March 2008 - Darwin-kernel - lists.apple.com

Re: task_info
by Timur Alperovich 12 Mar '08

12 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; bh=MliK9srUhWC8CMQjikHdo017G25RDwU2QYqFUcgezEA=; b=Sir3hKg82s1cLsn0zh+EB02Y3/7kT1SMXLnIHtasO+AJiNix/AVdLWPzK4dxXbPhplD/ccL2tocyO1YNz3aQ0zyfYqZTeNRscVqLrAqTfhSXRCbzd3Hmp0PIcWE4OQGXEHvtb5gnqOx7wKVFAbmJQL7hx/3srF+OmNfQ7mIe8Ss= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=ncci/cMOPAlnsnSZ864zFz7IqIAT6pxKuS2WkQlOqKv4Le6Ed/SMdiZPws9jhICPrsaUsMF0qnjaGqB7pyU2wE24lQijQvxhzAhD8h3IxmGipDHHz9/T1wQCeAmTBrjMx90sk2eWZBFdhnmiLmP98YWNTY3VQkTGKpTQxXfMj68= Ok, swap the vm and hw size :) Is it reporting kernel memory? because then 867 MB vmem would make sense I guess and same goes for the 41 MB of physical mem. Timur On Tue, Mar 11, 2008 at 8:37 PM, Timur Alperovich <timur.alperovich(a)gmail.com> wrote: > I tried using the task_info call to get memory usage information on > all the processes, but for a number of them it returns the same value: > 867000000 for the hw size and 40965 for the vm size. Are these magic > numbers of some sort? because 867 MB for syslogd seems just a tad too > much. > > Thank you, > Timur > _______________________________________________ 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

task_info
by Timur Alperovich 12 Mar '08

12 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; bh=Yz5/9BqiUc49ZYCUBMcat0a6idTw0oafY5fAIr9PNaU=; b=mpfNQdOxUFHXRhzmz69RGMsZAnTiRkglkAspzcxyX/5WWmtkPm39pLuVnGBuwupsF2iECxmIGZTWxebcw5n9vVZAWGoSXkGLVTj5icjHiG2W/zAr1A/GvRGHE9AmbbgbiCrCGpXnzh7jfvUMc3g02wk52Y2kyJZ3/YUCip0o0WU= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=nmdLqEwADRFPJLq7qb0lSplEB76CAPogiVTZ1AjU7SIoFq49SHZuKDXjXGu0RY2QNreG+iIPqj5HAQ1leg8/mgOsqYTUMtFDzdO/fDd8d++L49tR9/ZtYA4rOWAU79sIZOyUWLc03X34B8a34/kPInQXUc4MxQg1ujRbvoKETK0= I tried using the task_info call to get memory usage information on all the processes, but for a number of them it returns the same value: 867000000 for the hw size and 40965 for the vm size. Are these magic numbers of some sort? because 867 MB for syslogd seems just a tad too much. Thank you, Timur _______________________________________________ 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: 64 bit development -- Is mmap() of over 4GB possible on OS X? (errors shown)
by Terry Lambert 11 Mar '08

