October 2003 - Darwin-kernel

panicd & DB_KERN_DUMP_ON_PANIC
by Andrew Gallatin 29 Oct '03

29 Oct '03

When reading the Panther sources, I noticed that Panther has finally grown the ability to do a crashdump (hurray!!). However, there appears to be a daemon (panicd?) that needs to be run in order to collect the crashdump on a remote node. I can't seem to find this. A search on the ADC site turns up nothing. Can somebody point me at it, please? Thanks, Drew _______________________________________________ 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: Mach realtime thread with thread_policy_set
by Jim Magee 29 Oct '03

29 Oct '03

On Oct 29, 2003, at 3:01 AM, Greg Eigsti wrote: I am experimenting with Mach real time threads (thread_policy_set THREAD_TIME_CONSTRAINT_POLICY) and am having a problem where real time threads don't seem to be scheduled properly (far too often). I have a simple app that creates a pthread and then boosts the new pthread's priority via thread_policy_set; the pthread executes continuously and effectively stops the 'parent' thread from executing. I must note... that my real time pthread does some garbage calculation in a while(1) loop and calls YieldToAnyThread() to attempt to give up its time slice. Am I missing something obvious here? I would expect my real time thread to get regularly scheduled 'bits' of cpu time not 'continuous' cpu time. When you define a time-constrained thread, you are saying that "when it is able to run, I want it to run - at the expense of all other threads except possibly other time-constrained threads." Are you sure you really need a time-constrained thread here? The typical requirement for using one is that you have a real world time-sensitive deadline to meet (typically trying to provide data to a piece of hardware that is time-based). Specifically, garbage-[sp collection] should NEVER be done from a time-constrained thread. I am using 1ms for my computation value (workload takes .5ms) and 1000ms for my period value. I calculate computation and value with NanosecondsToAbsolute(). These values are simply to determine which of the time-constrained threads to run next/first if there is more than one available to run. They are not to define when the thread is to supposed to revert to "normal" scheduling again. At the moment, the values aren't used at all and all time-constrained threads are simply run round-robin (the deadlock avoidance mechanisms are not yet in place to use those values effectively - but they are coming so keep trying to be true to the vaues). Is there an API that I can use to signal that my real time thread is done executing for this period? I have scoured the internet and have not found a good answer. Actually "blocking" is the only way a time-constrained thread is expected to give up the CPU. --Jim _______________________________________________ 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

Mach realtime thread with thread_policy_set
by Greg Eigsti 29 Oct '03

29 Oct '03

I am experimenting with Mach real time threads (thread_policy_set THREAD_TIME_CONSTRAINT_POLICY) and am having a problem where real time threads don't seem to be scheduled properly (far too often). I have a simple app that creates a pthread and then boosts the new pthread's priority via thread_policy_set; the pthread executes continuously and effectively stops the 'parent' thread from executing. I must note... that my real time pthread does some garbage calculation in a while(1) loop and calls YieldToAnyThread() to attempt to give up its time slice. Am I missing something obvious here? I would expect my real time thread to get regularly scheduled 'bits' of cpu time not 'continuous' cpu time. I am using 1ms for my computation value (workload takes .5ms) and 1000ms for my period value. I calculate computation and value with NanosecondsToAbsolute(). Is there an API that I can use to signal that my real time thread is done executing for this period? I have scoured the internet and have not found a good answer. Thanks Greg _______________________________________________ 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: Stackable file systems: is it or isn't it?
by Quinn 28 Oct '03

28 Oct '03

At 13:29 -0600 28/10/03, Steven Bytnar wrote: First hit when you search for "stackable" on www.apple.com: http://www.apple.com/macosx/features/darwin/ *sigh* I did a purge of this on Apple's developer web site but didn't even think to look on the customer site. Oh well, here we go again. S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ 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: Stackable file systems: is it or isn't it?
by Steven Bytnar 28 Oct '03

28 Oct '03

First hit when you search for "stackable" on www.apple.com: http://www.apple.com/macosx/features/darwin/ Direct quote: >>>>> Since Mac OS X is designed to excel in a heterogeneous computing environment, Darwin also offers support for multiple file systems. Based on extensions to BSD and an enhanced Virtual File System (VFS) design, the file system component of Darwin uses a layered architecture in which file systems are stackable. <<<<< --Steve On Tue, Oct 28, 2003 at 10:33:47AM -0800, Quinn wrote: > At 18:36 +1100 28/10/03, Richard wrote: > >However now Apple is boasting a stackable file system > >for Panther. > > We are? Where? AFAIK the following document on our web site is > still valid. If there are statements that contradict this, I want to > know. > > <http://developer.apple.com/qa/qa2001/qa1242.html> > > S+E > -- > Quinn "The Eskimo!" <http://www.apple.com/developer/> > Apple Developer Technical Support * Networking, Communications, Hardware > _______________________________________________ > 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: Stackable file systems: is it or isn't it?
by Quinn 28 Oct '03

28 Oct '03

At 18:36 +1100 28/10/03, Richard wrote: However now Apple is boasting a stackable file system for Panther. We are? Where? AFAIK the following document on our web site is still valid. If there are statements that contradict this, I want to know. <http://developer.apple.com/qa/qa2001/qa1242.html> S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ 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

