Skip to contentSkip to conversation

More on the 760MP

DISCUSSION Willem Riede 5 participants 6 messages Oct 14, 2001 – Oct 16, 2001 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

I'm not having much luck with this :-(

I also have a Tyan Tiger MP, and have built a 2.4.10-ac12 based kernel for it. Actually, I modified the Rawhide kernel SRPM to use linux-2.4.10 and the ac12 patch, as I run ext3 and software raid installed by RedHat's Roswell beta, and I need to stay compatible with that; then made and installed the athlon-smp RPM.

It contains 2.6.1 for i2c, and I generated the patch for lm_sensors 2.6.1 to replace the older patch for lm_sensors in the SRPM.

The i2c-amd756 module hangs my kernel too, but you seem to have it work without it? I don't get anything detected without i2c-amd756. This is what the log messages say (but I have to reboot to read them):

Oct 14 16:17:16 linnie kernel: i2c-amd756.o version 2.6.1 (20010830)
Oct 14 16:17:16 linnie kernel: i2c-dev.o: Registered 'SMBus AMD7X6 adapter
at 80e0' as minor 1
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D chip] registered
to adapter [SMBus AMD7X6 adapter at 80e0](pos. 0).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]
registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 1).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]

registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 2).

Or is it this module (i2c-amd756) that you say must be included in the kernel to make 'it' work (English is so ambiguous :-))?

One more question if you don't mind: do you (still) have to first read the sensors in the BIOS before booting? If I do that, for some reason, I don't get my grub boot screen after exiting setup, instead the board tries to boot off the lan; I ctrl-atl-del out of that, and then things are back to normal, but that soft reset probably undid whatever reading the sensors in the bios changed :-(
Any other settings in the bios that matter?

Thanks, Willem Riede.

Quoted 73 lines · On 2001.10.09 18:20 German Gomez Garcia wrote:
Quoted 5 lines · Dan Hollis
Quoted 4 lines · German Gomez Garcia
Quoted 3 lines · Dan Hollis
Quoted 2 lines · German Gomez Garcia

it appears in the /proc/cmdline that message stills apears. Also I'm unable to get correct readings with lm_sensors (CVS),

I've been

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

enable to detect the w83781d chip using the i2c-amd756 SMbus but

half of

Quoted 5 lines
Quoted 3 lines
Quoted 2 lines
Quoted 1 line

the times the kernel hangs up when loading this module.

1) You need to enable ACPI in bios for sensors to work.

What about the kernel? must I enable it there too?

No. You don't need it in kernel.

Well, I must include it in the kernel to make it work, if I don't
it detects a w83627hf instead of the w83782d, and it is unable to set it up on the bus:

i2c-proc.o version 2.6.1 (20010825)
w83781d.o version 2.6.1 (20010830)
i2c-core.o: driver W83781D sensor driver registered. i2c-core.o: client [W83627HF chip] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 0).
i2c-core.o: client [W83627HF subclient] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 1).
i2c-core.o: client [W83627HF subclient] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 2).
i2c-core.o: client [W83782D chip] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 3).
w83781d.o: Subclient 0 registration at address 0x49 failed.

then I get incorrect readings for the every sensors 1 goes up to
77º, 2 goes down to 12º, and 3 goes really down to 0º

If I include it in the kernel I get the following:

i2c-proc.o version 2.6.1 (20010825)
w83781d.o version 2.6.1 (20010830)
i2c-core.o: driver W83781D sensor driver registered. i2c-core.o: client [W83782D chip] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 0).
i2c-core.o: client [W83782D subclient] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 1).
i2c-core.o: client [W83782D subclient] registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 2).

And after changing the type of the sensor to 2 (3904 transistor)
it works perfectly, no temp offset, at least it reports the same than the BIOS (well I cannot be sure of that, but it doesn't differ more than one or two degrees).

Yeaahh!! that's outstanding online support, anybody still think
that paid support is better than friendly penguins? :-D
I'll start burn-testing this box now that I have some way to
check temperature.

Regards and great thanks!

- german

-------------------------------------------------------------------------
German Gomez Garcia | Send email with "SEND GPG KEY" as subject
<[email protected]> | to receive my GnuPG public key.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#2 lore raw
Quoted 23 lines · Willem Riede

I'm not having much luck with this :-(

I also have a Tyan Tiger MP, and have built a 2.4.10-ac12 based kernel for it. Actually, I modified the Rawhide kernel SRPM to use linux-2.4.10 and the ac12 patch, as I run ext3 and software raid installed by RedHat's Roswell beta, and I need to stay compatible with that; then made and installed the athlon-smp RPM.

It contains 2.6.1 for i2c, and I generated the patch for lm_sensors 2.6.1 to replace the older patch for lm_sensors in the SRPM.

The i2c-amd756 module hangs my kernel too, but you seem to have it work without it? I don't get anything detected without i2c-amd756. This is what the log messages say (but I have to reboot to read them):

Oct 14 16:17:16 linnie kernel: i2c-amd756.o version 2.6.1 (20010830)
Oct 14 16:17:16 linnie kernel: i2c-dev.o: Registered 'SMBus AMD7X6 adapter
at 80e0' as minor 1
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D chip] registered
to adapter [SMBus AMD7X6 adapter at 80e0](pos. 0).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]
registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 1).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]
Quoted 4 lines

registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 2).

Or is it this module (i2c-amd756) that you say must be included in the kernel to make 'it' work (English is so ambiguous :-))?

