Hardy Griech

Yes, I understand the advantage of DAP_PACKET_COUNT and I also understand that RAM operation more or less tests only USB interface and SWD speed. Concerning effects of DAP_PACKET_COUNT: I'm looking...

Erasing takes time. Heavily depending on the chip. For the nRF52840 a single 4K page needs around 85ms, while the nRF52832 requires around 6ms. Total chip erasure is faster compared...

Could someone please post one of the log files (e.g. /home/hkuhn/.local/share/probe-rs/1699213010907.log)?

Hmmm, interesting... if I'm compiling YaPicoprobe as a CMSIS-DAPv1 probe, then it seems to work: ``` hardy@ntbox:/mnt/z/mirror/audio$ probe-rs info --protocol swd Probing target via SWD ARM Chip: Debug Port: Version...

It seems that the probe-rs timeout for CMSIS commands is to small for CMSIS-DAPv2. I have released a new version of YaPicoprobe (1.20) where RTT access can be disabled via...

The other probes do not have anything like RTT access. That is more or less a unique feature of YaPicoprobe.

BTW neither pyocd nor openocd shows this timeout behavior.

Thanks for the hint. I will check my build chain if it outputs similar image sizes without the additional llc/llvm-link/opt steps. Nevertheless for completeness could you please add the three...

Ok, i have played around with link time optimization and found out that it does not equal my previous buildchain which includes llvm-link, opt and llc. It actually breaks it....

Thank you Peter for the hint. Adding ``-flto`` did the trick: lld does now LTO (and the options from above became obsolete). But... there's always a but... LTO of lld...