Add support for IAR version 8.x #4938

deepikabhavnani · 2017-08-18T19:02:03Z

Added support to IAR 8.x, with the reference of release document

__aeabi_read_tp - Weak empty function is added to support the build. It should be implemented with TLS support in IAR Dlib.

0xc0170

What about testing? M0/M3/M4 with IAR all OK with these fixes (also backward compatible if possible. I assume yes as we are now using version < 8 macro in some changes here).

0xc0170 · 2017-08-21T11:13:59Z

rtos/mbed_boot.c

@@ -574,9 +576,14 @@ void pre_main(void)
    singleton_mutex_attr.cb_mem = &singleton_mutex_obj;
    singleton_mutex_id = osMutexNew(&singleton_mutex_attr);

+#if (__IAR_SYSTEMS_ICC__ >= 8)
+    __iar_Initlocks();


we are having exit function here defined as well? As the docs say, we should invoke it (as initlocks in multithreaded env, added here.

Info: http://netstorage.iar.com/SuppDB/Public/UPDINFO/012120/arm/doc/infocenter/iccarm.ENU.html

(This info is useful to know, could be part of the commit message to include details what this initlocks is fixing).

As per the release document, void __iar_clearlocks(void); should be called to destroy the locks. It should be called after atexit and _Close_all.
In mbed applications, we never do close all we just wait for system to die (in case of issue). Other option was to call it on main thread exit, but than timer thread and background thread will still be running.

0xc0170 · 2017-08-21T11:15:28Z

platform/mbed_retarget.cpp

@@ -880,6 +880,9 @@ extern "C" WEAK void __iar_file_Mtxinit(__iar_Rmtx *mutex) {}
 extern "C" WEAK void __iar_file_Mtxdst(__iar_Rmtx *mutex) {}
 extern "C" WEAK void __iar_file_Mtxlock(__iar_Rmtx *mutex) {}
 extern "C" WEAK void __iar_file_Mtxunlock(__iar_Rmtx *mutex) {}
+#if (__IAR_SYSTEMS_ICC__ >= 8)
+extern "C" WEAK void *__aeabi_read_tp (void) {}


why is this suddenly required, and what about the other requirements? As for any thread, there is more to implement than just read_tp ? See http://netstorage.iar.com/SuppDB/Public/UPDINFO/012120/arm/doc/infocenter/iccarm.ENU.html

Thread Local Storage requirement has changed in this version of IAR, and requires RTOS to allocate memory for secondary threads as part of TLS support. read_tp() is to point to the TLS section.

We do not have support of TLS in current RTOS, hence weak function is added to assist build. Actual support will be added in future with changes in RTOS.

theotherjimmy · 2017-08-21T13:46:35Z

nit: should the title be "Add support for IAR 8"? or better yet: "Update IAR to version 8".

0xc0170 · 2017-08-22T10:04:47Z

@studavekar Are we still waiting for IAR 8.x update in CI?

@deepikabhavnani - all these changes are compatible with the exporter for IAR (templates we have are for v7.x and are forward compatible , all builds with 8.x) ? What version have you been testing with?

deepikabhavnani · 2017-08-22T13:49:15Z

@0xc0170 - I have verified exporter and build for K64F and Cortex-M23 with IAR 7.8/8.x @studavekar will verify few devices with IAR 8.x in single CI machine. We can label it to "needs: CI" but that would be manually triggered.

We have not updated the template files to version 3. Version 2 is compatible, we will be missing few advantages of version 3, till we support new template version.

studavekar · 2017-08-22T20:13:32Z

@deepikabhavnani @0xc0170 have setup the CI machines with IAR 8.11.x

Triggered a build for current PR

00:01:04.984 [DEBUG] Output:    IAR ELF Linker V8.11.2.13589/W32 for ARM
00:01:04.984 [DEBUG] Output:    Copyright 2007-2017 IAR Systems AB.

00:01:04.984 [DEBUG] Output:    IAR ELF Tool V10.1.5.191 [BUILT at IAR]
00:01:04.984 [DEBUG] Output:    Copyright 2007-2017 IAR Systems AB.

