Cortex-M target reset unify by cederom · Pull Request #371 · ARMmbed/DAPLink

cederom · 2017-12-14T22:37:47Z

This PR brings the unification to target reset routines trying the software reset first and then hardware reset if the first fails.

Both swd_set_target_state_hw() and swd_set_target_state_sw() are replaced with target_set_state() which also removes set state from target implementations unless reset really requires some unique reset quirk. New functions will be placed in dedicated target.{c,h} file also separate from swd_host.{c,h}.

Still testing, please merge only after label is removed and code is confirmed after in depth testing..

cederom · 2017-12-14T22:38:06Z

/morph test

cederom · 2017-12-14T23:13:08Z

Boards test status:

Vendor	HIC	Target	Status	Previous Reset Implementation and Remarks
Freescale	K20DX	FRDM-K64F	💚 OK	HARDWARE + BEFORE_INIT + QUIRK + TRICKY (UN)LOCK + SECURITY_BITS
Onsemi	SAM3U2C	NCS36510-RFv1.0	💚 OK	SOFTWARE RESET
Nordic	SAM3U2C	nRF51, nRF52	💚 OK	SOFTWARE + SWD_SET_TARGET_RESET() QUIRK MIXING HW GPIO AND SW RESET
Microbit	KL26	nRF51822	💚 OK	NO HARDWARE LINES?
LPC	LPC4322	LPC54114	💚 OK
NXP	11U35	`*`	💚 OK	HARDWARE RESET
NXP	11U35	LPC4088	💚 OK	QUIRK
Renesas	11U35	R7S721001	💚 OK	HARDWARE, GR-PEACH, CORTEX-A, SW RST DOES NOT WORK!
SiliconLabs	`*`	`*`	💛 TODO	SOFTWARE
STM	`*`	`*`	🔴 DAPLink on STM	HARDWARE
STM	NZ32SC151	`*`	🔴 DAPLink on STM	HARDWARE
STM	xDOT-L151	`*`	🔴 DAPLink on STM	HARDWARE
Toshiba	`*`	`*`	🔴 HARDWARE UNAVAILABLE	SOFTWATE + HARDWARE_PRE_INIT
Wiznet	`*`	`*`	🔴 HARDWARE UNAVAILABLE	QUIRK

mbed-bot · 2017-12-14T23:19:14Z

Build finished.

Details and progress at ARMmbed#371 Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

c1728p9 · 2017-12-15T00:34:06Z

source/target/freescale/target_reset_Kseries.c


+/**
+ * Unlock Sequence clears out the Security Bits and Flash Memory.
+ * Hardware Reset line MUST be asserted due implementation issues.


Here are the details to add in the place of my todo:

During the unlock sequence the hardware reset must be asserted to ensure device code cannot execute and interfere with the unlocking procedure.

Added, thanks! :-)

c1728p9 · 2017-12-15T00:36:26Z

source/daplink/interface/swd_host.c


        case RESET_PROGRAM:
+            // Initialize SWD DAP Port so we have a known state of the Target.
+            // DAP is essential for both softwaRESET_HOLDre/hardware halt-on-reset.


I think this is a typo - softwaRESET_HOLDre

thanks! :-)

Details and progress at ARMmbed#371 Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom · 2017-12-16T00:46:31Z

Need to rewrite / separate nRF51822 code that de-asserts the nRST line because it also disables DAP which breaks RESET_PROGRAM (it has de-assert in the middle).. except that existing (de)asserting can be used in RESET_HOLD and RESET_RUN.. introduced swd_debug_deinit() for that :-)

Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state) * Removed return swd_set_target_state_sw(state) * Removed return swd_set_target_state_hw(state) * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom · 2017-12-19T00:08:53Z

Redesigned target_set_state():

RESET_HOLD and RESET_PROGRAM now works the same way and is the same code (init debug, reset, halt on reset using vector catch).
RESET_RUN first reset and halt just as above then reset again to run the code.

Remarks for NORDIC implementation and changes:

Unified target_set_state() now seems to work without specific swd_set_target_reset() implementation that was quirky and was causing problems.
NORDIC target probably still needs a quirk to disable debug logic.

swd_set_target_reset() is not __attribute__((weak)) anymore. This way only hardware reset is handled by those calls if implemented.

Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state) * Removed return swd_set_target_state_sw(state) * Removed return swd_set_target_state_hw(state) * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom · 2017-12-19T01:51:38Z

New reset code seems to work on nRF51822, nRF52832 (flashing, cdc break). Even on nRF52840 except flash algo cannot yet erase the memory but the cdc break worked..
Introduced changes did not break K64F and NCS36510 :-)

cederom · 2017-12-19T01:51:53Z

/morph test

mbed-bot · 2017-12-19T02:43:47Z

Build finished.

cederom · 2017-12-19T21:25:30Z

Microbit KL26 / nRF51822 works fine :-)

cederom · 2017-12-19T22:03:42Z

LPCXpressso5411x HIC_LPC4322 / TGT_LPC54114 works fine :-)

cederom · 2017-12-20T21:27:43Z

Embedded Artists 11U35_HIC / LPC4088_TGT works fine :-)

TODO: Rewrite ISP GPIO toggling wrapped around reset calls..

Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state) * Removed return swd_set_target_state_sw(state) * Removed return swd_set_target_state_hw(state) * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. * Updated some pre/post reset routine quirks. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

Derrick45 · 2018-01-13T15:49:19Z

LPC11U35 plus nrf51822 can't reset automatically using msd,why?

cederom · 2018-01-15T09:51:51Z

Hey @wzt19910506 this thread is about DAPLink RESET REWORK ONLY not the LPC+nRF Drag-n-Drop as stated in #375, please talk about LPC+nRF over there.

However, you can try new RESET code from this Pull Request on your LPC HIC with external nRF Target and will should work even with no nRST line attached as we first try the software reset by default!

Note: without custom HIC+Target firmware you only get HID Debug and CDC UART working, both can reset your target, so you can verify the new code.

WARNING: BREAKS NCS FLASHING! Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state) * Removed return swd_set_target_state_sw(state) * Removed return swd_set_target_state_hw(state) * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. * Updated some pre/post reset routine quirks. * Introducing pre/post reset flash/run quirks. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom · 2018-01-18T12:26:20Z

4.0 adds pre/post reset run/flash quirks as required by LPC4088DM board.

Some problems with NCS flashing are observed but not caused by last change.. BIN files flashing works fine, HEX flashing does not work over UMS. Is that a known issue?

cederom · 2018-01-18T12:26:28Z

/morph test

mbed-bot · 2018-01-18T13:26:53Z

Build finished.

cederom · 2018-01-18T19:54:41Z

Current (copy-paste) updated reset code for Cortex-M does not work on R7S Renesas GR-PEACH as this is the Cortex-A Target.
Cortex-M has its SWD implementation in swd_host.c while Cortex-A has its implementation in swd_host_ca.c. They both share swd_host.h but they are complete separate implementations compiled into the firmware based on TARGET_MCU_CORTEX_A define and source files.
On the GR-PEACH board, the old Cortex-A swd_set_target_state_hw(state) works fine, while swd_set_target_state_sw(state) does not. Therefore Cortex-A version of target_set_state() simply wraps the swd_set_target_state_hw(state).
Reset unify is be done for Cortex-M only. If that works fine, the target_ code could be taken out to the separate source file and use underlying (alternate) Cortex-M or Cortex-A SWD Host routines. Cortex-A could be rewritten then in a similar manner.

Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state) * Removed return swd_set_target_state_sw(state) * Removed return swd_set_target_state_hw(state) * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. * Updated some pre/post reset routine quirks. * Introducing pre/post reset flash/run quirks. * Leaves Cortex-A rework for Target code separation. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

Details and progress at ARMmbed#371 Summary of changes: * Introducing global unified target_set_state(TARGET_RESET_STATE state). * Removed swd_set_target_state_sw(state). * Removed swd_set_target_state_hw(state). * Function swd_set_target_reset(uint8_t asserted) is not weak anymore. * Updated some pre/post reset routine quirks. * Introducing pre/post reset flash/run quirks. * Leaves Cortex-A rework for Target code separation. Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom · 2018-01-22T00:28:05Z

@c1728p9 could you please take a look in a free moment if this is okay? :-)

Do we want to test the STM targets? If so how can I replace ST-LINK with DAPLink?
Cortex-A seems to have separate implementation of SWD Host so reset unify will have to be done similar way there too, preferably moving out the Target code from swd_host.c to target.c and using provided underlying Cortex-A/M swd_ functions..
I have also utilised the DEBUG and NO_DEBUG states using existing code + switch() syntax trick.

40Grit · 2019-05-22T11:24:40Z

What is status of this?

0xc0170 · 2019-05-22T11:38:20Z

@40Grit What's your interest in this PR? I believe it's stalled and might be just closed now. If there's anything that should be picked up again, let us know.

0Grit · 2019-05-22T13:17:57Z

@0xc0170
Wondering:

Has it been superseded?
Is it still relevant?
What was its benefit?

cederom · 2019-05-22T14:00:15Z

Hey @loverdeg-ep / @40Grit :-) Developers had different idea for reset mechanism and machine state design for DAPLink, and the whole architecture, if any, in general, so probably this is kinda dead end sorry. The master went ahead. You can make use of it as you like :-)

cederom force-pushed the cederom_reset_unify branch 2 times, most recently from 95cf63a to 2fd637d Compare December 14, 2017 23:39

cederom added a commit to cederom/DAPLink that referenced this pull request Dec 14, 2017

EXPERIMENTAL! Target reset unify. rev 3.3.

