Date: Sat, 2 Dec 1995 16:05:57 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: dyson@freefall.freebsd.org (John Dyson) Cc: terry@lambert.org, alexandr@louie.udel.edu, freebsd-questions@freebsd.org Subject: Re: Help!!!! Message-ID: <199512022305.QAA06536@phaeton.artisoft.com> In-Reply-To: <199512020314.TAA14907@freefall.freebsd.org> from "John Dyson" at Dec 1, 95 07:14:29 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > I don't think you can do anything for now. Apparently some patches for > > async mounts adversely affected non-async mounts. > > Please Terry, if you review the code changes for the Async operation, > and apparently you did not -- you will find that they are totally > benign (for sync mounts) -- in fact they are not fully implemented yet. I > respectfully request that you review the target of your criticism. The > problem lies elsewhere and it is still being researched. I apologise > in advance for the tone of my statements, but time is/has been wasted > by assertions as above that have not been carefully reviewed. > > If you are astute enough to confidently make the above statement, then > I suggest that a solution would be just as easy.... At least, as soon > as I have an analysis -- I usually have a solution... First, I said "apparently" intentionally. It is not my claim that the patches broke anything. It is the result of some 20-30 messages following the date the patches were committed complaining about erratic behaviour. Check the -hackers and -current list archives. I *know* the patches didn't affect the non-async case directly, having looked at them. If the circumstantial relationship were not so widely supported, or if I could say I haven't personally had the problems myself (and fixed them by reverting the code to the pre-"fix" code), I would have said nothing. My problem with this whole thing is that the "async 'fixes'" appear to be related to the problem (who knows? It might be tickling a dormant compiler bug, etc.), and are really not rationally justified except as a method of putting some bogus Linux benchmarks to rest. They were (and remain) reactive to what I believe to be a non-problem, at best, and they are circumstantially related to a large number of problems that didn't exist prior to their commit date, at worst. I'm not drawing peoples conclusions for them, it's all there in the list archives for anyone to see. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199512022305.QAA06536>