Home » Mailing list archives » linux-kernel » 2008 » April » 28

Re: i2c or eeprom enumeration problem

Previous thread: [patch 1/2] Immediate Values - jump liveliness by Mathieu Desnoyers on Sunday, April 27, 2008 - 11:34 pm. (1 message)

Next thread: [POWERPC] Use __always_inline for xchg* and cmpxchg* by Paul Mackerras on Monday, April 28, 2008 - 12:44 am. (1 message)
[view original message]
From: Ulrich Drepper <drepper@...>
To: Linux Kernel <linux-kernel@...>
Subject: i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 12:20 am

-----BEGIN PGP SIGNED MESSAGE-----

Hash: SHA1


On a NUMA system the eeprom interface in /sys


  /sys/bus/i2c/drivers/eeprom/*


do not list all the DRAM eeproms.  Only the DRAM from one node is

listed.  I have a 4 socket system with all 16 banks filled and see only

4 entries.


My suspicion (without looking at any code) is that the list of i2c

devices with eeproms is collected once and this happens only on one CPU.

 I sees not to be dynamic since when I read the files with taskset

restricting execution on certain sockets and cores the result doesn't

change.


Is this known or expected (I hope not the latter)?  Where should I look

at?  I assume that's in the i2c code?


- --

➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖

-----BEGIN PGP SIGNATURE-----

Version: GnuPG v1.4.7 (GNU/Linux)


iD8DBQFIFVCa2ijCOnn/RHQRAhA6AJwN9Mc7HzmLM4EgWunC9eucPsDcTwCgtExK

rsz/RJMFHQMB+zIX/WuE2M8=

=OQfT

-----END PGP SIGNATURE-----

--
[view original message]
From: Andrew Morton <akpm@...>
To: Ulrich Drepper <drepper@...>
Cc: Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 8:05 am

Let's cc the i2c list.

--
[view original message]
From: Jean Delvare <khali@...>
To: Andrew Morton <akpm@...>
Cc: Ulrich Drepper <drepper@...>, Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: [i2c] i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 8:15 am

Motherboard manufacturer and model?


Most likely the SMBus is multiplexed and each CPU gets it's own SMBus

segment, as was seen on the Tyan S4882. The EEPROMs you see are the

ones on the segment which happens to be active at boot time.


The i2c subsystem (still) doesn't support that properly, although we

have a quirk for the Tyan S4882 (see

drivers/i2c/busses/i2c-amd756-s4882.c).


--

Jean Delvare

--
[view original message]
From: Ulrich Drepper <drepper@...>
To: Jean Delvare <khali@...>
Cc: Andrew Morton <akpm@...>, Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: [i2c] i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 10:16 am

-----BEGIN PGP SIGNED MESSAGE-----

Hash: SHA1


Tyan S4985.  It's a quad socket, quad core AMD system.


- --

➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖

-----BEGIN PGP SIGNATURE-----

Version: GnuPG v1.4.7 (GNU/Linux)


iD8DBQFIFdxS2ijCOnn/RHQRAnnuAJ4oMLsIW3U1cWZuHCubPZVWvOTcNwCfY4NV

A24BWrZf19VkKnX3QKcS3qc=

=E3Vp

-----END PGP SIGNATURE-----

--
[view original message]
From: Jean Delvare <khali@...>
To: Ulrich Drepper <drepper@...>
Cc: Andrew Morton <akpm@...>, Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 10:21 am

OK. Maybe it uses the same trick as the S4882... Can you get the

information from Tyan?


Please provide the output of i2cdetect for all the SMBus channels.


--

Jean Delvare

--
[view original message]
From: Ulrich Drepper <drepper@...>
To: Jean Delvare <khali@...>
Cc: Andrew Morton <akpm@...>, Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 11:27 am

-----BEGIN PGP SIGNED MESSAGE-----

Hash: SHA1


I don't have a special link to Tyan but I can try.  What information do


[Strange, the lm_sensors package in F9 doesn't have i2cdetect anymore.

Anything replacing it?]


I see two busses:


  Installed I2C busses:

    i2c-0       smbus           SMBus nForce2 adapter at a000

    i2c-1       smbus           SMBus nForce2 adapter at a040


Using -a to get all the info on those busses I see this:


Bus 0:


     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f

00: -- -- -- -- -- -- -- -- 08 -- -- -- -- -- -- --

10: -- -- -- -- -- -- -- -- 18 -- -- -- -- -- -- --

20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

50: UU UU UU UU -- -- -- -- -- -- -- -- -- -- -- --

60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --


Bus 1:


     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f

00: -- -- -- -- -- -- -- -- 08 -- -- -- -- -- -- --

10: -- -- -- -- -- -- -- -- -- 19 -- -- -- -- -- --

20: -- -- -- -- -- -- -- -- -- -- -- -- -- 2d -- --

30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

40: -- -- -- -- -- -- -- -- 48 49 -- -- -- -- -- --

50: -- UU -- -- -- -- -- -- -- -- -- -- -- -- -- --

60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --


This motherboard uses the NVidia nForce n4250QE chipset.


- --

➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖

-----BEGIN PGP SIGNATURE-----

Version: GnuPG v1.4.7 (GNU/Linux)


iD8DBQFIFezP2ijCOnn/RHQRAmd2AJ4gAQWzPhjBylM8g0Tzrw4L88xy0ACghI9q

Za0TJOrHDSlMzq8S2DShh4o=

=aCCS

-----END PGP SIGNATURE-----

--
[view original message]
From: Jean Delvare <khali@...>
To: Ulrich Drepper <drepper@...>
Cc: Andrew Morton <akpm@...>, Linux Kernel <linux-kernel@...>, <i2c@...>
Subject: Re: i2c or eeprom enumeration problem
Date: Monday, April 28, 2008 - 12:53 pm

Hi Ulrich,


Basically, whether the SMBus is multiplexed or not, and if it is, what

is the SMBus topology (which mux chip is used and how to control it,


Yes, the i2c tools (i2cdetect, i2cdump, the eeprom decoding scripts,

etc.) now live in their separate package named i2c-tools:

http://www.lm-sensors.org/wiki/I2CTools

As far as I know, Hans de Goede packaged it in Fedora already.


SPD for 4 memory modules at 0x50 - 0x53. 0x18 could be a multiplexer,

that's the address that was used by the 8-channel multiplexer on the


Apparently you have some Winbond hardware monitoring chip at 0x2d +

0x48 + 0x49. You might try sensors-detect.


There seems to be an extra EEPROM at 0x51, I'm curious what it is...

Presumably not an SPD.


No idea what is at 0x19, it could be a multiplexer but I'd be

surprised, as it doesn't make much sense to multiplex both SMBus


We won't be able to re-use the S4882 code as is, as that one was using a

completely different chip (AMD).


If you are willing to do some experiments with the board, I can give

you commands to test (you'll need i2cdetect and i2cset). But maybe

you'll prefer to wait to have additional information from Tyan first.

Let me know.


--

Jean Delvare

--
Previous thread: [patch 1/2] Immediate Values - jump liveliness by Mathieu Desnoyers on Sunday, April 27, 2008 - 11:34 pm. (1 message)

Next thread: [POWERPC] Use __always_inline for xchg* and cmpxchg* by Paul Mackerras on Monday, April 28, 2008 - 12:44 am. (1 message)