Re: intel 537 with kernel 3.1.x

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Patrick,

Thank you so much for this detailled call stack. If you read http://www.linuxgrill.com/anonymous/fire/netfilter/kernel-hacking-HOWTO-5.html, the spinlock chapter, one driver's spinlock being involved in your case, would it happen you run an SMP Fedora 16 distribution (as per indicated by uname -a) on a single CPU configuration ?

From your previous $ cat /proc/interrupts, IRQ 20 looks to be free and there should not be any longer an IRQ conflict.

Please provide me with the output of a command like $ sensors. Here is what it tells for me:

$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0:       +56.0 C  (high = +86.0 C, crit = +100.0 C)
Core 1:       +56.0 C  (high = +86.0 C, crit = +100.0 C)

So I am sure my processor is a dual core. I need to know with a simple command output fact whether your processor is single or multicore. Well note I am just in the preliminary phase of my investigations.

I fear we may also need efax to satisfy you. Can you download http://vouters.dyndns.org/Intel/efax.zip ?

$ unzip -aa efax.zip
$ cd efax
$ make -f Makefile_orig
$ ./efax -vewinchmart

After efax run, do you still have in your Call Trace:
[<f9e07942>] linux_MT_ReleaseLock_timer_task+0x12/0x20 [Intel537]
???

For the moment I prefer you run efax on your computer as I shall be able to perform the very same test on my Dual Core processor. If we can compare comparible things, it much ease troubleshooting.

Philippe

Le 05/01/2012 20:01, Patrick Boyadjian a écrit :
Well I tried make uninstall and install again and it still fails...here is my dmesg output

[  694.046650] sound enabled
[  694.046664] Intel 537 card found
[  727.461869] pci 0000:01:00.0: PCI INT A ->  GSI 20 (level, low) ->  IRQ 20
[  727.483870] 537: Loaded
[  744.163496] ------------[ cut here ]------------
[  744.163511] WARNING: at kernel/softirq.c:159 local_bh_enable_ip+0x60/0x90()
[  744.163514] Hardware name:
[  744.163517] Modules linked in: Intel537 fuse 8021q garp bridge stp llc it87 hwmon_vid nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack joydev ppdev microcode snd_hda_codec_realtek netjet isdnhdlc crc_ccitt mISDNipac snd_hda_intel snd_hda_codec r8169 snd_hwdep mISDN_core mii snd_seq snd_seq_device serio_raw iTCO_wdt iTCO_vendor_support i2c_i801 snd_pcm parport_pc parport snd_timer snd soundcore snd_page_alloc uinput usb_storage firewire_ohci firewire_core crc_itu_t i915 drm_kms_helper drm i2c_algo_bit i2c_core video [last unloaded: scsi_wait_scan]
[  744.163582] Pid: 2168, comm: wvdial Not tainted 3.1.6-1.fc16.i686 #1
[  744.163585] Call Trace:
[  744.163592]  [<c090411b>] ? printk+0x2d/0x2f
[  744.163599]  [<c044c122>] warn_slowpath_common+0x72/0xa0
[  744.163604]  [<c0452950>] ? local_bh_enable_ip+0x60/0x90
[  744.163608]  [<c0452950>] ? local_bh_enable_ip+0x60/0x90
[  744.163613]  [<c044c172>] warn_slowpath_null+0x22/0x30
[  744.163618]  [<c0452950>] local_bh_enable_ip+0x60/0x90
[  744.163623]  [<c090d182>] _raw_spin_unlock_bh+0x12/0x20
[  744.163740]  [<f9e07942>] linux_MT_ReleaseLock_timer_task+0x12/0x20 [Intel537]
[  744.163831]  [<f9e06c73>] RTS_Disable+0x33/0x40 [Intel537]
[  744.163921]  [<f9e06565>] clm_terminate+0x15/0x70 [Intel537]
[  744.163928]  [<c06cafb9>] ? tty_ldisc_flush+0x39/0x50
[  744.164030]  [<f9e0540d>] close+0xd/0x30 [Intel537]
[  744.164125]  [<f9e088f7>] softserial_close+0x77/0xb0 [Intel537]
[  744.164131]  [<c06c4f8f>] tty_release+0x10f/0x4e0
[  744.164137]  [<c0434285>] ? __wake_up+0x45/0x60
[  744.164141]  [<c06c5ab0>] ? no_tty+0x30/0x30
[  744.164146]  [<c0537396>] ? do_vfs_ioctl+0x86/0x5a0
[  744.164151]  [<c0473f51>] ? ktime_get_ts+0xe1/0x110
[  744.164156]  [<c05293ea>] fput+0xba/0x220
[  744.164162]  [<c0525eb4>] filp_close+0x54/0x80
[  744.164167]  [<c0525f3c>] sys_close+0x5c/0xa0
[  744.164171]  [<c090d4f4>] syscall_call+0x7/0xb
[  744.164174] ---[ end trace c8b0b4915f88a90c ]---
[pboyadjian@pb3000 ~]$


