Skip to content
This repository was archived by the owner on Mar 7, 2026. It is now read-only.

New platform: WeActStudio.BluePill-Plus-GD32#2071

Draft
ALTracer wants to merge 332 commits into
blackmagic-debug:mainfrom
ALTracer:feature/bluepillplus-platform
Draft

New platform: WeActStudio.BluePill-Plus-GD32#2071
ALTracer wants to merge 332 commits into
blackmagic-debug:mainfrom
ALTracer:feature/bluepillplus-platform

Conversation

@ALTracer

@ALTracer ALTracer commented Feb 9, 2025

Copy link
Copy Markdown
Contributor

Detailed description

  • This is a whole new cleaned up platform.
  • The existing problems are mismatching LED pins and incomplete mappings of board features between stlink/swlink and bluepillplus. In-tree platforms default to 72 MHz and do not leverage full potential of Gigadevices. Unmapping own SWJ-DP is not always acceptable, and stlink pinout is just weird.
  • This PR solves that by providing all necessary bindings in a dedicated platform, with chip detection and PLL configuration.

I do not expect this to be merged, because in-tree -Dprobe=bluepill already does provide some functionality. I do not implement vanilla PC13 bluepills, it's impossible for me to discern these from bluepillplus. This platform has feature parity with blackpill-f4, except I'm dissatisfied with its USBOTG-FS DWC2 bugs.
Choice of USART for Aux and SWO was influenced by existing 2x128 aux_serial DMA ping-pong buffers, which are insufficient for 1024 byte Zmodem packet operation I need in DUT firmware. So SWO is faster for burst operation. When that's refactored, I'll swap them back such that 6Mbaud VCP is possible.

TODO: I'd like to use platform_ident to indicate autodetection results, and to leverage meson options to pick linkerscripts which influence flash size for drivers and RAM size for heap/buffers/SWO.

Your checklist for this pull request

Closing issues

@dragonmux dragonmux added this to the v2.1 release milestone Feb 9, 2025
@dragonmux dragonmux added the Foreign Host Board Non Native hardware to runing Black Magic firmware on label Feb 9, 2025
@ALTracer
ALTracer force-pushed the feature/bluepillplus-platform branch from 1d39692 to fb6cb65 Compare March 6, 2025 04:41
@ALTracer
ALTracer force-pushed the feature/bluepillplus-platform branch from fb6cb65 to b84edbf Compare March 23, 2025 13:17
@ALTracer

ALTracer commented Jul 13, 2025

Copy link
Copy Markdown
Contributor Author

This now supports 96 MHz APM32F103CB.
As well as classic 72MHz STM32F103CB, 96MHz GD32F103CB and 120MHz GD32F303CC (Cortex-M4), differentiating on CoreSight ROM table PIDR value. Unique ID in DESig is not exactly helpful, but serialno.c still works.
All four WeActStudio.BluePill-Plus-* boards have identical PCB, with LED on PB2 active-high, and button on PA0, USB FS device on type-C port, and 8 MHz crystal. All four chips lack USB DFU bootloader capability, only USART1 is supported for initial flashing, if you don't have another USB-SWD/JTAG adapter.
CH32F103C8 not considered due to insufficient Flash capacity.

