Hi Paul!
I have updated to 2.9p9 and it works fine again. Thank you :)
Markus
On Jan 6, Paul Alfille <paul.alfi...@gmail.com> wrote:
I looked closely at the error. Near as I can tell, it's the utility, like 'cat'
that's supplying the
131072 (128K) read request size.
The actual size is 12 bytes and that number is returned in the preceding fstat
call. I think the
ulitities just ask for a generic large read, and I presume that size has
increased from 64k to 128k
recently which is over owfs's arbitrary internal limit.
I added a size limiter in the owfs read code and everything now works. A new
release 2.9p9 now
includes the fix.
Thank you Markus and Stefano for finding and investigating this problem.
Paul Alfille
On Mon, Jan 5, 2015 at 2:54 PM, Markus Gaugusch <mar...@gaugusch.at> wrote:
Hi Stefano and Paul!
Thank you very much for your help!
I had migrated my scripts to use owget, etc. But it's good to know that
there is another
solution until the bug is fixed. I also noticed that my php script (using
read access to
owfs files, not owphp) is also still working fine.
Thanks again,
Markus
On Jan 5, Stefano Miccoli <mo...@icloud.com> wrote:
The present bug apparently affects only shell programs (and only if cat or
cp use a big buffer size). Waiting for a patch a quick workaround for bash
users could be substituting
temp=$(cat /mnt/1wire/XXXX/temp)
with
temp=$(< /mnt/1wire/XXXX/temp)
and
cat /mnt/1wire/XXXX/temp
with
echo $(< /mnt/1wire/XXXX/temp)
Of course another option is to avoid the FUSE driver, and access directly
owserver (via owshell/owdir from the shell, or one of the language
bindings
in module/ownet, not to speak of my own pyownet ;-))
Stefano
On 05 Jan 2015, at 05:43, David Lazarou <dlazaro...@gmail.com>
wrote:
I ran across this same problem using a Raspberry Pi.
http://owfs-developers.1086194.n5.nabble.com/quot-Message-too-long-quot-err
ors-with-owfs-tt10946.html
I never was never able to discover what the problem was.
--
David
On Mon, Jan 5, 2015 at 10:58 AM, Stefano Miccoli <mo...@icloud.com>
wrote:
Somehow I was able to replicate the problem in Python.
When I type in Python
>>> f = open(‘/mnt/1wire/10.67C6697351FF/temperature')
>>> f.read(1024)
'56.5153'
>>> f.close()
I get from owfs —debug
DEBUG: ow_read.c:(231) About to read
</10.67C6697351FF/temperature> extension=0 size=4096 offset=0
DEBUG: ow_read.c:(231) About to read
</10.67C6697351FF/temperature> extension=0 size=4096 offset=7
Increasing the input buffer size
>>> fip = open(‘/mnt/1wire/10.67C6697351FF/temperature')
>>> fip.read(131072)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
IOError: [Errno 90] Message too long
and
DEBUG: ow_read.c:(231) About to read
</10.67C6697351FF/temperature> extension=0 size=127292 offset=0
My conjecture: given the system call
ssize_t read(int fd, void *buf, size_t count);
the ‘count' parameter is passed via FUSE up
to ow_server_message.c and causes the error. (and presumably cat
and cp use a bigger buffer size than head, less and tail.)
Stefano
On 04 Jan 2015, at 23:05, Paul Alfille
<paul.alfi...@gmail.com> wrote:
I don't have the answer yet, but it happens only with cat
or cp, not less,head or tail.It also happens only with
owfs->owserver->bus, since owfs->bus is ok.
There must be something different in the tcp protocol
handling. I'll investigate further.
Paul
On Sun, Jan 4, 2015 at 1:20 PM, Jacob Joseph
<ja...@jjoseph.org> wrote:
I recently moved my OWFS setup to a new
installation (on Funtoo,
should it matter), and have noticed that I do
have the issue Markus
mentioned.
I haven't spent much time debugging quite
yet. Curiously, though,
while 'cat' doesn't work, other methods of
reading such as 'head' or
opening the file in Python seem unaffected.
