Hi Daniel, thanks for your response! > Don’t be, I’m sorry. This came out of a SONiC-related project at work. Work > got fixed and moved on. I was originally going to keep this an upstream project > after hours, then fell off the rails. I understand, happened to me as well. Is it okay to CC you on patches? Or are you too far away from it meanwhile? > As for v2 — I still believe the direction is good. The main issue I > had was that I lacked insight into one or more popular-enough > commodity bus drivers (amd? nvidia? intel?), for further verification, > as opposed to our proprietary accels around here. Real HW setups for testing are a problem, yes. I noticed this, too, when I said I fix the existing drivers. I reconsidered. Fixing them would result in non-trivial changes which need testing SMBUS_BLOCK_DATA transfers. This is too rare, so I'll just focus on the systems I have at hand. Digging in deeper, I found there is enough to fix at the core level anyhow. I won't be bored. > That 1->2 succession you outline below, starting with kernel and kernel-clients, > sounds a lot like what I was missing. Yeah, I hope it grows the pool of interested people. All the best, Wolfram
Attachment:
signature.asc
Description: PGP signature
