Problem with XP scheduler?
in [Win32 Kernel]
|
From: tanix on 19 Dec 2009 14:50 In article <OBp#vYNgKHA.6096(a)TK2MSFTNGP02.phx.gbl>, "Alexander Grigoriev" <alegr(a)earthlink.net> wrote: >If you run the process under debugger, then Pause/Break (and also F12) will >break into it. >You can also attach a remote debugger to it. Thanx. Well, I am not sure I'd even want to go into investigation of THIS depth and spend days on it as a result. I don't feel like spending that much time and effort. At this point, what I am looking for is for input from other people and see if someone is either aware of the issue or have some idea/suggestions. From my estimates, even if I do find the exact reason, no one is going to bother fixing it because everybody is into win 7 nowadays. XP is non priority item I bet. So, does it worth the bother to try to run firefox under debugger control? >"tanix" <tanix(a)mongo.net> wrote in message >news:hghrt5$4fe$2(a)news.eternal-september.org... >> In article <371FB587-D396-4499-AEF4-C7A12514B925(a)microsoft.com>, "Pavel >> A." <pavel_a(a)12fastmail34.fm> wrote: >>>If you can easily repro this, try to setup a ctrl/scroll dump as described >>>here: >>>http://msdn.microsoft.com/en-us/library/cc266483.aspx >>>and get a full dump. Or connect windbg to debug it on the site. >> >> Well, from what I see, it does not look like a system crash. >> It is simply some bug in the code getting into some dead loop >> and tying up the processor. >> >> Secondly, since I loose the keyboard, there is nothing I can do >> from what I see to cause any dumps. I don't even know at which >> exact point the box is going to hang. >> >>>That it does not occur on win7 looks like a good thing :) >> >> Well, depends. If it IS the XP scheduler or something of a kind, >> it is one thing. If it is firefox proper, it is another thing. >> If it is firefox bad plugings, it is a totally different thing. >> Except I am still having a hard time comprehending how firefox >> can lock the mouse, unless it runs in a kernel mode. >> Mouse is a highest priority interrrupt at least for things we >> are talking about here. >> >>>--pa >>> >>>"tanix" <tanix(a)mongo.net> wrote in message >>>news:hgf04f$33v$2(a)news.eternal-september.org... >>>> In article <#yIP7T5fKHA.1824(a)TK2MSFTNGP04.phx.gbl>, "Alexander >>>> Grigoriev" >>>> <alegr(a)earthlink.net> wrote: >>>>>I'd say it's a problem with Firefox. Send a bug to them. >>>> >>>> Well, I did post to a forum and its been 2 days with no answer. >>>> I am not sure you can file bugs as a user. >>>> >>>> Secondly, what I do not like about this is that it runs fine under win 7 >>>> on the same box. >>>> >>>> Secondly, I do not see how it could possibly lock up the entire box. >>>> Does it have some kernel mode drivers? >>>> >>>> This looks like a software bug. Some dead loop somewhere. >>>> Except I don't even see how to figure out what is doing it >>>> since the whole box is frozen. >>>> >>>>>"tanix" <tanix(a)mongo.net> wrote in message >>>>>news:hgdirp$t0o$4(a)news.eternal-september.org... >>>>>>I am having firefox 3.5.5 and 3.5.6 freeze my box when I update >>>>>> several tabs in a rapid succession on Win XP SP3. >>>>>> >>>>>> What happens is box totally freezes. Task manager -> >>>>>> Performance tab shows one of cores going up to 100% CPU time, >>>>>> mouse freezes, keyboard does not work (not even ctl-alt-del) >>>>>> and disk stops flashing every few seconds as it normally does >>>>>> under XP. >>>>>> >>>>>> Interestingly enough, it does not happen under Windows 7 on >>>>>> the same box no matter how hard I try. >>>>>> >>>>>> Some say it is a problem with crappy XP scheduler. >>>>>> Is there anything I can do under XP to fix this issue? >>>>>> >>>>>> Hardware: >>>>>> Asus M4A78 motherboard, >>>>>> Athlon 2, X4 620 CPU, >>>>>> 4 Gigs of DDR2 ram by Team Extreme. >>>>>> Running on Windows XP SP3. >> >> -- >> 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, >> organized by major topics of language, tools, methods, techniques. >> > > -- 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, organized by major topics of language, tools, methods, techniques.
