Dual Flash and Hex file support on NCS36510 #4813

jacobjohnson-ON · 2017-07-26T00:00:49Z

Description

Dual flash and Hex file support added. This is a separate pull request to replace #4397. The changes are the same as the latest revisions on the previous PR, including the changes requested by @maclobdell.

Status

READY

Related PRs

#4397

Deploy notes

Build output will be hex file. Post build script will insert FIB, User trim and trims gap.
Post script generates hex and binary file. Binary file required if user uses boot loader to download.

Steps to test or reproduce

Clone latest branch of mbed-os, and modify NCS.py, ncs36510.icf, and targets.json
Clone mbed-os-example-blinky, and delete main.cpp
Update mbed-os.lib to point to local mbed-os library.
mbed deploy ( or copy library into project) make sure changes are present in blinky project.
mbed test -m NCS36510 -t IAR --compile (Also tested with GCC and ARM)
mbed test -m NCS36510 -t IAR --run

All tests passed.

Also tested 640k operation by restoring the top level main.cpp file with standard blinky example, and adding the line:
const uint8_t array[12000] = {0};

This resulted in the compiled flash section being 327kB. Hex file was loaded to the device and it runs properly.

Also of note, the board is using the latest version of bootloader (V15) and the latest version of DAPLink (see links on original PR)

…the proposed changes by Mac. 7/25/17

0xc0170 · 2017-07-26T09:10:31Z

targets/targets.json

@@ -3132,7 +3132,7 @@
        "device_name": "NUC472HI8AE",
        "bootloader_supported": true
    },
-    "NCS36510": {
+    "36510": {


I believe this is unintentional change. If we ever rename a target, it should be backward compatible.

Correct, that was unintentional and has been fixed. Thanks for catching it.

I agree. This looks like an unintentional change. It causes everything to break.

@jacobjohnson-ON Please push an additional commit on your branch that sets the target name back to "NCS36510".

@maclobdell I have already pushed this, and on my end it shows up in the commits to this PR.

0xc0170 · 2017-07-26T09:11:44Z

tools/targets/NCS.py

        start, end = start_end_pairs[0]
+        print("Memory start 0x%08X, end 0x%08X" % (start, end))


are these debug logs? Should not print anything by default?

Previously accepted versions of this file have used print statements by default. Is that not allowed? I'm fine removing these print statements if needed.

@0xc0170 Functionally, the print statements do not hurt anything, but for consistent user experience perhaps in the future they could be turned off by default. There can be a macro in the post build script that could enable / disable them.

Can this PR be merged as-is and then the issue of extra prints be addressed this later?

@jacobjohnson-ON I recommend using the toolchain notification functions instead. Let's do this: Get this merged with the prints in for now, and you submit a PR to clean them up later.

Optimally, you could use the "notify" method of the toolchains. This would look something like:

t_self.info("Merory stat 0x%08X, end 0x%08X" % (start, end))

This has two main benifits:

It heeds the -v and -vv flags.

It can print in the online compiler.

0xc0170 · 2017-07-26T09:12:38Z

@maclobdell @c1728p9 @theotherjimmy Please review

@jacobjohnson-ON Thanks for updating the dual flash support

theotherjimmy

Tools changes look fine. I recommend improving the post-binary script in another PR.

theotherjimmy · 2017-07-26T14:53:44Z

tools/targets/NCS.py

@@ -177,7 +204,7 @@ def add_fib_at_start(arginput):
    output_hex_file[trim_area_start + 1] = (mac_addr_low >> 8)  & 0xFF
    output_hex_file[trim_area_start + 2] = (mac_addr_low >> 16) & 0xFF
    output_hex_file[trim_area_start + 3] = (mac_addr_low >> 24) & 0xFF
-    
+


Thanks for getting this. I suppose it might be editor settings though 👍

theotherjimmy · 2017-07-26T15:00:15Z

tools/targets/NCS.py

@@ -29,17 +32,17 @@ def ranges(i):


 def add_fib_at_start(arginput):
-    input_file = arginput + ".bin"
-    file_name_hex = arginput + "_fib.hex"
+    input_file = arginput + ".hex"


For best user experience, you should avoid assung the location of this file. This change can come with another PR.

If you change the code that calls this function (in tools/targets/init.py) to:

class NCS36510TargetCode: @staticmethod def ncs36510_addfib(t_self, resources, elf, binf): from tools.targets.NCS import add_fib_at_start add_fib_at_start(binf)

You could correspondingly change this function prototype (and first few lines) to:

def add_fib_at_start(input_file): basename, ext = os.path.splitext(input_file) assert(ext is ".hex", "Error: Post-bulid script input must be an intel-Hex format file") file_name_hex = basename + ".hex" file_name_hex = basename + ".bin"

This would be a bit safer, and a bit easier to understand when it fails.

maclobdell · 2017-07-26T15:03:01Z

For reference, I manually ran all the mbed-os tests with IAR, ARM, and GCC_ARM and they all passed on the fork submitted in this PR. Looks good to me. Thanks!

theotherjimmy · 2017-07-26T15:04:52Z

Let's get this tested!

