Fix build tool with ARMC6/ARMv8M #6644

ccli8 · 2018-04-16T08:36:43Z

Description

This PR tries to fix build tool issues with ARMC6/ARMv8M:

Add linking time preprocessor macro __DOMAIN_NS for non-secure build
For output .hex format, combine multiple .hex files (for multiple load regions) into one. This can help for Greentea test.

Fix None build_dir error with cmse_lib.o on Greentea test. None build_dir is passed to ARMC6 > __init__ and causes error from:
mbed-os/tools/test_api.py:

 def find_tests(base_dir, target_name, toolchain_name, app_config=None):
 ......
     # Temporary structure: tests referenced by (name, base, group, case) tuple
     tests = {}
     # List of common folders: (predicate function, path) tuple
     commons = []

     # Prepare the toolchain
     toolchain = prepare_toolchain([base_dir], None, target_name,

I move cmse_lib.o from ARMC6 > __init__ to ARM > link.

Pull request type

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

1. Add linking time preprocessor macro __DOMAIN_NS for non-secure build 2. For output .hex format, combine multiple .hex files (for multiple load regions) into one This can help for Greentea test. 3. Fix 'None' build_dir with cmse_lib.o on Greentea test

theotherjimmy · 2018-04-16T14:11:35Z

tools/toolchains/arm.py

@@ -233,6 +233,11 @@ def link(self, output, objects, libraries, lib_dirs, scatter_file):
        cmd_pre = self.ld + args
        cmd = self.hook.get_cmdline_linker(cmd_pre)

+        # Create Secure library
+        if self.target.core == "Cortex-M23" or self.target.core == "Cortex-M33":


Moving this here breaks exporters: The exporters have to know to append these options somehow, and this location prevents them from being able to.

theotherjimmy · 2018-04-16T14:19:15Z

tools/toolchains/arm.py

@@ -252,7 +257,8 @@ def archive(self, objects, lib_path):
    @hook_tool
    def binary(self, resources, elf, bin):
        _, fmt = splitext(bin)
-        bin_arg = {".bin": "--bin", ".hex": "--i32"}[fmt]
+        # On .hex format, combine multiple .hex files (for multiple load regions) into one 
+        bin_arg = {".bin": "--bin", ".hex": "--i32combined"}[fmt]


Good catch. This was always the indented behavior. I guess nobody combined multiple load regions with hex output before.

theotherjimmy · 2018-04-16T14:22:42Z

@ccli8 I'm familiar with the interaction between the toolchains and the exporters and the toolchains and the test builder. Let me know if you would like me to correct the "no build dir" bug.

ccli8 · 2018-04-17T01:48:07Z

@theotherjimmy Yes. Please correct the "no build dir" bug and I can check Grrentea flow on NUMAKER_PFM_M2351 with the corrected version. NUMAKER_PFM_M2351 is Cortex-M23 based and its port is on-going.

theotherjimmy · 2018-04-17T17:03:03Z

@ccli8 I changed the implementation on your branch. Good news: it reduced the diff by quite a bit.

deepikabhavnani

Looks good to me 👍

ccli8 · 2018-04-18T01:49:52Z

@theotherjimmy With your update, I run Greentea test on NUMAKER_PFM_M2351 with ARMC6 and it works.

0xc0170 · 2018-04-18T09:35:53Z

/morph build

mbed-ci · 2018-04-18T09:58:18Z

Build : SUCCESS

Build number : 1784
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6644/

Triggering tests

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

0xc0170 · 2018-04-18T12:30:11Z

/morph uvisor-test

mbed-ci · 2018-04-18T17:36:54Z

Test : FAILURE

Build number : 1594
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/6644/1594

mbed-ci · 2018-04-18T19:12:37Z

Exporter Build : SUCCESS

Build number : 1429
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/6644/

0xc0170 · 2018-04-19T07:45:59Z

/morph test
/morph uvisor-test

mbed-ci · 2018-04-19T07:47:40Z

Test : FAILURE

Build number : 1602
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/6644/1602

0xc0170 · 2018-04-19T15:18:52Z

@kegilbert @theotherjimmy Can you review the logs from the test failures?

I could find 02:46:31 mbedgt: no platform/target matching tests were found! , but why does it fail? Is it related to this change? No tests run.

cmonr · 2018-04-23T21:39:06Z

Did some link hopping,

and noticed a couple of things:

The build log (3) has it's mem table filled out, whereas compilation for the test (4) looks like this:

Memory map breakdown for built projects (values in Bytes):
+------+--------+-----------+------------+-------------+
| name | target | toolchain | static_ram | total_flash |
+------+--------+-----------+------------+-------------+
+------+--------+-----------+------------+-------------+

This is the command used to compile the test:

mbed test --compile -m K64F -t GCC_ARM -DMBED_HEAP_STATS_ENABLED=1 -DMBED_STACK_STATS_ENABLED=1 -DMBED_TRAP_ERRORS_ENABLED=1 -DMBED_TEST_BLOCKDEVICE=SPIFBlockDevice -DMBED_TEST_BLOCKDEVICE_DECL=SPIFBlockDevice bd(D11,D12,D13,D10)

MBED_TEST_BLOCKDEVICE_DECL should not have a space...

ccli8 · 2018-04-24T02:24:36Z

@theotherjimmy I find something wrong with the commit bc8b983. Test items are not listed with the command:

mbed test -m NUMAKER_PFM_NUC472 -t ARM --compile-list

If I roll back, it is OK.

