Re: [PATCH v2 3/3] isimodem: header updates for ISI2.5

2011-02-08 Thread Aki Niemi 
Hi,

2011/2/1 Aki Niemi aki.ni...@nokia.com:
 I'm starting to think me just saying this out loud is not concrete
 enough, so I guess one way forward here could be for me to go ahead and
 adapt e.g. the netreg driver to work with N900 and later ISI modems. The
 netreg driver would make a good example, as it is one of those where the
 resource ID changes post N900, and the messages do vary based on it.

I went ahead and pushed some patches refactoring the isimodem netreg
driver to work with both the old modem API as well as the new wgmodem
2.5 version.

Please take a look if it works for you.

Cheers,
Aki
___
ofono mailing list
ofono@ofono.org
http://lists.ofono.org/listinfo/ofono

Re: [PATCH v2 3/3] isimodem: header updates for ISI2.5

2011-02-04 Thread Aki Niemi 
Hi Andreas,

2011/2/3 Andreas WESTIN andreas.wes...@stericsson.com:
 I don't see what that has to do with the device supporting the PC
 suite USB profile, which allows modem access using cdc_phonet.

 Currently our platform does not support this.

Well, I know of platforms that do, and have different versions of ISI modems.

 That is exactly what I meant. This version allows building quirks inside
 the drivers. An example in pseudo-code:

 if GIsiVersion.major == 15 and GIsiVersion.minor= 3:
        send( { FOO, value, filler, filler}, callback1 )
 else:
        send( { FOO, value }, callback2 )


 Yes, we use a similar solution in gisi, but Rémi pointed out that it would
 be a maintenance problem once there are more versions to support and I tend
 to agree. A solution could be to use a lookup function, but this function
 would need to return something to compare with, an enum perhaps.
 Also are you certain that e.g major == 15 and minor == 25 is still the same
 ISI version, if not then the lookup function could become messy.

The PNS_SUBSCRIBED_RESOURCES_EXTEND_IND is a bit of a special case, as
it requires quirking inside GIsiModem. But see my other response in
that thread.

Cheers,
Aki
___
ofono mailing list
ofono@ofono.org
http://lists.ofono.org/listinfo/ofono

Re: [PATCH v2 3/3] isimodem: header updates for ISI2.5

2011-02-02 Thread Aki Niemi 
On Wed, 2011-02-02 at 10:03 +0100, ext Andreas WESTIN wrote:
 On 2011-02-01 20:32, Aki Niemi wrote:
  Like I replied earlier, I don't see the need for this. Also, it will not
  work well with the isiusb plugin, which has the potential to work with
  any ISI modem. I'm not willing to break that potential until we just
  absolutely have to.
 
 U8500 uses shrm driver and will not be recognized via isiusb.

I don't see what that has to do with the device supporting the PC
suite USB profile, which allows modem access using cdc_phonet.

  So, my stance holds: we quirk based on resource ID and based on the
  server's ISI version.
 
  From the information we have regarding ISI version from each server is 
 that it is not consistent, meaning the value you get is not really the 
 ISI version but the version of the server.

That is exactly what I meant. This version allows building quirks inside
the drivers. An example in pseudo-code:

if GIsiVersion.major == 15 and GIsiVersion.minor = 3:
send( { FOO, value, filler, filler}, callback1 )
else:
send( { FOO, value }, callback2 )

Cheers,
Aki

___
ofono mailing list
ofono@ofono.org
http://lists.ofono.org/listinfo/ofono

Re: [PATCH v2 3/3] isimodem: header updates for ISI2.5

2011-02-01 Thread Marcel Holtmann 
Hi Jessica,

  drivers/isimodem/debug.c   |  132 
 
  drivers/isimodem/debug.h   |9 +++
  drivers/isimodem/mtc.h |   38 +
  drivers/isimodem/version.h |   26 +
  4 files changed, 205 insertions(+), 0 deletions(-)
  create mode 100644 drivers/isimodem/version.h

snip

 +enum isi_version {
 + ISI_20 = 0,
 + ISI_25 = 1
 +};

so what is the final decision here now. ISI doing auto-detection and
this gets not exposed to the modem drivers or we do.

Aki, I remember you saying that this is not needed at all.

Regards

Marcel


___
ofono mailing list
ofono@ofono.org
http://lists.ofono.org/listinfo/ofono

Re: [PATCH v2 3/3] isimodem: header updates for ISI2.5

2011-02-01 Thread Aki Niemi 
Hi Marcel,

On Tue, 2011-02-01 at 16:43 +0100, ext Marcel Holtmann wrote:
  +enum isi_version {
  +   ISI_20 = 0,
  +   ISI_25 = 1
  +};
 
 so what is the final decision here now. ISI doing auto-detection and
 this gets not exposed to the modem drivers or we do.

Like I replied earlier, I don't see the need for this. Also, it will not
work well with the isiusb plugin, which has the potential to work with
any ISI modem. I'm not willing to break that potential until we just
absolutely have to.

So, my stance holds: we quirk based on resource ID and based on the
server's ISI version.

I'm starting to think me just saying this out loud is not concrete
enough, so I guess one way forward here could be for me to go ahead and
adapt e.g. the netreg driver to work with N900 and later ISI modems. The
netreg driver would make a good example, as it is one of those where the
resource ID changes post N900, and the messages do vary based on it.

Cheers,
Aki

___
ofono mailing list
ofono@ofono.org
http://lists.ofono.org/listinfo/ofono