I don't have efax installed.

The recurring error message seems always to be "no info from serial port"

Patrick

On Thu, Jan 05, 2012 at 09:35:09AM +0100, Philippe Vouters wrote:
Patrick,

With your modem board at IRQ 20 and the modem driver can gives us
the output of each of the commands below:
1/ cd intel536-537/
2/ sudo make uninstall
3/ dmesg | tail
4/ sudo make install
5/ dmesg | tail
6/ sudo cat /proc/interrupt
7/ efax -vewinchmart
8/ dmesg | tail

The more facts you can provide us is the best.
Philippe

Le 05/01/2012 04:55, Patrick Boyadjian a écrit :
One more piece of info:

My dmesg on boot says:

92.709152]  [<f9df78f7>] softserial_close+0x77/0xb0 [Intel537]

So I can't really do a setserial for irq 20 and that i/o port.


Patrick

On Wed, Jan 04, 2012 at 08:22:38PM +0100, Philippe Vouters wrote:
Hi Patrick,

Thank so much for your long reply.

First of all about my write up. I do first suggest to use efax to
troubleshoot for the driver's installation correctness. It does help
much highlighting most of the problems before wvdial or kppp times
attemtps to connect to Internet. And the very advantage of the efax
command I provide is that it is at no financial cost but an analog
phone line subscription. The efax command just engage your phone
line without making a real call. So, except the analog phone
subscription line, the efax test is at strictly NO additional cost
for the end-user and provides lots of very informative data,
enabling to best troubleshoot in case of problems.

As for /dev/ttyS*, I never felt the necessity, the modem code
creating a soft link between /dev/modem and the /dev/[537 | 536ep]
real device. /dev/modem is sufficient for both wvdial and kppp. With
an example on how to set up /etc/wvdial.conf on my Web site, you are
the first one to mention wvdialconf among all the driver end-users
since I inhierd this code back in Fedora Core 4. In fact, with all
my experience with the end-users of my work, it appears that the
most delicate thing is to correctly specify the suitable AT modem
commands.

Regarding the other technical items of your skills, could I suggest
you if possible for you to open up a public Web site on your
computer or ask for Web space to your Internet provider in order to
share with others all your acquired knowledge. You would be
surprised like I am always delighted to notice it with my Web site
that the specific work you do for someone in peticular may also
interest a vast population over the planet. This is also true with
any knowledge whichever its nature and topic. Your efforts would
then be very rewarded the very same as mines.

Back to your IRQ problem ! Did my suggestions regarding either a
possible jumper enabling IRQ selection on your modem board or
pluging in the modem into another PCI slot help you ?