11 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -- Terry ---------------------------------------------------------------------------------------------------------- #include <stdio.h> #include <sys/mman.h> #include <fcntl.h> #include <unistd.h> #include <string.h> #include <stdlib.h> #include <sys/errno.h> #include <sys/types.h> #define MEMSIZE ((long)4*1024*1024*1024) int main (void) { int fd1, err; char *mem1; printf ("sizeof(void *) = %d\n", sizeof(void *)); printf ("sizeof(int) = %d\n", sizeof(int)); printf ("sizeof(long) = %d\n", sizeof(long)); printf ("memsize = %ld\n", MEMSIZE); err = shm_unlink ("bank1"); printf ("shm_unlink bank1 = %d\n", err); //exit (EXIT_SUCCESS); fd1 = shm_open ("bank1", O_RDWR|O_CREAT); printf ("fd1 = %d\n", fd1); err = ftruncate (fd1, MEMSIZE); printf ("ftruncate bank1 = %d\n", err); printf ("Error = %s\n", strerror(errno)); memset (mem1, 0xAA, MEMSIZE); } ---------------------------------------------------------------------------------------------------------- Output is: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 4294967296 shm_unlink bank1 = 0 fd1 = 3 ftruncate bank1 = 0 mmap bank1 = 0xffffffffffffffff Error = Invalid argument Segmentation fault ---------------------------------------------------------------------------------------------------------- If I change MEMSIZE to ((long)3*1024*1024*1024) The Output is: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 3221225472 shm_unlink bank1 = 0 fd1 = 3 ftruncate bank1 = 0 mmap bank1 = 0x101000000 Error = Unknown error: 0 ---------------------------------------------------------------------------------------------------------- Send Darwin-kernel mailing list submissions to darwin-kernel(a)lists.apple.com To subscribe or unsubscribe via the World Wide Web, visit http://lists.apple.com/mailman/listinfo/darwin-kernel or, via email, send a message with subject or body 'help' to darwin-kernel-request(a)lists.apple.com You can reach the person managing the list at darwin-kernel-owner(a)lists.apple.com When replying, please edit your Subject line so it is more specific than "Re: Contents of Darwin-kernel digest..." Today's Topics: 1. 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? (Jamil J. Weatherbee) 2. Re: 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? (William Kucharski) ---------------------------------------------------------------------- Message: 1 Date: Mon, 10 Mar 2008 00:59:06 -0700 From: "Jamil J. Weatherbee" <jamil(a)weatherbeecorporation.com> Subject: 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? To: darwin-kernel(a)lists.apple.com Message-ID: <84933EFA-C325-4F81-ABB6-D28DA113180D(a)weatherbeecorporation.com> Content-Type: text/plain; charset="us-ascii" Gentlemen, compiled with "gcc -arch ppc64 -o openmem openmem.c" #include <stdio.h> #include <sys/mman.h> #include <fcntl.h> #include <unistd.h> #include <string.h> #include <stdlib.h> #include <sys/errno.h> #include <sys/types.h> #define MEMSIZE ((long)2*1024*1024*1024) int main (void) { int fd1, fd2, fd3, fd4, fd5, err; char *mem1, *mem2, *mem3, *mem4, *mem5; printf ("sizeof(void *) = %d\n", sizeof(void *)); printf ("sizeof(int) = %d\n", sizeof(int)); printf ("sizeof(long) = %d\n", sizeof(long)); printf ("memsize = %ld\n", MEMSIZE); err = shm_unlink ("bank1"); printf ("shm_unlink bank1 = %d\n", err); err = shm_unlink ("bank2"); printf ("shm_unlink bank2 = %d\n", err); err = shm_unlink ("bank3"); printf ("shm_unlink bank3 = %d\n", err); err = shm_unlink ("bank4"); printf ("shm_unlink bank4 = %d\n", err); err = shm_unlink ("bank5"); printf ("shm_unlink bank5 = %d\n", err); fd1 = shm_open ("bank1", O_RDWR|O_CREAT); printf ("fd1 = %d\n", fd1); fd2 = shm_open ("bank2", O_RDWR|O_CREAT); printf ("fd2 = %d\n", fd2); fd3 = shm_open ("bank3", O_RDWR|O_CREAT); printf ("fd3 = %d\n", fd3); fd4 = shm_open ("bank4", O_RDWR|O_CREAT); printf ("fd4 = %d\n", fd4); fd5 = shm_open ("bank5", O_RDWR|O_CREAT); printf ("fd5 = %d\n", fd5); err = ftruncate (fd1, MEMSIZE); printf ("ftruncate bank1 = %d\n", err); err = ftruncate (fd2, MEMSIZE); printf ("ftruncate bank2 = %d\n", err); err = ftruncate (fd3, MEMSIZE); printf ("ftruncate bank3 = %d\n", err); err = ftruncate (fd4, MEMSIZE); printf ("ftruncate bank4 = %d\n", err); err = ftruncate (fd5, MEMSIZE); printf ("ftruncate bank5 = %d\n", err); mem1 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd1, 0); printf ("mmap bank1 = %p\n", mem1); mem2 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd2, 0); printf ("mmap bank2 = %p\n", mem2); mem3 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd3, 0); printf ("mmap bank3 = %p\n", mem3); mem4 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd4, 0); printf ("mmap bank4 = %p\n", mem4); mem5 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd5, 0); printf ("mmap bank5 = %p\n", mem5); memset (mem1, 0xAA, MEMSIZE); memset (mem2, 0xBB, MEMSIZE); memset (mem3, 0xCC, MEMSIZE); memset (mem4, 0xDD, MEMSIZE); memset (mem5, 0xEE, MEMSIZE); } Output: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 2147483648 shm_unlink bank1 = 0 shm_unlink bank2 = 0 shm_unlink bank3 = 0 shm_unlink bank4 = 0 shm_unlink bank5 = 0 fd1 = 3 fd2 = 4 fd3 = 5 fd4 = 6 fd5 = 7 ftruncate bank1 = 0 ftruncate bank2 = 0 ftruncate bank3 = 0 ftruncate bank4 = 0 ftruncate bank5 = 0 mmap bank1 = 0x6008000 mmap bank2 = 0x8ff4e000 mmap bank3 = 0x10ff4e000 mmap bank4 = 0x18ff4e000 mmap bank5 = 0x20ff4e000

