Ubuntu Server: Auto Update to Linux 2.6.32.25.27 broke Digitemp

[NDLunchbox]

https://bugs.launchpad.net/ubuntu/+source/digitemp/+bug/650731

I reported the above bug to Ubuntu. The Munin plugins I wrote for my
DS9097U host and sensor stopped working shortly after my logs show the
kernel was upgraded from 2.6.32.24.25 to 2.6.32.25.27.

I tried downloading the latest version and recompiling, but same issue.

[Marc MERLIN]

Looks to me like your serial port isn't working in your new kernel, more a
kernel issue than a digitemp issue.

Can you confirm you still have a working serial port and you can get minicom
to talk to things plugged into it? If not, you'll have to solve that first.

Marc
--
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems & security ....
.... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/

[NDLunchbox]

I'm not using serial, unless this has a built-in serial to usb
converter: http://www.ibuttonlink.com/linkusbsd.aspx
I've been playing around with minicom based on a thread I found here:
http://comments.gmane.org/gmane.comp.file-systems.owfs.devel/6840
And was able to access it by setting the serial port to /dev/ttyUSB0
at 9600baud with hardware flow control - it responds to the " " and
"h" commands. Like the poster above it seems to have locked up after
quitting cleanly... I will have to unplug when I get home since a
reboot didn't seem to help...

[Marc MERLIN]

On Thu, Sep 30, 2010 at 03:42:43PM -0700, NDLunchbox wrote:
> I'm not using serial, unless this has a built-in serial to usb
> converter: http://www.ibuttonlink.com/linkusbsd.aspx

DS9097U is a serial interface, but indeed it's possible to use it behind a
USB to serial converter.

It may be that the usb support changed slightly in a way that the serial
converter is working quite like it did before?

[NDLunchbox]

So I got home and unplugged, reset and reconnected the device and
tried communicating with minicom and it seems to be working:

Welcome to minicom 2.4

OPTIONS: I18n
Compiled on Jan 25 2010, 07:02:36.
Port /dev/ttyUSB0

Press CTRL-A Z for help on special keys

R<?LinkUSB V1.44
0LinkUSB V1.4
0LinkUSB V1.4
??LinkUSB V1.4
PLinkUSB V1.4
LinkUSB Single Character Commands

f=First \=Expanded Command Set
n=Next
r=Reset l=Level
(=Extend
b=Byt(NN+) ^=57600
j=Bit(N+) `=38400
p=BytW/P ,=19200
~=BitW/P x=BusLo
d=Aux+ $=Scan
z=Aux- *=List
&=Aux? h=Help
t=Search Type
LinkUSB V1.4
+,F3000002A159CC28
-,3C000012D5A97401


So it looks like this may be Digitemp after all - do you guys know how
I can help narrow down what / where the issue is?

[Tom]

My system is Fedora, so I'm not really up on Ubuntu.
But... I can boot onto the previous kernel with grub.
If you can do that, and things work OK, then it's a USB or USB to
Serial driver problem.

> > Home page:http://marc.merlins.org/ - Hide quoted text -
>
> - Show quoted text -

[kastrolis]

hi!
what kernels, distros you run?
today I noticed that digitemp_DS9097 V3.5.0 doesn't run on all
kernels
newer than approx. 2.6.30
I also compiled 3.6.0 but binary shows me version 3.5.0, don’t
understand why.
tested kernels:
2.6.18 - working, debian stable
2.6.26 - working, debian testing, debian stable
2.6.29 - working, slackware
2.6.32 - not working, debian testing
2.6.33 - not working, debian testing
2.6.36 - not working, debian testing
I was used passive adapter with usb serial adapter and serial port
directly.
k.
p.s. sorry for my english

[NDLunchbox]

I run Ubuntu Server 10.04 AMD64. For me it broke when an auto-update
upgraded my kernel from 2.6.32.24.25 to 2.6.32.25.27. Prior to that
it worked fine.

