December 2002 - Darwin-kernel

Re: howto compile kernel for ktrace
by Justin C. Walker 12 Dec '02

12 Dec '02

On Thursday, December 12, 2002, at 08:58 AM, Michael Marti wrote: Thanks again Justin Again, if you use those tags, verbatim, they won't work. CVS only likes '-', not '.'. I just was not careful when I wrote my last post. I tried the tags using '-' and not '.' Still the tags do not exist. But no problem, meanwhile Quinn told me (on MacNetProg List <macnetworkprog(a)lists.apple.com>) that: " If you go to the CVS Web interface for xnu, you'll find a popup menu that lists all of the tags. " Good to know... I just make this post to promote useful information (and to try to be polite). Thanks. It is good to do this, so the information ends up in the archives. 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: howto compile kernel for ktrace
by Michael Marti 12 Dec '02

12 Dec '02

Thanks again Justin > Again, if you use those tags, verbatim, they won't work. CVS only > likes '-', not '.'. I just was not careful when I wrote my last post. I tried the tags using '-' and not '.' Still the tags do not exist. But no problem, meanwhile Quinn told me (on MacNetProg List <macnetworkprog(a)lists.apple.com>) that: >> " If you go to the CVS Web interface for xnu, you'll >> find a popup menu that lists all of the tags. " And this helps a lot. I just make this post to promote useful information (and to try to be polite). Best regards Michael. _______________________________________________ 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: 64-bit PCI transfers
by Tim Seufert 11 Dec '02

11 Dec '02

On Wednesday, December 11, 2002, at 01:38 PM, Michael George wrote: According to specs, the PCI bus in the G4s can do 64-bit transfers, at least for DMA transfers. GCC and the include files on OS X have the "long long" data type, which is 64-bits. However, one of our hardware designers said that when he tested 64-bit IO from the processor to the device on the PCI bus, that the high-order bits were all cleared and data was only in the bottom long. I am wondering if it is indeed the case that CPU <--> PCI bus transfers are limited to 4 bytes, or if we're overlooking something. G4 general purpose registers are all 32-bit. When GCC emits code for a 32-bit target, 64-bit 'long long' variables and operations on them are decomposed into multiple 32-bit ops. So, writing a 'long long' to memory/PCI/whatever will probably compile to two 32-bit store instructions, not one 64-bit (the only true 64-bit store operation available on 32-bit PPC is for floating point doubles). GCC could be clever and use the FPRs when copying 'long long' variables from one memory location to another without doing any computation on them in the middle. I don't know whether it does this. I'd look at the disassembly to find out for sure what it is doing. If it is doing 32-bit writes, that's not the end of the story. Some recent processors and/or PCI bridges may be capable of coalescing multiple 32-bit writes into 64-bit operations on the PCI bus. But if that didn't happen, it's very likely that the two 32-bit writes made it onto the PCI bus as individual transactions, and therefore one would expect the high order bits to be unused. To get maximum processor -> PCI write performance you really need to get the CPU to do burst writes (which of course are inherently 64-bit) instead of single-beat writes. PCI multiplexes address and data on the same pins, so the fewer addresses you send over PCI, the more bus cycles there are which get used for useful data transfer. A burst write is usually the maximum amount of contiguous data the processor can send in one operation, and thus is useful for getting the PCI bridge to do longer PCI bursts. This will generally require marking the destination for the writes as cacheable, as most PPCs cannot generate burst writes for anything but cacheable memory. Marking some PCI memory as cacheable means you really have to be careful in the driver, since cache coherence is NOT maintained in hardware for PCI memory. _______________________________________________ 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

64-bit PCI transfers
by Michael George 11 Dec '02

11 Dec '02

According to specs, the PCI bus in the G4s can do 64-bit transfers, at least for DMA transfers. GCC and the include files on OS X have the "long long" data type, which is 64-bits. However, one of our hardware designers said that when he tested 64-bit IO from the processor to the device on the PCI bus, that the high-order bits were all cleared and data was only in the bottom long. I am wondering if it is indeed the case that CPU <--> PCI bus transfers are limited to 4 bytes, or if we're overlooking something. Is anyone doing 64-bit CPU <--> PCI transfers? Thanks! -Michael _______________________________________________ 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: reading a file in the kernel
by Jim Magee 11 Dec '02

11 Dec '02

If you go to this length, just pass the file contents in, rather than the file descriptor number. Trying to convert from a filedescriptor number to real file data from within a KEXT will require you to access all sorts of kernel-private data structures which are not guaranteed against release-to-release changes. On Wednesday, December 11, 2002, at 02:39 PM, Jeffrey D. Chung wrote: If you must absolutely read a file from the kernel... the easiest thing to do is to have a userland application open the file and pass the file descriptor into the kernel... Then use the regular kernel reads and writes. I believe the vn driver is a good place to copy code from. _______________________________________________ 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: reading a file in the kernel
by Jeffrey D. Chung 11 Dec '02