1 0

Re: lbpcap
by Andreas Fink 11 Mar '08

11 Mar '08

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

lbpcap
by cristian pancià 11 Mar '08

11 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type; bh=ZFlXr1M9OhVE700FPJDZsSPYjJRw5MYJlurvp220Dvk=; b=eaJN+/cnhObs+JE+rnbGkj9z/fn0wQ1KhlmLtBnKEOOyfw+GZGp8QiDX6HY5PXT6IJ2NNPrEPB41ptyHU2eLwa2aG26UF6NMLIghoKRapi1SwG0+ENWknzUoYvnp4ENr3ckfUsOaa/VIah6/Olv4ScsZxp/SiB4aBxfkpa6tOV8= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:mime-version:content-type; b=PChgIh/GNgCAzPiItBIOEwqkD5WoPLbmtct7XgnKhS5C/qQ/mqRTcafYOuXByGUh/DpqNDJ+s/a+JpgVywcasKqT03L36774oyK5X4hcwtaOPADsOQH/ROmko22Prv2O0LbJZZVcAPH8uGZVEYQeJkLqKuhKCiefShNWgZdIqHk= _______________________________________________ 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: dyld 2 questions
by Terry Lambert 11 Mar '08

11 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com no it's really not at boottime imagine in a syslog loaded mach-o index 0 foo prog array {x86_64,i386} (mach-o [index 0 x86_64 code index 1 i386 code]) my question was related to runtime, if I could force to load i386 code rather than x86_64 code section Either of the two commands: ps gaxO flags open -a "Activity Monitor" Best Regards Per-process: arch -i386 foobar -- Terry -Dave On Mar 8, 2008, at 2:16 PM, mm w wrote: ha thank you, something that maches 8) my question was not related to gcc or compilation time it was related to runtime, so there is no way to force the OS to run in i386 mode, but as far I know, the mach-o loader is able to deal with it, i just wanted to know if it was possible to force it. thank you for your answers System-wide? No. Not without recompiling your kernel. Shantonu Sent from my MacBook On Mar 8, 2008, at 11:08 AM, mm w wrote: So, to make simple, imagine a wonderful OS running on 32 extend 64 arch, all the sources are fat i386/ x86_64 is it to possible to influence the prefered mode -- -mmw -- -mmw _______________________________________________ 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.… On Mar 8, 2008, at 12:31 PM, mm w wrote: Yes. That is precisely what the "arch" command does. If you are exec'ing the binary yourself, there are also preference order extensions to posix_spawn() to allow you to cause a particular slice to be spawned. That's the underlying primitive system call which the arch command is based on, and it is how the testing checkbox in XCode is implemented. That method is less portable, since CPU IDs and families tend to evolve over time, so you are unlikely to be able to write a program now using that that will continue working correctly in the future. You would be better served to invoke "arch" from within your program, if you plan on doing this; the reason the command exists is that it gets revised when the OS is revised, so it can be used portably in scripts. I don't care to get the process, I wanted to know which section has been loaded yes and sure I know well everything that you try to express but it's realy out the box of my question Will report 64 bit processes. If the flags field contains 0x0004, then it's a 64 bit process. Will run the GUI tool Activity Monitor, which reports whether a process is 32 or 64 bit. You can also use libproc to introspect your process (or you can call the sysctl() directly, but that is not the recommended API, going forward). Or if you're the one doing the logging to the system log, you can use #ifdef and log different strings. On Sat, Mar 8, 2008 at 11:54 AM, David Elliott <dfe(a)tgwbd.org> wrote: System-wide: On x86-64 you can add -legacy to the boot-args nvram property (using the nvram command-line tool). That will force the kernel to run in i386 mode instead of x86-64 mode so you will not be able to run 64- bit programs at all and everything including the kernel itself will run in true i386 mode. Be careful with this. I have been led to believe by the author of the flag that it doesn't work on the newest machines. In this situation, trying something like arch -x86_64 foobar will totally fail since the CPU will not ever be in 64-bit mode. If that's what you want, you can have it. It's also useful if you are trying to run Darwin on certain CPUs like late-model Pentium 4 that pass the startup checks in tsc.c, or at least did until Leopard's xnu removed the check for CPU family 0xf (extended), but fail later on in kernel startup because of a missing bcopy 64 commpage routine (which requires Supplemental SSE3 which is not present on Pentium 4). On Sat, Mar 8, 2008 at 11:11 AM, Shantonu Sen <ssen(a)apple.com> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Some possibly useful hacks for building xnu
by Larry Meadows 11 Mar '08