I was running V3.5.0 that was in the Ubuntu apt repository, I tried
compiling digitemp_DS9097U from source - same thing, shows ver.
3.5.0... I think it's just a bug.

Does anyone have digitemp running on a kernel > 2.6.32.24.25?

[Marc MERLIN]

On Fri, Oct 01, 2010 at 04:55:26PM -0700, NDLunchbox wrote:
> I run Ubuntu Server 10.04 AMD64. For me it broke when an auto-update
> upgraded my kernel from 2.6.32.24.25 to 2.6.32.25.27. Prior to that
> it worked fine.
>
> I was running V3.5.0 that was in the Ubuntu apt repository, I tried
> compiling digitemp_DS9097U from source - same thing, shows ver.
> 3.5.0... I think it's just a bug.
>
> Does anyone have digitemp running on a kernel > 2.6.32.24.25?

Now that you mention it, I hadn't realized that it doesn't work for me
on 2.6.34.1 either:
digitemp_DS9097 -s /dev/ttyS1 -w
DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane
GNU Public License v2.0 - http://www.digitemp.com
Turning off all DS2409 Couplers

Devices on the Main LAN

(nothing)

I just didn't realize since I switched to owfs a while ago since I have ow
devices that only owfs supports.

[NDLunchbox]

Since it looks like you're running a serial device directly (or it
would be a ttyUSBn, right?) - can we rule out an issue with the
kernel's handling of USB devices?

I have never dealt with the Linux kernel before - is there a change
log somewhere?
-B

On Oct 1, 8:51 pm, Marc MERLIN <m...@merlins.org> wrote:
> On Fri, Oct 01, 2010 at 04:55:26PM -0700, NDLunchbox wrote:
> > I run Ubuntu Server 10.04 AMD64.  For me it broke when an auto-update
> > upgraded my kernel from 2.6.32.24.25 to 2.6.32.25.27.  Prior to that
> > it worked fine.
>
> > I was running V3.5.0 that was in the Ubuntu apt repository, I tried
> > compiling digitemp_DS9097U from source - same thing, shows ver.
> > 3.5.0... I think it's just a bug.
>
> > Does anyone have digitemp running on a kernel > 2.6.32.24.25?
>
> Now that you mention it, I hadn't realized that it doesn't work for me
> on 2.6.34.1 either:
> digitemp_DS9097 -s /dev/ttyS1  -w
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

[NDLunchbox]

https://launchpad.net/ubuntu/+source/linux/2.6.32-25.44

I would guess something in here broke it...

[kastrolis]

it is not usb problem, as I reported problem persists in both cases.
k.

p.s. googlegroups sux, I can post only thru web interface :(

[Marc MERLIN]

On Fri, Oct 01, 2010 at 05:59:10PM -0700, NDLunchbox wrote:
> Since it looks like you're running a serial device directly (or it
> would be a ttyUSBn, right?) - can we rule out an issue with the
> kernel's handling of USB devices?

Correct, I used a real honest to god serial port :)

It's surprising that a kernel change would break just digitemp though.
Was digitemp maybe doing something wrong with the serial port but it just
happened to work until now?

> I have never dealt with the Linux kernel before - is there a change
> log somewhere?

Yes, and it's very long :)

On Fri, Oct 01, 2010 at 06:38:16PM -0700, kastrolis wrote:
> it is not usb problem, as I reported problem persists in both cases.
> k.
>
> p.s. googlegroups sux, I can post only thru web interface :(

What's the problem? Your posting address doesn't match your subscription
address?

[kastrolis]

address is the same.
I think there is too much anti-spam :(
google delivery notification:
Hello kast...@gmail.com,

We're writing to let you know that the group you tried to contact
(digitemp) may not exist, or you may not have permission to post
messages to the group. A few more details on why you weren't able to
post:

* You might have spelled or formatted the group name incorrectly.
* The owner of the group may have removed this group.
* You may need to join the group before receiving permission to post.
* This group may not be open to posting.

