Re: Calling Verizon from a GOBI [Was: Trying to get a GOBI modem [back] to working]

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 17:26 -0600, Dan Williams wrote:
> On Thu, 2012-11-29 at 16:57 -0500, Adam Tauno Williams wrote:
> > On Thu, 2012-11-29 at 16:28 -0500, Adam Tauno Williams wrote:
> > > On Thu, 2012-11-29 at 15:12 -0500, Adam Tauno Williams wrote:
> > > > Quoting Adam Tauno Williams :
> > > > > Quoting Dan Williams :
> > > > >> On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:
> > > > >>> On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> > > >  openSUSE 12.2 i686
> > > >  NetworkManager-0.9.4.0-5.13.1.i586
> > > >  ModemManager-0.5.2.0-2.4.1.i586
> > > >  3.4.11-2.16-desktop #1 SMP PREEMPT
> > > > >> If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS 
> > > > >> firmware.
> > > > >> Which is why you get the wrong configuration in the dialog that asks 
> > > > >> for
> > > > >> a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
> > > > >> 06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
> > > > >> 88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
> > > > >> e4d8d4dbd0
> > > > > So... with these firmware files the Network Settings still shows the  
> > > > > broadband connection.
> > > Anyone know the secret sauce to get a 3G GOBI 2000 modem to connect to
> > > Verizon's CDMA network.
> > > In the past I believe I have used just a dial string of #777 and an APN
> > > name of "vzwinternet", but it isn't working.
> > > I have one connection on number "*99#" which is what it wants by
> > > default.  Also have a profile #777.  Both have an APN of "vzwinternet"
> > > and a Type of "any".
> > > I can only edit this using the old nm-connection-editor as Network
> > > Settings has the Option button from "Mobile broadband" grayed out.
> > 
> > Hmmm.  Captured a modem-manager log file.  That doesn't look healthy.
> 
> The log indicates that the Gobi still has GSM firmware loaded, not
> Verizon EVDO firmware.  Once the EVDO firmware gets loaded, you'll want
> to go into nm-connection-editor, delete any existing WWAN/3G connection,
> and start from scratch.  For EVDO connections, no configuration is
> necessary, just pick "New connection..." item from GNOME Shell's menu,
> follow the wizard's steps and pick US -> Verizon, then Accept.  No
> numbers are required, no APN necessary.
> 
> ModemManager figures out the right number to dial (if needed, this is
> only required for PPP) in all cases.  Also, APNs are a GSM/UMTS/LTE
> thing, and CDMA/EVDO does not use them at all.
> 
> But in the end, you still have GSM firmware loaded.  We can see this
> because ModemManager sends the "AT+GCAP" request and sees this:
> 
> +GCAP: +CGSM,+DSOK
> 
> If there was CDMA/EVDO firmware loaded, you'd see "IS856" or "IS707"
> instead of CGSM.

Oddly, I can't get gobi_loader to even finish loading firmware on my
system right now (gobi_loader 0.7, kernel 3.6.7) so I can't double-check
that the GCAP info is supposed to be IS707/IS856, but I'm pretty sure it
is.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: [MM] [PATCH v2] serial-port: avoid opening a serial port that has been disposed

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 11:37 -0800, Ben Chan wrote:
> The crash happens to be a reuse of an already disposed MMSerialPort
> object. I'm now tracing the new/dispose paths.

Can you toss a mm_info() into the mm-base-modem.c's dispose() function?
Also one into mm_device_remove_modem().

It looks like the modem is getting disposed but then something is
causing it to get enabled again after wakeup:

ModemManager[2858]:  [1354146967.597474] [mm-base-modem.c:1252]
finalize(): Modem (Samsung) '/sys/devices/s5p-ehci/usb1/1-2/1-2.1'
completely disposed
ModemManager[2858]:   [1354147018.549702]
[mm-broadband-modem.c:7341] enabling_step(): Modem enabling...
ModemManager[2858]:   [1354147018.549838] [mm-iface-modem.c:1176]
mm_iface_modem_update_state():
Modem /org/freedesktop/ModemManager1/Modem/1: state changed (disabled ->
enabling)

Problems actually start earlier here:

ModemManager[2858]:   [1354146961.534481] [mm-iface-modem.c:1176]
mm_iface_modem_update_state():
Modem /org/freedesktop/ModemManager1/Modem/1: state changed (disabling
-> disabled)
ModemManager[2858]:   [1354146967.576974] [mm-serial-port.c:857]
mm_serial_port_open(): (ttyACM1) opening serial port...
ModemManager[2858]:  [1354146967.580958] [mm-serial-port.c:926]
mm_serial_port_open(): (ttyACM1) device open count is 1 (open)

(ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
`G_IS_OBJECT (object)' failed

would be nice to know what object that is, but the only good way of
debugging this is to be in gdb and break on 'g_log' and then get a
backtrace.  not very fun, though you could run without --debug and
reduce the number of times you have to hit "c".

In any case, we have to figure out why the modem is getting disposed,
but apparently isn't fully disposed, and also apparently isn't
un-exported from the bus.  That's what mm_device_remove_modem() should
be doing for us.

