December 2003 - Darwin-kernel

Re: questions abort kernel extension.
by linuxbest linuxbest 10 Dec '03

10 Dec '03

On Wed, 10 Dec 2003 23:44:09 +0800 "linuxbest linuxbest" <huganglinux(a)hotmail.com> wrote: now compile passed, First I load Library.ext first, In my point, Load HelloKernel will work, In fact, it still report unresoved message. Now, The problem has been resolved. Most work is on Info.Plist setup. Not need extra setup. Thanks for help. -- Hu Gang / Steve Email : hugang(a)soulinfo.com, steve(a)soulinfo.com GPG FinePrint : 4099 3F1D AE01 1817 68F7 D499 A6C2 C418 86C8 610E GPG Public Key: http://soulinfo.com/~hugang/HuGang.asc MSN# : huganglinux(a)hotmail.com [9:00AM - 5:30PM +8:00] RLU# : 204016 [1999] (Register Linux User) _________________________________________________________________ SkA*;z5DEsSQ=xPP=;Aw#,GkJ9SC MSN Messenger: http://messenger.msn.com/cn _______________________________________________ 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: questions abort kernel extension.
by linuxbest linuxbest 10 Dec '03

10 Dec '03

On Wed, 10 Dec 2003 14:49:05 +0000 Quinn <eskimo1(a)apple.com> wrote: Check out the IOKitWithLibrary sample. <http://developer.apple.com/samplecode/Sample_Code/Devices_and_Hardware/IOKi…> This example shows how to export a C interface from one kernel extension (the library) to be picked up by another (in this case an I/O Kit driver, but the technique should work for other types of kernel extension). Yes, This example can work. But when I change the languge from C to c, it not work any more. In example LibraryPort export a fuction 'com_apple_dts_library_IOKitWithLibrary_LibraryFunction', I'm change HelloKernel ---- extern void com_apple_dts_library_IOKitWithLibrary_LibraryFunction(void); kern_return_t HelloKernel_start (kmod_info_t * ki, void * d) { printf("Kext has loaded!\n"); com_apple_dts_library_IOKitWithLibrary_LibraryFunction(); return KERN_SUCCESS; } --- now compile passed, First I load Library.ext first, In my point, Load HelloKernel will work, In fact, it still report unresoved message. Where am i wrong? -- Hu Gang / Steve Email : hugang(a)soulinfo.com, steve(a)soulinfo.com GPG FinePrint : 4099 3F1D AE01 1817 68F7 D499 A6C2 C418 86C8 610E GPG Public Key: http://soulinfo.com/~hugang/HuGang.asc MSN# : huganglinux(a)hotmail.com [9:00AM - 5:30PM +8:00] RLU# : 204016 [1999] (Register Linux User) _________________________________________________________________ SkA*;z5DEsSQ=xPP=;Aw#,GkJ9SC MSN Messenger: http://messenger.msn.com/cn _______________________________________________ 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: questions abort kernel extension.
by Quinn 10 Dec '03

10 Dec '03

At 22:36 +0800 10/12/03, linuxbest linuxbest wrote: I am writing two kernel extension, the name is OSIFS and ROMFS, ROMFS use OSIFS function, so ROMFS dependent with OSIFS. My question is How I can export functions/values from OSIFS to other kext. Check out the IOKitWithLibrary sample. <http://developer.apple.com/samplecode/Sample_Code/Devices_and_Hardware/IOKi…> This example shows how to export a C interface from one kernel extension (the library) to be picked up by another (in this case an I/O Kit driver, but the technique should work for other types of kernel extension). 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