If you have questions related to this or any other Google Group, visit
the Help Center at http://groups.google.com/support/?hl=en_US.

Thanks,

Google Groups

----- Original message -----

Received: by 10.204.73.205 with SMTP id r13mr257272bkj.
10.1285974129877;
Fri, 01 Oct 2010 16:02:09 -0700 (PDT)
Received: by 10.204.73.205 with SMTP id r13mr257271bkj.
10.1285974129851;
Fri, 01 Oct 2010 16:02:09 -0700 (PDT)
Return-Path: <kast...@gmail.com>
Received: from smtp.aquapro.lv (smtp.aquapro.lv [83.241.14.66])
by gmr-mx.google.com with ESMTP id z17si870537bkb.
7.2010.10.01.16.02.09;
Fri, 01 Oct 2010 16:02:09 -0700 (PDT)
Received-SPF: neutral (google.com: 83.241.14.66 is neither permitted
nor denied by domain of kast...@gmail.com) client-ip=83.241.14.66;
Authentication-Results: gmr-mx.google.com; spf=neutral (google.com:
83.241.14.66 is neither permitted nor denied by domain of
kast...@gmail.com) smtp.mail=kast...@gmail.com
Received: from [192.168.0.4] (mpe-25-196.mpe.lv [83.241.25.196])
by smtp.aquapro.lv (Postfix) with ESMTPA id 2E3EB38053
for <digi...@googlegroups.com>; Sat, 2 Oct 2010 02:02:09 +0300
(EEST)
Message-ID: <4CA66870...@gmail.com>
Date: Sat, 02 Oct 2010 02:02:08 +0300
From: kastrolis <kast...@gmail.com>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.12) Gecko/
20100913 Icedove/3.0.7
MIME-Version: 1.0
To: digi...@googlegroups.com
Subject: Re: Ubuntu Server: Auto Update to Linux 2.6.32.25.27 broke
Digitemp
References:
<853768b0-0b3a-4c19...@n7g2000vbo.googlegroups.com>
<20100930215...@merlins.org> <087682d9-af0e-4ac0-b500-
bc483c...@f26g2000vbm.googlegroups.com>
<20100930231...@merlins.org>
<60ae9af9-0b75-4158...@a19g2000yql.googlegroups.com>
<5a83fe9f-6e99-4c1d...@e14g2000yqe.googlegroups.com>
In-Reply-To: <5a83fe9f-6e99-4c1d-b07e-
b22185...@e14g2000yqe.googlegroups.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit

[kastrolis]

I just searched bugzilla.kernel.org and found nothing about digitemp.
may be you can report it?
I think it will be more professional than my attempt ;)
or we must go into digitemp code analyzing?
k.
p.s. Brian seems to be offline...

On Oct 2, 3:51 am, Marc MERLIN <m...@merlins.org> wrote:
> On Fri, Oct 01, 2010 at 04:55:26PM -0700, NDLunchbox wrote:
> > I run Ubuntu Server 10.04 AMD64.  For me it broke when an auto-update
> > upgraded my kernel from 2.6.32.24.25 to 2.6.32.25.27.  Prior to that
> > it worked fine.
>
> > I was running V3.5.0 that was in the Ubuntu apt repository, I tried
> > compiling digitemp_DS9097U from source - same thing, shows ver.
> > 3.5.0... I think it's just a bug.
>
> > Does anyone have digitemp running on a kernel > 2.6.32.24.25?
>
> Now that you mention it, I hadn't realized that it doesn't work for me
> on 2.6.34.1 either:
> digitemp_DS9097 -s /dev/ttyS1  -w
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

[NDLunchbox]

There wouldn't be anything about Digitemp, I'm not a developer - but
my scientific wild-ass guess would be some module or API that digitemp
relied on was either changed or deprecated. I've seen this happen
with a gps interface for instance - the program needed to be updated
with the the new command format and switches.