Dan

> 
> Ben
> 
> 
> /* Don't read any input if the current command isn't done being
> sent yet */
> info = g_queue_peek_nth (priv->queue, 0);
>49f60:   6940ldr r0, [r0, #20]
> <- 0x151F60
>49f62:   f7c9 ebf6   blx 13750 <_init+0x1254>
> 
> 
> ==20047== Invalid read of size 4
> ==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
> ==20047==  Address 0x14 is not stack'd, malloc'd or (recently) free'd
> ==20047== 
> ==20047== 
> ==20047== Process terminating with default action of signal 11
> (SIGSEGV)
> ==20047==  Access not within mapped region at address 0x14
> ==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
> ==20047==  If you believe this happened as a result of a stack
> ==20047==  overflow in your program's main thread (unlikely but
> ==20047==  possible), you can try to increase the size of the
> ==20047==  main thread stack using the --main-stacksize= flag.
> ==20047==  The main thread stack size used in this run was 8388608.
> 
> 
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
> `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
> `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref:
> assertion `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref:
> assertion `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-WARNING **: gsignal.c:2576: instance
> `0x78624028' has no handler with id `148'
> (ModemManager:2858): GLib-GObject-WARNING **: invalid unclassed
> pointer in cast to `MMSerialPort'
> (ModemManager:2858): GLib-GObject-CRITICAL **:
> g_type_instance_get_private: assertion `instance != NULL &&
> instance->g_class != NULL' failed
> 
> On Thu, Nov 29, 2012 at 11:26 AM, Dan Williams 
> wrote:
> On Tue, 2012-11-27 at 13:19 -0800, Ben Chan wrote:
> > Yes, it's related to the data_available
> (mm-serial-port.c:767) crash
> > (crosbug.com/35391).  I'm running suspend/resume stress test
> with
> > ModemManager under valgrind.
> 
> 
> Any way to get MM running with --debug in those logs?  Also,
> in your
> sources, does line 767 match up with:
> 
> info = g_queue_peek_nth (priv->queue, 0);
> 
> or some other line?
> 
> One other thing to do, put an
> 
> mm_info ("(%s): disposing", mm_port_get_device (MM_PORT
> (self)));
> 
> into dispose() and lets see when it gets disposed and if
> anything tries
> to open it during/after dispose.  Drop another of these into
> mm_serial_port_new().
> 
> Dan
> 
> >
> > Thanks,
> > Ben
> >
> > On Tue, Nov 27, 2012 at 1:13 PM, Aleksander Morgado
> >  wrote:
> > On 11/27/2012 09:39 PM, Ben Chan wrote:
> > > ---
> > >  src/mm-serial-port.c |   10 ++
> > >

Re: [MM] [PATCH v2] serial-port: avoid opening a serial port that has been disposed

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 11:37 -0800, Ben Chan wrote:
> The crash happens to be a reuse of an already disposed MMSerialPort
> object. I'm now tracing the new/dispose paths.

Total aside; lucky you, T-Mobile has PCS HSPA in your location.  Please
share it with the rest of us.

Sincerely yours,
Dan

> 
> Ben
> 
> 
> /* Don't read any input if the current command isn't done being
> sent yet */
> info = g_queue_peek_nth (priv->queue, 0);
>49f60:   6940ldr r0, [r0, #20]
> <- 0x151F60
>49f62:   f7c9 ebf6   blx 13750 <_init+0x1254>
> 
> 
> ==20047== Invalid read of size 4
> ==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
> ==20047==  Address 0x14 is not stack'd, malloc'd or (recently) free'd
> ==20047== 
> ==20047== 
> ==20047== Process terminating with default action of signal 11
> (SIGSEGV)
> ==20047==  Access not within mapped region at address 0x14
> ==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
> ==20047==  If you believe this happened as a result of a stack
> ==20047==  overflow in your program's main thread (unlikely but
> ==20047==  possible), you can try to increase the size of the
> ==20047==  main thread stack using the --main-stacksize= flag.
> ==20047==  The main thread stack size used in this run was 8388608.
> 
> 
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
> `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
> `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref:
> assertion `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref:
> assertion `G_IS_OBJECT (object)' failed
> (ModemManager:2858): GLib-GObject-WARNING **: gsignal.c:2576: instance
> `0x78624028' has no handler with id `148'
> (ModemManager:2858): GLib-GObject-WARNING **: invalid unclassed
> pointer in cast to `MMSerialPort'
> (ModemManager:2858): GLib-GObject-CRITICAL **:
> g_type_instance_get_private: assertion `instance != NULL &&
> instance->g_class != NULL' failed
> 
> On Thu, Nov 29, 2012 at 11:26 AM, Dan Williams 
> wrote:
> On Tue, 2012-11-27 at 13:19 -0800, Ben Chan wrote:
> > Yes, it's related to the data_available
> (mm-serial-port.c:767) crash
> > (crosbug.com/35391).  I'm running suspend/resume stress test
> with
> > ModemManager under valgrind.
> 
> 
> Any way to get MM running with --debug in those logs?  Also,
> in your
> sources, does line 767 match up with:
> 
> info = g_queue_peek_nth (priv->queue, 0);
> 
> or some other line?
> 
> One other thing to do, put an
> 
> mm_info ("(%s): disposing", mm_port_get_device (MM_PORT
> (self)));
> 
> into dispose() and lets see when it gets disposed and if
> anything tries
> to open it during/after dispose.  Drop another of these into
> mm_serial_port_new().
> 
> Dan
> 
> >
> > Thanks,
> > Ben
> >
> > On Tue, Nov 27, 2012 at 1:13 PM, Aleksander Morgado
> >  wrote:
> > On 11/27/2012 09:39 PM, Ben Chan wrote:
> > > ---
> > >  src/mm-serial-port.c |   10 ++
> > >  1 files changed, 10 insertions(+), 0 deletions(-)
> >
> >
> >
> > Don't think it's the correct approach.
> >
> > I don't think we ever run g_object_run_dispose()
> ourselves for
> > port
> > objects, and that means that whenever we got the
> port object
> > disposed it
> > was because it was the last valid reference.
> >
> > And that means that you won't be able to read
> priv->disposed
> > afterwards
> > as that would mean reading already freed memory.
> >
> > There clearly is a missing reference around, or a
> timeout or
> > other
> > source which has the port as user_data and is not
> being
> > properly cleaned
> > up (e.g. removing the timeout or event source when
> the port is
> > disposed). Running it under valgrind should confirm
> this. Is
> > this
> > related to the data_available() crashes during
> suspend/resume?
> >
> >
> > >
> > > diff --git a/src/mm-serial-port.c
> b/src/mm-serial-port.c
> > > index 0a8820d..a33c745 100644
> > > --- a/src/mm-serial-port.c
> > > +++ b/src/mm-serial-port.c
> > > @@ -69,6 +69,7 @@ static guint
> signals[LAST_SIGN

Re: Calling Verizon from a GOBI [Was: Trying to get a GOBI modem [back] to working]

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 16:57 -0500, Adam Tauno Williams wrote:
> On Thu, 2012-11-29 at 16:28 -0500, Adam Tauno Williams wrote:
> > On Thu, 2012-11-29 at 15:12 -0500, Adam Tauno Williams wrote:
> > > Quoting Adam Tauno Williams :
> > > > Quoting Dan Williams :
> > > >> On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:
> > > >>> On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> > >  openSUSE 12.2 i686
> > >  NetworkManager-0.9.4.0-5.13.1.i586
> > >  ModemManager-0.5.2.0-2.4.1.i586
> > >  3.4.11-2.16-desktop #1 SMP PREEMPT
> > > >> If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS 
> > > >> firmware.
> > > >> Which is why you get the wrong configuration in the dialog that asks 
> > > >> for
> > > >> a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
> > > >> 06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
> > > >> 88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
> > > >> e4d8d4dbd0
> > > > So... with these firmware files the Network Settings still shows the  
> > > > broadband connection.
> > Anyone know the secret sauce to get a 3G GOBI 2000 modem to connect to
> > Verizon's CDMA network.
> > In the past I believe I have used just a dial string of #777 and an APN
> > name of "vzwinternet", but it isn't working.
> > I have one connection on number "*99#" which is what it wants by
> > default.  Also have a profile #777.  Both have an APN of "vzwinternet"
> > and a Type of "any".
> > I can only edit this using the old nm-connection-editor as Network
> > Settings has the Option button from "Mobile broadband" grayed out.
> 
> Hmmm.  Captured a modem-manager log file.  That doesn't look healthy.

