Add target definition: nrf51 16k on MCU_NRF51_UNIFIED platform #6178

andrewleech · 2018-02-23T00:59:07Z

The flash base address in the linker (APP_CODE_BASE) for nrf51_16k_s130 is incorrect, as per the official documentation for S130: http://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.s130.sds%2Fdita%2Fsoftdevices%2Fs130%2Fmem_usage%2Fmem_resource_reqs.html

Without this change applications do not boot when running on a 16K nrf51 (not very common any more) as the softdevice jumps to the wrong location to start the application.

I've been running with this fix in place using a Raytac MDBT40 module with only 16K ram.

Pull request type

[x] Fix
[ ] Refactor
[ ] New Target
[ ] Feature
[ ] Breaking Change

cmonr · 2018-03-06T03:24:29Z

@marcuschangarm Ping.

marcuschangarm · 2018-03-06T05:09:05Z

@cmonr I'm not familiar with the NRF51 code base anymore.

cmonr · 2018-03-06T20:30:49Z

@marcuschangarm Thanks for the update. Feel free to remove yourself from the reviewers list (oddly enough, not sure how to do that on my end...).

cmonr · 2018-03-06T20:32:26Z

@donatieng If you're not the right person to review, we're open to suggestions.

donatieng

Looks like there is no 16K version of the "TARGET_MCU_NRF51822_UNIFIED" target in targets.json.

The other NRF51 16K targets use TARGET_MCU_NRF51822 which uses the S130 v1 Softdevice for which the linker scripts are correct.

Although there's quite a bit of cleanup that should be done around NRF51 targets.

cmonr · 2018-03-21T02:47:04Z

@donatieng So what are the side effects of this PR noy being able to go in? Does this mean that the NRF51 cannot support Flash properly, or that it can not use mbed-os bootloader features?

andrewleech · 2018-03-21T04:21:57Z

Yes I didn't entirely realise there were no targets currently configured that use these linker files.
I'm running with a custom_targets.json to create nRF51 targets using nrf unified platform, which then hit this linker file.
I currently have to .mbedignore the built in ones and have this fixed copy locally in my project.

This does not relate to bootloader, but to regular apps running on nrf51 16k with unified platform.

0xc0170 · 2018-04-09T15:44:05Z

Yes I didn't entirely realise there were no targets currently configured that use these linker files.
I'm running with a custom_targets.json to create nRF51 targets using nrf unified platform, which then hit this linker file.
I currently have to .mbedignore the built in ones and have this fixed copy locally in my project.

This does not relate to bootloader, but to regular apps running on nrf51 16k with unified platform.

@donatieng What do you think? Please review

cmonr · 2018-05-08T01:30:03Z

@donatieng Any thoughts?

@andrewleech Sorry, but this now needs a rebase...

donatieng

Thanks @andrewleech, please see my comments. ~~We're also missing linker files for IAR and ARMCC.~~ Update: (sorry those are already in and correct)

donatieng · 2018-05-08T09:16:27Z

...5/TARGET_MCU_NRF51822_UNIFIED/device/TOOLCHAIN_GCC_ARM/TARGET_MCU_NRF51_16K_S130/NRF51822.ld

@@ -2,7 +2,7 @@

 MEMORY
 {
-  FLASH (rx) : ORIGIN = 0x0001C000, LENGTH = 0x24000
+  FLASH (rx) : ORIGIN = 0x0001B000, LENGTH = 0x24000


The length should be updated too, to 0x25000 to use the full 256K of flash.

Not on mine, no. I've only got 128K flash.
Yes we probably should have linker files for all different variants of the nrf51 and a convenient way to choose which one is in use (I guess it's just a different target for each).

Indeed it'd be much better to have separate targets definitions - if we're targeting a device with 128K of flash, length should be 0x5000

Actually no, I must have the 256K part. After you take off the softdevice's chunk the 128K part only has 20K left for user code... on an arm that's nothing. I doubt there's many people actually using them.

