March 2006 - Darwin-kernel - lists.apple.com

Re: Commpage, Copy on Write
by Shashi Rao 19 Mar '06

19 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mutt/1.5.11 * Mike Smith <drivers(a)mu.org> [2006-03-18 13:23:32-0800]: | | On Mar 18, 2006, at 12:41 PM, Shashi Rao wrote: | | >Why can't I write to the commpage on my G5 running Tiger? Does it | >not have COW semantics? | | Obviously, no. | | Communication via the commpage is one-way (kernel -> global userspace). How do I breakpoint a child process in the commpage then? How does gdb get away with it? (Does it?) Is there a reason the kernel engineers decided to not allow COW on these pages? Thanks, Shashi -- Shashi Rao _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Commpage, Copy on Write
by Mike Smith 18 Mar '06

18 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 18, 2006, at 12:41 PM, Shashi Rao wrote: Why can't I write to the commpage on my G5 running Tiger? Does it not have COW semantics? Obviously, no. Communication via the commpage is one-way (kernel -> global userspace). = Mike _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Commpage, Copy on Write
by Shashi Rao 18 Mar '06

18 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mutt/1.5.11 Hello, Why can't I write to the commpage on my G5 running Tiger? Does it not have COW semantics? Thanks, Shashi -- Shashi Rao _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: FS KPI vnop_blockmap
by Terry Lambert 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 17, 2006, at 1:54 PM, Ron Aldrich wrote: -- Terry _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Is it legal for a filesystem's blockmap vnode operation to be called for a file which hasn't been opened? I'm working on some optimizations, and need to know whether I should return an error in this case, or just handle it silently. If your FS has vended a vp, then your FS must honor operations on that vp until the vp has been vrele's (the vended vp reference is returned to the FS normally) or the vp has been deadfs'ed (e.g. by a forced unmount or other operation, in which case it's not your FS's vnode any more). Some FS's like to pretend that this isn't an issue for them, and that the FS's idea of "open" and the VFS layer's idea of open must somehow be congruent. An example of this is SMB client, which likes to have the file "open" on the server in order to be able to read it. In the case of execve(), we use namei() to get the FS to vend a vp for a path, and then explicitly map/read pages out of the vp, without ever implementing a formal open notification to the FS. There are several good reasons for permitting this type of access; in particular, for intermittent access, the open/close operations can be implied. This allows an FS that's handicapped (e.g. like SMB, which is incapable of deleting an open file) to more closely echo the expected UNIX FS semantics. The alternative to this would be to return ETXTBSY (for example) on executable files on FS's where the operations could not be accomplished - IFF the FS was notifid that an executing file was in fact "open" from the FS's perspective. This is problematic, since it would mean that some FS's would exhibit the error code, while other FS's would not exhibit the error code. To accomplish this, there would have to be a specific way of internally marking a vnode. Further, this would constitute a KPI change, since it's impossible to update existing programs that expect to be able to delete/rename open files. Even more problematic is that updates of running systems could no longer be accomplished, if this was made a requirement, or made universal to all FS's by implementing it without a KPI change at the VFS layer (consider trying to update "tar" while you are using "tar" to extract the new "tar" from a tar archive, or attempting to update "launchd" without killing "launchd" and effectively shutting down the OS to do it, etc.). The inability to rename/replace/delete running programs is one of the reasons why installing software on Windows requires a system reboot, and why, if you hit <esc> to get the text console, instead of things being hidden behind the splash screen, you see autoexec.bat effectively running a postinstall script in single user mode before the system comes back up (.bat files are tokenized into memory by command.com, so they are immune from this requirement: they are treated as cached data). So in general, you will want to honor all vnop requests on vnodes vended by your FS implementation to avoid behaving differently from other FS's, or damaging the expected UNIX FS semantics, except where absolutely necessary due to FS design limitations. This email sent to site_archiver(a)lists.apple.com

