November 2004 - Darwin-kernel

Re: change boot-file doesn't work
by Jim Magee 14 Nov '04

14 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com I successfully built the latest kernel, xnu-517.9.4. Before actually using it , I need to test whether it works or not. boot-file variable to boot the Mac with my built kernel, but it seems not works. Here is what I did: boot-device hd:,\\:tbxi sudo nvram boot-file='hd:, mach_kernel.gcc' sudo reboot /System/Library/Extensions.mkext /System/Library/Extensions.kextcache /System/Library/Caches/com.apple.kernelcaches/* _______________________________________________ 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 On Nov 14, 2004, at 9:54 AM, Gong Michael wrote: >From some previous threads, I learn that I can change sudo cp <path_to_my_built_kernel>/mach_kernel /mach_kernel.gcc sudo nvram boot-device There are typically a LOT of extensions to process during boot. Each one has to be relocated against the current kernel and their dependent extensions. That can take quite a bit of time - and yet the result rarely changes. So, the kernel booting mechanism keeps a few caches around: If you just replace the kernel (through a boot-arg assignment) BootX can sometimes pick up the wrong Extension.kextcache file and have a failed boot. Therefore, you should remove all these files booting an alternate kernel. Alternatively, you can just "touch /System/Library/Extensions," but you have to make sure your date is correct, etc.... I find it easier just to remove the cache files. --Jim smime.p7s

1 0

change boot-file doesn't work
by Gong Michael 14 Nov '04

14 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi, all, I successfully built the latest kernel, xnu-517.9.4. Before actually using it , I need to test whether it works or not. >From some previous threads, I learn that I can change boot-file variable to boot the Mac with my built kernel, but it seems not works. Here is what I did: >sudo cp <path_to_my_built_kernel>/mach_kernel /mach_kernel.gcc >sudo nvram boot-device boot-device hd:,\\:tbxi >sudo nvram boot-file='hd:, mach_kernel.gcc' >sudo reboot #the system doesn't come up. reboot the machine, press crt-option-o-f keys, >set-defaults /* this reset the boot-file varabile back to its default, which is mach_kernel */ reboot the machine, the system comes up. >sudo cp /mach_kernel /mach_kernel_good /* save the currently working kernel */ >sudo cp /mach_kernel.gcc /mach_kernel /* rename my built kernel with the default name: mach_kernel */ >sudo reboot #the system comes up, so it proves that my built kernel is working. As a conclusion, it looks like only by changing the boot-file variable name can't boot the system with a kernel having a different name other than mach_kernel. I am wondering has anyone successfully booted a system with a kernel, whose name is not mach_kernel ? If so, could you please provide some insights ? Thanks a lot. Michael ______________________________________________________________________ Post your free ad now! http://personals.yahoo.ca _______________________________________________ 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

IPOPT_RA
by baskara 12 Nov '04

12 Nov '04

site_archiver(a)lists.apple.com Delivered-To: Darwin-kernel(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=DTXfOf1e5fi6WhALEtU65jIRUOY5glx2doyo4ukzaU5f/waygEynnGIlRB+2f+6XdsTOtsGNh46Rf+c1YU0TImriOaDhHL5wSG8wnXCGZmofEOOuO37fUFFFAg/pU5TmbYdLxnpoQWbDKC4gw2PpiKTpNpx/gOiBO8nV+Q1gyVQ= Hi, Currently, I am making a program to intercept UDP packet which contains IP Router Alert Option (IPOPT_RA). I would like to ask whether Mac OS X kernel (Darwin 7.5) supports this IP option (in IPv4)? I can see the option's number is already defined in ip.h. But, I'm not so sure because it's not implemented yet in FreeBSD and the man pages of the IP options inside my PowerBook G4 only contain Record Route, Source Route and TimeStamp. Thank you. _______________________________________________ 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: Darwin, G5 and TimeBase sync
by Benjamin Herrenschmidt 09 Nov '04

