June 2002 - Darwin-kernel - lists.apple.com

mach.sym format description
by Luke Nihlen 11 Jun '02

11 Jun '02

Hi, I'm looking at the mach.sym file and I'm curious about it. How exactly is it used within the kernel, and where could I find some information about it's structure? Does anyone know? I can't find anything on the ADC website. Thanks, Luke Nihlen _______________________________________________ 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: VFS Stack Development
by Zack Angelo 08 Jun '02

08 Jun '02

Wow. That's discouraging. Does Apple even have plans in place to document the technology? -Zack > -----Original Message----- > From: darwin-kernel-admin(a)lists.apple.com [mailto:darwin-kernel- > admin(a)lists.apple.com] On Behalf Of Justin C. Walker > Sent: Thursday, June 06, 2002 10:47 PM > To: darwin-kernel(a)lists.apple.com > Subject: Re: VFS Stack Development > > On Thursday, June 6, 2002, at 04:14 , Justin C. Walker wrote: > > > On Wednesday, June 5, 2002, at 07:41 , Zack Angelo wrote: > > > >> Hi all, > >> > >> I'm looking for information/code on developing VFS stack components. > >> I've checked all over ADC and the mailing-list archives and the only > >> thing I could find is the 3 or 4 pages in the Kernel developer doc. If > >> anyone could point me to any information I'd greatly appreciate it. > >> Thanks in advance. > > > > There's precious little information on this, in the form of > > documentation. You can check source at > > Not much help, was it? What the author meant to say was "check the > source at > http://www.opensource.apple.com/cgi-bin/registered/cvs/ > Look for projects with 'fs' in the name. (Sadly, Apple decided that > cacaphony was the answer to people's difficulties finding source for > specific components, and scattered all the projects at one level). Note > that 'hfs' is only the user-mode components (mkfs, newfs, mount). The > file system implementation for this one is in the kernel (xnu). > > Regards, > > Justin > > -- > Justin C. Walker, Curmudgeon-At-Large * > Institute for General Semantics | Men are from Earth. > | Women are from Earth. > | Deal with it. > *--------------------------------------*-------------------------------* > _______________________________________________ > 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

How do I compile the OS X 10.1.5 kernel with KTRACE?
by John Kessler 08 Jun '02

08 Jun '02

So, I've spent a few days trying to get this working and have just come up with a handful of kernel panics. Luckily I found this newgroup today so I feel I've got a fighting chance of nailing this down. Forgive me if this topic has been brought up before. I've downloaded the latest XNU 201-91 from Apple's CVS server and tried to compile the kernel to support KTRACE, but apparently I'm going about it all wrong and am in need of a HowTo or a FAQ on the step-by-step processes required to get it up and running. I can compile the kernel just fine, but I can't figure out how to get the KTRACE options to work. I've looked at various NetBSD documentation on the web about precisely this issue, but the same rules don't seem to apply as there is no conf directory or /usr/src directory under Darwin/OS X. Basically I'm running in circles now trying things here and there that end up making now difference whatsoever in my final results so I'd like to get some help with this if anyone has been down this road before. _______________________________________________ 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: X Mass Storage - Subclassing Logical Unit Drivers
by Shawn Erickson 07 Jun '02

07 Jun '02

I have a project started on opendarwin that I intend to populate with common IOKit driver examples/models and I plan for them to include PB projects. I am hoping that folks would be interested in pruning out their IP and submit examples to this project and/or submit requests for examples. Also any improvements or fixes to examples would be greatly appreciated. I hope to have one or two up in the next week or so... <http://www.opendarwin.org/cgi-bin/cvsweb.cgi/proj/IOKitExamples/> -Shawn On Friday, June 7, 2002, at 08:25 AM, Kevin Elliott wrote: While we're on the subject, is that skeleton driver available online anywhere? Recalling my early experience with IOKit the hardest part was getting all the various pieces set correctly in ProjectBuilder, rather than understanding IOKit itself. I think it would be very helpful if people could get their hands on the project directly, rather than just the write up in the mass storage book. Just my 2 cents... -Kevin Elliott Software Engineer CharisMac Engineering _______________________________________________ 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

Remote Debugging
by Sean Morrell 07 Jun '02

07 Jun '02

I am currently getting a kernel panic caused by my audio driver, but the only module listed on the call stack is IOAudioFamily. Is it possible to compile a debug version of IOAudioFamily to be used for remote debugging? If so, how or when do I generate the symbol file needed by the remote debugging machine? The system will automatically load IOAudioFamily during boot for the built-in audio controller. So I wouldn't be able to use kmodload to generate the symbol file. Thanks Sean _______________________________________________ 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?: Kernel Memory Mapped to User Space
by Andrew Gallatin 07 Jun '02

07 Jun '02

Shawn Erickson writes: > Why don't you have your API provide memory allocation and deallocation > functions that do the needed driver interaction for them? Clients of you > API would be required to use your methods and all would be well. > > If they don't want to use your methods that is ok but then they must > deal with the issue themselves. > All of that is done as you say. They can choose between letting us allocate memory for them, or they can choose to "register" (wire) memory they've allocated themselves. For our most popular middleware (MPI) we have a shim layer that overrides malloc and free in the library they link with & makes sure to "deregister" (unwire) memory before calling the real free to help prevent erorrs. (it also provides a registration cache for platforms where pinning memory is expensive) But there is a huge base of existing code that our users port to our system; using the api correctly is sometimes more than they can handle. Even though its rare, when they shoot themselves in the foot our QA people think a crashed application with a bus error is acceptable notification, but an uninterruptible hang is not. 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: How To?: Kernel Memory Mapped to User Space
by Shawn Erickson 07 Jun '02