/morph test

theotherjimmy · 2017-07-26T15:04:59Z

/morph export-build

pradeep-gr · 2017-07-27T08:26:38Z

CC @danish-iftikhar @danclement @maclobdell @jacobjohnson-ON Changes looks fine to me and all gcc_arm/arm/iar tests are passing on latest mbed OS as on yesterday plus this PR.

theotherjimmy · 2017-07-31T15:21:09Z

@studavekar One non-reporter, one assembler bug. Could you retrigger after the release.

maclobdell · 2017-08-04T14:23:56Z

@theotherjimmy @studavekar Please retest this, so it can be merged and part of the next release.

studavekar · 2017-08-04T14:39:36Z

/morph test

studavekar · 2017-08-04T14:39:55Z

/morph export-build

mbed-bot · 2017-08-04T16:16:33Z

Result: FAILURE

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

/morph export-build

Output

mbed Build Number: 109

Exporter Build failed!

jacobjohnson-ON · 2017-08-04T16:23:08Z

I cannot access the "Details" link for the tests. Can someone explain the latest test failure?

mbed-bot · 2017-08-04T17:37:41Z

Result: SUCCESS

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

/morph test

Output

mbed Build Number: 937

All builds and test passed!

0xc0170 · 2017-08-07T13:19:50Z

/morph export-build

0xc0170 · 2017-08-07T13:28:35Z

I cannot access the "Details" link for the tests. Can someone explain the latest test failure?

The failures were not related to this patch, I restarted the CI.

mbed-bot · 2017-08-07T20:49:09Z

Result: FAILURE

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

/morph export-build

Output

mbed Build Number: 113

Exporter Build failed!

0xc0170 · 2017-08-09T11:46:30Z

/morph export-build

maclobdell · 2017-08-09T15:11:01Z

@0xc0170 @theotherjimmy This PR is high priority. Thanks for helping get it in.

studavekar · 2017-08-10T13:39:10Z

/morph export-build

mbed-bot · 2017-08-10T21:59:54Z

Result: SUCCESS

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

/morph export-build

Output

mbed Build Number: 118

All exports and builds passed!

mbed OS 5.5.3, including the changes needed to enable 640k. Includes …

41a0941

…the proposed changes by Mac. 7/25/17

jacobjohnson-ON mentioned this pull request Jul 26, 2017

Dual flash and Hex file support on NCS36510 #4397

Closed

0xc0170 reviewed Jul 26, 2017

View reviewed changes

0xc0170 added the needs: review label Jul 26, 2017

Changed 36510 -> NCS36510 in targets.json

bd60add

maclobdell assigned maclobdell and unassigned maclobdell Jul 26, 2017

theotherjimmy approved these changes Jul 26, 2017

View reviewed changes

theotherjimmy added needs: CI and removed needs: review labels Jul 26, 2017

adbridge merged commit 8a1c7fa into ARMmbed:master Aug 11, 2017

adbridge added release-version: 5.5.5 and removed needs: CI labels Aug 11, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dual Flash and Hex file support on NCS36510 #4813

Dual Flash and Hex file support on NCS36510 #4813

jacobjohnson-ON commented Jul 26, 2017 •

edited

0xc0170 Jul 26, 2017

jacobjohnson-ON Jul 26, 2017

maclobdell Jul 26, 2017

jacobjohnson-ON Jul 26, 2017

0xc0170 Jul 26, 2017

jacobjohnson-ON Jul 26, 2017

maclobdell Jul 26, 2017

theotherjimmy Jul 26, 2017 •

edited

0xc0170 commented Jul 26, 2017 •

edited

theotherjimmy left a comment

theotherjimmy Jul 26, 2017

theotherjimmy Jul 26, 2017 •

edited

maclobdell commented Jul 26, 2017

theotherjimmy commented Jul 26, 2017

theotherjimmy commented Jul 26, 2017

pradeep-gr commented Jul 27, 2017

theotherjimmy commented Jul 31, 2017

maclobdell commented Aug 4, 2017

studavekar commented Aug 4, 2017

studavekar commented Aug 4, 2017

mbed-bot commented Aug 4, 2017

jacobjohnson-ON commented Aug 4, 2017

mbed-bot commented Aug 4, 2017

0xc0170 commented Aug 7, 2017

0xc0170 commented Aug 7, 2017

mbed-bot commented Aug 7, 2017

0xc0170 commented Aug 9, 2017

maclobdell commented Aug 9, 2017

studavekar commented Aug 10, 2017

mbed-bot commented Aug 10, 2017

		start, end = start_end_pairs[0]
		print("Memory start 0x%08X, end 0x%08X" % (start, end))

Dual Flash and Hex file support on NCS36510 #4813

Dual Flash and Hex file support on NCS36510 #4813

Conversation

jacobjohnson-ON commented Jul 26, 2017 • edited

Description

Status

Related PRs

Deploy notes

Steps to test or reproduce

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

theotherjimmy Jul 26, 2017 • edited

Choose a reason for hiding this comment

0xc0170 commented Jul 26, 2017 • edited

theotherjimmy left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

theotherjimmy Jul 26, 2017 • edited

Choose a reason for hiding this comment

maclobdell commented Jul 26, 2017

theotherjimmy commented Jul 26, 2017

theotherjimmy commented Jul 26, 2017

pradeep-gr commented Jul 27, 2017

theotherjimmy commented Jul 31, 2017

maclobdell commented Aug 4, 2017

studavekar commented Aug 4, 2017

studavekar commented Aug 4, 2017

mbed-bot commented Aug 4, 2017

Output

jacobjohnson-ON commented Aug 4, 2017

mbed-bot commented Aug 4, 2017

Output

0xc0170 commented Aug 7, 2017

0xc0170 commented Aug 7, 2017

mbed-bot commented Aug 7, 2017

Output

0xc0170 commented Aug 9, 2017

maclobdell commented Aug 9, 2017

studavekar commented Aug 10, 2017

mbed-bot commented Aug 10, 2017

Output

jacobjohnson-ON commented Jul 26, 2017 •

edited

theotherjimmy Jul 26, 2017 •

edited

0xc0170 commented Jul 26, 2017 •

edited

theotherjimmy Jul 26, 2017 •

edited