09 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com > On my side, I've tried implementing the timebase synchronisation > in linux, on a CY28508 based machine (older dual 2Ghz), and it doesn't > seem to work. From what I see, register 1 of the chip seem to contain > 0xff, and the write of 0xf3 to it (clearing 0x0c) doesn't seem to have > any effect... though If I turn the I2C interface from combined mode to > stdsub mode for the write (the read stays in combined mode of course), > then I get a lockup, which could mean some clock went off ... Ok, I can confirm now that I have a working implementation in Linux on both Pulsar and Cypress chip, I _think_ (again, I didn't have a chance to actually test with the actual Darwin code, but I think the problem is indeed there) you have a bug. Here is my a resume of my analysis: NOTE: This may have gone unnoticed due to the rather slow timebase (33Mhz) used, the flags + wait loops using for handshacking the 2 CPUs while passing the timebase around may end up beeing "good enough" to give a good synchronisation without the HW freeze actually working. - If the code in Darwin actually froze the timebase as it's supposed to, I suspect it would cause a lockup as soon as the i2c command stopping the TB clock reaches the clock chip. The i2c driver does IODelay() which uses the timebase. - From my experiments, the reason it doesn't work is that the combined mode used for reading from the i2c chip isn't suitable for writing to it. In that case, an stdsub write should be used, that is the write of the register address (| 0x80) and the actual data byte write should only be separated by an ACK, not by a new start + address cycle. I found that from the CY28510 datasheet which sounds similar to the 28508 used by Apple. Experiment (reading back the register and testing if the TB is actually frozen) confirms this. The Pulsar seem to behave the same way. - Finally, a note about the IODelay in the I2C driver... I wonder if this delay is actually necessary at all. I've never had it in the Linux driver for the keywest/uni-n i2c cell without noticeable problem, and the OF forth driver doesn't have any delay neither. Switching explicitely to combined when reading and stdsub when writing in MacRISC4CPU.cpp would then work with just the removal of that delay. Of course, there may be a reason I don't know to have this delay here but if that is the case, then your whole scheme cannot work, unless that IODelay is replaced by some other implementation, for example using the Keylargo timer. Hope this is useful, Regards, Ben. _______________________________________________ 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

Darwin, G5 and TimeBase sync
by Benjamin Herrenschmidt 09 Nov '04

09 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi ! I'm wondering if the CPU TimeBase synchronisation in Darwin works at all, more specifically, if enableCPUTimeBase() using i2c works at all... The reason is that if it worked, the timebase would be stopped (disabled), which means that a call to IODelay() -> delay() would, according to the implementation which uses the CPU timebase, loop forever when this timebase is stopped. However, the PPCI2CInterface.cpp driver (AppleI2C.kext) which seem to be used to communicate with the clock chip ... uses IODelay() on every register write ! On my side, I've tried implementing the timebase synchronisation in linux, on a CY28508 based machine (older dual 2Ghz), and it doesn't seem to work. From what I see, register 1 of the chip seem to contain 0xff, and the write of 0xf3 to it (clearing 0x0c) doesn't seem to have any effect... though If I turn the I2C interface from combined mode to stdsub mode for the write (the read stays in combined mode of course), then I get a lockup, which could mean some clock went off ... If the CY28508 is similar to the 28510 for which the datasheet is available, than stdsub is what you want for writes anyway, not combined... I may be totally wrong, of course, but I have some doubts that the TB sync in darwin actually works, at least on this machine, but feel free to contradict me ! Regards, Ben. _______________________________________________ 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

kqueue panics (was Re: single machine debugging?...)
by Eric R Johnson 09 Nov '04

09 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Eric, Thanks for the tip. I found that in the JVM 1.4.2 release notes a couple of days ago and we're currently testing the property change. So far, things look good. :) Might be a good idea for someone to write a kbase article about this one. By the way, it appears that the issue is not restricted to a JVM-kqueues interaction. Annecdotal reports indicate that other non-Java, i/o-intensive apps see panics. For example, this user saw panics using a python-based bittorrent client (Tomato Torrent): http://www.macfixit.com/article.php?story=20040915081217505 Regards, --EJ On Nov 08, 2004, at 7:13 PM, Eric Albert wrote: Azureus is written in Java and uses direct memory access for some of its networking and disk functions. Try running it with the java.nio.preferSelect property set to true. When that property is false (the default), Java uses kqueue rather than select for certain operations and may (rarely) run into problems like the one you're seeing. Hope this helps, Eric _______________________________________________ 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: single machine debugging?/advice for noob on OS X panic analysis
by Eric Albert 09 Nov '04

