[Wio 3G] Adding platform #7098

ytsuboi · 2018-06-04T07:12:58Z

Description

Adding new platform Wio 3G, STM32F439VI with Quectel UG96 on-board 3G modem.

I don't want to use preprocessor macro to switch AT command. But since member function of cellular feature such as set_context_to_be_activated() are private, I decided to use macro to reduce shared code changes ( especially changing code structure ).

Test results:
GCC
armcc
EWARM

Pull request type

[ ] Fix
[ ] Refactor
[x] New target
[ ] Feature
[ ] Breaking change

0xc0170 · 2018-06-04T10:02:38Z

@ytsuboi only GCC ARM test results provided?

ytsuboi · 2018-06-05T04:39:13Z

@0xc0170 Added on description.
@toyowata Thanks for your help!

ashok-rao · 2018-06-05T10:34:16Z

targets/targets.json

+        "config": {
+            "clock_source": {
+                "help": "Mask value : USE_PLL_HSE_EXTC | USE_PLL_HSE_XTAL (need HW patch) | USE_PLL_HSI",
+                "value": "USE_PLL_HSE_XTAL|USE_PLL_HSI",


Question: What is the HW patch? Is this explained somewhere in the docs?

Oops, no need hardware patch. XTAL is mounted on the board. Will remove "need HW patch" on next commit.

ashok-rao · 2018-06-05T10:36:01Z

targets/TARGET_STM/TARGET_STM32F4/TARGET_STM32F439xI/TARGET_WIO_3G/PinNames.h

+    USB_OTG_HS_ULPI_DIR = PC_2,
+    USB_OTG_HS_ULPI_NXT = PC_3,
+*/
+


I think all of these un-used USB pins can be removed..

ashok-rao · 2018-06-05T10:37:03Z

targets/TARGET_STM/TARGET_STM32F4/TARGET_STM32F439xI/TARGET_WIO_3G/PinNames.h

+    LED_RED     = LED2,
+    USER_BUTTON = PC_13,
+    // Standardized button names
+//    BUTTON1 = USER_BUTTON,


This one is required for consistency.

ashok-rao

LGTM. Thanks, nice work!

0xc0170 · 2018-06-05T13:48:32Z

@ARMmbed/mbed-os-wan Can you review cellular additions here?

AnttiKauppila · 2018-06-05T15:01:55Z

features/cellular/framework/AT/AT_CellularNetwork.cpp

@@ -429,11 +429,21 @@ nsapi_error_t AT_CellularNetwork::set_context_to_be_activated()

    // if user has defined user name and password we need to call CGAUTH before activating or modifying context
    if (_pwd && _uname) {
+#if defined(TARGET_WIO_3G)


This is not a correct approach. Now I understand the context of your earlier question about set_context_to_be_activated accessibility. I agree it does not make sense to copy paste activate_context method to your own class.
Please propose a solution where set_context_to_be_activated is changed to protected and virtual method, so you can override that in your class.

What about introducing new protected virtual method for "user authentication" only and override that one?

I would like @mirelachirica approach, clear method for authentication. Does not need to be in API level.

I can solve this issue with @AnttiKauppila approach soon, but @mirelachirica approach is too heavy for me. ( I don't want to change structures of shared/common code, since it could affect others platform. ) In addition, I don't want to support separating a function from AT command to response.

So, can I change set_context_to_be_activated() to protected and virtual method? @mirelachirica @jarvte

I don't see any risk in changing the shared code something like this:

nsapi_error_t AT_CellularNetwork::set_context_to_be_activated() { // try to find or create context with suitable stack if (!get_context()) { return NSAPI_ERROR_NO_CONNECTION; } // if user has defined user name and password we need to call CGAUTH before activating or modifying context nsapi_error_t ret_val = do_user_authentication(); return ret_val; } nsapi_error_t AT_CellularNetwork::do_user_authentication() { if (_pwd && _uname) { _at.cmd_start("AT+CGAUTH="); _at.write_int(_cid); _at.write_int(_authentication_type); _at.write_string(_uname); _at.write_string(_pwd); _at.cmd_stop(); _at.resp_start(); _at.resp_stop(); if (_at.get_last_error() != NSAPI_ERROR_OK) { return NSAPI_ERROR_AUTH_FAILURE; } } return _at.get_last_error(); }

But with this approach, what is the need of set_context_to_be_activated function?
Also calling inheritable function from private method is not a good design.
I think the best would be to rename set_context_to_be_activated() to do_user_authentication()
We can do this without breaking anything as it was a private method.
Also is get_context() check really needed in this function or should it be in a place where this function is called (activate_context)? Usually private functions are only called when the internal status is ok. And checking what activate_context does it makes perfect sense to move this check there.

