May 2007 - Darwin-kernel - lists.apple.com

devfs_make_node() bug in 9.0.0d8
by Nick Blievers 16 May '07

16 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Thunderbird 2.0.0.0 (Macintosh/20070326) Hi, What I end up with is: /dev/xv/<edgename> Hope this helps, Nick _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I noticed something while porting our driver to leopard today. It seems that devfs_make_node() is truncating characters from the fmt specifier when creating subdirectories. Eg if I make a call: devfs_hdl = devfs_make_node(dev, DEVFS_CHAR, 0, 0, 0777, "%sxvm/%s", prefix, edgename); A much simpler test for what (i assume) is the same problem can be illustrated here: : [cxfsmac11{cxfs-dev}]kern ; sudo ln -s /dev/xv /dev/xvm : [cxfsmac11{cxfs-dev}]kern ; ls -la /dev/xvm lrwxr-xr-x 1 root wheel 0 May 16 16:58 /dev/xvm -> /dev/x Note that the link points to /dev/x rather than /dev/xv as the command line suggested. Adding a bogus extra character does actually work in creating the correct link. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Finding PCI device register values through KEXT and user App.
by Jason Seng 14 May '07

14 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Finding PCI device register values through KEXT and user App.
by rohit dhamija 13 May '07

13 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=mBYkWCWwdvA4eNLkY+Ur6n4LLEI7Wshshdx8nIik5yG77mMM3L4fC7mvpPDBmgwQEIFyS7SF3g1mbP2K+jdYGxB7eWzet4JrwtUqRWk2w1BZYSZH3bQwX2deWBoaQyw0Rew8UYcRkkhWy1414D+OuxRaJAOpRcOOYQpBxaZK3BA= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=dhMaa15QmMkrWdL3NunlRnZ54hiQheqqsr53cVZwoaCFul5n9yk/LePcdKmwUqDkDPdjFb45Y+dn8kjL2LGqKtYbtJTgc/G8sqjupRlogXtl2MHWVEl2XABQPuoD27xv13gKpXDDlTK7Ou6chPQYQ6upLzahfkLMOfYAeSvS+AM= Dear all, I have developed a kernel extension module for a monitor calibration tool. The prime responsibilty of this module is to communicate to graphics card which in turn communicates to the monitor through i2c communication. Video graphics card have set of registers for i2c communication - data and clock. I need to communicate with a particular card, (ATI Rage 128 pro DVI) card but I do not have the base offsets and masking values for the same. So my query is to "how to find base offsets of i2c in a video graphics card ?" Below is one of the method I planned to find the i2c data and clock offsets and want your expert comments for the same. I would be thankful if you could please review this idea and if possible share some better alternate ideas /////////////////////////////// Step 1) MAP PCI DEVICE MEMORY Map the PCI device memory and get the virtual address and size of the address. /* code snippet */ fRegisterMap = fPCIDevice->mapDeviceMemoryWithRegister(uladdr); deviceRegisters = (ati_reg *)fRegisterMap->getVirtualAddress(); /********************/ Step 2) MAP MEMORY TO USER LEVEL APPLICATION To map i plan to use clientMemoryForType() in my kext /* code snippet */ fMemoryToShare->retain(); *memory = fMemoryToShare; /*********************/ Step 3) USER LEVEL APPLICATION PROCESS The user level application will open the driver and communicate with the driver and get the mapped memory region. See the memory region using Xcode debugger, it will print the address and memory location /* code snippet */ IOConnectMapMemory() // to get the mapped memory at use level /*****************/ Step 4) I2C COMMUNICATION THROUGH COMMANDS In the application use i2c api;s to read EDID data. This will write/read parameters from the graphics card for i2c communication with monitor by writting/reading at some registers on the graphics card. /*code for EDID can be taken from*/ http://darwinsource.opendarwin.org/Current/IOGraphics-179.2/tools/i2cexampl… Step 4) Check the memory changes The address whose memory location changes will be the offset of the i2c for data/clock /////////////////////////////// Can you please comment if this approcah feasible and correct way to find offset in a video graphics card of a particular register ? Or there is any alternate methods for the same ? Any ideas? The main objective is to find the offset and masking values of i2c registers in a video graphics card, all suggestions are most welcome. Please comment Thanks and Regards, Rohit -- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