I'll fix the linker for 256K, if anyone really needs the 128K part that could be a separate commit (with new target definitions and a new full set of linker scripts)

Thanks, still need updating to 0x25000 for the 256K part :)

Oh what, I did fix that. Must have rebased an old copy or something :-(

donatieng · 2018-05-08T09:20:59Z

...5/TARGET_MCU_NRF51822_UNIFIED/device/TOOLCHAIN_GCC_ARM/TARGET_MCU_NRF51_16K_S130/NRF51822.ld

@@ -2,7 +2,7 @@

 MEMORY
 {
-  FLASH (rx) : ORIGIN = 0x0001C000, LENGTH = 0x24000
+  FLASH (rx) : ORIGIN = 0x0001B000, LENGTH = 0x24000
  RAM (rwx) :  ORIGIN = 0x20002800, LENGTH = 0x1800


RAM start should be 0x20002ef8 and length adjusted accordingly ~~- this will need to be tested, do we still have enough RAM to run a program reliably?~~ (already in for ARMCC and IAR so assuming that works!)

Where did you get 0x20002ef8 from?

The minimum is 0x200013C8 according to the official memory requirements:
http://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.s130.sds%2Fdita%2Fsoftdevices%2Fs130%2Fmem_usage%2Fmem_resource_reqs.html&anchor=mem_resource_reqs__table_n5z_2lp_3r

We don't have the configuration at minimum so more is needed that this, but I'm not sure if it does need to be as high as 0x20002ef8?

Obviously this all changes with my other PR #5980 but I haven't finished rebasing / trimming
/ re-testing that and I've well and truly run out of work time to keep adding more than taking back on that one.

And no there's not much ram to work with (hence wanting to keep the linker value low). You really can't fit the rtos but it's working well to build an interrupt driven custom uart <-> ble bridge.

Ok I've done some more testing at default softdevice settings and the reported required ram start is 0x20002EF0 which is pretty much the same as 0x20002ef8

cmonr · 2018-05-21T16:25:29Z

@andrewleech Please take a look at the travis issues.

For details see: http://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.s130.sds%2Fdita%2Fsoftdevices%2Fs130%2Fmem_usage%2Fmem_resource_reqs.html Set the ram start to 0x20002ef8 to be consistent with ATM_STD and IAR linker scripts

…orm: MCU_NRF51_16K_UNIFIED_S130 Add target definition for Raytac MDBT40 module to use the MCU_NRF51_16K_UNIFIED_S130 target.

cmonr · 2018-05-22T21:34:16Z

@donatieng @pan- @marcuschangarm Would y'all mind re-reviewing?

andrewleech · 2018-05-22T21:54:45Z

@cmonr, @donatieng did (in hidden comment thread above) my rebase managed to lose my previous change fixing the flash end address. I'll have to fix that again first

cmonr · 2018-05-22T22:32:02Z

@andrewleech Thanks for the heads up!

cmonr · 2018-05-31T16:29:23Z

@donatieng A quick re-review?
@pan- @marcuschangarm Thoughts?

donatieng

Great stuff, thanks @andrewleech!

cmonr · 2018-06-01T22:07:08Z

/morph build

mbed-ci · 2018-06-01T23:20:35Z

Build : SUCCESS

Build number : 2223
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6178/

Triggering tests

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

mbed-ci · 2018-06-02T05:38:16Z

Test : SUCCESS

Build number : 2013
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/6178/2013

mbed-ci · 2018-06-02T13:04:00Z

Exporter Build : FAILURE

Build number : 1851
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/6178/

0xc0170 · 2018-06-07T09:07:06Z

/morph build

mbed-ci · 2018-06-07T14:00:49Z

Build : SUCCESS

Build number : 2270
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6178/

Triggering tests

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

cmonr · 2018-06-07T16:44:47Z

Halting CI builds until RC3 PRs are completed. Will resume after.

0xc0170 · 2018-06-09T06:40:15Z

/morph build

mbed-ci · 2018-06-09T09:09:21Z

Build : SUCCESS

Build number : 2292
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6178/

Triggering tests

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

mbed-ci · 2018-06-09T14:00:40Z

Exporter Build : SUCCESS

Build number : 1916
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/6178/

mbed-ci · 2018-06-09T17:03:19Z

Test : SUCCESS

Build number : 2070
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/6178/2070