07 Jun '02

Why don't you have your API provide memory allocation and deallocation functions that do the needed driver interaction for them? Clients of you API would be required to use your methods and all would be well. If they don't want to use your methods that is ok but then they must deal with the issue themselves. With that said a feature like you describe would be helpful. -Shawn On Thursday, June 6, 2002, at 11:09 AM, Andrew Gallatin wrote: Godfrey van der Linden writes: As Jim said it is usually better to map user memory into the kernel and have the driver wire and unwire it. You don't really have to worry about the app aborting without your being aware of it as the act of wiring means that the memory stays valid until the DRIVER is ready to get rid of it. Unless you're wiring things for short bounded amounts of time, then you *do* need to worry about it. If the app attempts to free the wired memory, it will deadlock because the VM system is waiting for the driver to unwire the memory. But the driver has no idea the VM system wants it to unwire the memory. [...] The problem only happens when users violate our API which says you must "deregister" (unwire) memory before freeing it. But its still a nasty situation & nearly impossible for a hapless user to debug. (They'll just complain their app "works fine on Linux" & potentially not use OS-X.). I really wish I could get somebody at Apple interested in fixing this by providing some way for a driver to detect this situation. (Setting a flag you can get to off the proc struct like Solaris does would be nice) _______________________________________________ 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: VFS Stack Development
by Justin C. Walker 07 Jun '02

07 Jun '02

On Thursday, June 6, 2002, at 04:14 , Justin C. Walker wrote: On Wednesday, June 5, 2002, at 07:41 , Zack Angelo wrote: Hi all, I'm looking for information/code on developing VFS stack components. I've checked all over ADC and the mailing-list archives and the only thing I could find is the 3 or 4 pages in the Kernel developer doc. If anyone could point me to any information I'd greatly appreciate it. Thanks in advance. There's precious little information on this, in the form of documentation. You can check source at Not much help, was it? What the author meant to say was "check the source at http://www.opensource.apple.com/cgi-bin/registered/cvs/ Look for projects with 'fs' in the name. (Sadly, Apple decided that cacaphony was the answer to people's difficulties finding source for specific components, and scattered all the projects at one level). Note that 'hfs' is only the user-mode components (mkfs, newfs, mount). The file system implementation for this one is in the kernel (xnu). Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | Men are from Earth. | Women are from Earth. | Deal with it. *--------------------------------------*-------------------------------* _______________________________________________ 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?: Kernel Memory Mapped to User Space
by Andrew Gallatin 07 Jun '02

07 Jun '02

Scott Taggart writes: > yea, but that does not satisfy my requirements: > > 1) My driver have the buffer available before any apps start > 2) The buffer be persistent when apps come and go. > > Maybe something is far easier, but it is not of use if it doesn't meet the > requirements. So, is what I ask for doable or not? We do the following to map a few pages of board memory (memory mapped PCI) to user space. It should work for normal physical memory I think. "kva" is the kernel virtual address we are mapping out to user space. Bounds checking, etc, is omitted so you can just see the calls you need: md = IOMemoryDescriptor::withAddress (kva, len, kIODirectionOutIn); if (!md) { IOLog ("IOMemoryDescriptor::withAddress failed for %p len %d\n", kva, len); goto abort_with_entry; } /* create a region len bytes long in this process's address space */ user_map = md->map (current_task(), (unsigned int) 0, kIOMapAnywhere, 0, len); if (!user_map) { IOLog ("IOMemoryDescriptor::withAddress failed for %p len %d\n", kva, len); goto abort_with_md; } va = (void *) user_map->getVirtualAddress (); copyout (&va, arg, sizeof (va) != 0) { goto abort_with_user_map; } Then we save off md and user_map. When the application closes our device or exits, we do: user_map->unmap (); user_map->release (); md->release (); I hope this helps, 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: How To?: Kernel Memory Mapped to User Space
by Scott Taggart 07 Jun '02

07 Jun '02

yea, but that does not satisfy my requirements: 1) My driver have the buffer available before any apps start 2) The buffer be persistent when apps come and go. Maybe something is far easier, but it is not of use if it doesn't meet the requirements. So, is what I ask for doable or not? Scott At 10:19 AM 6/6/2002, you wrote: G'day Scott, At 18:08 -0700 02-6-5, Scott Taggart wrote: Hi, I have seen various posts, responses, etc. on this subject, but quite frankly, I got lost. I have the need to take a slug of kernel memory in my driver and make it available to my user app so that my app can access it directly (I'll take care of arbitration and protection with my driver). Can someone give me the simplest set of calls I need to do this? I am willing to research the calls but just need the general set of calls required and order they should be called along with the appropriate caveats. The memory I need to map to user land is allocated by my driver at startup and is wired. As Jim said it is usually better to map user memory into the kernel and have the driver wire and unwire it. You don't really have to worry about the app aborting without your being aware of it as the act of wiring means that the memory stays valid until the DRIVER is ready to get rid of it. By far the easiest way of doing this sort of thing is to pass a buffer in with an open style routine. Wrap it in a memory descriptor and register it with your driver. Your driver can then prepare() (i.e. wire) it and map() it to get a virtual address to access the memory. Now is also a good time to grab the physical addresses for any future DMA you may want to do. Godfrey _______________________________________________ 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