<sb0>
marmelada, you should install the openocd version that goes with your artiq version (see conda/artiq-dev/meta.yaml), not (necessarily) the latest
<marmelada>
sb0: I do have it
<marmelada>
in meta.yaml: - openocd 0.10.0 6
<sb0>
are you loading the correct and up-to-date proxy bitstream?
<marmelada>
yes, everything was built today
<marmelada>
I run artiq_flash --srcbuild ~/artiq-dev/artiq/artiq_sayma -t sayma
<marmelada>
I also verified that all files in artiq_sayma are modified today
rohitksingh has joined #m-labs
<whitequark>
1.8V bug?
<marmelada>
nope, 1v8 led works
<whitequark>
IIRC that doesn't help diagnose it
<whitequark>
but I'm not sure
<whitequark>
try reapplying power to sayma and quickly flashing it
<marmelada>
didn't help
X-Scale has joined #m-labs
<sb0>
the FPGAs are detected correctly, so it's not 1v8 bug
<sb0>
it looks like the incorrect proxy bitstream is getting loaded
<marmelada>
flash device id is changing from run to run
<sb0>
marmelada, the proxy bitstream is not part of artiq
<cr1901_modern>
Error looks like the "old style" proxy bitstream was loaded
<cr1901_modern>
but the openocd expects a "new style"
<marmelada>
"The following packages will be SUPERSEDED by a higher-priority channel: bscan-spi-bitstreams: 0.10.0-2 m-labs/label/main --> 0.10.0-2 m-labs"
<marmelada>
seems like it'll install same version
<cr1901_modern>
Any chance to do a diff of the proxy bitstream openocd actually loaded and the package on conda?
rohitksingh has quit [Quit: Leaving.]
<marmelada>
can we continue on github #913? I ahve to leave now
<GitHub-m-labs>
[artiq] jordens commented on issue #913: Iirc you didn't have this problem before. That code hasn't changed over the last months. What's different? Is that a new board? Has it worked before? What's your bscan-spi-bitstreams version? https://github.com/m-labs/artiq/issues/913#issuecomment-378997787
<hartytp>
marmelada: what's the plan for the EEPROMS on the EEMs/Kasli when they are produced by CreoTech?
<hartytp>
Will they be programmed? If so, have you discussed the header format with M-Labs? Would be great to have board type, hw-rev and serial number there.
<hartytp>
Thinking ahead a little bit to things like generating gateware by scanning the i2C tree to see what devices are connected. Or, checking on startup that we have the right devices connected to our Kasli (e.g. we haven't loaded the wrong bitstream(
<hartytp>
would be nice if we could do all that without having to program the EEPROM ourselves
hartytp has quit [Client Quit]
mumptai has quit [Remote host closed the connection]