August 2002 - Darwin-kernel

Re: Loading KEXTs dynamically
by Sean McBride 19 Aug '02

19 Aug '02

Sean Morrell (jseanm(a)cablelynx.com) wrote: >Can I load and use a KEXT dynamically? If so, what is the API and where >would I find a source example? man kextload You really should read the various docs about IOKit and kexts, etc. _______________________________________________ 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

Loading KEXTs dynamically
by Sean Morrell 19 Aug '02

19 Aug '02

Can I load and use a KEXT dynamically? If so, what is the API and where would I find a source example? Thank You, 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

Notification of boot progress?
by Sean Morrell 19 Aug '02

19 Aug '02

Is there a variable or notification indicating the progress of system boot? I am writing a driver that has dependencies that will not allow it to load at boot time. In other words, the driver's dependencies do not have the OSBundleRequired="Root". However I may be able to work around this if I knew when the system reached the point where a driver with such dependencies could be loaded. Is there a variable or event accessible from a driver indicating the progress of system boot? Is there a way to obtain this information from user space? 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: TrackPad driver
by David A. Gatwood 14 Aug '02

14 Aug '02

On Wed, 14 Aug 2002, Julien Guimont wrote: > Ok but it should be possible to say "intercept" the event of a tap, > double tap, drap with appropriate coordinates and do something usefull > with them. I'm thinking of something like logitech wheel mouse drivers > for OS X. I know they use USB which is fairly more simple than low level > firmware stuff but eh! i'm just beginning my investigation;) Ah. I assumed you were trying to do the whole scroll wheel on the edge of the trackpad thing that someone asked about on one of those online sites. :-) Yeah, I suspect you should be able to subclass the appropriate ADB driver and intercept taps and relative coordinate motion. A better place to ask this would probably be darwin-drivers or darwin-development, though, as they'd be more likely to know the ins and outs of the ADB stack and the way that it ties into the rest of the system. Later, David --------------------------------------------------------------------- David A. Gatwood dgatwood(a)gatwood.net Developer Docs Writer dgatwood(a)apple.com Apple Computer dgatwood(a)mklinux.org Check out my weekly web comic: http://www.techmagazine.org _______________________________________________ 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: TrackPad driver
by Julien Guimont 14 Aug '02

14 Aug '02

Ok but it should be possible to say "intercept" the event of a tap, double tap, drap with appropriate coordinates and do something usefull with them. I'm thinking of something like logitech wheel mouse drivers for OS X. I know they use USB which is fairly more simple than low level firmware stuff but eh! i'm just beginning my investigation;) thank you. Julien On Wednesday, Aug 14, 2002, at 03:55PM, David A. Gatwood <dgatwood(a)gatwood.net> wrote: >On Wed, 14 Aug 2002, Julien Guimont wrote: > >> I would be interested to implement new functions for the trackpad of >> ibook, Tibook trackpad. Can you point me out where I can find such >> drivers in darwin, or is it just proprietary stuff from apple. > >To my knowledge, you can't add functionality to the trackpad hardware >using drivers. To do that, you would have to add capabilities to the PMY >firmware, AFAIK, a procedure that is undocmented. The problem is if you >screw that firmware up (which is highly likely, since it's undocumented >bytecode), you won't be able to boot your system to undo the damage, and >you'd basically have to replace the main logic board in your laptop.... >In other words, it's not a very good idea to try this. > > >Later, >David > >--------------------------------------------------------------------- >David A. Gatwood dgatwood(a)gatwood.net >Developer Docs Writer dgatwood(a)apple.com >Apple Computer dgatwood(a)mklinux.org > > Check out my weekly web comic: > http://www.techmagazine.org > > > Julien Guimont juggy(a)mac.com www.juggysoft.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: TrackPad driver
by David A. Gatwood 14 Aug '02

14 Aug '02

On Wed, 14 Aug 2002, Julien Guimont wrote: > I would be interested to implement new functions for the trackpad of > ibook, Tibook trackpad. Can you point me out where I can find such > drivers in darwin, or is it just proprietary stuff from apple. To my knowledge, you can't add functionality to the trackpad hardware using drivers. To do that, you would have to add capabilities to the PMY firmware, AFAIK, a procedure that is undocmented. The problem is if you screw that firmware up (which is highly likely, since it's undocumented bytecode), you won't be able to boot your system to undo the damage, and you'd basically have to replace the main logic board in your laptop.... In other words, it's not a very good idea to try this. Later, David --------------------------------------------------------------------- David A. Gatwood dgatwood(a)gatwood.net Developer Docs Writer dgatwood(a)apple.com Apple Computer dgatwood(a)mklinux.org Check out my weekly web comic: http://www.techmagazine.org _______________________________________________ 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

TrackPad driver
by Julien Guimont 14 Aug '02

14 Aug '02

Hello I would be interested to implement new functions for the trackpad of ibook, Tibook trackpad. Can you point me out where I can find such drivers in darwin, or is it just proprietary stuff from apple. I looked at the CVS and I didn't find anything related to trackpad or tap, double tap functionnalities. Thank you for your help, Julien Guimont juggy(a)mac.com www.juggysoft.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: darwin-kernel: convert pid to a task right?
by Jim Magee 13 Aug '02