I agree with @AnttiKauppila proposals. One note still: not only rename set_context_to_be_activated() to do_user_authentication() but make it also protected virtual.

And the usage like:

nsapi_error_t AT_CellularNetwork::do_user_authentication() { if (_pwd && _uname) { _at.cmd_start("AT+CGAUTH="); _at.write_int(_cid); _at.write_int(_authentication_type); _at.write_string(_uname); _at.write_string(_pwd); _at.cmd_stop(); _at.resp_start(); _at.resp_stop(); if (_at.get_last_error() != NSAPI_ERROR_OK) { return NSAPI_ERROR_AUTH_FAILURE; } } return NSAPI_ERROR_OK; } nsapi_error_t AT_CellularNetwork::activate_context() { _at.lock(); nsapi_error_t err = NSAPI_ERROR_OK; // try to find or create context with suitable stack if(get_context()) { // try to authenticate user before activating or modifying context err = do_user_authentication(); } else { err = NSAPI_ERROR_NO_CONNECTION; } if (err != NSAPI_ERROR_OK) { .....

@mirelachirica @AnttiKauppila I also needed to change accessibility of get_context(). Applied suggested changes.

AnttiKauppila

Please modify AT_CellularNetwork class as I suggested in my comment

Changed accessibility cellular features member functions, fixed minor target issues

AnttiKauppila · 2018-06-09T16:14:55Z

features/cellular/framework/AT/AT_CellularNetwork.cpp

@@ -420,7 +429,7 @@ void AT_CellularNetwork::ppp_status_cb(nsapi_event_t event, intptr_t parameter)
 }
 #endif

-nsapi_error_t AT_CellularNetwork::set_context_to_be_activated()
+nsapi_error_t AT_CellularNetwork::do_user_authentication()
 {
    // try to find or create context with suitable stack
    if (!get_context()) {


This check is not needed anymore, it is done in activate_context already

AnttiKauppila · 2018-06-09T16:15:20Z

features/cellular/framework/AT/AT_CellularNetwork.h

@@ -180,6 +178,8 @@ class AT_CellularNetwork : public CellularNetwork, public AT_CellularBase
    AuthenticationType _authentication_type;
    int _cell_id;
    nsapi_connection_status_t _connect_status;
+    virtual nsapi_error_t do_user_authentication();


No need for error value anymore

Do you mean no error return value? What about NSAPI_ERROR_AUTH_FAILURE?

I think this topic is not about adding target, so I didn’t make any change about this. Please avoid this topic in this PR.

AnttiKauppila · 2018-06-09T16:15:47Z

features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.cpp

+nsapi_error_t QUECTEL_UG96_CellularNetwork::do_user_authentication()
+{
+    // try to find or create context with suitable stack
+    if (!get_context()) {


This check is not needed anymore, it is done in activate_context already

AnttiKauppila

do_user_authentication does not need return value anymore, please remove that

AnttiKauppila · 2018-06-09T16:22:32Z

features/cellular/framework/AT/AT_CellularNetwork.h

@@ -180,6 +178,8 @@ class AT_CellularNetwork : public CellularNetwork, public AT_CellularBase
    AuthenticationType _authentication_type;
    int _cell_id;
    nsapi_connection_status_t _connect_status;
+    virtual nsapi_error_t do_user_authentication();
+    bool get_context();


I don't see a reason why you changed the visibility of this method. Could you explain?

get_context() was called from do_user_authentication(), so there were compile error. This is why I made change.