11 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Cheers. -- Larry <lfm(a)easystreet.net> _______________________________________________ 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.… Hi, I've been lurking for a while. There was a post last year on building the xnu kernel, missing ctfconvert, compress-man-pages, etc. I got a (relatively) clean build of xnu-1228.3.13 with a bootable (debug) kernel just recently and thought you all might find some of this useful. The compress-man-pages problem just needed a little hack in makedefs/ MakeInc.dir: compress-man-pages.pl seems to want a '-d' flag before the directory name. ctfconvert was a little more of a challenge. Get the dtrace-48 tarball from the darwin source at http:// www.opensource.apple.com/darwinsource/. However, several files are missing. Get the libdwarf package from somewhere (I got version libdwarf20060614 from an RPM somewhere). Unpack it into libdwarf at the same level as libelf in the dtrace-48 package. Get libibierty from somewhere (the apple libgcc8.1 package seems OK). Unpack it somewhere and extract libiberty and include directories. I stuck those under a directory dtrace-48/libiberty. Configure libiberty but make sure getopt isn't included (by editing the Makefile) and build libiberty.a Configure libdwarf; I ran the autoconfig and then edited config.h by hand. You want HAVE_ELF_H and HAVE_LIBELF_H. Edit the Makefile and and includes to ../head and ../sys and a define -D_MIPS_SZLONG=64 and build libdwarf.a Then you need to build ctfconvert and ctfdump. I tried to do this with xcode but it was a hassle with the new stuff that I needed. So I created makefiles in libelf and in tools/ctfconvert. You need the same flags that are in the xcode project for those directories plus make sure you #define -D_MIPS_SZLONG=64 if including libdwarf.h . I also put some symlinks in the head include directory for decl.h, msg.h (in libelf) and dwarf.h, libdwarf.h (in libdwarf). Then link ctfconvert.o and other objs to get ctfconvert, and dump.o and other objs to get ctfdump. This works well enough to build xnu without any of the ctf related failures, and the files that come out seem to be well formed -- dwarfdump likes them, etc. I have no idea if any of this was worthwhile but I hate obscure error messages. If anyone wants a tarball of the whole mess let me know; I wasn't too careful in documenting all the steps. Now the only major annoyance in the build is a bunch of messages from the final link about "can't find atom for N_GSYM stabs... dtrace...". I think these are benign; the linker seems to expect debug information for some dtrace probes that is of course not generated since they are inserted with some hacky asm, not by the compiler.. It probably doesn't show up in a non-debug build. This email sent to site_archiver(a)lists.apple.com

1 0

64 bit development -- Is mmap() of over 4GB possible on OS X? (errors shown)
by Jamil Weatherbee 10 Mar '08

