Initial support for Cortex M-23/M-33 devices. #4875

deepikabhavnani · 2017-08-08T19:38:49Z

Added RTX + tools support for GCC only.

IAR 7.8 Cortex-23 does not support exclusive and secure commands. Move to IAR 8.0 requires changed in either Thread class or CMSIS (Ideal would be CMSIS). Hence dropping IAR in this PR

Required for : Rebased Nuvoton code for M2351 - PR #4892
Dependent on IAR 8.0 support

0xc0170 · 2017-08-09T08:19:09Z

Look at travis failure, looks like there's assembly file not using uppercase suffix.

Would this be easier if this patch would be split into two (generic M-23/M-33, and then nuvoton on top of it) ?

IAR 7.8 Cortex-23 seems to have Non-secure commands supported only.

Will IAR 8.x help there?

0xc0170 · 2017-08-14T07:54:23Z

tools/profiles/debug.json

@@ -42,7 +42,7 @@
            "--diag_suppress=Pa050,Pa084,Pa093,Pa082", "-On", "-r", "-DMBED_DEBUG",
            "-DMBED_TRAP_ERRORS_ENABLED=1"],
        "asm": [],
-        "c": ["--vla"],


why suddenly are not compatible ? This flag is only for C compiler ,thus should not cause any errors (we used to have it in common, and then I recall we faced an error with compatilibity , probably the same one you are seeing?) . See #1985 why it was enabled.

This commit should be a separate pull request

Yeah, you are correct. I tried using them in common location at *.json files and got errors. Will verify once.

My bad, I was getting compilation issues with exporters when enabling "--c++", "--no_rtti", "--no_exceptions" flags in *.json files. Compile flags and exporter flags didn't seemed synchronized.

"--no_rtti", "--no_exceptions" - Not working with export tool. Enabling those flags directly in project settings work, but updating the template files didn't help as well. Verified now, the *.ewp file version for IAR8 is 3 and we are using version 2. Migration required.

Will remove this commit and add bug to mbed-os

0xc0170 · 2017-08-14T07:55:19Z

platform/mbed_critical.c

@@ -23,12 +23,6 @@
 #include "platform/mbed_assert.h"
 #include "platform/mbed_toolchain.h"

-#if !defined (__CORTEX_M0) && !defined (__CORTEX_M0PLUS)


+1 , can this be send separately? Should be quick to integrate, and then just rebase your branch to get it in

PR #4902, will remove and rebase shortly

0xc0170 · 2017-08-17T14:32:56Z

cc @theotherjimmy

0xc0170 · 2017-08-17T14:35:50Z

IAR 7.8 Cortex-23 does not support exclusive and secure commands. Move to IAR 8.0 requires changed in either Thread class or CMSIS (Ideal would be CMSIS). Hence dropping IAR in this PR

Can you elabore this dependency? What is expected to be delivered and who is taking an action item to do so?

theotherjimmy

Tools changes look fine.

deepikabhavnani · 2017-08-17T15:23:20Z

Rebased on top of #4902

deepikabhavnani · 2017-08-17T15:27:06Z

@0xc0170 - I am looking into IAR changes with CMSIS team, but might get delayed due to vacation time.

deepikabhavnani · 2017-08-18T19:44:14Z

Rebased on top of IAR_fix branch

sg- · 2017-08-19T20:00:59Z

platform/CThunk.h

@@ -37,7 +37,8 @@
 #define CTHUNK_ADDRESS 1
 #define CTHUNK_VARIABLES volatile uint32_t code[2]

-#if (defined(__CORTEX_M3) || defined(__CORTEX_M4) || defined(__CORTEX_M7) || defined(__CORTEX_A9))
+#if (defined(__CORTEX_M3) || defined(__CORTEX_M4) || defined(__CORTEX_M7) || defined(__CORTEX_A9) \
+    || defined(__CORTEX_M23) || defined(__CORTEX_M33))


Does this work? I would have guessed that the M23 needed to be aligned with the M0/P version

Armv8-M architecture reference manual has new term now for Thumb instructions. It says v8-M variant supports T32 instruction set and does not distinguish between mainline (M-33) and baseline (M-23) architecture.
All instructions are decoded as T32 (not as 16-bit Thumb has to be aligned to half word).

