ESP32S3 USB JTAG cable maybe possible

[emard]

Currently I'm experimenting to get ESP32S3 USB JTAG which is supposedly
able to route its native JTAG to external GPIO, if that is possible then we
can use it as "cable" to program FPGA.

https://github.com/emard/esp32s3-jtag

Miniature ESP32S3 with USB-C are available for 7$ like
XIAO ESP32S3 MINI Mouser P/N: 713-113991114

So here is my wish/idea for openFPGALoader to support it if possible.

[emard]

openocd driver esp_usb_jtag is fixed here's my source
https://github.com/emard/openocd/blob/master/src/jtag/drivers/esp_usb_jtag.c

with this esp jtag it is very fast, speed is comparable to ft2232.

So I tried to make openfpgaloader driver
https://github.com/emard/openFPGALoader/blob/master/src/esp_usb_jtag.cpp

but it doesn't work, esp protocol is simple but has details
which I don't understant so it's not so easy =>expert needed

[trabucayre]

I have ordered one waveshare boards based on the same devices: will try as soon as I have received samples.
Thanks to point this feature!

[emard]

Great, I recommend they are so cheap and so fast. Supported should be esp32 S2, S3, C3, C6 and maybe few others. "Magic" iomux values that route jtag to external pins for s3 are undocumented but on the forums obviously I found required info :)

…

On Wed, Jan 1, 2025 at 6:30 PM Gwenhael Goavec-Merou < ***@***.***> wrote: I have ordered one waveshare boards based on the same devices: will try as soon as I have received samples. Thanks to point this feature! — Reply to this email directly, view it on GitHub <#501 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AC4K2R7WQM3N4ZWY6SWA6SL2IQQ4RAVCNFSM6AAAAABUKC25CSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRXGA4DONJYG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[trabucayre]

I have received my board, I suppose firmware is correctly flashed. I observe issues for read command. Could you provides a dump to compare?

[emard]

hardware runs this JTAG so firmware is here only to initialize registers, assigns signals to pins and wait. openocd was only that worked (after I fixed) I haven't tried any longer reads but now I can pull your source and check if I can see something strendge

…

On Sun, Jan 12, 2025 at 9:07 AM Gwenhael Goavec-Merou < ***@***.***> wrote: I have received my board, I suppose firmware is correctly flashed. I observe issues for read command. Could you provides a dump to compare? — Reply to this email directly, view it on GitHub <#501 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AC4K2R3C6MQOKVYSSI2HZY32KIPEJAVCNFSM6AAAAABUKC25CSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOBVGYZTGMRXHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[emard]

Here is compiled dump of my firmware small 14-pin S3 board pinout is on my git https://github.com/emard/esp32s3-jtag

…

On Sun, Jan 12, 2025 at 11:28 AM D EMARD ***@***.***> wrote: hardware runs this JTAG so firmware is here only to initialize registers, assigns signals to pins and wait. openocd was only that worked (after I fixed) I haven't tried any longer reads but now I can pull your source and check if I can see something strendge On Sun, Jan 12, 2025 at 9:07 AM Gwenhael Goavec-Merou < ***@***.***> wrote: > I have received my board, I suppose firmware is correctly flashed. I > observe issues for read command. Could you provides a dump to compare? > > — > Reply to this email directly, view it on GitHub > <#501 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AC4K2R3C6MQOKVYSSI2HZY32KIPEJAVCNFSM6AAAAABUKC25CSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOBVGYZTGMRXHA> > . > You are receiving this because you authored the thread.Message ID: > ***@***.***> >

[emard]

I have also some old devkit-1 ESP32S3 board https://docs.espressif.com/projects/esp-dev-kits/en/latest/esp32s3/esp32-s3-devkitc-1/user_guide_v1.0.html If you have this board and can't get it working I can compile and see what happens

…

On Sun, Jan 12, 2025 at 11:37 AM D EMARD ***@***.***> wrote: Here is compiled dump of my firmware small 14-pin S3 board pinout is on my git https://github.com/emard/esp32s3-jtag On Sun, Jan 12, 2025 at 11:28 AM D EMARD ***@***.***> wrote: > hardware runs this JTAG so firmware is here only > to initialize registers, assigns signals to pins and wait. > > openocd was only that worked (after I fixed) > I haven't tried any longer reads but now I can pull > your source and check if I can see something strendge > > > On Sun, Jan 12, 2025 at 9:07 AM Gwenhael Goavec-Merou < > ***@***.***> wrote: > >> I have received my board, I suppose firmware is correctly flashed. I >> observe issues for read command. Could you provides a dump to compare? >> >> — >> Reply to this email directly, view it on GitHub >> <#501 (comment)>, >> or unsubscribe >> <https://github.com/notifications/unsubscribe-auth/AC4K2R3C6MQOKVYSSI2HZY32KIPEJAVCNFSM6AAAAABUKC25CSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOBVGYZTGMRXHA> >> . >> You are receiving this because you authored the thread.Message ID: >> ***@***.***> >> >

[trabucayre]

Sorry for the delay to answer: no needs to prepare code/firmware. I'm now able to read IDCODE \o/.

I have to cleanup my local directly before pushing modifications in your branch. After that I see some improvements.

Thanks!

[emard]

eheheeee wonderful! One notice while I was experimienting with driver in openocd, this usb-jtag works good only if repeated commands are repeated using its special usb command for repeating RUNTEST IDLE 2 TCK 1.00E-02 SEC; here is this function used for repeating see function esp_usb_jtag_write_rlestream() https://github.com/espressif/openocd-esp32/blob/6bcb287a9deec4f9f68f6a8939b60792f55e9927/src/jtag/drivers/esp_usb_jtag.c#L482 I if replace it this CMD_REP command with simple for-loop that repeats "cmd" command "ct" number of times, then something doesn't work and usb timeout error appears.

…

On Tue, Jan 14, 2025 at 7:50 AM Gwenhael Goavec-Merou < ***@***.***> wrote: Sorry for the delay to answer: no needs to prepare code/firmware. I'm now able to read IDCODE \o/. I have to cleanup my local directly before pushing modifications in your branch. After that I see some improvements. Thanks! — Reply to this email directly, view it on GitHub <#501 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AC4K2R4OEOEWF7GLMQOEFUL2KSXS7AVCNFSM6AAAAABUKC25CSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOBZGE2TONZZGE> . You are receiving this because you authored the thread.Message ID: ***@***.***>