Also, while I am a newb to Linux bug reporting, I doubt the kernel
devs would much care - especially since they are the much larger
project... I would think digitemp has a few hundred or thousand users
vs. the millions that rely on the Linux kernel. I think the push
would be to make digitemp compatible with the kernel, not visa-versa.

[Marc MERLIN]

On Fri, Oct 01, 2010 at 07:32:00PM -0700, kastrolis wrote:
> I just searched bugzilla.kernel.org and found nothing about digitemp.
> may be you can report it?
> I think it will be more professional than my attempt ;)
> or we must go into digitemp code analyzing?

To be honest, I'm not the right person, I don't use digitemp anymore since
as I explained it did not support my devices properly, so I switched to
owfs.

I guess I should unsubscribe from this list :)

[Tom]

I run Fedora release 13.
My kernel version is 2.6.34.7-56.
I just did a fresh recompile of digitemp 3.6.0

The header still says:
DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

but it works fine.


[tom@kahuna2 digitemp-3.6.0]$ ./digitemp_DS9097U -i -s /dev/ttyS0

DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

GNU Public License v2.0 - http://www.digitemp.com

Turning off all DS2409 Couplers

.......
Searching the 1-Wire LAN
10B47F050208009A : DS1820/DS18S20/DS1920 Temperature Sensor
108E720502080054 : DS1820/DS18S20/DS1920 Temperature Sensor
285C8F8F020000E5 : DS18B20 Temperature Sensor
28C63418020000C3 : DS18B20 Temperature Sensor
28C9648F02000015 : DS18B20 Temperature Sensor
28958E8F02000021 : DS18B20 Temperature Sensor
28EB628F02000051 : DS18B20 Temperature Sensor
ROM #0 : 10B47F050208009A
ROM #1 : 108E720502080054
ROM #2 : 285C8F8F020000E5
ROM #3 : 28C63418020000C3
ROM #4 : 28C9648F02000015
ROM #5 : 28958E8F02000021
ROM #6 : 28EB628F02000051
Wrote .digitemprc

[NDLunchbox]

Tom,
Any tricks to compiling it? Did you tweak the makefile or anything?
Did it stop working for you and then a recompile fixed it?

On Oct 2, 10:03 am, Tom <tombro...@gmail.com> wrote:
> I run Fedora release 13.
> My kernel version is 2.6.34.7-56.
> I just did a fresh recompile of digitemp 3.6.0
> The header still says:
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane
> but it works fine.
>
> [tom@kahuna2 digitemp-3.6.0]$ ./digitemp_DS9097U -i -s /dev/ttyS0
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

[Tom]

I didn't do anything special.
It has always worked for me.
I haven't compiled it for a longtime, so this recompile was just a
test to see if anything changed.

The makefile is a little screwed-up, so I can't use the install
option.
I just execute make ds9097u to create the digitemp_DS9097U
file in the base directory.
I then use ln -s digitemp_DS9097U /usr/local/bin/digitemp to
create a link in my /usr/local/bin directory.

[kastrolis]

so...
for now it seems like only debian kernels have problems.
detailed report later.
k.

[NDLunchbox]

Possibly, or some of the people reporting it working have much newer
kernels (Ubuntu freezes their kernels at certain points), perhaps some
change broke it and was fixed or reverted in a future release? It's
like the Microsoft world - cross your fingers and hope a patch in the
near future fixes it!

