Problem with XP scheduler?
in [Win32 Kernel]
|
From: tanix on 18 Jan 2010 08:31 In article <ujZoeGEmKHA.5520(a)TK2MSFTNGP06.phx.gbl>, "Paul Baker [MVP, Windows Desktop Experience]" <paulrichardbaker(a)community.nospam> wrote: > >Boba, > >I don't recall my mouse events being eaten during times of high CPU usage, They are not eaten. They are misrouted from what I see. >but I suppose it's possible. What I usually see is that when whatever is >using a lot of CPU is complete, user events suddenly all take effect. Correct. This is an exact issue I am talking about. The events are synchronized and that is bad news. I suspect this is optimization for performance purposes. Sorry I have not time for this right now > This >is not necessarily a good thing for the click happy user who goes nuts >retrying or clicking something else whenever it does not respond immediately >(as tanix probably does). The good news is that it remembers the mouse >position for each message and so as long as the message handler is written >correctly, it won't think you clicked on something else. I think that was >mentioned earlier in this thread. > >Bottom line: if your computer is too slow you cannot expect it to do >everything you want perfectly. > >Paul > >"Boba" <Boba(a)somewhere.net> wrote in message >news:ezFEReplKHA.5520(a)TK2MSFTNGP06.phx.gbl... >> "Paul Baker [MVP, Windows Desktop Experience]" >> <paulrichardbaker(a)community.nospam> wrote in >> message news:uGhoWsglKHA.4408(a)TK2MSFTNGP04.phx.gbl... >> >>>... You should be happy that it buffered all those events without losing >>>them (presumably interrupt based?), if it didn't that would bad. >> | ^ >> | 6 mouse events got lost here >> | >>>...doing the processing you expect *and* taking key/mouse events promptly. >>>It just can't. Something has to give. >> | ^ >> | and at least 6 more here >> | >> d'u use oe or u'r texting while driving :) >> >> > > -- Programmer's Goldmine collections: http://preciseinfo.org Tens of thousands of code examples and expert discussions on C++, MFC, VC, ATL, STL, templates, Java, Python, Javascript, PHP, organized by major topics of language, tools, methods, techniques.
From: Paul Baker [MVP, Windows Desktop Experience] on 18 Jan 2010 08:32 tanix, You have a lot to read, my friend. I am more of a user application kind of a person myself, as you can imagine from my MVP category. But I like to know a little bit about everything, and everyone should have a decent understanding of the kernel if they are a serious Windows developer. Although I am not one myself, it is obvious that some, perhaps all, of the people who responded to you are experts in a related area. They have all answered your questions fully and patiently. If you didn't want to hear the details, you should not have followed up almost every one with an argument. If all you want to know is whether there is someone who is competent in kernel stuff here that is of the opinion that the Windows scheduler has a problem on a multi-core system, this is my take: There are always going to be differing opinions and on what an OS should do, which Microsoft and others involved had to turn into design decisions at some point. And I am sure noone would be naive enough to suggest that it is "perfect", nothing is. However, what I hear is a resounding "NO". So please be happy in the knowledge that there is nothing "wrong" with the kernel on a multi-core system and pay attention to the real cause of your issues. Paul "tanix" <tanix(a)mongo.net> wrote in message news:hiss3m$tmn$1(a)news.eternal-september.org... > In article <uRdv63slKHA.1648(a)TK2MSFTNGP05.phx.gbl>, "Alexander Grigoriev" > <alegr(a)earthlink.net> wrote: >> >>You've already heard opinions of competent kernel guys, who write kernel >>drivers and usermode code for living since Windows 3.1. You decided to >>ignore them. > > Just relax, willya? > It is not an end of the world just as yet. > >>"tanix" <tanix(a)mongo.net> wrote in message >>news:hisik1$vnd$1(a)news.eternal-september.org... >>>I would still like to hear an opinion of competent kernel guys >>> on whether there are issues with XP scheduler on a multi-core >>> system. >>> >>> What surprises me is that so far I recall seeing the system hangs >>> only with firefox. Do not recall seeing these hangs with any other >>> app so far. This new box of mine has been running for almost a month >>> now. >>> > > -- > Programmer's Goldmine collections: > > http://preciseinfo.org > > Tens of thousands of code examples and expert discussions on > C++, MFC, VC, ATL, STL, templates, Java, Python, Javascript, PHP, > organized by major topics of language, tools, methods, techniques. >
From: Paul Baker [MVP, Windows Desktop Experience] on 18 Jan 2010 08:54 tanix, We would have to discuss specific cases, but most likely the explanation is an incorrectly written application. For example, it is opening a context menu at the current mouse pointer position instead of the message position. Also, if you are not getting "mouse over" feedback, like flat toolbar buttons adding/removing a border or buttons changing color, it is possible you are clicking on something you didn't intend to. Paul "tanix" <tanix(a)mongo.net> wrote in message news:hj1nri$fee$1(a)news.eternal-september.org... > In article <ujZoeGEmKHA.5520(a)TK2MSFTNGP06.phx.gbl>, "Paul Baker [MVP, > Windows Desktop Experience]" <paulrichardbaker(a)community.nospam> wrote: >> >>Boba, >> >>I don't recall my mouse events being eaten during times of high CPU usage, > > They are not eaten. They are misrouted from what I see.
From: Boba on 18 Jan 2010 11:35 "Paul Baker [MVP, Windows Desktop Experience]" <paulrichardbaker(a)community.nospam> wrote in message news:ujZoeGEmKHA.5520(a)TK2MSFTNGP06.phx.gbl... > Boba, > > I don't recall my mouse events being eaten during times of high CPU usage, > but I suppose it's possible. i was just kidding about short words missing when typing in a hurry. have a good week!
From: Paul Baker [MVP, Windows Desktop Experience] on 19 Jan 2010 13:31
:-D Okay, in the words of tanix, "enough". You won't hear from me again (poof). Really. Paul "Boba" <Boba(a)somewhere.net> wrote in message news:u8iRKxFmKHA.2680(a)TK2MSFTNGP04.phx.gbl... > "Paul Baker [MVP, Windows Desktop Experience]" > <paulrichardbaker(a)community.nospam> wrote in message > news:ujZoeGEmKHA.5520(a)TK2MSFTNGP06.phx.gbl... >> Boba, >> >> I don't recall my mouse events being eaten during times of high CPU >> usage, but I suppose it's possible. > > i was just kidding about short words missing > when typing in a hurry. have a good week! > > |