debug kernel
by cristian pancià 13 May '07

13 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type; b=FcpcTcFhEBq8n4MWWKQh7DoU/3XphdYeqoDCsBHoh4qjzV76T91I3y4luni4S/p+SqgltC7fzirOV4DSNvCarUrBN1ODpTSzZqyU7rOYmgmnHNocaMl3CCJSeb2M6jIWp1878VaWfJ1KqrPFLIW6+wl2wOHWeOV0El7t0nkLLhE= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type; b=Xa8Iz5y4EDuELMFxLYj2vEx/Bm42fFiS6CkpLLVq5OaM68IaUFDl5X0WppfmBq2A/SnqOeZWqszIymtWUKTzBsBEfS2CANOwEU/NYLrzFzc17WNHbZIwc/rXS4S+XHJD7INHweR/ZujBPphz/ExGTuFSd36zRkAeljlg9AIoXOs= _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: sharing custom filesystem
by Quinn 11 May '07

11 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Relations, Developer Technical Support, Core OS/Hardware _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… At 17:09 -0400 10/5/07, Allen Briggs wrote: [...] is it hard-coded to using a list of known filesystem types? Yes, at least in current public implementations. Hopefully we'll have some good news on this front soon. For those following along at home, the bug tracking this is <rdar://problem/3168703>. This email sent to site_archiver(a)lists.apple.com

1 0

sharing custom filesystem
by Allen Briggs 10 May '07

10 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mutt/1.5.13 (2006-08-11) Hi, I was looking through the archives and found basically the question that I currently have. In 2002, but the answer seems to be a mystery. The last message I see in the archives is http://lists.apple.com/archives/darwin-kernel/2002/Oct/msg00024.html Does anyone have contact information for Marek Kozubal or have an answer to the question: * How do I make a custom file system sharable through AppleFileServer ? Or is it hard-coded to using a list of known filesystem types? The documentation is quite limited, the source doesn't appear to be available, and I can't find much of any information available on its behavior. Thanks, -allen -- Allen Briggs | http://www.ninthwonder.com/~briggs/ | briggs(a)ninthwonder.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

IODTPlatformExpert
by apic | puredarwin.org 10 May '07

10 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com PGP.sig

1 0

Re: Anything wrong with x86_64 and SYSV shared memory allocation?
by Terry Lambert 10 May '07

10 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -- Terry Actually, this seems to be incorrect: #include <semaphore.h> #include <stdio.h> int main() { sem_t *sem1, *sem2; if (sem_close(sem1)) { perror("failed closing sem1"); } if (sem_close(sem2)) { perror("failed closing sem2"); } } bigmac:~/fpc/tests/test jonas$ gcc -o tt tt.c .bigmac:~/fpc/tests/test jonas$ ./tt sem2 with same name as sem1 failed before unlink sem2(4) with same name as sem1(3) succeeded after unlink bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ cat tt.c #include <semaphore.h> #include <stdio.h> int main() { sem_t *sem1, *sem2; if (sem_close(sem1)) { perror("failed closing sem1"); } if (sem_close(sem2)) { perror("failed closing sem2"); } } bigmac:~/fpc/tests/test jonas$ gcc -o tt tt.c bigmac:~/fpc/tests/test jonas$ ./tt sem2 with same name as sem1 failed before unlink sem2(4) with same name as sem1(3) succeeded after unlink bigmac:~/fpc/tests/test jonas$ sw_vers ProductName: Mac OS X ProductVersion: 10.4.9 BuildVersion: 8P135 Jonas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/tlambert%40apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… You end up with a different semaphore; that's what I meant by "fail". The unlink destroys the namespace entry for it. On May 10, 2007, at 2:39 AM, Jonas Maebe wrote: On 24 apr 2007, at 14:47, Terry Lambert wrote: Of course, POSIX semaphores can be forced to be resource tracked using sem_unlink() after the last sem_open(), but once they are unlinked, subsequent sem_open()'s fail until everyone has drained out their references through either _exit() or sem_close(), so they are more or less in the same boat. sem1 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem1 == (sem_t*)SEM_FAILED) { printf("Failed to create sem1\n"); return 1; } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed before unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded before unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem1"); } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed after unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded after unlink \n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem2"); } sem1 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem1 == (sem_t*)SEM_FAILED) { printf("Failed to create sem1\n"); return 1; } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed before unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded before unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem1"); } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed after unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded after unlink \n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem2"); } This email sent to tlambert(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: Anything wrong with x86_64 and SYSV shared memory allocation?
by Jonas Maebe 10 May '07