[NDLunchbox]

2.6.18 - working, debian stable
2.6.26 - working, debian testing, debian stable
2.6.29 - working, slackware

2.6.32 - broken, debian testing
2.6.32.24.25 - Working, Ubuntu 10.04
2.6.32.25.27 - broken, Ubuntu 10.04
2.6.33 - broken, debian testing
2.6.34.1 - broken, ???
2.6.34.7.56 - working, Fedora 13
2.6.36 - broken, debian testing

[NDLunchbox]

So for what it's worth, I can replicate the problem. Yesterday I
accidentally hosed my server with a careless recursive permissions
change... dumb dumb dumb.
I rebuilt it today remotely after doing a default Ubuntu 10.04 LAMP +
SSH Server install. I ran all patches and then remembered to try
digitemp. Here is what happened:

$ uname -r
2.6.32-21-server

$ digitemp_DS9097U -i -a -s /dev/ttyUSB0

DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

GNU Public License v2.0 - http://www.digitemp.com

Turning off all DS2409 Couplers
..

Searching the 1-Wire LAN
28CC59A1020000F3 : DS18B20 Temperature Sensor
ROM #0 : 28CC59A1020000F3
Wrote .digitemprc
Oct 06 19:41:06 Sensor 0 C: 26.38 F: 79.47

So that works! I noticed a few patches were being held back, so I
figured out the kernel updates were actually part of an OS Service
Pack (or whatever the incremental releases of Ubuntu LTS distros are
called). So I ran apt-get dist-upgrade to go to 10.04.1:

The following NEW packages will be installed:
bc linux-headers-2.6.32-25 linux-headers-2.6.32-25-server linux-
image-2.6.32-25-server
The following packages will be upgraded:
landscape-common linux-headers-server linux-image-server linux-
server
4 upgraded, 4 newly installed, 0 to remove and 0 not upgraded.
Need to get 42.5MB of archives.
After this operation, 214MB of additional disk space will be used.

And test digitemp again:

$ uname -r
2.6.32-25-server

$ digitemp_DS9097U -i -a -s /dev/ttyUSB0

DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane
GNU Public License v2.0 - http://www.digitemp.com

Error 5: DS2480B Adapter Not Detected
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed
Error 10: Read COM Failed

So the upgrades and / or new packages listed above are what is killing
digitemp on Debian systems.

[NDLunchbox]

No one's made any progress I guess... how about we start a pot that
goes to whoever fixes this bug. I pledge $10 via paypal...

On Oct 6, 8:46 pm, NDLunchbox <bmali...@gmail.com> wrote:
> So for what it's worth, I can replicate the problem.  Yesterday I
> accidentally hosed my server with a careless recursive permissions
> change... dumb dumb dumb.
> I rebuilt it today remotely after doing a default Ubuntu 10.04 LAMP +
> SSH Server install.  I ran all patches and then remembered to try
> digitemp.  Here is what happened:
>
> $ uname -r
> 2.6.32-21-server
>
> $ digitemp_DS9097U -i -a -s /dev/ttyUSB0
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

> GNU Public License v2.0 -http://www.digitemp.com

[berland]

On Oct 27, 9:47 pm, NDLunchbox <bmali...@gmail.com> wrote:
> No one's made any progress I guess... how about we start a pot that
> goes to whoever fixes this bug.  I pledge $10 via paypal...

Count me in on that pledge. I will happily donate $10.

[berland]

On Oct 27, 9:47 pm, NDLunchbox <bmali...@gmail.com> wrote:

> No one's made any progress I guess... how about we start a pot that
> goes to whoever fixes this bug.  I pledge $10 via paypal...

I am in, I will happily donate $10 to whoever fixes the bug.

[Danny]

Hey if we are talking money. I am willing to try. I need to update
my ubuntu box to that kernel.

[Danny]

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/554099

There is a kernel patch for a usb serial issue. I wonder if this is
it.

[xinie]

I think I have found the kernel change that caused this (on ubuntu).
See https://bugs.launchpad.net/ubuntu/+source/linux/+bug/655868?comments=all.

It's unclear to me whether digitemp has to change to match this or
whether the kernel change should be reverted. It may just be a
question of rebuilding the digitemp package from source.

