July 2004 - Darwin-kernel - lists.apple.com

Re: debugging kernel paniced system
by Scott Judd 26 Jul '04

26 Jul '04

Try ktrace. For example: bash$ ktrace Foo This command will execute Foo and create a binary file in the current working directory called ktrace.out. To read the file, just type kdump. This may not be what you're looking for, but it's the first thing I do when I'm tracing a kernel panic. I hope it helps you out. scott On Jul 25, 2004, at 8:22 PM, Scott Judd wrote: Try ktrace. For example: bash$ ktrace Foo This command will execute Foo and create a binary file in the current working directory called ktrace.out. To read the file, just type kdump. This may not be what you're looking for, but it's the first thing I do when I'm tracing a kernel panic. I hope it helps you out. scott On Jul 25, 2004, at 7:57 PM, Matt Jaffa wrote: Hi, I have successfully tracked down how to repeat my KEXT causing kernel panics, and I have successfully connected to it remotely to debug, but once I am connected and attached to the remote machine, what must I type/do to figure where in the code was the problem? Matt _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: debugging kernel paniced system
by Shawn Erickson 26 Jul '04

26 Jul '04

On Jul 25, 2004, at 5:57 PM, Matt Jaffa wrote: Hi, I have successfully tracked down how to repeat my KEXT causing kernel panics, and I have successfully connected to it remotely to debug, but once I am connected and attached to the remote machine, what must I type/do to figure where in the code was the problem? Not sure someone can help you much via the medium of email, it is a big complex subject... I suggest reviewing what documentation you can find on GDB and any example of using it. Have you looked over the following that Apple provides...? <http://developer.apple.com/documentation/Darwin/Conceptual/ KernelProgramming/build/chapter_18_section_5.html> <http://developer.apple.com/technotes/tn2002/tn2063.html> -Shaw _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

debugging kernel paniced system
by Matt Jaffa 26 Jul '04

26 Jul '04

Hi, I have successfully tracked down how to repeat my KEXT causing kernel panics, and I have successfully connected to it remotely to debug, but once I am connected and attached to the remote machine, what must I type/do to figure where in the code was the problem? Matt _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: How to get CPU cycles per sec?
by Jay A. Kreibich 25 Jul '04

25 Jul '04

On Sun, Jul 25, 2004 at 02:46:18PM -0700, Shawn Erickson scratched on the wall: > >>pbg3:ksh$ sysctl hw.cpufrequency > >>hw.cpufrequency = 300000000 > > > > Are there any known problems with this call? On my 800MHz TiBook, it > > always returns "hw.cpufrequency: 667000000". > > > > And, yes, it is really an 800MHz TiBook. At least that's what the > > box said (and what the bill said!). System Profiler also says > > 800MHz. But sysctl always says 667MHz. > > What is your energy save setting for CPU clock speed cycling? Sorry, I should have been more specific. When I said "always" I meant that regardless of the energy saver settings, it always returns the same value. I've played with this quite a bit trying to figure it out, since the energy saver settings were also my first guess as to why the number didn't seem to make sense. This also seems strange... does anyone know if the cpufreq value is really dependent on the energy savings settings? $ sysctl -A | grep cpufreq hw.cpufrequency: 667000000 hw.cpufrequency_min: 667000000 hw.cpufrequency_max: 667000000 -j -- Jay A. Kreibich | Integration & Software Eng. jak(a)uiuc.edu | Campus IT & Edu. Svcs. <http://www.uiuc.edu/~jak> | University of Illinois at U/C _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: How to get CPU cycles per sec?
by Shawn Erickson 25 Jul '04

25 Jul '04

On Jul 25, 2004, at 1:30 PM, Jay A. Kreibich wrote: On Sun, Jul 25, 2004 at 10:32:24PM +1000, Paul Ripke scratched on the wall: On Sunday, Jul 25, 2004, at 00:08 Australia/Sydney, Lalit Chandivade wrote: Hi As per the following discussion http://lists.apple.com/archives/darwin-kernel/2004/Mar/31/ gettingsystemperfo rmance.001.txt clock_get_uptime gives system up time in CPU cycles. Is there a way in kernel or IOKit to know the CPU cycles per second? A little digging through xnu source should show where this is coming from: pbg3:ksh$ sysctl hw.cpufrequency hw.cpufrequency = 300000000 Are there any known problems with this call? On my 800MHz TiBook, it always returns "hw.cpufrequency: 667000000". And, yes, it is really an 800MHz TiBook. At least that's what the box said (and what the bill said!). System Profiler also says 800MHz. But sysctl always says 667MHz. What is your energy save setting for CPU clock speed cycling? -Shawn _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: How to get CPU cycles per sec?
by Jay A. Kreibich 25 Jul '04