0xc0170 · 2018-06-10T17:37:11Z

/morph mbed2-build

No files to build - should not be in targets Reverts part of the ARMmbed#6178

No files to build - should not be in targets Reverts part of the #6178

No files to build - should not be in targets Reverts part of the ARMmbed#6178

cmonr requested a review from marcuschangarm February 23, 2018 03:13

cmonr added the needs: review label Feb 23, 2018

cmonr requested a review from pan- March 6, 2018 03:19

cmonr requested a review from donatieng March 6, 2018 20:31

donatieng suggested changes Mar 7, 2018

View reviewed changes

andrewleech changed the title ~~nrf51_16k_s130 Fix flash origin for APP_CODE_BASE~~ Add target definition: nrf51 16k on MCU_NRF51_UNIFIED platform Mar 21, 2018

donatieng suggested changes May 8, 2018

View reviewed changes

andrewleech force-pushed the nrf51_16k_s130_linker branch from 9e21c61 to afbd51e Compare May 14, 2018 05:34

cmonr added needs: work and removed needs: review labels May 14, 2018

pi-anl added 2 commits May 22, 2018 11:40

Add target definitions for using 16k nrf51 on the NRF51_UNIFIED platf…

71b2caf

…orm: MCU_NRF51_16K_UNIFIED_S130 Add target definition for Raytac MDBT40 module to use the MCU_NRF51_16K_UNIFIED_S130 target.

andrewleech force-pushed the nrf51_16k_s130_linker branch from 46252a9 to 71b2caf Compare May 22, 2018 01:47

cmonr added needs: review and removed needs: work labels May 22, 2018

Fix the flash length to use all of 256K part.

78d31cf

0xc0170 added the scope: new-target label May 23, 2018

donatieng approved these changes Jun 1, 2018

View reviewed changes

cmonr added needs: CI and removed needs: review labels Jun 1, 2018

kjbracey added ready for merge and removed needs: CI labels Jun 11, 2018

cmonr approved these changes Jun 11, 2018

View reviewed changes

cmonr merged commit b989afa into ARMmbed:master Jun 11, 2018

0xc0170 removed the ready for merge label Jun 11, 2018

0xc0170 mentioned this pull request Jul 10, 2018

Raytac MDBT40 - files are missing - fails to build #7460

Closed

0xc0170 added a commit to 0xc0170/mbed-os that referenced this pull request Jul 10, 2018

Raytac: target removal

396c88a

No files to build - should not be in targets Reverts part of the ARMmbed#6178

0xc0170 mentioned this pull request Jul 10, 2018

Raytac: target removal #7461

Merged

cmonr pushed a commit that referenced this pull request Jul 15, 2018

Raytac: target removal

b456430

No files to build - should not be in targets Reverts part of the #6178

cmonr pushed a commit that referenced this pull request Jul 16, 2018

Raytac: target removal

032051e

No files to build - should not be in targets Reverts part of the #6178

cmonr pushed a commit that referenced this pull request Jul 16, 2018

Raytac: target removal

3950834

No files to build - should not be in targets Reverts part of the #6178

pan- pushed a commit to pan-/mbed that referenced this pull request Aug 22, 2018

Raytac: target removal

dde0930

No files to build - should not be in targets Reverts part of the ARMmbed#6178

Add target definition: nrf51 16k on MCU_NRF51_UNIFIED platform #6178

Add target definition: nrf51 16k on MCU_NRF51_UNIFIED platform #6178

Conversation

andrewleech commented Feb 23, 2018 • edited by cmonr Loading

Pull request type

cmonr commented Mar 6, 2018

marcuschangarm commented Mar 6, 2018

cmonr commented Mar 6, 2018

cmonr commented Mar 6, 2018

donatieng left a comment

Choose a reason for hiding this comment

cmonr commented Mar 21, 2018

andrewleech commented Mar 21, 2018

0xc0170 commented Apr 9, 2018

cmonr commented May 8, 2018

donatieng left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

donatieng May 16, 2018 • edited Loading

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

donatieng May 8, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cmonr commented May 21, 2018

cmonr commented May 22, 2018

andrewleech commented May 22, 2018

cmonr commented May 22, 2018

cmonr commented May 31, 2018

donatieng left a comment

Choose a reason for hiding this comment

cmonr commented Jun 1, 2018

mbed-ci commented Jun 1, 2018

Triggering tests

mbed-ci commented Jun 2, 2018

mbed-ci commented Jun 2, 2018

0xc0170 commented Jun 7, 2018

mbed-ci commented Jun 7, 2018

Triggering tests

cmonr commented Jun 7, 2018

0xc0170 commented Jun 9, 2018

mbed-ci commented Jun 9, 2018

Triggering tests

mbed-ci commented Jun 9, 2018

mbed-ci commented Jun 9, 2018

0xc0170 commented Jun 10, 2018

andrewleech commented Feb 23, 2018 •

edited by cmonr

Loading

donatieng left a comment •

edited

Loading

donatieng May 16, 2018 •

edited

Loading

donatieng May 8, 2018 •

edited

Loading