The log indicates that the Gobi still has GSM firmware loaded, not
Verizon EVDO firmware.  Once the EVDO firmware gets loaded, you'll want
to go into nm-connection-editor, delete any existing WWAN/3G connection,
and start from scratch.  For EVDO connections, no configuration is
necessary, just pick "New connection..." item from GNOME Shell's menu,
follow the wizard's steps and pick US -> Verizon, then Accept.  No
numbers are required, no APN necessary.

ModemManager figures out the right number to dial (if needed, this is
only required for PPP) in all cases.  Also, APNs are a GSM/UMTS/LTE
thing, and CDMA/EVDO does not use them at all.

But in the end, you still have GSM firmware loaded.  We can see this
because ModemManager sends the "AT+GCAP" request and sees this:

+GCAP: +CGSM,+DSOK

If there was CDMA/EVDO firmware loaded, you'd see "IS856" or "IS707"
instead of CGSM.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Calling Verizon from a GOBI [Was: Trying to get a GOBI modem [back] to working]

2012-11-29 Thread Adam Tauno Williams 
On Thu, 2012-11-29 at 16:28 -0500, Adam Tauno Williams wrote:
> On Thu, 2012-11-29 at 15:12 -0500, Adam Tauno Williams wrote:
> > Quoting Adam Tauno Williams :
> > > Quoting Dan Williams :
> > >> On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:
> > >>> On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> >  openSUSE 12.2 i686
> >  NetworkManager-0.9.4.0-5.13.1.i586
> >  ModemManager-0.5.2.0-2.4.1.i586
> >  3.4.11-2.16-desktop #1 SMP PREEMPT
> > >> If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS firmware.
> > >> Which is why you get the wrong configuration in the dialog that asks for
> > >> a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
> > >> 06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
> > >> 88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
> > >> e4d8d4dbd0
> > > So... with these firmware files the Network Settings still shows the  
> > > broadband connection.
> Anyone know the secret sauce to get a 3G GOBI 2000 modem to connect to
> Verizon's CDMA network.
> In the past I believe I have used just a dial string of #777 and an APN
> name of "vzwinternet", but it isn't working.
> I have one connection on number "*99#" which is what it wants by
> default.  Also have a profile #777.  Both have an APN of "vzwinternet"
> and a Type of "any".
> I can only edit this using the old nm-connection-editor as Network
> Settings has the Option button from "Mobile broadband" grayed out.

Hmmm.  Captured a modem-manager log file.  That doesn't look healthy.

Log @


-- 
Adam Tauno Williams  GPG D95ED383
Systems Administrator, Python Developer, LPI / NCLA

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Calling Verizon from a GOBI [Was: Trying to get a GOBI modem [back] to working]

2012-11-29 Thread Adam Tauno Williams 
On Thu, 2012-11-29 at 15:12 -0500, Adam Tauno Williams wrote:
> Quoting Adam Tauno Williams :
> > Quoting Dan Williams :
> >> On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:
> >>> On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
>  openSUSE 12.2 i686
>  NetworkManager-0.9.4.0-5.13.1.i586
>  ModemManager-0.5.2.0-2.4.1.i586
>  3.4.11-2.16-desktop #1 SMP PREEMPT
> >> If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS firmware.
> >> Which is why you get the wrong configuration in the dialog that asks for
> >> a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
> >> 06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
> >> 88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
> >> e4d8d4dbd0
> > So... with these firmware files the Network Settings still shows the  
> > broadband connection.

Anyone know the secret sauce to get a 3G GOBI 2000 modem to connect to
Verizon's CDMA network.

In the past I believe I have used just a dial string of #777 and an APN
name of "vzwinternet", but it isn't working.

I have one connection on number "*99#" which is what it wants by
default.  Also have a profile #777.  Both have an APN of "vzwinternet"
and a Type of "any".

I can only edit this using the old nm-connection-editor as Network
Settings has the Option button from "Mobile broadband" grayed out.

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: [MM] [PATCH v2] serial-port: avoid opening a serial port that has been disposed

2012-11-29 Thread Ben Chan 
The crash happens to be a reuse of an already disposed MMSerialPort object.
I'm now tracing the new/dispose paths.

Ben

/* Don't read any input if the current command isn't done being sent
yet */
info = g_queue_peek_nth (priv->queue, 0);
   49f60:   6940ldr r0, [r0, #20]
  <- 0x151F60
   49f62:   f7c9 ebf6   blx 13750 <_init+0x1254>

==20047== Invalid read of size 4
==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
==20047==  Address 0x14 is not stack'd, malloc'd or (recently) free'd
==20047==
==20047==
==20047== Process terminating with default action of signal 11 (SIGSEGV)
==20047==  Access not within mapped region at address 0x14
==20047==at 0x151F60: ??? (in /usr/sbin/ModemManager)
==20047==  If you believe this happened as a result of a stack
==20047==  overflow in your program's main thread (unlikely but
==20047==  possible), you can try to increase the size of the
==20047==  main thread stack using the --main-stacksize= flag.
==20047==  The main thread stack size used in this run was 8388608.

(ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
`G_IS_OBJECT (object)' failed
(ModemManager:2858): GLib-GObject-CRITICAL **: g_object_ref: assertion
`G_IS_OBJECT (object)' failed
(ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref: assertion
`G_IS_OBJECT (object)' failed
(ModemManager:2858): GLib-GObject-CRITICAL **: g_object_unref: assertion
`G_IS_OBJECT (object)' failed
(ModemManager:2858): GLib-GObject-WARNING **: gsignal.c:2576: instance
`0x78624028' has no handler with id `148'
(ModemManager:2858): GLib-GObject-WARNING **: invalid unclassed pointer in
cast to `MMSerialPort'
(ModemManager:2858): GLib-GObject-CRITICAL **: g_type_instance_get_private:
assertion `instance != NULL && instance->g_class != NULL' failed

On Thu, Nov 29, 2012 at 11:26 AM, Dan Williams  wrote:

> On Tue, 2012-11-27 at 13:19 -0800, Ben Chan wrote:
> > Yes, it's related to the data_available (mm-serial-port.c:767) crash
> > (crosbug.com/35391).  I'm running suspend/resume stress test with
> > ModemManager under valgrind.
>
> Any way to get MM running with --debug in those logs?  Also, in your
> sources, does line 767 match up with:
>
> info = g_queue_peek_nth (priv->queue, 0);
>
> or some other line?
>
> One other thing to do, put an
>
> mm_info ("(%s): disposing", mm_port_get_device (MM_PORT (self)));
>
> into dispose() and lets see when it gets disposed and if anything tries
> to open it during/after dispose.  Drop another of these into
> mm_serial_port_new().
>
> Dan
>
> >
> > Thanks,
> > Ben
> >
> > On Tue, Nov 27, 2012 at 1:13 PM, Aleksander Morgado
> >  wrote:
> > On 11/27/2012 09:39 PM, Ben Chan wrote:
> > > ---
> > >  src/mm-serial-port.c |   10 ++
> > >  1 files changed, 10 insertions(+), 0 deletions(-)
> >
> >
> >
> > Don't think it's the correct approach.
> >
> > I don't think we ever run g_object_run_dispose() ourselves for
> > port
> > objects, and that means that whenever we got the port object
> > disposed it
> > was because it was the last valid reference.
> >
> > And that means that you won't be able to read priv->disposed
> > afterwards
> > as that would mean reading already freed memory.
> >
> > There clearly is a missing reference around, or a timeout or
> > other
> > source which has the port as user_data and is not being
> > properly cleaned
> > up (e.g. removing the timeout or event source when the port is
> > disposed). Running it under valgrind should confirm this. Is
> > this
> > related to the data_available() crashes during suspend/resume?
> >
> >
> > >
> > > diff --git a/src/mm-serial-port.c b/src/mm-serial-port.c
> > > index 0a8820d..a33c745 100644
> > > --- a/src/mm-serial-port.c
> > > +++ b/src/mm-serial-port.c
> > > @@ -69,6 +69,7 @@ static guint signals[LAST_SIGNAL] = { 0 };
> > >
> > >  typedef struct {
> > >  guint32 open_count;
> > > +gboolean disposed;
> > >  gboolean forced_close;
> > >  int fd;
> > >  GHashTable *reply_cache;
> > > @@ -849,6 +850,12 @@ mm_serial_port_open (MMSerialPort
> > *self, GError **error)
> > >
> > >  device = mm_port_get_device (MM_PORT (self));
> > >
> > > +/* If the MMSerialPort object has been disposed, just
> > return an error. */
> > > +if (priv->disposed) {
> > > +mm_info ("(%s) skipped opening serial port that has
> > been disposed", device);
> > > +return FALSE;
> > > +}
> > > +
> > >  if (priv->open_count) {
> > >  /* Already open */
> > >  goto success;
> > > @@ -1537,6 +1544,9 @@ dispose (GOb

Re: Trying to get GOBI modem [back] to working

2012-11-29 Thread Adam Tauno Williams 

Quoting Adam Tauno Williams :


Quoting Dan Williams :


On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:

On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:

openSUSE 12.2 i686
NetworkManager-0.9.4.0-5.13.1.i586
ModemManager-0.5.2.0-2.4.1.i586
3.4.11-2.16-desktop #1 SMP PREEMPT
I have an HP Netbook Pavilion dm1.
Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless
It worked on a previous install, at least at some point.  Now the device
does not appear to be available [we appear to have at least the same
version of ModemManager - I do have
a /usr/lib/ModemManager/libmm-plugin-gobi.so].
Does NetworkManager / ModemManager initialize this device or does that
still fall to the gobi_loader?
Previously in /var/log/message we'd see -
Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
registered for Qualcomm USB modem
Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB0
Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB1
Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB2
The device configured as wwan0 "HP un2420 Mobile Broadband Module".
Now these additional devices do not appear [assuming the firmware is not
initialized].


Nevermind, I now have NetworkManager recognizing the device [at least].



It was a firmware issue.

Product: HP un2420 Mobile Broadband Module
/lib/firmware/gobi # ls -l
total 14236
-rw-r--r-- 1 root root 11333684 Nov 29 13:40 amss.mbn
-rw-r--r-- 1 root root  3223596 Nov 29 13:40 apps.mbn
-rw-r--r-- 1 root root10748 Nov 29 13:40 UQCN.mbn
pc02813:/lib/firmware/gobi # md5sum *
80fcfbb41a7d4331d4b7145972f5f3c4  amss.mbn
00cbd411048cdadc3e4caf0d89d14fca  apps.mbn
bdf27325ebb63251c1310cd3a8f7bab6  UQCN.mbn

If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS firmware.
Which is why you get the wrong configuration in the dialog that asks for
a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
e4d8d4dbd0
So... with these firmware files the Network Settings still shows the  
broadband connection.


But turning the connection On immediately results in a  
Activation-of-network-failed message.


It doesn't log anything to /var/log/messages or the ring buffer.

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Trying to get GOBI modem [back] to working

2012-11-29 Thread Adam Tauno Williams 

Quoting Dan Williams :


On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:

On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> openSUSE 12.2 i686
> NetworkManager-0.9.4.0-5.13.1.i586
> ModemManager-0.5.2.0-2.4.1.i586
> 3.4.11-2.16-desktop #1 SMP PREEMPT
> I have an HP Netbook Pavilion dm1.
> Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless
> It worked on a previous install, at least at some point.  Now the device
> does not appear to be available [we appear to have at least the same
> version of ModemManager - I do have
> a /usr/lib/ModemManager/libmm-plugin-gobi.so].
> Does NetworkManager / ModemManager initialize this device or does that
> still fall to the gobi_loader?
> Previously in /var/log/message we'd see -
> Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
> registered for Qualcomm USB modem
> Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB0
> Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB1
> Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB2
> The device configured as wwan0 "HP un2420 Mobile Broadband Module".
> Now these additional devices do not appear [assuming the firmware is not
> initialized].