From: Pavel A. on 19 Dec 2009 16:23 Then, let me explain this again. Since the system becomes not responsible, it looks like a kernel mode problem. Try to obtain a crash dump as per this KB article http://msdn.microsoft.com/en-us/library/cc266483.aspx Note that these instructions work only with PS/2 keyboard (not USB). This is not because the system crashes. This is so that you could show the dump to someone who understands. Instead of getting a dump, you could break into the machine with a kernel debugger, but this is more complex. This NG is not a good place to ask Firefox questions. It is not a MS product. If you get evidence (dumps, traces) on your hands, somebody from the Firefox fans may be able to help further. Good luck, --pa "tanix" <tanix(a)mongo.net> wrote in message news:hghsd8$4fe$3(a)news.eternal-september.org... > In article <edhcj0EgKHA.5792(a)TK2MSFTNGP05.phx.gbl>, "m" <m(a)b.c> wrote: >>What process is using that CPU time? > > Well, I tried to use the task manager -> processes tab to see if > it displays me which process is eating up the cpu time. I was hoping > to see the firefox doing it. > > But the problem is. Once it locks up, it locks up so bad, that even > task manager -> processes is not updated. So you just sit there zombified. > >> If it is not directly ascribable to a >>process, you can bet it is a driver causing the problem. Most likely, the >>difference in Windows 7 has nothing to do with Firefox, but a better >>driver >>for some piece of HW in your system. > > Well, it is not clear it has anything to do with any kind of driver. > What kind of driver in firefox can possibly do that? > The network driver is fine cause I am running some pretty heavy duty > processes that do multi-threaded downloads from the network at a peak > rate. > The display driver is fine cause I don't have any kind of problem with > any other app so far. > The disk drivers are obviosly fine. > I just downloaded the latest version of BIOS. So, I doubt that is the > problem. > > I just can not think what kind of driver could be doing it. > I have no clue what kind of drivers firefox might use for example for > optimizing the rendering performance. > > So, it is a wild goose chase. > > The latest thing I did is to disable all firefox addons exept of imacros. > So far, the box behaves MUCH better. In fact, I have not seen a single > hang since and i tried to update 6 of my firefox tabs in a rapid > succession for 3 times and there is no problems so far. > > Too early to say though. I need to run it for few days to see if it > pops up again. > >>You can test this theory by altering the HW config and trying to reproduce >>the problem. BIOS settings can have important impact on driver and IO >>path >>selection too. > >>"tanix" <tanix(a)mongo.net> wrote in message >>news:hgdirp$t0o$4(a)news.eternal-september.org... >>> I am having firefox 3.5.5 and 3.5.6 freeze my box when I update >>> several tabs in a rapid succession on Win XP SP3. >>> >>> What happens is box totally freezes. Task manager -> >>> Performance tab shows one of cores going up to 100% CPU time, >>> mouse freezes, keyboard does not work (not even ctl-alt-del) >>> and disk stops flashing every few seconds as it normally does >>> under XP. >>> >>> Interestingly enough, it does not happen under Windows 7 on >>> the same box no matter how hard I try. >>> >>> Some say it is a problem with crappy XP scheduler. >>> Is there anything I can do under XP to fix this issue? >>> >>> Hardware: >>> Asus M4A78 motherboard, >>> Athlon 2, X4 620 CPU, >>> 4 Gigs of DDR2 ram by Team Extreme. >>> Running on Windows XP SP3. >
