![]() ![]() Toughbook series) and see this bug or one almost exactly like it on those Tool and we run it on Intel Core Mobile systems (Sony UX series, Panasonic "Now, as far as this bug being AMD64 only. These bugs have been outstanding about a year now!! And it's not just AMD64 anymore either. The kernel folks are missing the obvious and utter brokenness of the IO scheduling. Mathematically", the method of resolution is quite different, if you think The concepts are related, but they are definitively not "the same thing To calculate some of the key CFS variables for the debug output. The basic concept is quite different enough, one can e.g. Wild claims tell me something completely different.īTW who is "we" and how is it possible that this meta mind can come to That, which would mean you actually understood what I wrote, OTOH your ![]() :-(Įspecially interesting is that you don't need to ask a single question for > utilization for a synthetic corner-case)Įverytime I'm amazed how quickly you get to your judgements. > mathematically, expressed without a "/weight" divisor, resulting in no ![]() > something fundamentally different though - it's just the same thing We do disagree with this being positioned as > All in one, we dont disagree, this is an incremental improvement we are So I'm really amazed how you can make such claims. Peter's patches don't remove limit_wait_runtime() and AFAICT they can't, Rounding has been improved lately, so it's not as easy to trigger with I didn't say that it's a "common problem", it's a conceptual problem. Peter's queue worksĭid you even try to understand what I wrote? > any synthetic corner-case we could come up with. > with Peter's queue there are no underflows/overflows either anymore in > all - unless there's some testcase that proves the opposite. ![]() > of thousand of context switches per second with a yield-ing loop to be > scheduling and an artificially constructed workload: several hundreds > negative nice levels and even there it needs a very high rate of > problem, while in practice it's only a corner-case limited to extreme > ( your characterisation errs in that it makes it appear to be a common > over time, thus there are no more underflow/overflow anymore within > This model is far more accurate than CFS is and doesn't add an error Mentioning in an earlier mail, but the value is still rounded. It addresses the conversion error between the different units I was Nothing about the rounding error I was talking about. > normalizing/denormalizing overhead and rounding error.Īctually it changes wait-runtime to a normalized value and it changes > 'normalized' to 'non-normalized' wait-runtime, to avoid the > Peter's patches change the CFS calculations gradually over from > already for CFS (almost a month ago), in a finegrained way. So the most intrusive (math) aspects of your patch have been implemented ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
January 2023
Categories |