questions abort kernel extension.
by linuxbest linuxbest 10 Dec '03

10 Dec '03

Hello all: I am writing two kernel extension, the name is OSIFS and ROMFS, ROMFS use OSIFS function, so ROMFS dependent with OSIFS. My question is How I can export functions/values from OSIFS to other kext. The same thing in linux, is use EXPORT_SYMBOL(x), But I don't what's the replacment in darwin. I just write them, and do nothing, The OSIFS can load into kernel, I got unresolved message when I load ROMFS. Hope someone can hope me. thanks. -- Hu Gang / Steve Email : hugang(a)soulinfo.com, steve(a)soulinfo.com GPG FinePrint : 4099 3F1D AE01 1817 68F7 D499 A6C2 C418 86C8 610E GPG Public Key: http://soulinfo.com/~hugang/HuGang.asc MSN# : huganglinux(a)hotmail.com [9:00AM - 5:30PM +8:00] RLU# : 204016 [1999] (Register Linux User) _________________________________________________________________ OmSCJ@=gIOWn4s5D5gWSSJ<~O5M3!* MSN Hotmail!# http://www.hotmail.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: IPC: interprocess communications
by Quinn 07 Dec '03

07 Dec '03

At 1:31 -0700 7/12/03, Matt Jaffa wrote: I wrote this KEXT, a NKE, and I want to communicate with this daemon that i will have running, so that the daemon can do all the reading and writing to files since it is not smart to do it in the kernel, How do I communicate between the two, The best approach in this particular context is to use a socket. The reason for this choice is that your KEXT is already knee deep in sockets code, so it makes sense to use sockets for the user space communication. An alternative approach (Mach messages, I/O Kit user client, etc) might make sense if you're KEXT was plugged into the kernel in some other place. If you're looking for an example of this, grab the WebDAV file system code from Darwin. <http://www.opensource.apple.com/darwinsource/tarballs/apsl/webdavfs-148.2.t…> [This link requires an APSL <http://www.opensource.apple.com/apsl/> account.] I know that if you match the KEXT from within the daemon you can call the KEXT methods, but is there any way that the KEXT can start the communication. The best way to handle this is to have your user space daemon call into the kernel to request work from the KEXT. When the KEXT has work for the daemon to do, it sends a parameter block describing the request to the socket, and unblocks the user space daemon. The daemon does the work and calls back into the kernel with the results, and then blocks waiting for the next request. Most people new to Mac OS X kernel programming try to use some sort of callout mechanism, where the kernel actually tries to call user space code. This is not the right approach. When programming this sort of communication, you have to be careful about what happens when there's no user space daemon thread waiting to pick up your request. This may just be a temporary condition, or it may be that the daemon has crashed. For reliability reasons it is important that your kernel code has a sensible alternative strategy. For example, a firewall NKE might switch into "deny everything" or "allow everything" mode, depending on the level of user paranoia. 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

IPC: interprocess communications
by Matt Jaffa 07 Dec '03

07 Dec '03

Ok, I wrote this KEXT, a NKE, and I want to communicate with this daemon that i will have running, so that the daemon can do all the reading and writing to files since it is not smart to do it in the kernel, How do I communicate between the two, I know that if you match the KEXT from within the daemon you can call the KEXT methods, but is there any way that the KEXT can start the communication. I have tried using mach_msg within the KEXT, when i compile it, everything is fine, but when i try to load the KEXT it says it cant load because it can't find _mach_msg ->> which means it can't link it right at load time. Any suggestions? Thanks _______________________________________________ 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 to communicate with existing device driver from kernel module
by Girish 04 Dec '03

04 Dec '03

Hello Darwin Devloper, I want to write a kernel module to interact with existing device driver.I am trying to indentify the method by which kernel module can interact with existing device driver. If you know such method then please let me know. If you know any book or site releted to writing kernel module specially mention above then please let me now. Thanks in advance. Regards Girish S P Software Technologies (I) Pvt Ltd. India __________________________________ Do you Yahoo!? Free Pop-Up Blocker - Get it now http://companion.yahoo.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: Finder copy chunk size
by Sam Vaughan 04 Dec '03