From: tanix on 19 Dec 2009 19:00 In article <81878418-E87A-4C50-B6BA-EDBE053B03E2(a)microsoft.com>, "Pavel A." <pavel_a(a)12fastmail34.fm> wrote: >Then, let me explain this again. >Since the system becomes not responsible, it looks like a kernel mode >problem. Yep. That is what I thought first. Except Windows is not trully multi-tasking system and I saw plenty of times the system is freezing because of some not well behaved program so far. But the MOUSE? Well, that one is a stretch. Because mouse is one of the top priority items in Windows. >Try to obtain a crash dump as per this KB article >http://msdn.microsoft.com/en-us/library/cc266483.aspx >Note that these instructions work only with PS/2 keyboard (not USB). > >This is not because the system crashes. This is so that you could show the >dump >to someone who understands. >Instead of getting a dump, you could break into the machine with a kernel >debugger, >but this is more complex. > >This NG is not a good place to ask Firefox questions. It is not a MS >product. Well, except I am not sure what the issue is. As far as I can see, no application program should lock your box no matter what it does. That is why I asked if someone might know about this issue here. I know what this group is all about. Not to worry. >If you get evidence (dumps, traces) on your hands, somebody from the Firefox >fans may be able to help further. > >Good luck, >--pa > >"tanix" <tanix(a)mongo.net> wrote in message >news:hghsd8$4fe$3(a)news.eternal-september.org... >> In article <edhcj0EgKHA.5792(a)TK2MSFTNGP05.phx.gbl>, "m" <m(a)b.c> wrote: >>>What process is using that CPU time? >> >> Well, I tried to use the task manager -> processes tab to see if >> it displays me which process is eating up the cpu time. I was hoping >> to see the firefox doing it. >> >> But the problem is. Once it locks up, it locks up so bad, that even >> task manager -> processes is not updated. So you just sit there zombified. >> >>> If it is not directly ascribable to a >>>process, you can bet it is a driver causing the problem. Most likely, the >>>difference in Windows 7 has nothing to do with Firefox, but a better >>>driver >>>for some piece of HW in your system. >> >> Well, it is not clear it has anything to do with any kind of driver. >> What kind of driver in firefox can possibly do that? >> The network driver is fine cause I am running some pretty heavy duty >> processes that do multi-threaded downloads from the network at a peak >> rate. >> The display driver is fine cause I don't have any kind of problem with >> any other app so far. >> The disk drivers are obviosly fine. >> I just downloaded the latest version of BIOS. So, I doubt that is the >> problem. >> >> I just can not think what kind of driver could be doing it. >> I have no clue what kind of drivers firefox might use for example for >> optimizing the rendering performance. >> >> So, it is a wild goose chase. >> >> The latest thing I did is to disable all firefox addons exept of imacros. >> So far, the box behaves MUCH better. In fact, I have not seen a single >> hang since and i tried to update 6 of my firefox tabs in a rapid >> succession for 3 times and there is no problems so far. >> >> Too early to say though. I need to run it for few days to see if it >> pops up again. >> >>>You can test this theory by altering the HW config and trying to reproduce >>>the problem. BIOS settings can have important impact on driver and IO >>>path >>>selection too. >> >>>"tanix" <tanix(a)mongo.net> wrote in message >>>news:hgdirp$t0o$4(a)news.eternal-september.org... >>>> I am having firefox 3.5.5 and 3.5.6 freeze my box when I update >>>> several tabs in a rapid succession on Win XP SP3. >>>> >>>> What happens is box totally freezes. Task manager -> >>>> Performance tab shows one of cores going up to 100% CPU time, >>>> mouse freezes, keyboard does not work (not even ctl-alt-del) >>>> and disk stops flashing every few seconds as it normally does >>>> under XP. >>>> >>>> Interestingly enough, it does not happen under Windows 7 on >>>> the same box no matter how hard I try. >>>> >>>> Some say it is a problem with crappy XP scheduler. >>>> Is there anything I can do under XP to fix this issue? >>>> >>>> Hardware: >>>> Asus M4A78 motherboard, >>>> Athlon 2, X4 620 CPU, >>>> 4 Gigs of DDR2 ram by Team Extreme. >>>> Running on Windows XP SP3. -- Programmer's Goldmine collections: http://preciseinfo.org Tens of thousands of code examples and expert discussions on C++, MFC, VC, ATL, STL, templates, Java, Python, Javascript, organized by major topics of language, tools, methods, techniques.