13 Aug '02

On Tuesday, August 13, 2002, at 05:28 PM, Joshua LeVasseur wrote: I want a client program to send a mach message to a server program. But I know of no mechanism to communicate the server's port send name to the client. The typical paradigm uses a third party name server, to which the server registers its service port, and to which the client can query for port rights to the server. The client obtains access to the name server's port during the process bootstrap. Try task_get_bootstrap_port(mach_task_self(), &my_bootstrap_port) or just using the bootstrap_port global variable. Neither of these should be NULL. The array forms (mach_ports_register() and mach_ports_lookup()) are obsolete (which probably explains why they panic and we didn't notice). It seems like you may be working from an old Mach 2.5 documentation (forgetting to insert a send right is common problem when converting from Mach 2.5 to Mach 3.0). You may want to look at the later Darwin documentation, and specifically the "Bootstrap Contexts Unleashed" chapter: http://developer.apple.com/techpubs/macosx/Darwin/General/ KernelProgramming/index.html The book title is a bit misleading. It is more about using the kernel services than programming within the kernel. My Darwin install seems to operate without a name server. The bootstrap ports are all nil, if I inspect them via mach_ports_lookup(). I tried installing my own server ports in the bootstrap ports via mach_ports_register(), but that only causes a kernel panic. As I said in another message, I'm looking into the panic. But the mach_init process should be running on your Darwin system, and it is the bootstrap Mach port name server that handles bootstrap_register() and bootstrap_look_up() RPCs and should be the default bootstrap server for all tasks in the system. So for an alternative solution, if I want a server to give send rights to a client, I can make the server the parent of the client. Then the server can execute mach_port_insert_right() on the child task to transfer the send rights. But the parent needs to know the task port of the child process. The parent should get the task port when it calls task_create(). But in Darwin, task_create() is hidden behind fork(). And so the parent has the child's pid, but not the child's task port. Is there anyway to convert a pid to a task right? There is, it's called task_for_pid(), but you should restrict your usage of it if you can avoid it. You shouldn't need it in this case. --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

Re: how to cause a kernel panic
by Jim Magee 13 Aug '02

13 Aug '02

First, a gentle nudge. The registered ports are reserved for system library use. You should not be attempting to use them for your own services. We may likely use those slots for other system services in the future, and your software will not be compatible with that use. You should use the bootstrap namespace to pass ports between processes. If exposing the ports to other tasks is a problem, you can create a private bootstrap subset for passing ports shared between the parent and child only. At first, I was suspicious. That's too easy to cause a panic. But I get it too. I'll let you know when there is a fix. But it is triggered by a flaw in your logic. You created a receive right, and then made a call which assumed you have a send right as well (mach_ports_register() is defined to copy each of the send rights specified in the slots array into the kernel). You have to insert a send right (using mach_port_insert_right() or mach_port_extract_right()) for your newly created port before making the call. Of course, this should just return an error. As I said, I'm looking into why that isn't so. ;-( On Tuesday, August 13, 2002, at 02:04 PM, Joshua LeVasseur wrote: Hello. I run Darwin 1.4.1 on a Pismo. Just a plain Darwin install; not OS X. From sysctl: kern.osrelease = 1.4.1 kern.osrevision = 199506 kern.version = Darwin Kernel Version 1.4.1: Sun Sep 30 20:54:21 PDT 2001; root:xnu-201.roots/xnu-201.obj/RELEASE_PPC I cause a kernel panic when using mach_ports_register(). The offending code: --------------- mach_port_t *registered_ports; mach_msg_type_number_t count; mach_port_t my_service_port; mach_port_allocate( mach_task_self(), MACH_PORT_RIGHT_RECEIVE, &my_service_port ); mach_ports_lookup( mach_task_self(), &registered_ports, &count ); registered_ports[ ENVIRONMENT_SLOT ] = my_service_port; mach_ports_register( mach_task_self(), registered_ports, count ); ---------------- Sorry if this is an inappropriate forum for this info ... I'm new to Apple's Darwin resources and haven't figured out where to submit bugs. -jtl _______________________________________________ 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

Power management with driver dependencies
by Francis bouchard 12 Aug '02

12 Aug '02

Hi, I'm working an a display driver for a card(PCI) that does not only serves as a graphic card. So I have a service provider for my Framebuffer driver (declared in the bundle bits of FB and receive it in the start.) Now I'm worried about power management. We all know that diplay can go to sleep even if the system does not go to sleep. This is not a problem. The problem occurs if the provider receive a notification to go to sleep before the framebuffer does. In which case the framebuffer will talk to nobody when wnating to go sleep afterwards. Q1: will IOKit be smart emough to sleep my drivers in the dependency order ? I.e: Framebuffer first, provider second. Q: if no to Q1, can I use the registerInterrestedDriver mechanism in my framebuffer to make sure it will shut itself down (and notify the OS it has shut down) before the provider go down ? Francis _______________________________________________ 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