Amaury Pouly
Amaury Pouly
For what it is worth, I don't think it is necessary to read 1024 samples in the ISR. It seems perfectly fine to me to read, for example, 512 samples...
This is why the reading and writing paths need to be asynchronous in my opinion: there is one big 1024-sample buffer. The reading ISR regularly reads samples from the observe...
> Just so I'm understanding, are you saying to have the read ISR push small amounts of data into an in-memory buffer so that the observe FIFO doesn't overflow, then...
> LGTM, modulo whether this distribution doesn't lead to one excessively long job. > > If the execute_sival_fpga_tests_cw310 job becomes too large, it might make sense to put the cw310_sival_rom_ext...
I followed your suggestion.
> Looks like CW310 SiVal Tests is too long now. We might want to consider defaulting to ROM_EXT tests only, then only adding ROM-only environments for the tests that require...
I figured out a way to do it in a single query. I agree with @cfrantz that we should be running all of them in the nightly. I will update...
I managed to rewrite it using a single query (see the abomination in the diff). There are some failures which I suspect come from the fact that some tests have...
> I looks at a few failing tests, and it appears they are all failing at setting up keymgr? I don't think that's allowed in ROM ext. I'd suggest we...
So the situation is a bit worse than expected: it turns out that some test like `flash_ctrl_mem_protection_test` were never run in the CI SiVal to start with. This particular test...