[Error] AT_CellularNetwork.h@157,10: 'bool mbed::AT_CellularNetwork::get_context()' is private [Error] QUECTEL_UG96_CellularNetwork.cpp@49,10: within this context [Error] AT_CellularNetwork.h@157,10: 'bool mbed::AT_CellularNetwork::get_context()' is private [Error] QUECTEL_UG96_CellularNetwork.cpp@49,22: within this context [ERROR] In file included from ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.h:21:0, from ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.cpp:18: ./mbed-os/features/cellular/framework/AT/AT_CellularNetwork.h: In member function 'virtual nsapi_error_t mbed::QUECTEL_UG96_CellularNetwork::do_user_authentication()': ./mbed-os/features/cellular/framework/AT/AT_CellularNetwork.h:157:10: error: 'bool mbed::AT_CellularNetwork::get_context()' is private bool get_context(); ^ ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.cpp:49:10: error: within this context if (!get_context()) { ^ In file included from ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.h:21:0, from ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.cpp:18: ./mbed-os/features/cellular/framework/AT/AT_CellularNetwork.h:157:10: error: 'bool mbed::AT_CellularNetwork::get_context()' is private bool get_context(); ^ ./mbed-os/features/cellular/framework/targets/QUECTEL/UG96/QUECTEL_UG96_CellularNetwork.cpp:49:22: error: within this context if (!get_context()) { ^

So I'm guessing I can return get_context() to private, if I remove this check.

Now, get_context() backed to private.

Removed context check from do_user_authentication.

AnttiKauppila · 2018-06-11T12:43:59Z

0xc0170 This is fine from Cellular team perspective, ready to be tested

cmonr · 2018-06-12T02:32:36Z

/morph build

mbed-ci · 2018-06-12T03:57:20Z

Build : SUCCESS

Build number : 2323
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7098/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

mbed-ci · 2018-06-12T05:52:09Z

Exporter Build : SUCCESS

Build number : 1945
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7098/

mbed-ci · 2018-06-12T23:30:39Z

Test : FAILURE

Build number : 2103
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7098/2103

cmonr · 2018-06-13T02:53:45Z

Failed test does not appear to be caused by the PR, and mirrors an issue we observed earlier today.
/morph test

mbed-ci · 2018-06-14T00:16:10Z

Test : SUCCESS

Build number : 2116
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7098/2116

[Wio 3G] Adding platform

e5b8591

0xc0170 added needs: review scope: new-target labels Jun 4, 2018

0xc0170 requested a review from ashok-rao June 4, 2018 09:59

ashok-rao reviewed Jun 5, 2018

View reviewed changes

ashok-rao previously approved these changes Jun 5, 2018

View reviewed changes

0xc0170 requested a review from a team June 5, 2018 13:48

AnttiKauppila reviewed Jun 5, 2018

View reviewed changes

AnttiKauppila suggested changes Jun 5, 2018

View reviewed changes

0xc0170 added needs: work and removed needs: review labels Jun 6, 2018

[Wio 3G] Changed suggested points

16e706e

Changed accessibility cellular features member functions, fixed minor target issues

ytsuboi dismissed ashok-rao’s stale review via 16e706e June 8, 2018 11:14

AnttiKauppila reviewed Jun 9, 2018

View reviewed changes

AnttiKauppila suggested changes Jun 9, 2018

View reviewed changes

AnttiKauppila reviewed Jun 9, 2018

View reviewed changes

[Wio 3G] Changed suggested things

0751dd7

Removed context check from do_user_authentication.

mirelachirica approved these changes Jun 11, 2018

View reviewed changes

AnttiKauppila approved these changes Jun 11, 2018

View reviewed changes

cmonr added needs: CI and removed needs: work labels Jun 11, 2018

cmonr approved these changes Jun 12, 2018

View reviewed changes

cmonr added the release-version: 5.9.1 label Jun 12, 2018

cmonr added ready for merge and removed needs: CI labels Jun 14, 2018

cmonr merged commit 276588f into ARMmbed:master Jun 14, 2018

0xc0170 removed the ready for merge label Jun 14, 2018

[Wio 3G] Adding platform #7098

[Wio 3G] Adding platform #7098

Uh oh!

Conversation

ytsuboi commented Jun 4, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Pull request type

Uh oh!

0xc0170 commented Jun 4, 2018

Uh oh!

ytsuboi commented Jun 5, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ashok-rao left a comment

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Jun 5, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnttiKauppila left a comment

Choose a reason for hiding this comment

Uh oh!

AnttiKauppila Jun 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnttiKauppila left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ytsuboi Jun 10, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnttiKauppila commented Jun 11, 2018

Uh oh!

ytsuboi commented Jun 4, 2018 •

edited

Loading

AnttiKauppila Jun 9, 2018 •

edited

Loading

ytsuboi Jun 10, 2018 •

edited

Loading