No, you can compile it as a module, you need to insert manually it
(modprobe) or better modify modules.conf so it gets inserted automatically when loading i2c. The problem with the machine hanging up was resolved as Dan Hollis suggested comenting the initialization routine (I mean the content of the routine, the routine must exist).

Quoted 7 lines

One more question if you don't mind: do you (still) have to first read the sensors in the BIOS before booting? If I do that, for some reason, I don't get my grub boot screen after exiting setup, instead the board tries to boot off the lan; I ctrl-atl-del out of that, and then things are back to normal, but that soft reset probably undid whatever reading the sensors in the bios changed :-(
Any other settings in the bios that matter?

Uhmm, I don't know if reading the sensors in the BIOS would
make it work better, but I must make a sensors -s before reading, if not I'll get that 70ºC ... the same that you get when you enter the BIOS until you press a key.

It's important to change the type of thermal sensor to
3904 transistor ( set sensor1 2 ....) if not you won't get the right temperatures.

- german

-------------------------------------------------------------------------
German Gomez Garcia | Send email with "SEND GPG KEY" as subject
<[email protected]> | to receive my GnuPG public key.

#3 lore raw
In reply to Willem Riede
Quoted 5 lines · * German Gomez Garcia ([email protected]) wrote:

On Sun, 14 Oct 2001, Willem Riede wrote:

It's important to change the type of thermal sensor to
3904 transistor ( set sensor1 2 ....) if not you won't get the right temperatures.

Can you put up your sensors.conf file online somewhere so I/we can take a look at it? Perhaps your insmod line(s) as well and the patch done to lm-sensors... It'd be very useful to many of us I think as I am trying to go through the steps to get lm-sensors working for my on my Thunder S2462 board.

Thanks!

Stephen

#4 lore raw
Quoted 7 lines · Stephen Frost
Quoted 5 lines · * German Gomez Garcia ([email protected]) wrote:

On Sun, 14 Oct 2001, Willem Riede wrote:

It's important to change the type of thermal sensor to
3904 transistor ( set sensor1 2 ....) if not you won't get the right temperatures.

Quoted 5 lines

Can you put up your sensors.conf file online somewhere so I/we can take a look at it? Perhaps your insmod line(s) as well and the patch done to lm-sensors... It'd be very useful to many of us I think as I am trying to go through the steps to get lm-sensors working for my on my Thunder S2462 board.

You can get everything at:

http://www.piraos.com/~german/sensors

Sorry for the delay but the server is in fact in the dual athlon,
and I'm recompiling everything.

Regards,

- german

-------------------------------------------------------------------------
German Gomez Garcia | Send email with "SEND GPG KEY" as subject
<[email protected]> | to receive my GnuPG public key.

#5 lore raw
In reply to Stephen Frost

On Mon, 2001-10-15 16:01:34 +0200, German Gomez Garcia <[email protected]> wrote in message <[email protected]>:

[lmsensor config for Dual Athlon board by Tyan]

Quoted 3 lines

You can get everything at:

http://www.piraos.com/~german/sensors

Thanks, but sensors.conf seems to be broken. The file I downloaded ended with:
-----------------------------------------------

    set in6_min -5 * 0.95
    set in6_max -5 * 1.05
    set in7_min 5 * 0.95
    set in7_ma

-----------------------------------------------

Btw., where do you have the "compute" statements taken from?

MfG, JBG

Signature · 1 line
Jan-Benedict Glaw . [email protected] . +49-172-7608481
#6 lore raw
In reply to a message not in the archive
Quoted 28 lines · Willem Riede

I'm not having much luck with this :-(

I also have a Tyan Tiger MP, and have built a 2.4.10-ac12 based kernel for it. Actually, I modified the Rawhide kernel SRPM to use linux-2.4.10 and the ac12 patch, as I run ext3 and software raid installed by RedHat's Roswell beta, and I need to stay compatible with that; then made and installed the athlon-smp RPM.

It contains 2.6.1 for i2c, and I generated the patch for lm_sensors 2.6.1 to replace the older patch for lm_sensors in the SRPM.

The i2c-amd756 module hangs my kernel too, but you seem to have it work without it? I don't get anything detected without i2c-amd756. This is what the log messages say (but I have to reboot to read them):

Oct 14 16:17:16 linnie kernel: i2c-amd756.o version 2.6.1 (20010830)
Oct 14 16:17:16 linnie kernel: i2c-dev.o: Registered 'SMBus AMD7X6 adapter
at 80e0' as minor 1
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D chip] registered
to adapter [SMBus AMD7X6 adapter at 80e0](pos. 0).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]
registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 1).
Oct 14 16:17:16 linnie kernel: i2c-core.o: client [W83782D subclient]

registered to adapter [SMBus AMD7X6 adapter at 80e0](pos. 2).

Or is it this module (i2c-amd756) that you say must be included in the kernel to make 'it' work (English is so ambiguous :-))?

Quoted 7 lines

One more question if you don't mind: do you (still) have to first read the sensors in the BIOS before booting? If I do that, for some reason, I don't get my grub boot screen after exiting setup, instead the board tries to boot off the lan; I ctrl-atl-del out of that, and then things are back to normal, but that soft reset probably undid whatever reading the sensors in the bios changed :-(
Any other settings in the bios that matter?

I've turned on APIC in bios and kernel, set all parallel ports and serial ports to auto in BIOS and removed support parllel ports in the kernel. I'm now able to insert the drivers unpatched from lm_sensors cvs-version.

Seems like it's the autoprobing in the parallel driver that messes it up.

Signature · 1 line
carpe noctem