09 Nov '04

(Catching up on mailing lists....) In article <mailman.16700.1099335141.424.darwin-kernel(a)lists.apple.com>, Eric R Johnson <ericrjohnson(a)mac.com> wrote: > I'm trying to lend a hand in troubleshooting a persistent kernel panic > problem that's associated with the open source p2p bittorrent client, > Azureus. In the past, I've done a little crash analysis on Solaris but > I'm new to OS X and have alot of questions. If you're going to be > annoyed by noob questions, please delete. Thanks in advance to > anyone that can help. > > Azureus is written in Java and uses direct memory access for some > of its networking and disk functions. Try running it with the java.nio.preferSelect property set to true. When that property is false (the default), Java uses kqueue rather than select for certain operations and may (rarely) run into problems like the one you're seeing. Hope this helps, Eric _______________________________________________ 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

sysctl
by Matt Jaffa 08 Nov '04

08 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi, Any help is appreciated, Thanks, Matt _______________________________________________ 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 have defined my own sysctl variables in my NKE , and they are working great and everything, but I would like a non-root application to be able to set my variables, but it is not letting me even though I am doing this: SYSCTL_STRUCT(_kern, OID_AUTO, jaffaethernode, CTLFLAG_RW|CTLFLAG_ANYBODY, &somo2, EtherNode, "EtherNode"); I thought that setting the flag to CTLFLAG_ANYBODY would allow anyone to be able to set this structure no matter if it is a root app or regular user app. Is it because I am branching off of _kern that it will only allow a root app to set my structure. And if is this way, must I branch off of something else like _user to allow me to modify it with a non-root app. I have tried to tie my structure to the _user branch, but it doesn't register it with _user so that when I do sysctlbyname it returns ENOENT = 2. So that is why I am sticking with _kern, but would really like it if a non-root app can modify my variable jaffaethernode. This email sent to site_archiver(a)lists.apple.com

1 0

sysctlbyname returns 1
by Geoffrey Schmit 08 Nov '04

08 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com The following code, when made from within a kernel extension, results in the error variable being set to 1 (EPERM?): int error; int numberOfCPUs = 0; size_t numberOfCPUsSize = sizeof( int ); error = sysctlbyname( "hw.ncpu", &numberOfCPUs, &numberOfCPUsSize, NULL /* new value */, 0 /* new value length */ ); Any ideas? The same code works fine from a user-mode console application. I'm running Mac OS X 10.3.5. Thanks! geoff -- Geoffrey Schmit Senior Software Engineer National Instruments geoffrey.schmit(a)ni.com www.ni.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

Breakpointing the commpage crashes kernel: vm_protect, submaps, ...
by Chandra Khan 06 Nov '04

06 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=peIxTwDAZb159OAv2LigEfUJYw4l+B+frIrCbSXMH8Tsajh5aRnMSiQ9pxVwQtc7ikgxcWpjTMTb6dHSMXp0k9f1YMSULPWeJiWumZe1bE/0tPZ78C75+S727U7p2GHPpz6vb15QM5XeNzoLlenUGH83bxTOLC3ZBvND9nt0KOQ= Hi I have a need to momentarily modify instructions [from within a process's context in kernel space] in that process's address space. gdb does similar things when setting up breakpoints in gdb/src/gdb/macosx/macosx-nat-mutils.c:mach_xfer_memory() In my experiment, writing bytes in the 0xffff8000-0xffffefff (commpage) range ostensibly works, except that the vm maps of *other* processes somehow seems affected after my experiment. The symptom is that ntpd crashes the kernel, soon after I run my experiment, trying to update some timer value which resides in the commpage. Disabling ntpd allows me to run my experiment and do a vmmap on a boring unrelated pid, such as the login shell. Here are the results: Submap ffff8000-ffffefff r--/r-- process-only submap ffff8000 [ 4K] r--/r-- SM=COW commpage ffff9000 [ 4K] r--/r-- SM=SHM commpage I don't understand this -- using gdb to set up breakpoints on a PID X doesn't change the vmmap on some other PID Y. Why does doing it from the kernel have different effects? The calls I use from the process's kernel context are: vm_region(..., &info, ...) vm_protect(current_map(), user_addr, size, FALSE, // change "current", not "max" protection VM_PROT_READ | VM_PROT_WRITE | VM_PROT_COPY) followed by a copyout(), a pmap_sync_caches_phys(), and then a protection restoration: vm_protect(current_map(), user_addr, size, FALSE, // change "current", not "max" protection info.protection) In my effort at debugging this, I got lost trying to understand some submap nesting code. Oh, and I'm new to the whole Mach/Darwin/Xnu game. Any help is appreciated, Thanks, Chandra _______________________________________________ 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