[xinie]

Tried rebuilding digitemp-3.5.0ds1-1ubuntu from source, running under
linux-image-2.6.32-25-generic version 2.6.32-25.45. Doesn't change
anything.

[xinie]

Tried building 3.6.0 (which doesn't appear to have had the version
bumped in the source).
No change in behaviour, on 2.6.32-25-generic.

[xinie]

The digitemp_DS9097U that I built from 3.6.0 source works on kernel
2.6.32-24-generic.

[xinie]

On Nov 5, 12:07 am, xinie <vincent.mcint...@gmail.com> wrote:
> The digitemp_DS9097U that I built from 3.6.0 source works on kernel
> 2.6.32-24-generic.

I think I've got it.
Mainline kernel commit 677aeafe19e88c282af74564048243ccabb1c590
(http://git.kernel.org/?p=linux/kernel/git/torvalds/
linux-2.6.git;a=commit;h=677aeafe19e88c282af74564048243ccabb1c590)

USB: ftdi_sio: revert "USB: ftdi_sio: fix DTR/RTS line modes"

This reverts commit 6a1a82df91fa0eb1cc76069a9efe5714d087eccd.

RTS and DTR should not be modified based on CRTSCTS when calling
set_termios.

This is in conformance with the way serial ports work today and
many
applications depend on this behaviour to be able to talk to
hardware
implementing hardware flow control (without the applications
actually using
it).

Hardware which expects different behaviour on these lines can
always
use TIOCMSET/TIOCMBI[SC] after port open to change them.

[Danny]

I have it working on a pl2303 ubuntu10.10(2.6.35-22). I have found
that an stty option is not set correctly. I am trying to isolate it
to exactly which option so I can fix it in the source code.

I am able to reproduce the failure and then break it again.

[Danny]

Update:
Using this command stty -F /dev/ttyUSB0 9600 -cread -brkint -icrnl -
imaxbel -opost -onlcr -isig -icanon -iexten -echo -echoe -echok -
echoctl -echoke

I can get digitemp to work, although it will still fail sometimes. It
seems that setting the com port to some value other than what the
default is causes it work 90+percent of the time. I know
this doesnt make sense as within the digitemp software the com port is
setup. The only thing I can speculate is that the latest kernel sets
the com port defaults differently and setting all the values on the
command line forces a setting that digitemp needs.

Here is the output I derived:
Linux ubuntu 2.6.35-22-generic-pae #35-Ubuntu SMP Sat Oct 16 22:16:51
UTC 2010 i686 GNU/Linux

***********before tty changes

DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

GNU Public License v2.0 - http://www.digitemp.com

Turning off all DS2409 Couplers

Devices on the Main LAN

**tty changes here not sure which one helped yet.
$>#stty -F /dev/ttyUSB0 9600 -cread -brkint -icrnl -imaxbel -opost -
onlcr -isig -icanon -iexten -echo -echoe -echok -echoctl -echoke
$>#stty -F /dev/ttyUSB0 500:5:cbd:8a3b:3:1c:7f:15:4:0:1:0:11:13:1a:
0:12:f:17:16:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0



******after tty changes

DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

GNU Public License v2.0 - http://www.digitemp.com

Turning off all DS2409 Couplers
.

Devices on the Main LAN

1091ABED01080037 : DS1820/DS18S20/DS1920 Temperature Sensor

[berland]

On Nov 5, 1:11 pm, xinie <vincent.mcint...@gmail.com> wrote:
> On Nov 5, 12:07 am, xinie <vincent.mcint...@gmail.com> wrote:
>
> > The digitemp_DS9097U that I built from 3.6.0 source works on kernel
> > 2.6.32-24-generic.
>
> I think I've got it.
> Mainline kernel commit 677aeafe19e88c282af74564048243ccabb1c590
> (http://git.kernel.org/?p=linux/kernel/git/torvalds/
> linux-2.6.git;a=commit;h=677aeafe19e88c282af74564048243ccabb1c590)

I see that this particular commit is just a revert of a commit from
June 2010
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=6a1a82df91fa0eb1cc76069a9efe5714d087eccd

so I guess that if it is this commit that broke digitemp, then we
should have seen this problem earlier as well? (but there might be
many buts)

[Vincent McIntyre]

That depends when 6a1a82df91fa0eb1cc76069a9efe5714d087eccd got merged
into the ubuntu kernel. From the git log it looks like it was merged
for 2.6.32-25.43.