Also, callbacks in test were successful.

0xc0170 · 2017-08-21T14:39:12Z

LGTM, waiting for IAR 8.x support, will tag this accordingly.

mbed-bot · 2017-09-07T21:14:13Z

Result: FAILURE

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

/morph test

Output

mbed Build Number: 1250

Build failed!

studavekar · 2017-09-07T22:51:03Z

/morph test

mbed-bot · 2017-09-07T23:20:05Z

Result: FAILURE

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

/morph test

Output

mbed Build Number: 1254

Build Prep failed!

studavekar · 2017-09-07T23:23:08Z

Looks like git fetch error from github

0:00:01.528 using GIT_SSH to set credentials 
00:00:02.162  > git fetch --no-tags --progress [email protected]:ARMmbed/mbed-os.git +refs/heads/*:refs/remotes/origin/* # timeout=20
00:19:43.472 ERROR: Error cloning remote repo 'origin'
00:19:43.473 hudson.plugins.git.GitException: Command "git fetch --no-tags --progress [email protected]:ARMmbed/mbed-os.git +refs/heads/*:refs/remotes/origin/*" returned status code 1:
00:19:43.473 stdout:

studavekar · 2017-09-08T06:57:46Z

/morph test

mbed-bot · 2017-09-08T15:38:44Z

Result: FAILURE

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

/morph test

Output

mbed Build Number: 1259

Build failed!

deepikabhavnani · 2017-09-08T15:41:23Z

IAR license issue

studavekar · 2017-09-09T05:36:13Z

/morph test

mbed-bot · 2017-09-09T08:42:15Z

Result: SUCCESS

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

/morph test

Output

mbed Build Number: 1263

All builds and test passed!

0xc0170 · 2017-09-09T10:27:54Z

Testing Done, however as this has dependency on another PR, this will require rebase once that goes on, and retest again

If we can (me or someone) do that rebase (push to your branch), would be good in case you are not around

Patater

The irq_armv8mml and etc. files were copied from CMSIS, but to file paths that are different than in upstream CMSIS. Please be sure to update any file path mapping scripts or documentation (cc @bulislaw)

Patater · 2017-09-11T14:35:16Z

tools/toolchains/gcc.py

+        if target.core == "Cortex-M23" or target.core == "Cortex-M23-NS":
+            self.cpu.append("-march=armv8-m.base")
+        elif target.core == "Cortex-M33" or target.core == "Cortex-M33-NS":
+            self.cpu.append("-march=armv8-m.main")


Please consider adding:
self.cpu.append("-mcmse")

Patater · 2017-09-11T14:55:06Z

platform/mbed_retarget.cpp

@@ -884,6 +884,9 @@ 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) {}


Is an empty implementation really sufficient? Do the TLS tests still pass with this empty implementation?

TLS (Thread local storage support) is not yet added to mbed-os

adbridge · 2017-09-11T16:34:10Z

@Patater Could your review comments go on a separate PR that we get into RC2 ?

deepikabhavnani · 2017-09-11T16:49:01Z

Rebased after IAR changes were merged

deepikabhavnani · 2017-09-11T16:52:45Z

@Patater - Added security flag. Will check with @bulislaw regarding file path mapping scripts or documentation

deepikabhavnani · 2017-09-11T16:53:02Z

/morph test

Patater · 2017-09-11T17:04:48Z

@adbridge @bulislaw @deepikabhavnani

Yes, the "file path mapping scripts and documentation" changes can go into a separate PR.

Thanks!

mbed-bot · 2017-09-11T22:36:53Z

Result: SUCCESS

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

/morph test

Output

mbed Build Number: 1280

All builds and test passed!