Build : http://mbed-ci-master-2.austin.arm.com:8081/view/All/job/build_matrix_stage_iar_8/2/

studavekar · 2017-08-22T21:17:15Z

/morph test

mbed-bot · 2017-08-22T21:53:08Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1081

Build failed!

deepikabhavnani · 2017-08-23T14:05:25Z

/morph test

0xc0170 · 2017-08-23T17:03:43Z

/morph test

mbed-bot · 2017-08-24T02:04:59Z

Result: ABORTED

Your command has finished executing! Here's what you wrote!

deepikabhavnani · 2017-08-24T03:51:51Z

/morph test

studavekar · 2017-08-24T07:22:14Z

/morph test

mbed-bot · 2017-08-24T10:40:51Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1107

Build failed!

deepikabhavnani · 2017-08-25T14:32:51Z

/morph test

mbed-bot · 2017-08-25T15:35:24Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1115

Example Build failed!

deepikabhavnani · 2017-08-28T16:09:41Z

/morph test

mbed-bot · 2017-08-28T18:33:11Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1126

All builds and test passed!

theotherjimmy · 2017-08-29T18:43:07Z

@studavekar Could you trigger this manually with IAR8?

studavekar · 2017-08-30T05:06:40Z

This would run with IAR 8

/morph test

mbed-bot · 2017-08-30T05:11:38Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1140

Build failed!

studavekar · 2017-08-30T07:09:23Z

Build results IAR 8.XX :

http://mbed-ci-master-2.austin.arm.com:8081/view/All/job/build_matrix_stage_iar_8/7/

Failure :
http://mbed-ci-master-2.austin.arm.com:8081/view/All/job/build_matrix_stage_iar_8/7/target=UBLOX_EVK_ODIN_W2,toolchain=IAR/console

http://mbed-ci-master-2.austin.arm.com:8081/view/All/job/build_matrix_stage_iar_8/7/target=REALTEK_RTL8195AM,toolchain=IAR/console

Test results

http://mbed-ci-master-2.austin.arm.com:8081/job/test_matrix/880/

deepikabhavnani · 2017-09-10T15:02:15Z

MTS_DRAGONFLY_F411RE + IAR 8 + mbed-os-example-cellular - failed
http://mbedosci.cloudapp.net/results/pr/4938/5/FAIL/MTS_DRAGONFLY_F411RE/IAR/

[DEBUG] Output: bool success = at->send("AT+CNMI=2,"CTX) && at->recv("OK");
[DEBUG] Output: ^
[DEBUG] Output: "/home/jenkins/mbed_jenkins_node2/workspace/bm_wrap/1358/mbed-os/features/netsocket/cellular/generic_modem_driver/PPPCellularInterface.cpp",113 Error[Pe2486]: user-defined literal operator not found

deepikabhavnani · 2017-09-10T15:03:05Z

@geky @sarahmarshy : Please review the fix in cellular example.

studavekar · 2017-09-10T15:49:26Z

@0xc0170 IAR build failure : http://mbed-ci-master-2.austin.arm.com:8081/job/iar_8_temp_pr_pipeline/

With the fix added , triggering the test again

/morph test

studavekar · 2017-09-10T16:02:45Z

/morph test

0xc0170 · 2017-09-10T16:36:05Z

Looks fine to me. I had a look earlier this morning, syntax looked correct to me. Not sure why IAR 8 complained.

@hasnainvirk To be aware of this change, the latest commit with CTX update.

mbed-bot · 2017-09-10T21:05:06Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1271

All builds and test passed!

studavekar · 2017-09-11T03:55:51Z

Results for IAR 8 : http://mbedosci.cloudapp.net/results/pr/4938/6/

hasnainvirk · 2017-09-11T07:32:05Z

@0xc0170 @deepikabhavnani Are we updating now to C++14 ? What about other compilers , are they going to be switched to C++14 as well ?
Regarding change in cellular example: Before C++11, string literals were handled separated by tokens and if there was a macro being combined with a string literal, it was considered as yet another token and hence the code like "my"MACRO compiled. Since C++11, all tokens must be separated by a space , so the proper fix to cellular example is to add a space between string literal and macro.