Yours truly,


Le 04/01/2012 19:25, Patrick Boyadjian a écrit :
Hi Philippe,

First off, well done on this write up.  An area that you may like to improve on, as you have suggested to use wvdial as a test app after installation, would be the importance of wvdial needing to see the modem device on a /dev/ttyS* link for use by the wvdialconf tool.  As it stands, you have recommended to symbolically link /dev/modem to /dev/537ep which is good, however, for the sake of less tech savvy users that would simply like to run wvdialconf a symbolic link to /dev/ttyS15 (15 being the experimental one) would allow wvdialconf to run successfully - advantages would be init string validation and correctness as well as the creation of /etc/wvdial.conf for post editing of the ISP username and password of course.

As far as other technical subjects, you may guffaw at what I have been able to accomplish with my experience with Linux back when I first explored it back in late 1990's.  That being said here are my main uses:

- openvpn with bridging to allow me to vpn in to my linux machine and obtain an IP on my subnet to see other devices on my network
- squid to filter traffic from my kids PC on a whitelist to the internet
- httpd to host my archive of pics, tunes and movies for access by anyone in my subnet
- chntpw app to constantly help my friends who have forgotten their windows NT admin account passwords
- simple ffmpeg conversions such as:
	ffmpeg -i input.wav output.mp3
	ffmpeg -i input.avi -b 64k -s 320x240 -r 24 output.mp4
- sendmail.mc and access config for the mail command for my scripts below
- mbr backup and restore scripts (dd if=blah blah)
- samba for giving my archives access to any UNC connection on my subnet
- tcpdump commands to sniff my network
- I have a network attached IP camera that ftps motion detected images to my linux machine on vsftpd
- iptables of course to firewall the entire operation

scripts I use:

- a tclsh script that logs all incoming calls to a text file (i.e. date, number, name)
- current ip script that checks my external ip every 30 mins and if it changes to email me the new one
- script to email me the previous days proxy violations from squid
- script to email me the previous days tty login attempts
- script to email me the previous days call log
- lmsensors script to read the internal temperature of my PC including fan speeds and basement temperature via snmpget...typical output is:

[pboyadjian@pb3000 scripts]$ temp

Basement Temperature = 17 C

PC Internal Temperature = 34 C
PC Chip Fan =  672 RPM
PC Case Fan = 1132 RPM

[pboyadjian@pb3000 scripts]$



If you would like my technical documentation on any of the above I would be more than happy to offer it.


Regards,
Patrick


On Wed, Jan 04, 2012 at 03:23:19PM +0100, Philippe Vouters wrote:
Patrick,

If there is anything I can ever do to repay you please let me know!

Yes, indeed, you can repay us by attentively redeading my
litterature on the Intel modems subject at
http://vouters.dyndns.org/Intel/Intel-Readme.html and give us your
feedback upon it.

What we do expect are your detailled impressions about this document
clarity and informations completeness.

Post-scriptum : if you are looking for informations on either Linux,
Unix, OpenVMS or Windows, you may refer to another side of my Web
site at http://vouters.dyndns.org/tima/
If, even better, you mail me with technical subjects I can study and
document, you'll make the worldwide IT community happy.

Yours truly,
Philippe

Le 04/01/2012 14:46, Patrick Boyadjian a écrit :
Brilliant, it works!

Philippe and Antonio, thank you so much for your support in this regard!

If there is anything I can ever do to repay you please let me know!

Regards,
Patrick



On Wed, Jan 04, 2012 at 12:27:39PM +0100, Philippe Vouters wrote:
My work ought to support your modem.
[philippe@victor ~]$ grep e159:0001 intel-536-537/*
intel-536-537/makefile_537:     vendor_device_ids[0]="e159:0001";\

The alternative to modifying the bash files is to download
http://vouters.dyndns.org/Intel/intel-536EP-537EP_2012_01_04.tar.bz2

Please reply whether you are fully satisfied or not.

Regards,
Philippe

Le 04/01/2012 04:03, Patrick Boyadjian a écrit :
Thanks for the swift reply Antonio.

