04:16 <avis> It happens that I do not have enough documentation to properly write the full protocol; page programming also doesn't guarantee the execution of the bitfile on the board itself. Please dismiss the above question, thanks

11:16 <Fleck> https://pastebin.com/hNE0Ns4V << I have this config, but upload speed seems to be still 2MHz, from this line: Info : clock speed 2000 kHz

11:20 <karlp> you're probably hitting this sort of thing: https://review.openocd.org/gitweb?p=openocd.git;a=blob;f=tcl/target/stm32f4x.cfg;h=35d8275b553be1cc602a6668fb9262a9399b31b8;hb=refs/heads/master#l69

11:20 <karlp> but you should be getting line 142 as well?

11:23 <Fleck> ghmm, ok, added adapter speed line AFTER target, have now: Info : clock speed 4000 kHz

11:23 <Fleck> doesn't seem to affect upload speed tho :D

11:23 <Fleck> but I guess it's OK

11:24 <PaulFertser> Fleck: how are you uploading with what command?

11:25 <Fleck> openocd -f openocd_stm32f4xx.cfg -c "program build/project.elf verify reset exit"

11:25 <PaulFertser> Hm that's correct, it should be using reset init before actual flashing.

11:26 <Fleck> I tried adapter speed 500, same speed as 4000 :D

11:26 <PaulFertser> And reset init should give 8000.

11:26 <Haohmaru> what debugger is that?

11:26 <Fleck> aliexpress J-Link

11:26 <Haohmaru> could it just be shrugging on the given value?

11:26 <Fleck> but works fast in Ozone

11:26 <PaulFertser> Fleck: please pastebin result of running that command you use to flash with -d3 added.

11:27 <Fleck> Yes sir, stand by!

11:28 <Haohmaru> btw, PaulFertser, when the "program" (e.g. openocd) tells my debugger to switch to <xyz> Hz, i guess i should round down to the closest speed i can do, right?

11:28 <Fleck> wow, thats gonna need hosting service to paste :DDD

11:28 <Haohmaru> huhuhu

11:28 <Fleck> You have exceeded the maximum size of 512 kilobytes per Paste.

11:29 <Haohmaru> toss it in a .zip

11:29 <Fleck> pastebin knows .zip?

11:29 <PaulFertser> Fleck: don't need all of that, just first lines before it starts "flash write_image"

11:30 <Fleck> ok

11:30 <Haohmaru> ^

11:30 <PaulFertser> I mean between 0 and actual flash write_image execution

11:30 <Fleck> https://pastebin.com/9fQAZY28

11:32 <PaulFertser> Fleck: line 434 tells it's trying to boost the freq there.

11:32 <PaulFertser> So that should work...

11:35 <karlp> ozone does some smarter flashing that openocd though, it does a bunch of checks to only flash what's cahnged,

11:35 <Fleck> ohh I see

11:35 <karlp> speed of interface is unrelated to that.

11:36 <PaulFertser> Huh, that's smart.

11:37 <karlp> with tweaking workareasize and adapterspeed, on one project here, I could get oocd up to about 100=K/sec, but jlinkgdbserver could still get to 1500K/sec, with same hardware and connections

11:38 <karlp> (the out of the box configs for openocd were only ~40K/sec for comparison)

11:38 <karlp> here, make your own inferences: https://paste.centos.org/view/13c3393e

11:44 <Fleck> thanks

16:27 <antto> well, there it is: https://antonsavov.net/tmp/samd21_daplink.pdf https://i.imgur.com/nkXGVQM.png

16:30 <antto> with muchos test points

16:53 <borneoa___> Fleck: missing "-c init" before "-c program..." ?

16:57 <PaulFertser> borneoa___: but program runs init on its own?

17:09 <borneoa___> PaulFertser: you are right. I never noticed that

17:12 <PaulFertser> :)