From: Pavel A. on 19 Dec 2009 22:23 "tanix" <tanix(a)mongo.net> wrote in message news:hgjpef$el6$3(a)news.eternal-september.org... > In article <81878418-E87A-4C50-B6BA-EDBE053B03E2(a)microsoft.com>, "Pavel > A." <pavel_a(a)12fastmail34.fm> wrote: >>Then, let me explain this again. >>Since the system becomes not responsible, it looks like a kernel mode >>problem. > > Yep. That is what I thought first. > Except Windows is not trully multi-tasking system > and I saw plenty of times the system is freezing > because of some not well behaved program so far. Understood. Yes, this should not normally occur, even on Windows. Yet this can occur because of bugs in certain video drivers & BIOS, various utilities that install hooks in various places, malware and anti=malware. Windows still is a true multitasking system (unless abused). Regards, --pa
From: tanix on 20 Dec 2009 06:18
In article <DAB9CB7C-E4B5-4FC2-90DC-90736DF1D9F6(a)microsoft.com>, "Pavel A." <pavel_a(a)12fastmail34.fm> wrote: >"tanix" <tanix(a)mongo.net> wrote in message >news:hgjpef$el6$3(a)news.eternal-september.org... >> In article <81878418-E87A-4C50-B6BA-EDBE053B03E2(a)microsoft.com>, "Pavel >> A." <pavel_a(a)12fastmail34.fm> wrote: >>>Then, let me explain this again. >>>Since the system becomes not responsible, it looks like a kernel mode >>>problem. >> >> Yep. That is what I thought first. >> Except Windows is not trully multi-tasking system >> and I saw plenty of times the system is freezing >> because of some not well behaved program so far. > >Understood. Yes, this should not normally occur, even on Windows. >Yet this can occur because of bugs in certain video drivers & BIOS, >various utilities that install hooks in various places, malware and >anti=malware. That is why I asked: is there a low level kernel mode driver in firefox? Cause that is about the only thing I can think of that can possibly cause this kind of a behavior. To freeze the kernel so bad that it even stops executing its normal disk activity as you disk stops flashing the disk access light? I really do not see what can possibly be the reason for the app to cause such a behavior, unless firefox does some fast rendering via their own kernel mode driver. >Windows still is a true multitasking system (unless abused). Well, that is not my experience. But let is slide for now. I could care less. It is what it is and I have to live with it even though Linux Ubuntu seems MUCH more pleasant and stable to me. Plus the very fact that you do not have security related issues, such as trojans and viruses, is becoming one of the prime criterias for me. Plus the ability to have as many desktops as you like is also a big plus for me. Plus the ability to install just about anything I possibly want without all the hassle and without worrying about security issues is also a big plus. Even such issues as text editors are much better on Linux then on windows, even though notepad++ is quite a nice editor. This is one of the things I care about quite a bit. Plus the ability to do all sorts of installs on line and not needing to reboot the box, is quite a plus for me. Plus knowing that Linux IS in fact a true multitasking system and that no app, no matter what, can possibly freeze your box or make your system non responsive as I saw WAY too many times on windows, counts for me. Probably about the only reason I am forced to stay with win for now is very poor performance of the JVM (Java Virtual Machine), which translates in my case in doubling the time of processing some big job I regularly do that takes hours if not days to complete. Another thing is GUI is not as fine grained as Windows and it looks a bit bulky. Beyond that, I'd be willing to swith to Linux entirely and forget all these windows horror trips. Cause they do get under my skin to the point that just one more thing may make me switch to Linux forever. >Regards, >--pa -- 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, organized by major topics of language, tools, methods, techniques. |