Well originally I had a conexant chipset for a HCF modem that the linuxant group had drivers for, however those drivers ONLY worked on pre 3.x kernels so they broke when I upgraded.  Dev team from Linuxant has informed me they are aware of the API break in the driver for the newer kernel and say they should fix it sometime soon.  Meanwhile I dug up the Intel modem you see below to try and get it to work.

So short answer I never should have upgraded!! :)

Regards,
Patrick



On Tue, Jan 03, 2012 at 07:47:12PM -0600, Antonio Olivares wrote:
Patrick,

I am unsure if that modem in particular is served by Phillipe's
driver.  IT does not seem to be a true INTEL 537 modem :(
However, there was a recent update to the driver since it was failing
to build up in Ubuntu and it was working fine in Fedora 15.  Did you
make sure that you got the latest driver available from Phillipe's
website?

Oops sorry, you did get the latest one. Maybe Phillipe can comment
here to make sure that maybe something went wrong, or that it may be
specific to 537 modem, since the driver also works with 536 Intel
based modems.

Regards,


Antonio

BTW,
since you have experience with linux since the Red Hat days, did you
get to use this particular modem before?  If you did and the drivers
worked, I see no reason why the modem should not work with the newer
kernels.


On Tue, Jan 3, 2012 at 7:24 PM, Patrick Boyadjian
<patrick@xxxxxxxxxxxxxxxxxxxxxxx>      wrote:
Hi,

This is my first attempt to contact this group so apologize in advance if I=  don't provide sufficient information.

I'm not exactly a newbie to linux and Fedora since I've used it back when i= t was called red hat...I've upgraded to Fedora 16 recently and boy how thin= gs have changed!  Anyways, here is where I'm at:

lspci shows my modem as:
01:01.0 Communication controller: Tiger Jet Network Inc. Tiger3XX Modem/ISD= N interface

scanModem details as:
For candidate card in slot 01:01.0, firmware information and bootup diagnos= tics are:
PCI slot       PCI ID          SubsystemID     Name
----------     ---------       ---------       --------------
01:01.0        e159:0001       8086:0003       Communication controller: Ti=
ger Jet Network Inc. Tiger3XX Modem/ISDN interface

Modem interrupt assignment and sharing:
--- Bootup diagnostics for card in PCI slot 01:01.0 ----
[    0.109098] pci 0000:01:01.0: [e159:0001] type 0 class 0x000780
[    0.109120] pci 0000:01:01.0: reg 10: [io  0xa000-0xa0ff]
[    0.109133] pci 0000:01:01.0: reg 14: [mem 0xf1004000-0xf1004fff]
[    0.109200] pci 0000:01:01.0: supports D2
[    0.109203] pci 0000:01:01.0: PME# supported from D0 D2 D3hot D3cold
[    0.109209] pci 0000:01:01.0: PME# disabled

The PCI slot 01:01.0 of the modem card may be disabled early in a bootup process,  but then enabled later. If modem drivers load but the  modem is not responsive, read DOCs/Bootup.txt about possible fixes= .
Send dmesg.txt along with ModemData.txt to discuss@xxxxxxxxxxxxx if help is needed.



Predictive  diagnostics for card in bus 01:01.0:
        Modem chipset  detected on
NAME=3D"Communication controller: Tiger Jet Network Inc. Tiger3XX Modem/ISD= N interface"
CLASS=3D0780
PCIDEV=3De159:0001
SUBSYS=3D8086:0003
IRQ=3D11
HDA2=3D00:1b.0
IDENT=3DINTEL537

For candidate modem in:  01:01.0
   0780 Communication controller: Tiger Jet Network Inc. Tiger3XX Modem/ISD= N interface
      Primary device ID:  e159:0001
Support type needed or chipset:        INTEL537


So I visited http://vouters.dyndns.org/Intel/ and heeded Philippe's note to=
:


If you execute a 3.x.y kernel, download:

../Intel/intel-536EP-537EP_2011_12_11.tar.bz2<http://vouters.dyndns.org/Int=
el/intel-536EP-537EP_2011_12_11.tar.bz2>



...and continued to untar and try to build.