Nevermind, I now have NetworkManager recognizing the device [at least].



It was a firmware issue.

Product: HP un2420 Mobile Broadband Module
/lib/firmware/gobi # ls -l
total 14236
-rw-r--r-- 1 root root 11333684 Nov 29 13:40 amss.mbn
-rw-r--r-- 1 root root  3223596 Nov 29 13:40 apps.mbn
-rw-r--r-- 1 root root10748 Nov 29 13:40 UQCN.mbn
pc02813:/lib/firmware/gobi # md5sum *
80fcfbb41a7d4331d4b7145972f5f3c4  amss.mbn
00cbd411048cdadc3e4caf0d89d14fca  apps.mbn
bdf27325ebb63251c1310cd3a8f7bab6  UQCN.mbn

If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS firmware.
Which is why you get the wrong configuration in the dialog that asks for
a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:
06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
e4d8d4dbd0a10d17f01f7f3bbd2ea734  1/UQCN.mbn


So... with these firmware files the Network Settings still shows the  
broadband connection.


But with either firmwares the "Options" button is greyed out, except  
sometimes for few seconds during connection attempt.  That has got to  
be a bug???



It also doesn't appear to be possible to delete/recreate the mobile  
broadband connection.


<>___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: [MM] [PATCH v2] serial-port: avoid opening a serial port that has been disposed

2012-11-29 Thread Dan Williams 
On Tue, 2012-11-27 at 13:19 -0800, Ben Chan wrote:
> Yes, it's related to the data_available (mm-serial-port.c:767) crash
> (crosbug.com/35391).  I'm running suspend/resume stress test with
> ModemManager under valgrind.

Any way to get MM running with --debug in those logs?  Also, in your
sources, does line 767 match up with:

info = g_queue_peek_nth (priv->queue, 0);

or some other line?

One other thing to do, put an

mm_info ("(%s): disposing", mm_port_get_device (MM_PORT (self)));

into dispose() and lets see when it gets disposed and if anything tries
to open it during/after dispose.  Drop another of these into
mm_serial_port_new().

Dan

> 
> Thanks,
> Ben
> 
> On Tue, Nov 27, 2012 at 1:13 PM, Aleksander Morgado
>  wrote:
> On 11/27/2012 09:39 PM, Ben Chan wrote:
> > ---
> >  src/mm-serial-port.c |   10 ++
> >  1 files changed, 10 insertions(+), 0 deletions(-)
> 
> 
> 
> Don't think it's the correct approach.
> 
> I don't think we ever run g_object_run_dispose() ourselves for
> port
> objects, and that means that whenever we got the port object
> disposed it
> was because it was the last valid reference.
> 
> And that means that you won't be able to read priv->disposed
> afterwards
> as that would mean reading already freed memory.
> 
> There clearly is a missing reference around, or a timeout or
> other
> source which has the port as user_data and is not being
> properly cleaned
> up (e.g. removing the timeout or event source when the port is
> disposed). Running it under valgrind should confirm this. Is
> this
> related to the data_available() crashes during suspend/resume?
> 
> 
> >
> > diff --git a/src/mm-serial-port.c b/src/mm-serial-port.c
> > index 0a8820d..a33c745 100644
> > --- a/src/mm-serial-port.c
> > +++ b/src/mm-serial-port.c
> > @@ -69,6 +69,7 @@ static guint signals[LAST_SIGNAL] = { 0 };
> >
> >  typedef struct {
> >  guint32 open_count;
> > +gboolean disposed;
> >  gboolean forced_close;
> >  int fd;
> >  GHashTable *reply_cache;
> > @@ -849,6 +850,12 @@ mm_serial_port_open (MMSerialPort
> *self, GError **error)
> >
> >  device = mm_port_get_device (MM_PORT (self));
> >
> > +/* If the MMSerialPort object has been disposed, just
> return an error. */
> > +if (priv->disposed) {
> > +mm_info ("(%s) skipped opening serial port that has
> been disposed", device);
> > +return FALSE;
> > +}
> > +
> >  if (priv->open_count) {
> >  /* Already open */
> >  goto success;
> > @@ -1537,6 +1544,9 @@ dispose (GObject *object)
> >  {
> >  MMSerialPortPrivate *priv = MM_SERIAL_PORT_GET_PRIVATE
> (object);
> >
> > +/* Mark the MMSerialPort object as disposed to prevent
> it from being re-opened. */
> > +priv->disposed = TRUE;
> > +
> >  if (priv->timeout_id) {
> >  g_source_remove (priv->timeout_id);
> >  priv->timeout_id = 0;
> >
> 
> 
> --
> 
> Aleksander
> 
> 
> ___
> networkmanager-list mailing list
> networkmanager-list@gnome.org
> https://mail.gnome.org/mailman/listinfo/networkmanager-list


___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Trying to get GOBI modem [back] to working

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 14:12 -0500, Adam Tauno Williams wrote:
> On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> > openSUSE 12.2 i686
> > NetworkManager-0.9.4.0-5.13.1.i586
> > ModemManager-0.5.2.0-2.4.1.i586
> > 3.4.11-2.16-desktop #1 SMP PREEMPT
> > I have an HP Netbook Pavilion dm1.
> > Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless
> > It worked on a previous install, at least at some point.  Now the device
> > does not appear to be available [we appear to have at least the same
> > version of ModemManager - I do have
> > a /usr/lib/ModemManager/libmm-plugin-gobi.so].
> > Does NetworkManager / ModemManager initialize this device or does that
> > still fall to the gobi_loader?
> > Previously in /var/log/message we'd see -
> > Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
> > registered for Qualcomm USB modem
> > Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
> > USB modem converter detected
> > Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
> > converter now attached to ttyUSB0
> > Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
> > USB modem converter detected
> > Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
> > converter now attached to ttyUSB1
> > Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
> > USB modem converter detected
> > Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
> > converter now attached to ttyUSB2
> > The device configured as wwan0 "HP un2420 Mobile Broadband Module".
> > Now these additional devices do not appear [assuming the firmware is not
> > initialized].
> 
> Nevermind, I now have NetworkManager recognizing the device [at least].
> 
> 
> 
> It was a firmware issue.
> 
> Product: HP un2420 Mobile Broadband Module
> /lib/firmware/gobi # ls -l
> total 14236
> -rw-r--r-- 1 root root 11333684 Nov 29 13:40 amss.mbn
> -rw-r--r-- 1 root root  3223596 Nov 29 13:40 apps.mbn
> -rw-r--r-- 1 root root10748 Nov 29 13:40 UQCN.mbn
> pc02813:/lib/firmware/gobi # md5sum *
> 80fcfbb41a7d4331d4b7145972f5f3c4  amss.mbn
> 00cbd411048cdadc3e4caf0d89d14fca  apps.mbn
> bdf27325ebb63251c1310cd3a8f7bab6  UQCN.mbn