25 Jul '04

On Sun, Jul 25, 2004 at 10:32:24PM +1000, Paul Ripke scratched on the wall: > On Sunday, Jul 25, 2004, at 00:08 Australia/Sydney, Lalit Chandivade > wrote: > > >Hi > > > >As per the following discussion > >http://lists.apple.com/archives/darwin-kernel/2004/Mar/31/ > >gettingsystemperfo > >rmance.001.txt > > > >clock_get_uptime gives system up time in CPU cycles. > >Is there a way in kernel or IOKit to know the CPU cycles per second? > > A little digging through xnu source should show where this is coming > from: > > pbg3:ksh$ sysctl hw.cpufrequency > hw.cpufrequency = 300000000 Are there any known problems with this call? On my 800MHz TiBook, it always returns "hw.cpufrequency: 667000000". And, yes, it is really an 800MHz TiBook. At least that's what the box said (and what the bill said!). System Profiler also says 800MHz. But sysctl always says 667MHz. -j -- Jay A. Kreibich | Integration & Software Eng. jak(a)uiuc.edu | Campus IT & Edu. Svcs. <http://www.uiuc.edu/~jak> | University of Illinois at U/C _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: How to get CPU cycles per sec?
by Paul Ripke 25 Jul '04

25 Jul '04

On Sunday, Jul 25, 2004, at 00:08 Australia/Sydney, Lalit Chandivade wrote: Hi As per the following discussion http://lists.apple.com/archives/darwin-kernel/2004/Mar/31/ gettingsystemperfo rmance.001.txt clock_get_uptime gives system up time in CPU cycles. Is there a way in kernel or IOKit to know the CPU cycles per second? A little digging through xnu source should show where this is coming from: pbg3:ksh$ sysctl hw.cpufrequency hw.cpufrequency = 300000000 Cheers, -- Paul Ripke Unix/OpenVMS/TSM/DBA I love deadlines. I like the whooshing sound they make as they fly by. -- Douglas Adams _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: debugging kernel paniced system
by Nikita Danilov 25 Jul '04

25 Jul '04

Shawn Erickson writes: > On Jul 25, 2004, at 6:34 PM, Shawn Erickson wrote: > > > > > On Jul 25, 2004, at 5:57 PM, Matt Jaffa wrote: > > > >> Hi, > >> > >> I have successfully tracked down how to repeat my KEXT causing kernel > >> panics, and I have successfully connected to it remotely to debug, > >> but once I am connected and attached to the remote machine, what must > >> I type/do to figure where in the code was the problem? > > > > Not sure someone can help you much via the medium of email, it is a > > big complex subject... I suggest reviewing what documentation you can > > find on GDB and any example of using it. > > > > Have you looked over the following that Apple provides...? > > > > <http://developer.apple.com/documentation/Darwin/Conceptual/ > > KernelProgramming/build/chapter_18_section_5.html> > > > > <http://developer.apple.com/technotes/tn2002/tn2063.html> > > Oops... I meant to point out the use of the "trace" command in GDB is I think you meant the "backtrace" (or "bt") command that shows stack back-trace. It's usually too late to use tracepoints when kernel panicked. > likely the first step if you have it trapped during a panic in the > debugger. If you have symbols for your KEXTs loaded it should quickly > point you to the location of the issue. Additionally you can down load > the symbols for the kernel you are running [1] to be able to better > understand the Apple half of the equation. > > Also the Apple folks have a set of GDB macros to help debug IOKit, etc. > however I cannot recall how to get them at the moment. They are in kgmacros file in the top-level directory of XNU source. Load this file with "source kgmacros" (or, better yet, put this into your .gdbinit), and then "help kgm" provides an overview. By the way, I have related question. When kernel is under debugger control for enough time (for example, when I single-step through kernel code, or just inspect kernel data), fans start rotating faster and faster, emitting rather loud noise. It looks like CPU monitoring hardware is confused somehow (maybe thinks that CPU is in busy loop with interrupts off?), or just cannot work correctly without constant OS feedback. So, questions are: (a) is it possible to work around this, and (b) is it dangerous (like, can it ultimately damage fans)? > > -Shawn > Nikita. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