10 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On 24 apr 2007, at 14:47, Terry Lambert wrote: Actually, this seems to be incorrect: #include <semaphore.h> #include <stdio.h> int main() { sem_t *sem1, *sem2; if (sem_close(sem1)) { perror("failed closing sem1"); } if (sem_close(sem2)) { perror("failed closing sem2"); } } bigmac:~/fpc/tests/test jonas$ gcc -o tt tt.c .bigmac:~/fpc/tests/test jonas$ ./tt sem2 with same name as sem1 failed before unlink sem2(4) with same name as sem1(3) succeeded after unlink bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ bigmac:~/fpc/tests/test jonas$ cat tt.c #include <semaphore.h> #include <stdio.h> int main() { sem_t *sem1, *sem2; if (sem_close(sem1)) { perror("failed closing sem1"); } if (sem_close(sem2)) { perror("failed closing sem2"); } } bigmac:~/fpc/tests/test jonas$ gcc -o tt tt.c bigmac:~/fpc/tests/test jonas$ ./tt sem2 with same name as sem1 failed before unlink sem2(4) with same name as sem1(3) succeeded after unlink bigmac:~/fpc/tests/test jonas$ sw_vers ProductName: Mac OS X ProductVersion: 10.4.9 BuildVersion: 8P135 Jonas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Of course, POSIX semaphores can be forced to be resource tracked using sem_unlink() after the last sem_open(), but once they are unlinked, subsequent sem_open()'s fail until everyone has drained out their references through either _exit() or sem_close(), so they are more or less in the same boat. sem1 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem1 == (sem_t*)SEM_FAILED) { printf("Failed to create sem1\n"); return 1; } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed before unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded before unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem1"); } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed after unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded after unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem2"); } sem1 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem1 == (sem_t*)SEM_FAILED) { printf("Failed to create sem1\n"); return 1; } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed before unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded before unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem1"); } sem2 = sem_open("/mysemtest1",O_CREAT|O_EXCL,0,0); if (sem2 == (sem_t*)SEM_FAILED) { printf("sem2 with same name as sem1 failed after unlink\n"); } else { printf("sem2(%d) with same name as sem1(%d) succeeded after unlink\n",(int)sem2,(int)sem1); } if (sem_unlink("/mysemtest1")) { perror("failed unlinking sem2"); } This email sent to site_archiver(a)lists.apple.com

1 0

Re: Collecting Usage Statistics
by Graham J Lee 09 May '07

09 May '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On 9 May 2007, at 18:03, Thomas Inskip wrote: Cheers, Graham. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… By looking at the source code to top, I have learned how to do the CPU and VM stat collection, but the only ways I've found to do I/O accounting are on a per device (disk or network interface) basis, not per-process, task, or thread. Accidentally didn't post this to the list - http:// www.opensource.apple.com/darwinsource/10.4.9.x86/system_cmds-336.20/ fs_usage.tproj/fs_usage.c This email sent to site_archiver(a)lists.apple.com

1 0