@ALTracer
ALTracer force-pushed the feature/bluepillplus-platform branch from ed79933 to f6dab76 Compare July 13, 2025 14:53
@ALTracer
ALTracer force-pushed the feature/bluepillplus-platform branch from f6dab76 to 97627db Compare July 23, 2025 19:23
dragonmux and others added 23 commits March 14, 2026 15:43
@dragonmux
target: Simplified the Flash memory entry formatter, shaving some siz…
e524e57 
…e off the binary and fixing it being grumpy if the length parameter was 0
@dragonmux
target: Added some missing const to the parameters of the memory ma…
3f556b0 
…p formatting functions
@dragonmux
riscv_debug: Fixed some type conversions lints in the target descript…
de9c855 
…ion builder function
@dragonmux
hosted/probe_info: Fixed the allocation failure error in probe_info_a…
0b2d04f 
…dd_by_id() not being an actual error
@dragonmux
hosted/bmp_libusb: Fixed (at least partially) probe_info_add_by_id() …
37c2393 
…usage being able to loose the input pointers if the entry allocation fails
@dragonmux
remote: Fix a comment carrying the incorrect command characters for a…
94fc333 
… command (JTAG run cycles command)
@dragonmux
hosted/remote/protocol_v2: Avoid issuing Jc to the probe if there's…
65ef172 
… nothing to do
@dragonmux
common/jtagtap: Avoid running any cycles if we're asked to do nothing…
7fc79bf 
… with `jtagtap_cycle()`
@dragonmux
codeberg: Implemented steps for building the firmware for all the dif…
dd49a6c 
…ferent probe platforms and their variants in CI
@dragonmux
codeberg: Implemented steps for building BMDA on Linux with both Clan…
3699698 
…g and GCC
@dragonmux
codeberg: Install the pre-requisites for BMDA in the firmware flow so…
a267510 
… we aren't having to build libftdi and other components and getting spurious errors out the flow
@dragonmux
codeberg: Remove -Werror from the BMDA flow on Linux for now till o…
b712c7a 
…ur dependencies can be cleaned up a biti
@dragonmux
codeberg: Enabled LTO for the Linux BMDA builds
1725e19
@dragonmux
codeberg: Added a cross-file describing how to build BMDA for Windows…
29436d8 
… from Linux using UCRT64 tooling
@dragonmux
codeberg: Implemented steps for building BMDA for Windows from Linux
b1c1679
@dragonmux
meson: Allow BMDA to be cross-compiled when not building the firmware…
c4ea63b 
… so that CI can function correctly
@hardesk
mspm0: Flash target using on-device core loop stub
50a001a
@hardesk
mspm0: Expand the id list of supported mspm0 targets
db181ee
@hardesk
mspm0: Remove the FACTORY/NONMAIN register inspection
9007cab 
TI added multiple layouts for the registers depending on a part. Implementing and maintaining it all is too much effort (and code) with dubious value.
@hardesk
mspm0: Fix condition check whether flash erase command has finished
7d51c11
@dragonmux
ctxlink/WiFi_Server: Cleaned up in the GDB send interface code
ab56f9f
@dragonmux
gdb_packet: Converted consume_remote_packet() to using the new GDB pa…
003f96b 
…cket structure
@dragonmux
remote: Begun to switch the remote protocol to be expressed in terms …
ed378c8 
…of the GDB packet structure
dragonmux and others added 29 commits July 2, 2026 12:02
@dragonmux
hosted/remote/protocol_v0_swd: Corrected some cycle count checks bein…
4edae41 
…g missing to properly bound the requests to the probe
@dragonmux
hosted/remote/protocol_v0_jtag: Corrected a missing cycle count check…
7d9151b 
… to properly bound the requests to the probe in the JTAG TMS sequence impl
@dragonmux
hosted/remote: Fixed the copyright years in several of the license he…
d6174ff 
…aders to reflect the updates being made
@dragonmux
hosted/remote/protocol_v4_adiv5: Fixed the debug prints to properly h…
55595e8 
…andle 64-bit addresses
@dragonmux
hosted/remote/protocol_v4_riscv: Added a missing static to fix a li…
32fbf67 
…nkage issue
@dragonmux
hosted/remote: Fixed an issue in the error dispatches in a bunch of t…
be70c33 
…he remote protocol command functions that would result in garabge or non-sensical error messages being displayed when `platform_buffer_read()` failed with a negative error value
@HrMitrev
usb: fix UB in global variable usb_config
65007de
@dragonmux
remote: Fixed a crash where if a user initialises JTAG comms and then…
7fee3fa 
… SWD comms for ADIv5, `dp->ensure_idle` could be left stale pointing at the JTAG function and crash the probe
@dragonmux
meson: Made sure to enable RISC-V protocol accelerations when the RIS…
178120c 
…C-V support itself is enabled in the firmware
@dragonmux
samx5x: Fixed the target and flash nomenclature
02bd728
@dragonmux
samx5x: Removed the optional memory write commands as they're all ava…
899bb64 
…ilable via the normal GDB memory interfaces anyway
@dragonmux
samx5x: Fixed a parens placement to improve readability
b40c1ab
@dragonmux
samx5x: Fixed a format string issue
4048ccf
@dragonmux
samx5x: Normalised the code organisation
704a130
@dragonmux
stm32h5: Fixed a missing static on the H5 command list
070301a
@dragonmux
timing: Fixed the timeout nomenclature
31e0fde
@dragonmux
sam3x: Fixed a missing static on the command list
976bc21
@dragonmux
renesas_ra: Fixed a couple nomenclature and static issues
ead40c2
@dragonmux
kinetis: Fixed the missing static's on the two command lists
9afbbe3
@dragonmux
efm32: Fixed a missing static on the command list
b624393
@dragonmux
ch32vx: Fixed a missing static on the command list
7fc83d2
@esden
meson: Disabled building of support for TI parts by default for ST-Li…
631a5de 
…nk v1 firmware to make things fit again
@esden
misc: Bumped the meson out of Git version to v2.1.0-rc1
79ad43c
@ALTracer
bluepillplus: New platform for WeActStudio.BluePill-Plus
7938b62 
* Uses PB2 LED, USB FS device with external 1.5k pullup
* BMD bootloader also uses PA0 button and boot magic flags
@ALTracer
bluepillplus: Add a platform README
b7c621c
@ALTracer
bluepillplus: Add support for SPI1
e7cd78b 
* WeActStudio.BluePill-Plus boards have a DNF SOIC-8
  footprint for 25-series SPI NOR flash, wired to PA4/5/6/7.
* Add the pin mappings regardless of the platform flavour.
@ALTracer
bluepillplus: Add chip detection and faster PLL configs
49a3c94 
* Rely on DBGMCU_IDCODE and SCB_CPUID
@ALTracer
bluepillplus: Add chip detection via CoreSight ROM tables
574ed94
@ALTracer
bluepillplus: Add chip detection of APM32F103
581a69c
@ALTracer
ALTracer force-pushed the feature/bluepillplus-platform branch from 5e3cdaf to 581a69c Compare July 13, 2026 17:24
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Reviewers

No reviews

Assignees

No one assigned

Labels

Foreign Host Board Non Native hardware to runing Black Magic firmware on

Projects

None yet

Milestone

v2.1 release

Development

Successfully merging this pull request may close these issues.

10 participants

@ALTracer @dragonmux @hardesk @mdrssv @esden @lasutek @lethalbit @emeb @zhangjiance @HrMitrev