Issuing make clean and make 537 seem to work with a warning:

[root@pb3000 intel-536-537]# make clean
cd coredrv; make clean
make[1]: Entering directory `/root/intel-536-537/coredrv'
rm -f *.ko .*.o.cmd *.mod.c .*.ko.cmd *.o *~ core Module.* modules.* rm -rf .tmp_versions
make[1]: Leaving directory `/root/intel-536-537/coredrv'
rm -f *.o *.ko
[root@pb3000 intel-536-537]# make 537
cd coredrv; make clean
make[1]: Entering directory `/root/intel-536-537/coredrv'
rm -f *.ko .*.o.cmd *.mod.c .*.ko.cmd *.o *~ core Module.* modules.* rm -rf .tmp_versions
make[1]: Leaving directory `/root/intel-536-537/coredrv'
rm -f *.o *.ko
   Module precompile check
   Current running kernel is: 3.1.6-1.fc16.i686
   /lib/modules...   autoconf.h exists
diff: /boot/vmlinuz.autoconf.h: No such file or directory
   autoconf.h matches running kernel
diff: /boot/vmlinuz.version.h: No such file or directory
   version.h matches running kernel
make[1]: Entering directory `/root/intel-536-537'
Building for 537
3.1.6-1.fc16.i686
make[2]: Entering directory `/root/intel-536-537/coredrv'
make -C /lib/modules/3.1.6-1.fc16.i686/build M=3D/root/intel-536-537/coredr= v modules
make[3]: Entering directory `/usr/src/kernels/3.1.6-1.fc16.i686'
  CC [M]  /root/intel-536-537/coredrv/coredrv.o
  CC [M]  /root/intel-536-537/coredrv/clmmain.o
  CC [M]  /root/intel-536-537/coredrv/rts.o
  CC [M]  /root/intel-536-537/coredrv/task.o
  CC [M]  /root/intel-536-537/coredrv/uart.o
  CC [M]  /root/intel-536-537/coredrv/wwh_dflt.o
  CC [M]  /root/intel-536-537/coredrv/locks.o
  CC [M]  /root/intel-536-537/coredrv/softserial_io.o
  CC [M]  /root/intel-536-537/coredrv/softserial_ioctl.o
  CC [M]  /root/intel-536-537/coredrv/softserial.o
  CC [M]  /root/intel-536-537/coredrv/afedsp_int.o
  LD [M]  /root/intel-536-537/coredrv/Intel537.o
  Building modules, stage 2.
  MODPOST 1 modules
WARNING: could not find /root/intel-536-537/coredrv/.537core.lib.cmd for /r= oot/intel-536-537/coredrv/537core.lib
  CC      /root/intel-536-537/coredrv/Intel537.mod.o
  LD [M]  /root/intel-536-537/coredrv/Intel537.ko
make[3]: Leaving directory `/usr/src/kernels/3.1.6-1.fc16.i686'
make[2]: Leaving directory `/root/intel-536-537/coredrv'
make[1]: Leaving directory `/root/intel-536-537'
[root@pb3000 intel-536-537]#


However, make install fails complaining about kernel version:

[root@pb3000 intel-536-537]# make install
make[1]: Entering directory `/root/intel-536-537'
rm -f /usr/sbin/hamregistry.bin
bash 537_inst
running kernel 3.1.6-1.fc16.i686
unsupported kernel version: only 2.4.x and 2.6.x are supported
make[1]: *** [install] Error 1
make[1]: Leaving directory `/root/intel-536-537'
[root@pb3000 intel-536-537]#


Am I missing something?  Any help here is super appreciated.

Patrick




--
Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/


--
Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/


--
Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/


--
Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/


--
Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/


[Index of Archives]     [Linux Media Development]     [Asterisk]     [DCCP]     [Netdev]     [X.org]     [Xfree86]     [Fedora Women]     [Linux USB]

  Powered by Linux