0xc0170 added the needs: work label Aug 9, 2017

deepikabhavnani force-pushed the cm_23_support branch from fe0fb0a to 678e1e0 Compare August 9, 2017 20:56

theotherjimmy added needs: review and removed needs: work labels Aug 10, 2017

deepikabhavnani force-pushed the cm_23_support branch from b535ded to 4b68b5c Compare August 11, 2017 16:11

deepikabhavnani mentioned this pull request Aug 11, 2017

Add Nuvoton m2351device #4892

Closed

0xc0170 reviewed Aug 14, 2017

View reviewed changes

deepikabhavnani force-pushed the cm_23_support branch from 4b68b5c to c6eebfc Compare August 14, 2017 20:44

0xc0170 approved these changes Aug 17, 2017

View reviewed changes

theotherjimmy approved these changes Aug 17, 2017

View reviewed changes

deepikabhavnani force-pushed the cm_23_support branch from c6eebfc to 070f88d Compare August 17, 2017 15:21

deepikabhavnani requested a review from sg- August 17, 2017 15:24

deepikabhavnani force-pushed the cm_23_support branch from 070f88d to 98fcad6 Compare August 18, 2017 19:43

sg- reviewed Aug 19, 2017

View reviewed changes

0xc0170 added the needs: preceding PR label Aug 21, 2017

sg- approved these changes Aug 21, 2017

View reviewed changes

adbridge added needs: preceding PR and removed needs: review needs: preceding PR labels Aug 24, 2017

deepikabhavnani added the needs: CI label Aug 29, 2017

deepikabhavnani force-pushed the cm_23_support branch from 98fcad6 to 586479b Compare August 29, 2017 19:25

deepikabhavnani force-pushed the cm_23_support branch from 586479b to 43c8907 Compare September 6, 2017 15:24

Patater suggested changes Sep 11, 2017

View reviewed changes

Patater reviewed Sep 11, 2017

View reviewed changes

theotherjimmy removed the needs: preceding PR label Sep 11, 2017

Initial RTX and tools support for Cortex M-23/M-33 devices

9422c35

deepikabhavnani force-pushed the cm_23_support branch from 2d8e199 to 9422c35 Compare September 11, 2017 16:47

Add -mcmse flag for GCC and improved logic for cpu core

ca1f24c

Patater approved these changes Sep 11, 2017

View reviewed changes

theotherjimmy merged commit 4b98566 into ARMmbed:master Sep 11, 2017

theotherjimmy removed the needs: CI label Sep 11, 2017

deepikabhavnani deleted the cm_23_support branch September 12, 2017 02:13

ccli8 mentioned this pull request Sep 12, 2017

Fail to enter hal_deepsleep when LowPowerTimer is used #5076

Closed

This was referenced Sep 28, 2017

Update CMSIS/RTX docs ARMmbed/mbed-os-5-docs#273

Closed

Update CMSIS/RTX docs ARMmbed/mbed-os-5-docs#276

Closed

Rtos changes ARMmbed/mbed-os-5-docs#278

Merged

Initial support for Cortex M-23/M-33 devices. #4875

Initial support for Cortex M-23/M-33 devices. #4875

Uh oh!

Conversation

deepikabhavnani commented Aug 8, 2017 • edited by adbridge Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

0xc0170 commented Aug 9, 2017

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!

0xc0170 commented Aug 17, 2017

Uh oh!

0xc0170 commented Aug 17, 2017

Uh oh!

theotherjimmy left a comment

Choose a reason for hiding this comment

Uh oh!

deepikabhavnani commented Aug 17, 2017

Uh oh!

deepikabhavnani commented Aug 17, 2017

Uh oh!

deepikabhavnani commented Aug 18, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Aug 21, 2017

Uh oh!

mbed-bot commented Sep 7, 2017

Output

Uh oh!

studavekar commented Sep 7, 2017

Uh oh!

mbed-bot commented Sep 7, 2017

Output

Uh oh!

studavekar commented Sep 7, 2017

Uh oh!

studavekar commented Sep 8, 2017

Uh oh!

mbed-bot commented Sep 8, 2017

Output

Uh oh!

deepikabhavnani commented Sep 8, 2017

Uh oh!

studavekar commented Sep 9, 2017

Uh oh!

mbed-bot commented Sep 9, 2017

Output

Uh oh!

0xc0170 commented Sep 9, 2017

Uh oh!

Patater 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!

Choose a reason for hiding this comment

Uh oh!

adbridge commented Sep 11, 2017

Uh oh!

deepikabhavnani commented Sep 11, 2017

Uh oh!

deepikabhavnani commented Sep 11, 2017

deepikabhavnani commented Aug 8, 2017 •

edited by adbridge

Loading