If it's Verizon, you need to load CDMA/EVDO firmware, not UMTS firmware.
Which is why you get the wrong configuration in the dialog that asks for
a "Plan Name".  Try directory '1' perhaps; I have these MD5 sums:

06f76ed398458dad7b91c2d99a85a0a7  1/amss.mbn
88a60ed745d75fb1b92c539574ecc972  1/apps.mbn
e4d8d4dbd0a10d17f01f7f3bbd2ea734  1/UQCN.mbn

One way to check is to do something like:

cd /where/your/firmware/is
strings 0/* | grep -i vzw
strings 1/* | grep -i vzw
strings 2/* | grep -i vzw
strings 3/* | grep -i vzw
(etc)

until you get a hit for something, like @vzw3g.com.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Trying to get GOBI modem [back] to working

2012-11-29 Thread Adam Tauno Williams 
On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> openSUSE 12.2 i686
> NetworkManager-0.9.4.0-5.13.1.i586
> ModemManager-0.5.2.0-2.4.1.i586
> 3.4.11-2.16-desktop #1 SMP PREEMPT
> I have an HP Netbook Pavilion dm1.
> Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless
> It worked on a previous install, at least at some point.  Now the device
> does not appear to be available [we appear to have at least the same
> version of ModemManager - I do have
> a /usr/lib/ModemManager/libmm-plugin-gobi.so].
> Does NetworkManager / ModemManager initialize this device or does that
> still fall to the gobi_loader?
> Previously in /var/log/message we'd see -
> Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
> registered for Qualcomm USB modem
> Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB0
> Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB1
> Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB2
> The device configured as wwan0 "HP un2420 Mobile Broadband Module".
> Now these additional devices do not appear [assuming the firmware is not
> initialized].

Nevermind, I now have NetworkManager recognizing the device [at least].



It was a firmware issue.

Product: HP un2420 Mobile Broadband Module
/lib/firmware/gobi # ls -l
total 14236
-rw-r--r-- 1 root root 11333684 Nov 29 13:40 amss.mbn
-rw-r--r-- 1 root root  3223596 Nov 29 13:40 apps.mbn
-rw-r--r-- 1 root root10748 Nov 29 13:40 UQCN.mbn
pc02813:/lib/firmware/gobi # md5sum *
80fcfbb41a7d4331d4b7145972f5f3c4  amss.mbn
00cbd411048cdadc3e4caf0d89d14fca  apps.mbn
bdf27325ebb63251c1310cd3a8f7bab6  UQCN.mbn




___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Wifi sharing

2012-11-29 Thread Dan Williams 
On Tue, 2012-11-27 at 22:37 +0100, Michael Below wrote:
> Sorry, too much traffic for me, unsubscribing. I'd be happy if you
> contact me off-list about the Wifi sharing functionality.
> 
> (So far I haven't found a single person successfully using
> networkmanager for wifi sharing, so something like "yes, I have managed
> to configure an access point" would be a fine start...)

Full AP-mode (ie, Infrastructure) support was added after NetworkManager
0.9.6 and will be in the next release.  It also requires a reasonably
new version of wpa_supplicant, at least anything 1.0 or later should
work.

NetworkManager 0.9.6 and earlier only support ad-hoc mode "hotspot"
functionality, and WPA is disabled here because there are kernel driver
bugs that make your supposedly ad-hoc WPA hotspot actually be created as
an open/unencrypted hotspot, which isn't cool.  Thus only open and WEP
are supported.

There's ongoing work to also add RSN IBSS (ie, WPA2 Ad-Hoc) support for
hotspots in NetworkManager, which is useful if your hardware doesn't
support full AP mode which even a lot of recent cards don't.  That has
yet to land.

Dan

> Am Sonntag, den 25.11.2012, 16:44 +0100 schrieb Michael Below:
> > Hi,
> > 
> > I would like to use networkmanager for Wifi sharing. The situation:
> > 
> > I have a Debian testing AMD64 system, connected to the LAN, with an
> > additional D-Link System AirPlus G DWL-G122 Wireless Adapter(rev.C1)
> > [Ralink RT2571W] I want to use to occasionally provide Wifi for some
> > devices.
> > 
> > After trying this with networkmanager unsuccessfully for a first time, I
> > set up an access point using hostapd. This works fine in general, but
> > the GNOME apps like evolution, pidgin etc. believe the network is down,
> > so they don't connect to the network. Probably they ask networkmanager
> > for the status of eth0, which isn't up. I didn't find a way to tell
> > networkmanager to look for br0 instead.
> > 
> > Now I tried setting it up again with networkmanager. The trouble starts
> > in the basic setup: networkmanager doesn't remember my settings. I want
> > Infrastructure mode with WPA, not Ad hoc mode with WEP. Basically, I
> > just want networkmanager to do it like I set it up with hostapd.
> > networkmanager shouldn't forget my settings every time I disable the
> > network.
> > 
> > The networkmanager packages are version 0.9.4
> > 
> > Is there a way to use networkmanager for Wifi sharing? Either to get it
> > to notice that br0 is there, or to have it setup a stable access point
> > of its own? 
> > 
> > Cheers
> > 
> > Michael
> > 
> 
> 


___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: 0.9.6.4 and 0.9.7.0

2012-11-29 Thread Dan Williams 
On Wed, 2012-11-28 at 07:47 -0500, Gene Czarcinski wrote:
> I am more than a little confused.
> 
> There was an announcement on this list that 0.9.6.4 is the latest stable 
> version for NM and the applet and these updates are available for Fedora 17.
> 
> Fedora 18 has entered beta and includes 0.9.7.0 (git20121004) which can 
> be rebuilt for Fedora 17.  I also notice that Fedora has its own git 
> repository for NetworkManager.

The Fedora git thing is actually *package* git, which doesn't contain
NetworkManager source files, but really just stores the RPM specfile and
associated configuration for the Fedora distro.  So upstream
NetworkManager (which is distro-agnostic) has all the sources, while
downstream Fedora package git has all the Fedora-specific stuff in it.