Reverting it, then rebuilding the kernel where I first noticed the
problem, made it possible for digitemp to talk to the device again.

[NDLunchbox]

So long and short of it, does Digitemp have to be fixed or do you
think they'll revert the change to the kernel that is causing the
issue?
I don't understand why they'd go making changes to how serial devices
work now since most are for legacy hardware and applications and
haven't changed since the early 1990's!

On Nov 7, 5:40 pm, Vincent McIntyre <vincent.mcint...@gmail.com>
wrote:

> That depends when 6a1a82df91fa0eb1cc76069a9efe5714d087eccd got merged
> into the ubuntu kernel. From the git log it looks like it was merged
> for 2.6.32-25.43.
>
> Reverting it, then rebuilding the kernel where I first noticed the
> problem, made it possible for digitemp to talk to the device again.
>

> On 11/8/10, berland <havardberl...@gmail.com> wrote:
>
> > On Nov 5, 1:11 pm, xinie <vincent.mcint...@gmail.com> wrote:
> >> On Nov 5, 12:07 am, xinie <vincent.mcint...@gmail.com> wrote:
>
> >> > The digitemp_DS9097U that I built from 3.6.0 source works on kernel
> >> > 2.6.32-24-generic.
>
> >> I think I've got it.
> >> Mainline kernel commit 677aeafe19e88c282af74564048243ccabb1c590
> >> (http://git.kernel.org/?p=linux/kernel/git/torvalds/
> >> linux-2.6.git;a=commit;h=677aeafe19e88c282af74564048243ccabb1c590)
>
> > I see that this particular commit is just a revert of a commit from
> > June 2010

> >http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=co...

[NDLunchbox]

Danny - tried 'stty -F /dev/ttyUSB0 9600 -cread -brkint -icrnl -

imaxbel -opost -onlcr -isig -icanon -iexten -echo -echoe -echok -

echoctl -echoke' didn't work for me.

On Nov 7, 12:34 am, Danny <dannymul...@gmail.com> wrote:
> Update:
> Using this command stty -F /dev/ttyUSB0 9600 -cread -brkint -icrnl -
> imaxbel -opost -onlcr -isig -icanon -iexten -echo -echoe -echok -
> echoctl -echoke
>
> I can get digitemp to work, although it will still fail sometimes.  It
> seems that setting the com port to some value other than what the
> default is causes it work 90+percent of the time.  I know
> this doesnt make sense as within the digitemp software the com port is
> setup.  The only thing I can speculate is that the latest kernel sets
> the com port defaults differently and setting all the values on the
> command line forces a setting that digitemp needs.
>
> Here is the output I derived:
> Linux ubuntu 2.6.35-22-generic-pae #35-Ubuntu SMP Sat Oct 16 22:16:51
> UTC 2010 i686 GNU/Linux
>
> ***********before tty changes
>
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

> Turning off all DS2409 Couplers
>
> Devices on the Main LAN
>
> **tty changes here not sure which one helped yet.
> $>#stty -F /dev/ttyUSB0 9600 -cread -brkint -icrnl -imaxbel -opost -
> onlcr -isig -icanon -iexten -echo -echoe -echok -echoctl -echoke
> $>#stty -F /dev/ttyUSB0 500:5:cbd:8a3b:3:1c:7f:15:4:0:1:0:11:13:1a:
> 0:12:f:17:16:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0
>
> ******after tty changes
> DigiTemp v3.5.0 Copyright 1996-2007 by Brian C. Lane

> GNU Public License v2.0 -http://www.digitemp.com

[Vincent McIntyre]

On 11/8/10, NDLunchbox <bmal...@gmail.com> wrote:
> So long and short of it, does Digitemp have to be fixed or do you
> think they'll revert the change to the kernel that is causing the
> issue?