~Jacob
On Sun, 04 Jan 2015 10:59:08 +0100
Stefano Miccoli <mo...@icloud.com> wrote:
> It seems that your question is still
unanswered, so I will try to
> help you, although I___m not able to
replicate your setup (and I___m no
> expert of the internal workings of owfs).
>
> From your message it seems that the problem
is limited to owfs, and
> not in owserver, but let me double-check.
>
> 1) Do you have an owserver running and owfs
is quering owserver?
> (You can check by inspecting owfs.conf and
running ___sudo netstat
> -tlp___. If there is an owserver deamon
running there should be a line
> like
>
> Active Internet connections (only servers)
> Proto Recv-Q Send-Q Local Address
Foreign Address
> State PID/Program name tcp 0
0
> *:4304 *:*
LISTEN
> 31595/owserver
>
> 2) If you have an owserver running can you
confirm that
> ___owread /10.FE771D020800/temperature___
gives you a sensible answer
> while ___cat
/mnt/1wire/10.FE771D020800/temperature___
yields an error?
>
>
> If the problem is only in owfs a workaround
would be using
> owdir/owread/owwrite in the shell, and one
of the clients in
> module/ownet for other languages.
>
> Bye
>
> Stefano
>
> > On 30 Dec 2014, at 22:15, Markus Gaugusch
<mar...@gaugusch.at
> > <mailto:mar...@gaugusch.at>> wrote:
> >
> > Hi,
> >
> > Symptom:
> > monitor:/mnt/1wire/10.FE771D020800 # cat
temperature
> > cat: temperature: Message too long
> >
> > I've updated my monitoring server from
openSUSE 13.1 to 13.2, which
> > broke owfs. I had 2.9p5 before and also
tried to update to 2.9p8,
> > but the problem stayed the same. Access
using owdir, owget and
> > owwrite works fine. I've started owfs in
debug mode, leading to the
> > following output (error at the bottom).
> >
> > DEBUG: ow_tcp_read.c:(63) attempt 24
bytes Time: 2.000000 seconds
> > DEBUG: ow_tcp_read.c:(113) read: 24 - 0
= 24
> > DEBUG: ow_tcp_read.c:(63) attempt 8
bytes Time: 2.000000 seconds
> > DEBUG: ow_tcp_read.c:(113) read: 8 - 0 =
8
> > DEBUG: ow_buslock.c:(54)
pthread_mutex_lock 9A270CC begin
> > DEBUG: ow_buslock.c:(54)
pthread_mutex_lock 9A270CC done
> > DEBUG: ow_buslock.c:(56)
pthread_mutex_lock B779D9A0 begin
> > DEBUG: ow_buslock.c:(56)
pthread_mutex_lock B779D9A0 done
> > DEBUG: ow_buslock.c:(56)
pthread_mutex_unlock B779D9A0 begin
> > DEBUG: ow_buslock.c:(56)
pthread_mutex_unlock B779D9A0 done
> > DEBUG: ow_buslock.c:(76)
pthread_mutex_lock B779D9A0 begin
> > DEBUG: ow_buslock.c:(76)
pthread_mutex_lock B779D9A0 done
> > DEBUG: ow_buslock.c:(79)
pthread_mutex_unlock B779D9A0 begin
> > DEBUG: ow_buslock.c:(79)
pthread_mutex_unlock B779D9A0 done
> > DEBUG: ow_buslock.c:(81)
pthread_mutex_unlock 9A270CC begin
> > DEBUG: ow_buslock.c:(81)
pthread_mutex_unlock 9A270CC done
> > DEBUG: ow_presence.c:(269) Presence of
10 FE 77 1D 02 08 00 3A
> > FOUND on bus 127.0.0.1
> > DEBUG: ow_cache.c:(546) Adding device
location 10 FE 77 1D 02 08
> > 00 3A bus=0
> > DEBUG: ow_cache.c:(635) Add to cache sn
10 FE 77 1D 02 08 00 3A
> > pointer=0xb779d914 index=0 size=4
> > DEBUG: ow_cache.c:(667)
pthread_mutex_lock B779D9A0 begin
> > DEBUG: ow_cache.c:(667)
pthread_mutex_lock B779D9A0 done
> > DEBUG: ow_cache.c:(670)
pthread_mutex_unlock B779D9A0 begin
> > DEBUG: ow_cache.c:(670)
pthread_mutex_unlock B779D9A0 done
> > DEBUG: ow_cache.c:(336)
pthread_mutex_lock B779D9A0 begin
> > DEBUG: ow_cache.c:(336)
pthread_mutex_lock B779D9A0 done
> > DEBUG: ow_cache.c:(336)
pthread_mutex_unlock B779D9A0 begin
> > DEBUG: ow_cache.c:(336)
pthread_mutex_unlock B779D9A0 done
> > DEBUG: ow_read.c:(88) pthread_mutex_lock
B779D9A0 begin
> > DEBUG: ow_read.c:(88) pthread_mutex_lock
B779D9A0 done
> > DEBUG: ow_read.c:(95)
pthread_mutex_unlock B779D9A0 begin
