October 2002 - Darwin-kernel

MFS
by Matthias Kretschmer 03 Oct '02

03 Oct '02

Hello I was trying to compile the darwin kernel with MFS enabled. So enabled the option in MASTER file and started the compilation, but it always fails at compiling the mfs source in bsd/ufs/mfs. I am running Jaguar and want to have some sort of ramdisk. What I am wondering about is why there are changes at July 2002 (so I am guessing there is something done to it), but it is left broken. Is there some solution available? Maybe some patches? -- Greetings Matthias Kretschmer _______________________________________________ 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: Altivec in a driver?
by Jonas Maebe 03 Oct '02

03 Oct '02

On donderdag, okt 3, 2002, at 01:27 Europe/Brussels, Michel wrote: I have that painful conversion loop in a driver. Can I use altivec in that context? is it safe? are the vector register saved etc? Thats assuming my installer would install the relevant binary etc. I don't know whether it's safe, but if it is: please do *not* install one binary or another depending on whether altivec is available at installation time. Currently, an installation of Mac OS X can always be used to boot any machine supporting it, which is very handy. Doing this kind of stuff would break this (very useful) functionality. Just do a test a runtime and set a function pointer instead. Jonas _______________________________________________ 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

Altivec in a driver?
by Michel 02 Oct '02

02 Oct '02

I have that painful conversion loop in a driver. Can I use altivec in that context? is it safe? are the vector register saved etc? Thats assuming my installer would install the relevant binary etc. Michel _______________________________________________ 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

Kernel allocation in the first 256Mb of memory
by Michel 02 Oct '02

02 Oct '02

My device requires addresses in that range for DMA transfert. ie - I have only 28 bits available. However, it seems theres no way to ensure that IOMallocContiguous() returns 'safe' addresses in that range. When IOMallocContiguous works, the driver works as expected, but successives load/unload of the driver gets new addresses for that memory, even if I'm very careful to release the blocks properly. So after a few times it returns pointers > 256Mb and I have to reboot. Is there some other trick? I need rougly 100Kb. It's a major roadblock for me on that driver, if I can't guarantee a working behavior somehow, I can't release! Michel _______________________________________________ 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: fork/exec performance?
by Ragnar Sundblad 02 Oct '02

02 Oct '02

On Mon, 23 Sep 2002 18:06:53 -0400 Jim Magee <jmagee(a)apple.com> wrote: vfork() yields 100X (yes, 100 times, not percent more) the performance of fork() in most situations on Mac OS X. So, if you can live with vfork() requirements, please do so. Getting common tools to use it where appropriate appears to be the trick in your case. I don't think Linux has vfork(), so people tend to stick with just fork() in portable code. Is there any chance of improving fork+exec performance in the OS since, as you say, many programs currently use fork? (Just for fun I did a quick and dirty change to bash to make it use vfork+_exit, which included moving the fork up one stack frame and some other small changes, to see if I could improve automake, (gnu-)libtool and friends' performance (they currently run really slow on mac os x, it is sad when the compile time is just a fraction of the time spent to build things.) The program seems to work for simple interactive use, but crashes immediately on shellscripts, I haven't checked what the problem is yet. Anyway, as far as I can tell, with my very limited knowledge on both vfork world limitations and the bash code, it could possibly be pretty hairy to (correctly) change for example bash to use vfork instead of fork, it would be nice if there was a smaller performance hit. If anyone more knowledgeable would take on the bash-vfork task instead I for one would be most grateful. Maybe if I filed a radar... :-) /ragge _______________________________________________ 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: kld displacement too large error
by Godfrey van der Linden 02 Oct '02

02 Oct '02

You need to copy and paste the gcc flags that project builder uses by default for a kext into your static library target. Specifically in this case you are missing the -mlong-branch build flag. Godfrey At 12:28 -0400 2002-10-2, Michael Cashwell wrote: Greetings all, I've been trying to get my kext to work with a portion of it coming from a static library (containing C and C++ code). After considerable effort I finally have the kext building. But when I try to kextload it I get these obscure errors: [dhcp02:/System/Library/Extensions] cashwell% sudo kextload -t netShim.kext/ kextload: extension netShim.kext/ appears to be valid kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 32 in section (__TEXT,__text) (displacement too large) kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 53 in section (__TEXT,__text) (displacement too large) kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 58 in section (__TEXT,__text) (displacement too large) kextload: kld_load_from_memory() failed for module /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim kextload: a link/load error occured for kernel extension netShim.kext/ load failed for extension netShim.kext/ I've tried "otool -r -v" on the executable but don't know how to interpret the output. Does anyone have clue what kld is complaining about here? How would I map the "entry numbers" in the error message to the otool output? (Are the decimal? Do they count from 0 or 1?, etc.) Perhaps there's some compiler or libtool option that I should be using to specify "wide" displacements when building either the library or the kext? Any clues would be appreciated. Thanks, Mike _______________________________________________ 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

kld displacement too large error
by Michael Cashwell 02 Oct '02

02 Oct '02

Greetings all, I've been trying to get my kext to work with a portion of it coming from a static library (containing C and C++ code). After considerable effort I finally have the kext building. But when I try to kextload it I get these obscure errors: [dhcp02:/System/Library/Extensions] cashwell% sudo kextload -t netShim.kext/ kextload: extension netShim.kext/ appears to be valid kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 32 in section (__TEXT,__text) (displacement too large) kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 53 in section (__TEXT,__text) (displacement too large) kld(): /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim relocation overflow for relocation entry 58 in section (__TEXT,__text) (displacement too large) kextload: kld_load_from_memory() failed for module /System/Library/Extensions/netShim.kext/Contents/MacOS/netShim kextload: a link/load error occured for kernel extension netShim.kext/ load failed for extension netShim.kext/ I've tried "otool -r -v" on the executable but don't know how to interpret the output. Does anyone have clue what kld is complaining about here? How would I map the "entry numbers" in the error message to the otool output? (Are the decimal? Do they count from 0 or 1?, etc.) Perhaps there's some compiler or libtool option that I should be using to specify "wide" displacements when building either the library or the kext? Any clues would be appreciated. Thanks, Mike _______________________________________________ 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

About sip.c in SharedIP
by Stéphane Sudre 02 Oct '02

02 Oct '02

Just by curiosity, isn't there any cases where sipList is NULL when SIP_stop is called? Because when this is the case, the NFDescriptor is not unregistered. _______________________________________________ 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: What does "not authentic" mean?
by Stéphane Sudre 01 Oct '02

01 Oct '02

On Monday, September 30, 2002, at 07:35 PM, Pelle Johansson wrote: On Mon, 30 Sep 2002, Stiphane Sudre wrote: When the absolute path of the file contains a character with an accent (like a ` (or à if you prefer the HTML version)), then the kext is not loaded. ` is a special character in the shells and need to be quoted. I'd recommend using tab completion which will probably fix this automatically for you, just type the first few characters of the path and press tab. (see man tcsh or man bash depending on your shell). Alternatively, when pasting paths, alvays quote it in ' ' "a`" = à = ` = 0x88 (in 4D) = 0x85 here (http://www.cdrummond.qc.ca/cegep/informat/Professeurs/Alain/files/ ascii.htm) "a `" does not mean 1 ` but an a with a `. _______________________________________________ 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