04 Dec '03

Thanks for the great responses Brian and Quinn. I returned 32MB in the f_iosize field of the statfs structure and saw an immediate improvement. The Finder started using 512k chunks to copy data between our volume and HFS, which improved the transfer speed immensely. I also tried setting the MNT_LOCAL flag in the f_flags field of the statfs structure, but the chunk size didn't increase above 512k. I doubt HFS imposes this limit, so I'm guessing the local file system message isn't getting through. Finally, I tried copying between two of our volumes. The 100MB QT movie I copied transferred so fast that the progress window barely had a chance to draw. My tracing showed that it used 32MB chunks - what an improvement over 12k! And to answer your question Brian, no, not XFS but CXFS, SGI's clustered file system: <http://www.sgi.com/products/storage/tech/file_systems.html> Soon Mac users everywhere will be able to enjoy the chewy goodness of simultaneous direct fibre channel access to a central data store. :-) Sam _______________________________________________ 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: Finder copy chunk size
by Quinn 03 Dec '03

03 Dec '03

At 15:02 +1100 3/12/03, Sam Vaughan wrote: Does the Finder have any logic to decide what size transfers to issue? Does it call any VFS layer hooks that we could implement? I look at the Mac OS X 10.3 implementation. The Finder uses a 2 MB buffer unless you're copying between different physical devices or the source is a network volume. If either apply, the copy size in the minimum of the source transfer size and the destination transfer size. If this number is smaller than 4 KB, 4 KB is used instead. If the volume is a floppy, Finder uses 16 KB for the optimal transfer size. Otherwise it tests whether the volume is local. For a remote volume, it uses f_iosize. For a local volume, it uses either the greater of f_iosize or 512 KB. [There's also a bunch of logic that comes into play if statfs returns an error, but that shouldn't ever happen, right?] The Finder determines whether a volume is local via the standard Mac OS practice, which is to call PBHGetVolParms (a Carbon File Manager routine) and then test the vMServerAdr field of the resulting GetVolParmsInfoBuffer. Local volumes have a value of 0; non-local volumes have a non-zero value. The Carbon File Manager sets vMServerAdr based on the MNT_LOCAL flag in the f_flags field returned by statfs. 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: Finder copy chunk size
by Brian Bergstrand 03 Dec '03

03 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Finder uses (struct statfs).f_iosize as a hint for it's chunk size. A lot of other utils use this too (like ls), so make sure they work with the larger size. What I want to know, is how/why the Finder special cases HFS volumes, as the HFS driver sets f_iosize to it's block size (usually 4K), but somehow the Finder uses much larger chunks. Are you porting XFS to Mac OS X? That would be kind of cool. I've been thinking about trying to port Resier once I get Ext3 journaling support done (or should I say IF I get it done). HTH. On Tuesday, December 2, 2003, at 10:02 PM, Sam Vaughan wrote: > When doing a copy to/from our file system, the Finder issues the > writes/reads in tiny 12k chunks. Since it sets the VNOCACHE_DATA > flag, we do direct I/O to the disk to get the best performance. > > The problem is that we could provide vastly better performance if the > chunk sizes were much larger. Chunks of 4MB or more would be ideal, > but even increasing to 128k would improve performance by up to a > factor of 8. > > Does the Finder have any logic to decide what size transfers to issue? > Does it call any VFS layer hooks that we could implement? > > Thanks, > > Sam > Brian Bergstrand <http://www.classicalguitar.net/brian/> PGP Key: <http://www.classicalguitar.net/brian/misc/public_key.txt> I may not have gone where I intended to go, but I think I have ended up where I intended to be. - Douglas Adams -----BEGIN PGP SIGNATURE----- Version: PGP 8.0.2 iQA/AwUBP82DDXnR2Fu2x7aiEQJs6ACgzeeKZPvYBJrYSe/SADeplnbqNDQAn0IV nW0LNmmsbYsJHFlZt5gHLt36 =g5+r -----END PGP SIGNATURE----- _______________________________________________ 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