10 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com The following code compiled with "gcc -arch ppc64 -o openmem openmem.c": ---------------------------------------------------------------------------------------------------------- #include <stdio.h> #include <sys/mman.h> #include <fcntl.h> #include <unistd.h> #include <string.h> #include <stdlib.h> #include <sys/errno.h> #include <sys/types.h> #define MEMSIZE ((long)4*1024*1024*1024) int main (void) { int fd1, err; char *mem1; printf ("sizeof(void *) = %d\n", sizeof(void *)); printf ("sizeof(int) = %d\n", sizeof(int)); printf ("sizeof(long) = %d\n", sizeof(long)); printf ("memsize = %ld\n", MEMSIZE); err = shm_unlink ("bank1"); printf ("shm_unlink bank1 = %d\n", err); //exit (EXIT_SUCCESS); fd1 = shm_open ("bank1", O_RDWR|O_CREAT); printf ("fd1 = %d\n", fd1); err = ftruncate (fd1, MEMSIZE); printf ("ftruncate bank1 = %d\n", err); printf ("Error = %s\n", strerror(errno)); memset (mem1, 0xAA, MEMSIZE); } ---------------------------------------------------------------------------------------------------------- Output is: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 4294967296 shm_unlink bank1 = 0 fd1 = 3 ftruncate bank1 = 0 mmap bank1 = 0xffffffffffffffff Error = Invalid argument Segmentation fault ---------------------------------------------------------------------------------------------------------- If I change MEMSIZE to ((long)3*1024*1024*1024) The Output is: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 3221225472 shm_unlink bank1 = 0 fd1 = 3 ftruncate bank1 = 0 mmap bank1 = 0x101000000 Error = Unknown error: 0 ---------------------------------------------------------------------------------------------------------- Send Darwin-kernel mailing list submissions to darwin-kernel(a)lists.apple.com To subscribe or unsubscribe via the World Wide Web, visit http://lists.apple.com/mailman/listinfo/darwin-kernel or, via email, send a message with subject or body 'help' to darwin-kernel-request(a)lists.apple.com You can reach the person managing the list at darwin-kernel-owner(a)lists.apple.com When replying, please edit your Subject line so it is more specific than "Re: Contents of Darwin-kernel digest..." Today's Topics: 1. 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? (Jamil J. Weatherbee) 2. Re: 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? (William Kucharski) ---------------------------------------------------------------------- Message: 1 Date: Mon, 10 Mar 2008 00:59:06 -0700 From: "Jamil J. Weatherbee" <jamil(a)weatherbeecorporation.com> Subject: 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5? To: darwin-kernel(a)lists.apple.com Message-ID: <84933EFA-C325-4F81-ABB6-D28DA113180D(a)weatherbeecorporation.com> Content-Type: text/plain; charset="us-ascii" Gentlemen, I want to use shared memory with a very large data object that will have multiple unrelated readers (process wise) and a single writer. I will be synchronizing access using flock() when I need to write but most of the time the data will be read only. Here is some sample code I wrote to test the mmap() subsystem --- I am using shm_open() because I do not want backing store for this data it will be rebuilt from a separate data store upon reboot. Am I somehow misusing mmap() by wanting to have one large logically contiguous shared memory space to store this data? I can imagine this has come up in HPC and database engines before, so I am surprised that I cannot find any documents that address it. Here is some very quickly hacked together sample code and the output it generates for me on a powerPC G5 XServe with 8GB of RAM. (I know the actual physical memory doesn't matter except in the amount of time you are going to be waiting for pageouts while memset runs) Try increasing MEMSIZE until it breaks for you and you get 0xffffffffffffffff returned