The problem was the kernel change. It has already been reverted in the mainline.

As for when ubuntu get around to fixing it, I don't know.
The bug is ranked about 5700 out of about 6500 bugs against the 'linux' package
so it may take a few weeks to get some attention.
Once it gets triaged it should progress quickly, since we've provided the fix.

Meanwhile you could build your own kernel, just edit ftdi_sio.c and
rebuild following:
https://help.ubuntu.com/community/Kernel/Compile

> I don't understand why they'd go making changes to how serial devices
> work now since most are for legacy hardware and applications and
> haven't changed since the early 1990's!

It's a puzzle, indeed. This is the "joy" of Linux.

[berland]

On Nov 8, 10:17 am, Vincent McIntyre <vincent.mcint...@gmail.com>
wrote:

> As for when ubuntu get around to fixing it, I don't know.
> The bug is ranked about 5700 out of about 6500 bugs against the 'linux' package
> so it may take a few weeks to get some attention.
> Once it gets triaged it should progress quickly, since we've provided the fix.

I see. Is there a Launchpad entry for this particular bug? I couldn't
find it. I'll see if I compile my own kernel or wait for it to get
fixed.

[Danny]

Thanks for trying. I was unsure it would work as it was unstable for
me, but it did go from not working ever to working 90+ percent of the
time.

I think the correct answer is to wait for the kernel fix. I dont
think there is any fix required as this appears to be a kernel driver
issue.

[NDLunchbox]

So I noticed my server rebooted and I got mail - sure enough, a new
kernel was installed - 2.6.32-26. I got all excited, tried
DigiTemp... and... no luck :(
It looks like 10.4.2 will release 2.6.35-23 backported from 10.10...
guess we'll have to wait until Jan 27 / Feb 17 (I've seen both days)
and hope that fixes it.

I have three 4 Ubuntu computers (1 server, 1 laptop, 1 desktop and one
eSATA HDD with it installed) - all 10.04. If I find the time I may
try moving my DigiTemp install to one, see if it's broken and then try
upgrading to 10.10 and see if that fixes it...

[NDLunchbox]

FIXED!!! Ran an apt-get dist-upgrade to see if the new point release
was out (it's not) but there was a kernel upgrade to 2.6.32-28... and
now digitemp runs! Woot!

[Vincent McIntyre]

On 1/30/11, NDLunchbox <bmal...@gmail.com> wrote:
> FIXED!!! Ran an apt-get dist-upgrade to see if the new point release
> was out (it's not) but there was a kernel upgrade to 2.6.32-28... and
> now digitemp runs! Woot!

I would be very interested to see if it continues to run over a long
period of time, for you.

I have had one failure (could not read anything back from the device)
18h after boot that got fixed by pulling and reinserting the usb
device. This has not yet recurred but I think a patch or two more may
be required to really fix this.

Details at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/690798.

[NDLunchbox]

I'll keep an eye on it, since it finally works I spent a bit over the
last two days getting Munin and the custom plugin I wrote working
again. I think it's logging 5 minute intervals so you can check for
breaks / #NAN errors on the graphs I symlink to the public site:
Daily:
http://www.ndlunchbox.net/images/currenttemps-day.png
<img src="http://www.ndlunchbox.net/images/currenttemps-day.png">

Weekly:
http://www.ndlunchbox.net/images/currenttemps-week.png
<img src="http://www.ndlunchbox.net/images/currenttemps-week.png">

On Jan 30, 3:50 am, Vincent McIntyre <vincent.mcint...@gmail.com>
wrote: