On Mon, Nov 02, 2015 at 09:03:35PM -0500, Doug Ledford wrote: > No, one kernel consumer that never worked on iWARP before now works on a > different iWARP controller but doesn't work on the old iWARP controller. > Hardly the end of the world.
NFS is gone/going as well, and that used to work. No iWarp enabled ULP (now or in future) will work without at least FMRs.. I just disagree we should continue to hang on to unused hardware that isn't even capable of running the whole stack we are maintaining. That makes no sense to me. If it doesn't run our full stack then ditch it. > > *WHY* spend an ounce of time fixing up code that *NO-ONE* will ever > > even run? Wasted effort. Just delete them now. > > You don't just "delete them now" because it denies anyone advanced > warning of the change and denies them the chance to speak up in > opposition to the driver's removal. We've already gone through a cycle, 'now' is a fine time to do it. You are talking like deletion is irreversible. It isn't. If someone comes forward with a need and a desire to test that it still works then reverting a delete shouldn't be that difficult. On the other hand, the developer time wasted keeping it working while we wait is not recoverable. > distros proactively queried their customer's hardware usage. But > without either a query activity or a push back cycle, your assertion > that "*NO-ONE* will ever even run?" is mere assumption not a statement > of fact. No, it is a statement based on my knowledge of the history of these cards and our user base, combined with the responses from the vendors who do have a better sense. Ammasso went out of buisness before the driver was even in mainline. Tom got it in mainline after the fact as a early way to work on generic iWarp support. There were few users (<<100) and similarly few cards. As far as I'm concerned all users got the 'obsolete, do not use' message 10 years ago. ipath only supports the HTX cards, which where a short run engineering experiment that were not market successful (10 years ago), few were made. Those cards are unsuable in modern hardware. The vendor announced they were no longer supported when the qib driver was made, as far as I'm concerned all users got the 'obsolete, do no use' message 5 years ago. ehca was part of IBM's high end systems. A >= 4.4 kernel isn't going to be qualified or tested on that hardware. Nobody would run production like that. I actually bet some are still running somewhere.. Seriously, there are zero production users, it isn't even a question. If a hobby user appears, then they can restore the driver and fix it up, and explain why we should carry it when it doesn't run the full stack. > I suspect you are right and we *could* delete them now without a > problem. But without that push back cycle or a query to find out for > sure, you're asking me to take it on faith instead of doing due > diligence, and I don't have any intention of being called out later for > failing to do due diligence. Nobody is going to call you out for removing a driver, especially when the *vendor* said it was OK. On the other hand, people have been sending unhappy messages with the current arrangement... kernel.org isn't a distro - we can flip flop ... Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