from mmap. BTW - this leaves 10GB of garbage around in the VM if you run it successfully. What I ideally want is to map one shared 6GB area, for my purposes if I need more virtual memory then I will just increase the physical memory size as the XServes are now accepting up to 32GB of RAM. compiled with "gcc -arch ppc64 -o openmem openmem.c" #include <stdio.h> #include <sys/mman.h> #include <fcntl.h> #include <unistd.h> #include <string.h> #include <stdlib.h> #include <sys/errno.h> #include <sys/types.h> #define MEMSIZE ((long)2*1024*1024*1024) int main (void) { int fd1, fd2, fd3, fd4, fd5, err; char *mem1, *mem2, *mem3, *mem4, *mem5; printf ("sizeof(void *) = %d\n", sizeof(void *)); printf ("sizeof(int) = %d\n", sizeof(int)); printf ("sizeof(long) = %d\n", sizeof(long)); printf ("memsize = %ld\n", MEMSIZE); err = shm_unlink ("bank1"); printf ("shm_unlink bank1 = %d\n", err); err = shm_unlink ("bank2"); printf ("shm_unlink bank2 = %d\n", err); err = shm_unlink ("bank3"); printf ("shm_unlink bank3 = %d\n", err); err = shm_unlink ("bank4"); printf ("shm_unlink bank4 = %d\n", err); err = shm_unlink ("bank5"); printf ("shm_unlink bank5 = %d\n", err); fd1 = shm_open ("bank1", O_RDWR|O_CREAT); printf ("fd1 = %d\n", fd1); fd2 = shm_open ("bank2", O_RDWR|O_CREAT); printf ("fd2 = %d\n", fd2); fd3 = shm_open ("bank3", O_RDWR|O_CREAT); printf ("fd3 = %d\n", fd3); fd4 = shm_open ("bank4", O_RDWR|O_CREAT); printf ("fd4 = %d\n", fd4); fd5 = shm_open ("bank5", O_RDWR|O_CREAT); printf ("fd5 = %d\n", fd5); err = ftruncate (fd1, MEMSIZE); printf ("ftruncate bank1 = %d\n", err); err = ftruncate (fd2, MEMSIZE); printf ("ftruncate bank2 = %d\n", err); err = ftruncate (fd3, MEMSIZE); printf ("ftruncate bank3 = %d\n", err); err = ftruncate (fd4, MEMSIZE); printf ("ftruncate bank4 = %d\n", err); err = ftruncate (fd5, MEMSIZE); printf ("ftruncate bank5 = %d\n", err); mem1 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd1, 0); printf ("mmap bank1 = %p\n", mem1); mem2 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd2, 0); printf ("mmap bank2 = %p\n", mem2); mem3 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd3, 0); printf ("mmap bank3 = %p\n", mem3); mem4 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd4, 0); printf ("mmap bank4 = %p\n", mem4); mem5 = mmap (NULL, MEMSIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd5, 0); printf ("mmap bank5 = %p\n", mem5); memset (mem1, 0xAA, MEMSIZE); memset (mem2, 0xBB, MEMSIZE); memset (mem3, 0xCC, MEMSIZE); memset (mem4, 0xDD, MEMSIZE); memset (mem5, 0xEE, MEMSIZE); } Output: sizeof(void *) = 8 sizeof(int) = 4 sizeof(long) = 8 memsize = 2147483648 shm_unlink bank1 = 0 shm_unlink bank2 = 0 shm_unlink bank3 = 0 shm_unlink bank4 = 0 shm_unlink bank5 = 0 fd1 = 3 fd2 = 4 fd3 = 5 fd4 = 6 fd5 = 7 ftruncate bank1 = 0 ftruncate bank2 = 0 ftruncate bank3 = 0 ftruncate bank4 = 0 ftruncate bank5 = 0 mmap bank1 = 0x6008000 mmap bank2 = 0x8ff4e000 mmap bank3 = 0x10ff4e000 mmap bank4 = 0x18ff4e000 mmap bank5 = 0x20ff4e000

1 0

Darwin development - Academic research project on online communities
by Jonathan Bigler - EPFL 10 Mar '08

10 Mar '08

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

Re: 64 bit development -- Is mmap() of over 4GB possible on OS X 10.4/10.5?
by William Kucharski 10 Mar '08

10 Mar '08

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 10, 2008, at 1:59 AM, Jamil J. Weatherbee wrote: What is errno set to when this happens? mmap(2) is defined to return -1 to the caller in case of error. William Kucharski _______________________________________________ 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.… Try increasing MEMSIZE until it breaks for you and you get 0xffffffffffffffff returned from mmap. BTW - this leaves 10GB of garbage around in the VM if you run it successfully. This email sent to site_archiver(a)lists.apple.com

1 0