How to get CPU cycles per sec?
by Lalit Chandivade 24 Jul '04

24 Jul '04

Hi As per the following discussion http://lists.apple.com/archives/darwin-kernel/2004/Mar/31/gettingsystemperfo rmance.001.txt clock_get_uptime gives system up time in CPU cycles. Is there a way in kernel or IOKit to know the CPU cycles per second? Thanks, Lalit. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

RE: problem with real-time scheduling
by Michelle Munson 23 Jul '04

23 Jul '04

Continuing on this thread: I have now experimented with making both application threads non-timesharing with thread_policy_set(THREAD_EXTENDED_POLICY) but am still having similar problems. The app has a network receiving thread (I am calling the main thread), which uses select with a finite timeout, and a disk thread. Both are implemented with pthreads. The entire app is implemented using BSD APIs only. Results were that the application runs stable with a good portion of the CPU (about 40%) for 16 or 32 seconds (it seems to be a multiple of 8), and then is abruptly demoted to a small %, and never recovers. Do you think setting the network thread priority lower as you suggested will help? If so, what is the recommended way of doing this -- using pthread setschedparam API? Given the behavior, do you think I should just give up on this route and try to make the network thread explicitly yield? I thought the select call with fininte timeout should yield enough, but doesn't seem to. Thank you for any additional advice you might have. -----Original Message----- From: darwin-kernel-admin(a)lists.apple.com [mailto:darwin-kernel-admin@lists.apple.com]On Behalf Of Jim Magee Sent: Thursday, July 22, 2004 4:18 PM To: Michelle Munson Cc: darwin-kernel(a)lists.apple.com Subject: Re: problem with real-time scheduling On Jul 22, 2004, at 5:01 PM, Michelle Munson wrote: > Using the real-time API for the Mach kernel > (set_thread_priority(...)), we > are able to guarantee our application thread enough of the CPU to run > fairly > smoothly even when another process is competing. However, now the > behavior > we observe is that when our process has run smoothly for a while at > high CPU > usage (approx 70-80% for example), it will be suddenly "demoted" by the > scheduler to a very low CPU%. Piecing together background from the > kernel > programming documentation on Apple's developer site, I believe that > this > "demotion" effect happens because the scheduler automatically demotes > processes deemed to be taking too much of the CPU. > > What's surprising is that we are using "reasonable" > thread_time_constraint_policy settings. We have experimented with a > range of > values based on examples from Apple and other "real-time" applications > such > as audio and video players. We are taking a high but still reasonable > percentage of the CPU. It doesn't matter what you set for the scheduler settings. Unless you explicitly block or yield the CPU, the scheduler doesn't know that you haven't just overrun your defined computation and need more time to complete. So, it gives it to you (in fact, besides some sanity checks, those values are ignored in the current implementation of the scheduler - but will be used at some future date). It is your application/thread's behavior, and not its "supposed to work this way" scheduler definition - that determines whether it is hogging the CPU (today, that's defined as running for 8 seconds in realtime mode without ever once yielding the CPU away or blocking). > So the basic question is, is there a way that we can guarantee that our > process will not be completely stopped when competing processes launch > other > than using the real-time API? And if we need to use the real-time API, > is > there a way we can avoid the demotion effect, for example, yield to > other > processes when necessary? Because your thread is consuming tons of CPU time - and is running under a timeshare (priority adjusting) policy - it probably is degrading to just above an idle thread by the time the other process launches. It's no wonder it is then taking a big chunk of the CPU away from you - before the scheduler finally adjusts you back up to where you can fairly compete for the CPU. The method used to yield the CPU from time to time (and avoid losing your time-constraint policy setting) depends upon the programming mode of the application. If it is Carbon, WaitNextEvent(), or better yet, going back to the runloop does the trick (the latter works for Cocoa too). Using usleep() or select() from a Unix application will also do it. If you can't yield any time explicitly (or go back to the runloop), you can take the timeshare (adjusting priority) nature of your process away without boosting it into higher priorities of time-constrained threads. The typical solution is to use thread_policy_set(THREAD_EXTENDED_POLICY) setting the info->timeshare to FALSE. In addition, we have found the setting your priority lower by just a few notches helps when doing this. It allows a timeslice or two for other processes before they degrade below your priority - giving reasonable response time for them and relatively low latency for you (since you will always be near the top of the priority scale for "normal threads - and they don't stay at the top for very long if they are timeshare). --Jim [demime 0.98b removed an attachment of type application/pkcs7-signature which had a name of smime.p7s] _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0