Dan

> Why are there two such streams?  I do not understand.  Could someone 
> please explain?
> 
> Gene
> ___
> networkmanager-list mailing list
> networkmanager-list@gnome.org
> https://mail.gnome.org/mailman/listinfo/networkmanager-list


___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Sierra USB 306 - GetSignalQuality() fails when connected: "Could not parse the +CIND response"

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 11:18 -0600, Dan Williams wrote:
> On Tue, 2012-11-27 at 17:09 -0800, Kelsey Sigurdur wrote:
> > On Tue, Nov 27, 2012 at 2:29 AM, Bjørn Mork  wrote:
> > > Kelsey Sigurdur  writes:
> > >
> > >> I do have the sierra_net driver on my system but you are correct that
> > >> it's not getting bound to the device.
> > >>
> > >> For now I'm guessing it's a kernel issue. Google, and my rudimentary
> > >> understanding of C, suggests that the USB 306 is blacklisted in
> > >> sierra.c
> > >
> > > The sierra.c comment refers to blacklisting the DirectIP USB interfaces
> > > in the serial driver, giving the sierra_net driver a chance to handle
> > > them instead.  This is a prerequisite for using sierra_net at all.
> > >
> > >> but would work with a firmware version >= M3.0.
> > >> Unfortunately,  the Sierra Wireless site offers only M2_0_11_10AP as
> > >> the latest firmware for this device.
> > >>
> > >> See http://lwn.net/Articles/385096/ if you're interested in where I
> > >> came across the blacklist info.
> > >
> > > Right.  If the firmware version is the problem, then your device is
> > > probably failing this test in the sierra_net driver:
> > >
> > > /* verify fw attributes */
> > > status = sierra_net_get_fw_attr(dev, &fwattr);
> > > dev_dbg(&dev->udev->dev, "Fw attr: %x\n", fwattr);
> > >
> > > /* test whether firmware supports DHCP */
> > > if (!(status == sizeof(fwattr) && (fwattr & 
> > > SWI_GET_FW_ATTR_MASK))) {
> > > /* found incompatible firmware version */
> > > dev_err(&dev->udev->dev, "Incompatible driver and 
> > > firmware"
> > > " versions\n");
> > > kfree(priv);
> > > return -ENODEV;
> > > }
> > >
> > >
> > > You should be able to verify that easily by looking for the
> > > "Incompatible driver and firmware versions" message.  Is that present in
> > > your kernel logs?
> > >
> > 
> > Yes that error is present. That error is actually what prompted me to
> > start peeking at the source code.
> > 
> > 
> > > If so, then I don't think there is much anyone can do here.  We can only
> > > assume there is a valid reason for Sierra to put that test in the
> > > driver.  Did you try requesting newer firmware from Sierra Wireless?  In
> > > my experience they don't necessarily put all available firmwares on
> > > their public site.
> > >
> > >
> > >
> > > Bjørn
> > 
> > 
> > Requesting newer firmware is my next step. Hopefully citing the email
> > from Elina, and Dans' note about having version 3.0 on his 306, will
> > make acquiring the update a relatively painless process.
> 
> I received my 306 as an engineering sample directly from Sierra, but
> that was early 2010, and I've completely forgotten what firmware version
> it came with.  It currently has M3_0_10_1AP, but I can't find any
> references to that firmware version anywhere, which makes me think that
> perhaps my version wasn't released publicly.  None of the USB306
> variants have released upgrades to v3 firmware that I can see.
> 
> In any case, I'll go ahead and add the 306 to the PPP check.

Pushed to MM_06 and master branches.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Trying to get GOBI modem [back] to working

2012-11-29 Thread Dan Williams 
On Thu, 2012-11-29 at 13:12 -0500, Adam Tauno Williams wrote:
> openSUSE 12.2 i686
> NetworkManager-0.9.4.0-5.13.1.i586
> ModemManager-0.5.2.0-2.4.1.i586
> 3.4.11-2.16-desktop #1 SMP PREEMPT
> 
> I have an HP Netbook Pavilion dm1.
> Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless
> 
> It worked on a previous install, at least at some point.  Now the device
> does not appear to be available [we appear to have at least the same
> version of ModemManager - I do have
> a /usr/lib/ModemManager/libmm-plugin-gobi.so].
> 
> Does NetworkManager / ModemManager initialize this device or does that
> still fall to the gobi_loader?
> 
> Previously in /var/log/message we'd see -
> 
> Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
> registered for Qualcomm USB modem
> Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB0
> Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB1
> Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
> USB modem converter detected
> Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
> converter now attached to ttyUSB2
> 
> The device configured as wwan0 "HP un2420 Mobile Broadband Module".
> 
> Now these additional devices do not appear [assuming the firmware is not
> initialized].

Firmware is still handled by gobi_loader.  If you only see one ttyUSBx
port, then the device is waiting for firmware.  If you see more than
one, or you see wwan0 or cdc-wdm0, then the device's firmware is alive
and it should be picked up by ModemManager.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Trying to get GOBI modem [back] to working

2012-11-29 Thread Adam Tauno Williams 
openSUSE 12.2 i686
NetworkManager-0.9.4.0-5.13.1.i586
ModemManager-0.5.2.0-2.4.1.i586
3.4.11-2.16-desktop #1 SMP PREEMPT

I have an HP Netbook Pavilion dm1.
Device is 03f0:241d Hewlett-Packard Gobi 2000 Wireless

It worked on a previous install, at least at some point.  Now the device
does not appear to be available [we appear to have at least the same
version of ModemManager - I do have
a /usr/lib/ModemManager/libmm-plugin-gobi.so].

Does NetworkManager / ModemManager initialize this device or does that
still fall to the gobi_loader?

Previously in /var/log/message we'd see -

Sep 27 18:32:14 linux kernel: [3.303672] USB Serial support
registered for Qualcomm USB modem
Sep 27 18:32:14 linux kernel: [3.305245] qcserial 1-5:1.1: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.305360] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB0
Sep 27 18:32:14 linux kernel: [3.307537] qcserial 1-5:1.2: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.307810] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB1
Sep 27 18:32:14 linux kernel: [3.309392] qcserial 1-5:1.3: Qualcomm
USB modem converter detected
Sep 27 18:32:14 linux kernel: [3.309549] usb 1-5: Qualcomm USB modem
converter now attached to ttyUSB2

The device configured as wwan0 "HP un2420 Mobile Broadband Module".