hasnainvirk · 2017-09-11T07:37:44Z

features/netsocket/cellular/generic_modem_driver/PPPCellularInterface.cpp

@@ -110,7 +110,7 @@ static bool set_CNMI(ATCmdParser *at)
    // New SMS indication configuration
    // set AT+CMTI=[mode, index] , 0 PDU mode, 1 text mode
    // Multiple URCs for SMS, i.e., CMT, CMTI, UCMT, CBMI, CDSI as DTE could be following any of these SMS formats
-    bool success = at->send("AT+CNMI=2,"CTX) && at->recv("OK");
+    bool success = at->send("AT+CNMI=2,""1") && at->recv("OK");


Add a space string literal and macro, it will work. Like:

bool success = at->send("AT+CNMI=2," CTX) && at->recv("OK");

hasnainvirk · 2017-09-11T07:38:11Z

features/netsocket/cellular/generic_modem_driver/PPPCellularInterface.cpp

@@ -141,7 +141,7 @@ static void CMT_URC(ATCmdParser *at)

 static bool set_atd(ATCmdParser *at)
 {
-    bool success = at->send("ATD*99***"CTX"#") && at->recv("CONNECT");
+    bool success = at->send("ATD*99***""1""#") && at->recv("CONNECT");


same like above

hasnainvirk · 2017-09-11T07:38:22Z

features/netsocket/cellular/generic_modem_driver/PPPCellularInterface.cpp

@@ -469,7 +469,7 @@ nsapi_error_t PPPCellularInterface::setup_context_and_credentials()
 #endif
    success = _at->send("AT"
                          "+FCLASS=0;" // set to connection (ATD) to data mode
-                          "+CGDCONT="CTX",\"%s\",\"%s%s\"",
+                          "+CGDCONT=""1"",\"%s\",\"%s%s\"",


same like above

The previous fix to replace CTX with string literals was the wrong solution. All that was actually required was to insert a space before the macro.

deepikabhavnani · 2017-09-11T10:16:16Z

@0xc0170 - LP ticker and sleep manager tests are failing for all three toolchains. Is it expected for all 4 devices? NCS36510 | NRF51_DK | NRF52_DK | NUCLEO_F429ZI |

adbridge · 2017-09-11T10:16:33Z

@hasnainvirk I've pushed your change suggestions to Deepika's branch .Does this look better ?

adbridge · 2017-09-11T10:16:51Z

/morph test

adbridge · 2017-09-11T10:17:36Z

@studavekar Can we re-run with IAR 8 please and paste results once done?

adbridge · 2017-09-11T10:18:38Z

@deepikabhavnani yeah we are waiting on #5063 going in

adbridge · 2017-09-11T11:23:40Z

@SeppoTakalo @mikaleppanen @hasnainvirk pr_head has failed for some reason, any ideas ?
"ERROR: ARMmbed pipeline builds » mbed-os-cliapp » master #6025 completed with status ABORTED (propagate: false to ignore)
Finished: FAILURE"

hasnainvirk · 2017-09-11T11:58:05Z

@adbridge It seems CI timed out and couldn't complete jobs for many targets. Too many jobs in parallel pipe line.

SeppoTakalo · 2017-09-11T12:28:29Z

ARMmbed pipeline builds » mbed-os-cliapp » master is area of Systest team. @marhil01 Who is responsible of maintaining mbed-os-cliapp?

I have no better knowledge, so how about restarting the testcases?

adbridge · 2017-09-11T12:57:23Z

I've restarted pr-head, hopefully it won't fail in the same way again

mbed-bot · 2017-09-11T13:41:41Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

Output

mbed Build Number: 1277

All builds and test passed!

sg- · 2017-09-11T14:30:14Z

targets/TARGET_Realtek/TARGET_AMEBA/rtw_emac.cpp

@@ -217,7 +217,7 @@ emac_interface_t *wlan_emac_init_interface()
 {

    if (_emac == NULL) {
-        _emac = new emac_interface_t();
+        _emac = (emac_interface_t*) malloc(sizeof(emac_interface_t));


This should use new or error in the case it fails, not print

@sg- Everywhere in file print statements are used instead error, I can add that as bug and get it fix in new PR

new would have failed as assert but malloc won't. That is the change in behavior.

Oh, @deepikabhavnani, this was the issue with const members right?