Stackable file systems: is it or isn't it?
by Richard 28 Oct '03

28 Oct '03

A stackable file system would appear to be the solution to a project I may be associated with except the docs on Apples site (minimal as they are) say they will not be supporting this feature in the future due to performance reasons. This would not have affected me as I planned on inserting my layer into a HFS file system on a .dmg volume. However I was put off the idea. However now Apple is boasting a stackable file system for Panther. So what gives??? Regards from Richard _______________________________________________ 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: Disk icon for mounted file system in the Finder
by Sam Vaughan 28 Oct '03

28 Oct '03

We're skipping the getattrlist call currently simply because we're getting away with it. Looking at the other file systems' implementations it looks like too much work for too little reward at the moment. Your dummy /dev entry and single IOMedia object idea looks interesting though. What is the connection that DiskArb uses to associate one of your mounted disks with the dummy /dev entry and the published IOMedia object? I trawled through the DiskArb code the other day and couldn't see why it would particularly care that we don't have any IOMedia objects. Thanks Marek, Sam On Saturday, Oct 25, 2003, at 09:05 Australia/Melbourne, Marek Kozubal wrote: On Tue, 21 Oct 2003, Sam Vaughan wrote: But in our case we don't have any IOMedia objects. Although the physical disks we use are directly connected to the machine, the logical volumes we present to the user are mounted like a network file system. We've completely avoided any integration with the IOKit. Then you cannot offically set a default icon. Also DiskArb really hates file systems that are missing IOMedia's. You can probably fudge something with a custom icon deal however. In our project where we're doing something simular we ended up making a dummy /dev entry and IOMedia object so that DiskArb and such would be happy. Otherwise they do nothing. --- Marek Kozubal marek(a)portents.com _______________________________________________ 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: replaceOrCopyPacket() question
by Joe Liu 27 Oct '03

27 Oct '03

Hi Michael, On Oct 27, 2003, at 6:28 AM, Michael Cashwell wrote: Greetings, I've read the docs and done a cursory examination of the source but thought I'd ask here in case someone "just knows" the answer. I'm working on packet reception in an Ethernet driver and have questions about replaceOrCopyPacket() and friends. I originally allocate a set of packets large enough to hold an entire received frames in one segment. This allows me to use mbufs almost directly with the card's DMA hardware. As I understand it, when I get back one of these buffers I need to determine its true logical length and then call replaceOrCopyPacket() to handle what goes up the stack and what is returned back to the DMA engine. OK, so far. But for efficiency I want to set the packet's m_pkthdr.len length [and possibly the internal m_len member as well, I'm unsure here] before calling that method. There's no way it can efficiently do its job if I haven't supplied the logical length first. Perhaps I'm suppose to leave the member fields alone and supply the logical length only via the argument to replaceOrCopyPacket() and it magically does what I need. That would be great, but the docs don't say and I haven't taken the source apart enough to figure it out. When you allocate a packet, the length fields in the mbuf are initialized to the capacity requested. When it is replaced, its m_pkthdr.len is examined so that the replacement mbuf will have the same capacity. The length passed to replaceOrCopyPacket() will be used to truncate the mbuf destined for the networking stack. In essence, you shouldn't need to update the mbuf length fields directly. In any case, I'm worried about the packet I get back for reuse with the DMA hardware not having the same attributes as what I originally allocated. They need to have the same capacity and be single-segment just like the original. No need to worry. If the length + alignment padding is less than MCLBYTES (2048 bytes), then it is guaranteed that it will come back as a single mbuf, likely attached to a mbuf cluster. This mbuf memory can be described with no more than 2 physical memory segments (typically 1 will suffice). -Joe _______________________________________________ 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

replaceOrCopyPacket() question
by Michael Cashwell 27 Oct '03

27 Oct '03

Greetings, I've read the docs and done a cursory examination of the source but thought I'd ask here in case someone "just knows" the answer. I'm working on packet reception in an Ethernet driver and have questions about replaceOrCopyPacket() and friends. I originally allocate a set of packets large enough to hold an entire received frames in one segment. This allows me to use mbufs almost directly with the card's DMA hardware. As I understand it, when I get back one of these buffers I need to determine its true logical length and then call replaceOrCopyPacket() to handle what goes up the stack and what is returned back to the DMA engine. OK, so far. But for efficiency I want to set the packet's m_pkthdr.len length [and possibly the internal m_len member as well, I'm unsure here] before calling that method. There's no way it can efficiently do its job if I haven't supplied the logical length first. Perhaps I'm suppose to leave the member fields alone and supply the logical length only via the argument to replaceOrCopyPacket() and it magically does what I need. That would be great, but the docs don't say and I haven't taken the source apart enough to figure it out. In any case, I'm worried about the packet I get back for reuse with the DMA hardware not having the same attributes as what I originally allocated. They need to have the same capacity and be single-segment just like the original. I could just punt and always allocate a new packet myself, but I'd like to place nice and be efficient. Any comments welcome. Thanks, Michael Cashwell _______________________________________________ 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