Fix: Speed up bus turnaround for blackpill-f4 like other platforms#1690
Fix: Speed up bus turnaround for blackpill-f4 like other platforms#1690dragonmux merged 1 commit intoblackmagic-debug:mainfrom
Conversation
dragonmux
left a comment
There was a problem hiding this comment.
This looks good and provides some nice consistency with the other platforms. There are a couple of review items concerning signed-unsigned conversion to sort, but with the assumption that you've tested this with a Black Pill and it works, we've got no problems merging this once those are sorted.
ebcc14d to
c8252b7
Compare
dragonmux
left a comment
There was a problem hiding this comment.
LGTM, merging. Thank you for the contribution!
* Manipulate STM32F4 GPIO MODER registers directly (a read-modify-write) instead of going through function calls like `gpio_mode_setup()` * Add the two alternative pinouts to pre-cooked macros
c8252b7 to
7f1e88b
Compare
|
Adjusted per review items, squashed into one; and rebased to latest main. |
Detailed description
blackpill-f4family of platforms.Testing on fix/tap-timing branch with a blackpill-f411ce in max frequency (~8MHz) against a STM32MP157_CM4 shows an increase of performance of loading a 64 KiB binary into RETRAM from 156 to 180 KiB/s. Impact may be less noticeable when using Flash-based targets, or using this BMF with BMDA as proxy.
I'll mention that stlinkv3 also toggles a GPIO wired to direction pin of a bidirectional level translator in the respective macros.
Your checklist for this pull request
make PROBE_HOST=native) and should not affect itmake PROBE_HOST=hosted) and should not affect itClosing issues
Enhances #1688 (PR, not an Issue)