@sg-, the emac_interface_t as written can't be default constructed in c++ (here). The const member (ops) must be constructed but the emac_interface_t class doesn't have a constructor to initialize the member. Since the emac_api header is also included by C files, we can't add a constructor. C supports struct initializers for this case, but C++ relies on constructors.

We can either remove the const or allocate with malloc. We could change the error print to assert.

deepikabhavnani mentioned this pull request Aug 18, 2017

Add Nuvoton m2351device #4892

Closed

deepikabhavnani force-pushed the IAR_fixes branch from fcd51c1 to d98f9ca Compare August 18, 2017 21:38

0xc0170 reviewed Aug 21, 2017

View reviewed changes

0xc0170 added the needs: review label Aug 21, 2017

theotherjimmy changed the title ~~Add support to IAR 8.x~~ Add support for IAR 8.x Aug 21, 2017

deepikabhavnani changed the title ~~Add support for IAR 8.x~~ Update IAR to version 8 Aug 21, 2017

0xc0170 approved these changes Aug 22, 2017

View reviewed changes

0xc0170 added needs: CI and removed needs: review labels Aug 23, 2017

deepikabhavnani force-pushed the IAR_fixes branch from 3893a93 to 7e4b4fe Compare August 29, 2017 19:14

0xc0170 added needs: CI and removed ready for merge labels Sep 10, 2017

hasnainvirk suggested changes Sep 11, 2017

View reviewed changes

Remove string literal values and revert back to using CTX macro .

6bd28ec

The previous fix to replace CTX with string literals was the wrong solution. All that was actually required was to insert a space before the macro.

hasnainvirk approved these changes Sep 11, 2017

View reviewed changes

sg- reviewed Sep 11, 2017

View reviewed changes

adbridge added ready for merge and removed needs: CI labels Sep 11, 2017

adbridge merged commit cab660d into ARMmbed:master Sep 11, 2017

adbridge removed the ready for merge label Sep 11, 2017

deepikabhavnani mentioned this pull request Sep 14, 2017

Changed error print to assert #5069

Merged

deepikabhavnani changed the title ~~Update IAR to version 8~~ Add support for IAR version 8.x Jul 23, 2018

Add support for IAR version 8.x #4938

Add support for IAR version 8.x #4938

Uh oh!

Conversation

deepikabhavnani commented Aug 18, 2017

Uh oh!

0xc0170 left a comment

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!

theotherjimmy commented Aug 21, 2017

Uh oh!

0xc0170 commented Aug 22, 2017

Uh oh!

deepikabhavnani commented Aug 22, 2017

Uh oh!

studavekar commented Aug 22, 2017

Uh oh!

studavekar commented Aug 22, 2017

Uh oh!

mbed-bot commented Aug 22, 2017

Output

Uh oh!

deepikabhavnani commented Aug 23, 2017

Uh oh!

0xc0170 commented Aug 23, 2017

Uh oh!

mbed-bot commented Aug 24, 2017

Uh oh!

deepikabhavnani commented Aug 24, 2017

Uh oh!

studavekar commented Aug 24, 2017

Uh oh!

mbed-bot commented Aug 24, 2017

Output

Uh oh!

deepikabhavnani commented Aug 25, 2017

Uh oh!

mbed-bot commented Aug 25, 2017

Output

Uh oh!

deepikabhavnani commented Aug 28, 2017

Uh oh!

mbed-bot commented Aug 28, 2017

Output

Uh oh!

theotherjimmy commented Aug 29, 2017

Uh oh!

studavekar commented Aug 30, 2017

Uh oh!

mbed-bot commented Aug 30, 2017

Output

Uh oh!

studavekar commented Aug 30, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

deepikabhavnani commented Sep 10, 2017

Uh oh!

deepikabhavnani commented Sep 10, 2017

Uh oh!

studavekar commented Sep 10, 2017

Uh oh!

studavekar commented Sep 10, 2017

Uh oh!

0xc0170 commented Sep 10, 2017

Uh oh!

mbed-bot commented Sep 10, 2017

Output

Uh oh!

studavekar commented Sep 11, 2017

Uh oh!

hasnainvirk commented Sep 11, 2017

Uh oh!

studavekar commented Aug 30, 2017 •

edited

Loading

hasnainvirk Sep 11, 2017 •

edited

Loading

adbridge commented Sep 11, 2017 •

edited

Loading