> > DEBUG: ow_read.c:(95)
pthread_mutex_unlock B779D9A0 done
> > DEBUG: ow_read.c:(96)
/10.FE771D020800/temperature return -90
> > DEBUG: ow_parsename.c:(61)
/10.FE771D020800/temperature
> > unique: 140, error: -90 (Message too
long), outsize: 16
> > unique: 141, opcode: RELEASE (18), nodeid:
41, insize: 64, pid: 0
> > release[0] flags: 0x8000
> > CALL: owfs_callback.c:(135) RELEASE
> > path=/10.FE771D020800/temperature unique:
141, success, outsize: 16
> >
> >
> >
> > It might be related to the new kernel, but
I'm not sure ...
> > monitor:~ # uname -a
> > Linux monitor 3.16.7-7-default #1 SMP Wed
Dec 17 18:00:44 UTC 2014
> > (762f27a) i686 i686 i386 GNU/Linux
> >
> >
> > Help! :-)
> > (and many thanks for owfs!)
> >
> > kind regards,
> > Markus
> >
>
>---------------------------------------------------------------------------
---
> > Dive into the World of Parallel
Programming! The Go Parallel
> > Website, sponsored by Intel and developed
in partnership with
> > Slashdot Media, is your hub for all things
parallel software
> > development, from weekly thought
leadership blogs to news, videos,
> > case studies, tutorials and more. Take a
look and join the
> > conversation now.
http://goparallel.sourceforge.net
> > <http://goparallel.sourceforge.net/>
> > _______________________________________________
Owfs-developers
> > mailing list
Owfs-developers@lists.sourceforge.net
> > <mailto:Owfs-developers@lists.sourceforge.net>
> >
https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
---------------------------------------------------------------------------
---
Dive into the World of Parallel Programming! The Go
Parallel Website,
sponsored by Intel and developed in partnership with
Slashdot Media, is your
hub for all things parallel software development,
from weekly thought
leadership blogs to news, videos, case studies,
tutorials and more. Take a
look and join the conversation now.
http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
---------------------------------------------------------------------------
---
Dive into the World of Parallel Programming! The Go
Parallel Website,
sponsored by Intel and developed in partnership with
Slashdot Media, is your
hub for all things parallel software development, from
weekly thought
leadership blogs to news, videos, case studies, tutorials
and more. Take a
look and join the conversation
now.http://goparallel.sourceforge.net__________________________________________
_____
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
---------------------------------------------------------------------------
---
Dive into the World of Parallel Programming! The Go Parallel
Website,
sponsored by Intel and developed in partnership with Slashdot
Media, is your
hub for all things parallel software development, from weekly
thought
leadership blogs to news, videos, case studies, tutorials and
more. Take a
look and join the conversation now.
http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
---------------------------------------------------------------------------
---
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media,
is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more.
Take a
look and join the conversation
now.http://goparallel.sourceforge.net__________________________________________
_____
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
--
__________________ /"\
Markus Gaugusch \ / ASCII Ribbon Campaign
markus(at)gaugusch.at X Against HTML Mail
/ \
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
--
__________________ /"\
Markus Gaugusch \ / ASCII Ribbon Campaign
markus(at)gaugusch.at X Against HTML Mail
/ \------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