1 0

Re: how to get thread CPU time
by Terry Lambert 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 17, 2006, at 7:53 AM, info(a)yourkit.com wrote: Hello, is there any way to get time which particular thread is being executed on CPU. In other words how many milliseconds (or nonoseconds) of CPU time was consumed by thread since it has been started. I'm writing CPU profiler for Mac OS X and thread CPU times is very important for me. Please advise. -- Terry _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… IAIK, in Linux there is a clock_gettime(CLOCK_THREAD_CPUTIME_ID) system call (and Windows has similar API) but I cannot find any analog for Darwin. The best way is to do it by observation. This is what chud and shark do. Formal resource utilization accounting is only ever done on a per process or per task basis, not a per thread basis. If you want absolute values, there are two timers that are started for system and user time when threads are started that will provide this information from the scheduler. However, in order to access them, you will need to write a kext, and you will need to explicitly link against the mach_kernel instead of against one of the symbol sets, which will require that you recompile your application for each point release of the OS (e.g. 10.4.3->10.4.4, etc.). The closest hint I'm willing to give you is "look at the implementation of task_info()". How you link in such a way as to make your application not run except on a specific OS version has been discussed before (you will need to look at the list archives, or combine the results of two knowledge base articles from developer.apple.com) and for the identity of the timers themselves and how they are accessed, you will need to look at the source code on opendarwin.org. Basically, if you do this, you are on your own, since it touches on details of the microarchitecture, such as the relationship between tasks and processes, threads and uthreads, and how the scheduler works at the lowest level. We neither document nor guarantee not to change these things from point release to point release of the OS, and we steer internal tools developers away from doing this sort of thing as well. This email sent to site_archiver(a)lists.apple.com

1 0

FS KPI vnop_blockmap
by Ron Aldrich 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thanks, Ron Aldrich Software Architects, Inc. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Is it legal for a filesystem's blockmap vnode operation to be called for a file which hasn't been opened? I'm working on some optimizations, and need to know whether I should return an error in this case, or just handle it silently. This email sent to site_archiver(a)lists.apple.com

1 0

Re: NewWorld CD Booting?
by Dave Carlton 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: ifnet_output_raw() frees all packets?
by Vincent Lubet 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 17, 2006, at 3:02 AM, Brian Doyle wrote: Vincent _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I'm wondering if someone with a bit more hardware experience with ethernet controllers would mind sanity checking this observation: The docs for ifnet_output_raw() in kpi_interface.h say that the (mbuf_t) packet passed to the routine will be freed in the event of an error. Looking through IONetworkController::outputPacket() and BCM440X::outputPacket() it looks like even if there *isn't* an error the packet will be freed (in the case of noErr on a real controller the packet is put onto the ethernet ring buffer and then freed after transmission during serviceTxInterrupt()). Am I interpreting the source correctly and should the docs for kpi_interface.h be updated to reflect this? What the doc means is that the mbuf will be free in all cases, even if ifnet_output_raw() is passed a bad parameter or another error happens before the mbuf gets to the driver. This email sent to site_archiver(a)lists.apple.com

1 0

how to get thread CPU time
by info＠yourkit.com 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hello, is there any way to get time which particular thread is being executed on CPU. In other words how many milliseconds (or nonoseconds) of CPU time was consumed by thread since it has been started. IAIK, in Linux there is a clock_gettime(CLOCK_THREAD_CPUTIME_ID) system call (and Windows has similar API) but I cannot find any analog for Darwin. I'm writing CPU profiler for Mac OS X and thread CPU times is very important for me. Please advise. Best regards and thank you in advance, Vladimir Kondratyev YourKit, LLC http://www.yourkit.com "Don't get lost in data, get information!" _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

ifnet_output_raw() frees all packets?
by Brian Doyle 17 Mar '06

17 Mar '06

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thanks! Brian _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I'm wondering if someone with a bit more hardware experience with ethernet controllers would mind sanity checking this observation: The docs for ifnet_output_raw() in kpi_interface.h say that the (mbuf_t) packet passed to the routine will be freed in the event of an error. Looking through IONetworkController::outputPacket() and BCM440X::outputPacket() it looks like even if there *isn't* an error the packet will be freed (in the case of noErr on a real controller the packet is put onto the ethernet ring buffer and then freed after transmission during serviceTxInterrupt()). Am I interpreting the source correctly and should the docs for kpi_interface.h be updated to reflect this? This email sent to site_archiver(a)lists.apple.com

1 0