Scan resources ignores the build dir so we can't set it to something dumb like the directory we want to scan

theotherjimmy · 2018-04-24T15:08:16Z

Crap. I forgot that the resource scan ignores the assigned build directory. Sorry about that. Fixing now.

cmonr · 2018-04-24T18:20:12Z

/morph build

mbed-ci · 2018-04-24T19:02:29Z

Build : SUCCESS

Build number : 1844
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6644/

Triggering tests

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

mbed-ci · 2018-04-24T22:40:50Z

Exporter Build : SUCCESS

Build number : 1490
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/6644/

mbed-ci · 2018-04-25T07:55:42Z

Test : SUCCESS

Build number : 1653
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/6644/1653

ccli8 · 2018-04-26T02:34:00Z

@theotherjimmy The issue gets fixed with your update da69f01.

mbed test -m NUMAKER_PFM_NUC472 -t ARM --compile-list

Fix build tool with ARMC6/ARMv8M

b81c0e5

1. Add linking time preprocessor macro __DOMAIN_NS for non-secure build 2. For output .hex format, combine multiple .hex files (for multiple load regions) into one This can help for Greentea test. 3. Fix 'None' build_dir with cmse_lib.o on Greentea test

0xc0170 requested review from theotherjimmy and deepikabhavnani April 16, 2018 09:18

0xc0170 added the needs: review label Apr 16, 2018

theotherjimmy suggested changes Apr 16, 2018

View reviewed changes

cmonr added needs: work and removed needs: review labels Apr 17, 2018

theotherjimmy force-pushed the nuvoton_fix_armc6_armv8m branch 5 times, most recently from 4d1e335 to 5e83ae9 Compare April 17, 2018 17:01

Correct exporting with cortex-M23 and M33

bc8b983

theotherjimmy force-pushed the nuvoton_fix_armc6_armv8m branch 2 times, most recently from fef7297 to bc8b983 Compare April 17, 2018 17:02

theotherjimmy previously approved these changes Apr 17, 2018

View reviewed changes

deepikabhavnani previously approved these changes Apr 17, 2018

View reviewed changes

0xc0170 removed the needs: work label Apr 18, 2018

0xc0170 previously approved these changes Apr 18, 2018

View reviewed changes

0xc0170 added the needs: CI label Apr 18, 2018

cmonr added needs: work and removed needs: CI labels Apr 23, 2018

Correct test finding behavior

da69f01

Scan resources ignores the build dir so we can't set it to something dumb like the directory we want to scan

theotherjimmy dismissed stale reviews from 0xc0170, deepikabhavnani, and themself via da69f01 April 24, 2018 15:08

cmonr approved these changes Apr 24, 2018

View reviewed changes

0xc0170 approved these changes Apr 25, 2018

View reviewed changes

0xc0170 added needs: CI ready for merge and removed needs: work needs: CI labels Apr 26, 2018

cmonr added the release-version: 5.8.4 label Apr 27, 2018

cmonr merged commit 1f259a4 into ARMmbed:master Apr 27, 2018

cmonr removed the ready for merge label Apr 27, 2018

ccli8 deleted the nuvoton_fix_armc6_armv8m branch April 27, 2018 02:26

Fix build tool with ARMC6/ARMv8M #6644

Fix build tool with ARMC6/ARMv8M #6644

Uh oh!

Conversation

ccli8 commented Apr 16, 2018

Description

Pull request type

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

theotherjimmy commented Apr 16, 2018

Uh oh!

ccli8 commented Apr 17, 2018

Uh oh!

theotherjimmy commented Apr 17, 2018

Uh oh!

deepikabhavnani left a comment

Choose a reason for hiding this comment

Uh oh!

ccli8 commented Apr 18, 2018

Uh oh!

0xc0170 commented Apr 18, 2018

Uh oh!

mbed-ci commented Apr 18, 2018

Triggering tests

Uh oh!

0xc0170 commented Apr 18, 2018

Uh oh!

mbed-ci commented Apr 18, 2018

Uh oh!

mbed-ci commented Apr 18, 2018

Uh oh!

0xc0170 commented Apr 19, 2018

Uh oh!

mbed-ci commented Apr 19, 2018

Uh oh!

0xc0170 commented Apr 19, 2018

Uh oh!

cmonr commented Apr 23, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ccli8 commented Apr 24, 2018

Uh oh!

theotherjimmy commented Apr 24, 2018

Uh oh!

cmonr commented Apr 24, 2018

Uh oh!

mbed-ci commented Apr 24, 2018

Triggering tests

Uh oh!

mbed-ci commented Apr 24, 2018

Uh oh!

mbed-ci commented Apr 25, 2018

Uh oh!

ccli8 commented Apr 26, 2018

Uh oh!

Uh oh!

cmonr commented Apr 23, 2018 •

edited

Loading