11 Dec '02

If you must absolutely read a file from the kernel... the easiest thing to do is to have a userland application open the file and pass the file descriptor into the kernel... Then use the regular kernel reads and writes. I believe the vn driver is a good place to copy code from. Jeffrey ----- Original Message ----- From: "Eric Long" <ericsmac(a)aladdinsys.com> To: "Michael George" <george(a)auroravideosys.com>; "Darwin Kernel" <darwin-kernel(a)lists.apple.com>; <darwin-drivers(a)lists.apple.com> Sent: Wednesday, December 11, 2002 11:22 AM Subject: Re: reading a file in the kernel > > Looking into the archives, it seems that it is a bad idea to read a > > file from the filesystem from inside the kernel. However, I couldn't > > find anywhere that it's not possible. > > In my kext, DTS told me not to do file I/O. I was told to use a log-in app > to read configuration settings and pass them to the kext. > > > The communication between a Kext and an application process involves the > > implementation of a UserClient. > > <http://developer.apple.com/samplecode/Sample_Code/Devices_and_Hardware/IOKi t/ > > SimpleUserClient.htm> > > > Eric Long > > ___________________________________________________________________________ > Macintosh Software Engineer > > Aladdin Systems, Inc. http://www.aladdinsys.com > 245 Westridge Drive, Watsonville, CA 95076 (831) 761-6200 x628 > _______________________________________________ > 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: 3DFX Voodoo 4 Driver
by Sean McBride 11 Dec '02

11 Dec '02

Sam Hart (samh(a)granada-learning.com) wrote on Tue, 10 Dec 2002 09:40:49 +0000: >lot of .mac sites ... strange ;) ). What I am wanting to know is, does >anyone know anything about writing a video driver in Darwin / OS X or >writing a 3DFX driver in general, as I am not above getting my hands >dirty and attempting one ( how bad can it be ? ). It is likely that the card has an EEPROM with an OS 9 ndrv driver on it. OS X supports these (and only these) OS 9 drivers, with some caveats (search IOKitFundamentals for 'ndrv'). So its likely that the card works at all because of this ndrv. Unless there really is a OS X native IOFramebuffer for 3dfx cards, but I don't think so. Anyway, writing an IOFramebuffer is what you'd need to do. But without docs on the 3dfx hardware, it'd be near impossible I dare say. _______________________________________________ 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: reading a file in the kernel
by Eric Long 11 Dec '02

11 Dec '02

> Looking into the archives, it seems that it is a bad idea to read a > file from the filesystem from inside the kernel. However, I couldn't > find anywhere that it's not possible. In my kext, DTS told me not to do file I/O. I was told to use a log-in app to read configuration settings and pass them to the kext. > The communication between a Kext and an application process involves the > implementation of a UserClient. > <http://developer.apple.com/samplecode/Sample_Code/Devices_and_Hardware/IOKi… > SimpleUserClient.htm> Eric Long ___________________________________________________________________________ Macintosh Software Engineer Aladdin Systems, Inc. http://www.aladdinsys.com 245 Westridge Drive, Watsonville, CA 95076 (831) 761-6200 x628 _______________________________________________ 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

reading a file in the kernel
by Michael George 11 Dec '02

11 Dec '02

Looking into the archives, it seems that it is a bad idea to read a file from the filesystem from inside the kernel. However, I couldn't find anywhere that it's not possible. I want to read from a file on disk to initialize a device. It will be done infrequently, perhaps just once, so if it's a hassle I will probably live with it. There is then the issue of whether or not the next release is likely to make it impossible to read from the disk. I've searched the driver and kernel lists for "kernel file read" and "file open", but the posts I found didn't answer these questions. Could someone either give me a pointer to a thread that will give me the info, perhaps a related string to search for, or the answers to my questions, I'd appreciate it. -Michael _______________________________________________ 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: Float processing not working
by Heiko Panther 11 Dec '02

11 Dec '02

Sean McBride wrote: The technique apple suggests, which I have done successfully, is to limit your floating point work to a shared library. Then this lib can be built with appropriate project settings. I forget the name, but one of Apple's sample shows this. Its some coreaudio sample driver I think. Originally, it was recommended to save cpu cycles, I think (because the fpu emulation is a little expensive, especially if used to crunch numbers). Now, the reasoning could be extended to "because the compiler's buggy". However, if I have float calculations scattered across my code, it really makes no difference if I call a library to do the calculations. For now, I like the convenience better than the speed gain of having all my float code concentrated in one place. Plus, I'm hoping the soft-fpu will get fixed eventually. So, I still need to know how to build a .kext with working float calculations. Is it possible without the library? Heiko _______________________________________________ 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