >>>>>>> appear in your sensors-detect output. So maybe your board has a > >>>>>>> multiplexed SMBus too. > >>>>>>> > >>>>>>> If the trick doesn't work, you may want to try the new jc42 driver > >>>>>>> > >>>>>>> > >>>>>>> which can monitor the temperature of some DDR3 memory modules. I > >>>>>>> suspect your memory modules have such sensors. A more recent version > >>>>>>> of > >>>>>>> sensors-detect would confirm that: > >>>>>>> http://dl.lm-sensors.org/lm-sensors/files/sensors-detect > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> I also see that sensors-detect suggested to use the k10temp driver, > >>>>>>> which your kernel doesn't have. You can try the standalone flavor of > >>>>>>> the drivers which I maintain at: > >>>>>>> http://khali.linux-fr.org/devel/misc/k10temp/ > >>>>>>> > >>>>>>> > >>>>>>> The accuracy of the CPU internal sensors is limited, but that's still > >>>>>>> better than nothing. > >>>>>>> > >>>>>>> -- > >>>>>>> Jean Delvare > >>>>>>> http://khali.linux-fr.org/wishlist.html > >>>>>>> > >>>>>>> > >>>>> > >>>> > >> > > > > > > _______________________________________________ > lm-sensors mailing list > lm-sensors@xxxxxxxxxxxxxx > http://lists.lm-sensors.org/mailman/listinfo/lm-sensors _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