Now these additional devices do not appear [assuming the firmware is not
initialized].

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Sierra USB 306 - GetSignalQuality() fails when connected: "Could not parse the +CIND response"

2012-11-29 Thread Dan Williams 
On Tue, 2012-11-27 at 17:09 -0800, Kelsey Sigurdur wrote:
> On Tue, Nov 27, 2012 at 2:29 AM, Bjørn Mork  wrote:
> > Kelsey Sigurdur  writes:
> >
> >> I do have the sierra_net driver on my system but you are correct that
> >> it's not getting bound to the device.
> >>
> >> For now I'm guessing it's a kernel issue. Google, and my rudimentary
> >> understanding of C, suggests that the USB 306 is blacklisted in
> >> sierra.c
> >
> > The sierra.c comment refers to blacklisting the DirectIP USB interfaces
> > in the serial driver, giving the sierra_net driver a chance to handle
> > them instead.  This is a prerequisite for using sierra_net at all.
> >
> >> but would work with a firmware version >= M3.0.
> >> Unfortunately,  the Sierra Wireless site offers only M2_0_11_10AP as
> >> the latest firmware for this device.
> >>
> >> See http://lwn.net/Articles/385096/ if you're interested in where I
> >> came across the blacklist info.
> >
> > Right.  If the firmware version is the problem, then your device is
> > probably failing this test in the sierra_net driver:
> >
> > /* verify fw attributes */
> > status = sierra_net_get_fw_attr(dev, &fwattr);
> > dev_dbg(&dev->udev->dev, "Fw attr: %x\n", fwattr);
> >
> > /* test whether firmware supports DHCP */
> > if (!(status == sizeof(fwattr) && (fwattr & SWI_GET_FW_ATTR_MASK))) 
> > {
> > /* found incompatible firmware version */
> > dev_err(&dev->udev->dev, "Incompatible driver and firmware"
> > " versions\n");
> > kfree(priv);
> > return -ENODEV;
> > }
> >
> >
> > You should be able to verify that easily by looking for the
> > "Incompatible driver and firmware versions" message.  Is that present in
> > your kernel logs?
> >
> 
> Yes that error is present. That error is actually what prompted me to
> start peeking at the source code.
> 
> 
> > If so, then I don't think there is much anyone can do here.  We can only
> > assume there is a valid reason for Sierra to put that test in the
> > driver.  Did you try requesting newer firmware from Sierra Wireless?  In
> > my experience they don't necessarily put all available firmwares on
> > their public site.
> >
> >
> >
> > Bjørn
> 
> 
> Requesting newer firmware is my next step. Hopefully citing the email
> from Elina, and Dans' note about having version 3.0 on his 306, will
> make acquiring the update a relatively painless process.

I received my 306 as an engineering sample directly from Sierra, but
that was early 2010, and I've completely forgotten what firmware version
it came with.  It currently has M3_0_10_1AP, but I can't find any
references to that firmware version anywhere, which makes me think that
perhaps my version wasn't released publicly.  None of the USB306
variants have released upgrades to v3 firmware that I can see.

In any case, I'll go ahead and add the 306 to the PPP check.

Dan

___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Alternative way to set static ip for Ubuntu

2012-11-29 Thread Uwe Geuder 
On 29 November 2012 13:03, Pavel Simerda psimerda redhat com wrote:

>> From: "footkeong Chien" 
>> I would like to know whether there is any alternative way to set
>> static ip for Ubuntu using command lines without affecting the
>> etc/network/interfaces.
...
> Is this a NetworkManager question? 

I guess it is. Ubuntu uses NetworkManager and at least I would not try
to fiddle with a interface managed by NetworkManager with some basic
commands, because you never know what NetworkManager will override
(well "you" as in average user, not "you" as those on this list
who really know all NM internals)

Unfortunately the original poster did not write to what (working?)
approach he was looking for an alternative, why he wanted to use the
command line, does he want the static IP on eth0 or some other
interface, and does he want the assignment to survive a reboot.

Earlier this week I had the issue that I wanted to assign a permanent
static IP to my eth0 on a Kubuntu Lucid system. Intuitively I would
have done that by setting a static address on the "auto eth0"
connection in the KDE connections control module. However, that does
not work because at least in Kubuntu Lucid "auto eth0" is somehow
automagically hidden from the UI. I have the feeling that I might have
seen it in Ubuntu systems (using gnome connection manager), but I
cannot check at the moment. Yes, I also tried to create a new
connection, but that connected only manually on the first attempt and
caused a crash (in knetworkmanager I believe) in the second one, so I
did not pursue that approach any further.

On the net I found the same question many times, but no really good
answers, which seems to support the impression that people are
struggling with the issue. I found e.g. these instructions
http://en.opensuse.org/SDB:Knetworkmanager_static_ip_auto_eth0 too
complicated for my purpose.

So I just edited /etc/network/interfaces instead and defined eth0 as
auto and static. That leads to desired functionality. NetworkManager
regards the interface as unmanaged. However the knetworkmanager user
experience is non-intuitive in this case. It shows me a disconnected
Ethernet cable. Somewhat irritating at least the first week, but I can
live with it.

Still from the "naive" end user standpoint it is a bit weird that
NetworkManager cannot handle such simple use case. (Yes, technically
the limit might not be in NetworkManager proper but in KDE connection
editor. And static addresses is not the main reason why NetworkManager
exists.) But regardless of the reason, there is still a bit to go
until Linux Networking is made really easy... Or have things changed
completely since then? Lucid is of course 2.5 years old.


Regards,

Uwe
___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: Alternative way to set static ip for Ubuntu

2012-11-29 Thread Pavel Simerda 
- Original Message -
> From: "footkeong Chien" 
> I would like to know whether there is any alternative way to set
> static ip for Ubuntu using command lines without affecting the
> etc/network/interfaces.
> Thanks in advance.

Is this a NetworkManager question? Type 'man ip' in the commandline
(and install iproute/iproute2 package if missing).
___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Alternative way to set static ip for Ubuntu

2012-11-29 Thread footkeong Chien 

I would like to know whether there is any alternative way to set static ip for 
Ubuntu using command lines without affecting the etc/network/interfaces.
Thanks in advance.___
networkmanager-list mailing list
networkmanager-list@gnome.org
https://mail.gnome.org/mailman/listinfo/networkmanager-list