2fd637d

Details and progress at ARMmbed#371 Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom force-pushed the cederom_reset_unify branch from 2fd637d to d493edd Compare December 14, 2017 23:49

cederom added a commit to cederom/DAPLink that referenced this pull request Dec 14, 2017

EXPERIMENTAL! Target reset unify. rev 3.4.

d493edd

Details and progress at ARMmbed#371 Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

c1728p9 reviewed Dec 15, 2017

View reviewed changes

cederom force-pushed the cederom_reset_unify branch from d493edd to ea96943 Compare December 16, 2017 00:41

cederom added a commit to cederom/DAPLink that referenced this pull request Dec 16, 2017

EXPERIMENTAL! Target reset unify. rev 3.5 (Working on nRF51 RST)

ea96943

Details and progress at ARMmbed#371 Signed-off-by: CeDeROM Tomasz CEDRO <tomek@cedro.info>

cederom mentioned this pull request Dec 18, 2017

CORE / Unify HW and SW reset sequences #338

Open

cederom force-pushed the cederom_reset_unify branch from ea96943 to 48b75e2 Compare December 18, 2017 23:58

cederom force-pushed the cederom_reset_unify branch from 48b75e2 to 011dcb5 Compare December 19, 2017 00:10

cederom force-pushed the cederom_reset_unify branch from 011dcb5 to 66de826 Compare December 20, 2017 21:34

cederom force-pushed the cederom_reset_unify branch from 66de826 to c78bf50 Compare December 20, 2017 21:50

cederom mentioned this pull request Jan 14, 2018

Documentation is unclear / OSHChip_CMSIS_DAP_V1.0 + nRF51822 #375

Closed

cederom force-pushed the cederom_reset_unify branch from c78bf50 to fe222a5 Compare January 18, 2018 12:25

cederom force-pushed the cederom_reset_unify branch from fe222a5 to 8be221d Compare January 18, 2018 20:55

cederom force-pushed the cederom_reset_unify branch from 8be221d to acb0b7c Compare January 18, 2018 20:59

cederom changed the title ~~target reset unify - DO NOT YET MERGE -~~ Cortex-M target reset unify - DO NOT YET MERGE - Jan 19, 2018

cederom force-pushed the cederom_reset_unify branch from acb0b7c to bf51af5 Compare January 22, 2018 00:08

cederom mentioned this pull request Feb 1, 2018

Add support for Rapid-IoT & Hexiwear boards #385

Merged

cederom changed the title ~~Cortex-M target reset unify - DO NOT YET MERGE -~~ Cortex-M target reset unify Feb 9, 2018

This was referenced Apr 2, 2018

Does LPC11U35 DAPLink support drag and drop programming (with nRF52832 connected)? #415

Closed

Avoid calling swd_init() when invoking the RUN state #417

Merged

cederom mentioned this pull request Jan 4, 2019

Error: The transfer timed out. #440

Closed

cederom mentioned this pull request Mar 20, 2019

CDC break reset doesn't work with SW reset #611

Open

0xc0170 closed this Nov 8, 2019

cederom mentioned this pull request Apr 24, 2021

Add software auto reset to Atmel and stm32 #794

Open

Conversation

cederom commented Dec 14, 2017

Uh oh!

cederom commented Dec 14, 2017

Uh oh!

cederom commented Dec 14, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mbed-bot commented Dec 14, 2017

Uh oh!

c1728p9 Dec 15, 2017

Choose a reason for hiding this comment

Uh oh!

cederom Dec 16, 2017

Choose a reason for hiding this comment

Uh oh!

c1728p9 Dec 15, 2017

Choose a reason for hiding this comment

Uh oh!

cederom Dec 16, 2017

Choose a reason for hiding this comment

Uh oh!

cederom commented Dec 16, 2017

Uh oh!

cederom commented Dec 19, 2017

Uh oh!

cederom commented Dec 19, 2017

Uh oh!

cederom commented Dec 19, 2017

Uh oh!

mbed-bot commented Dec 19, 2017

Uh oh!

cederom commented Dec 19, 2017

Uh oh!

cederom commented Dec 19, 2017

Uh oh!

cederom commented Dec 20, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Derrick45 commented Jan 13, 2018

Uh oh!

cederom commented Jan 15, 2018

Uh oh!

cederom commented Jan 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cederom commented Jan 18, 2018

Uh oh!

mbed-bot commented Jan 18, 2018

Uh oh!

cederom commented Jan 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cederom commented Jan 22, 2018

Uh oh!

40Grit commented May 22, 2019

Uh oh!

0xc0170 commented May 22, 2019

Uh oh!

0Grit commented May 22, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cederom commented May 22, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

cederom commented Dec 14, 2017 •

edited

Loading

cederom commented Dec 20, 2017 •

edited

Loading

cederom commented Jan 18, 2018 •

edited

Loading

cederom commented Jan 18, 2018 •

edited

Loading

0Grit commented May 22, 2019 •

edited

Loading