November 2004 - Darwin-kernel

Re: single machine debugging?/advice for noob on OS X panic
by Eric R Johnson 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com David, ^I have the very same problem with my employer. :) --EJ _______________________________________________ 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.… Thanks for your work on the backtrace. I'll file a bug and pass the info along to the Azureus developers. To put it more in perspective, the odds of this being helpful to a virus (unless the goal of the virus were to cause a kernel panic) are probably roughly on the same order as the odds of Ralph Nader winning today's election by a landslide. :-) Actually, I *was* thinking that some malicious person might simply seek to cause panics on lots of Mac systems. Not a big deal, so long as the system isn't doing anything very important, comes back up and critical data hasn't been damaged. Alot of qualifiers there, though... I wasn't trying to suggest that the vulnerability could be used as the basis for self-replicating code. One could use spam/social engineering (eg, phishing), hacked websites or other means to get the code executing on people's systems. As for Nader, he's looking strong for 2012 or maybe 2016... Of course the machines will have taken over by then, but we can always send the governor of California back in time to straighten things out. ;) Just my $0.02. As always, my opinions rarely (if ever) reflect the views of my employer. :-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: single machine debugging?/advice for noob on OS X panic
by Markus Hitter 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Am 02.11.2004 um 21:57 schrieb David Gatwood: Sometime in the recent past, Eric R Johnson wrote: What else do you want to get the potential above "zero"? Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ 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.… Markus, to your point regarding the user-land Java app crashing the kernel... I couldn't agree more. There's been some bluster on the SF thread about writing a virus when they figure out the issue, but I think that's idle talk born of frustration. The potential is pretty ugly, though. The potential should be pretty much zero. A sudden kernel panic causes the loss of opened, but unsaved documents. It easily triggers the corruption of opened files. On a default OS X installation, the machine won't come up again without user interaction, resulting in a loss of all services hosted on this machine for a potentially long time. 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 Markus Hitter 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Am 02.11.2004 um 19:07 schrieb Eric R Johnson: Third, you make sure they are aware of the problem. Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ 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.… Anyone with a dual cpu mac and an extended time to let the app run (while downloading and seeding torrents) will probably see a panic, though. This should be already enough for a bug report. Apple engineers have few time but many Macs so they just have to set up one and wait (while working on other things ;-) ) for the panic to occur. Then, they have sources for the java engine so they can get a more meaningful backtrace. Hopefully, they agree they have to fix something in their engine. I'll see if I can push to get users to submit information to get a rough idea of MTBF or TDR trends. It's pretty hard to get good data out of your average user, though. Azureus always works fine here. Great work. But then, I have single-CPU Macs, only. :-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: Very poor PPP interface performance
by Dave MacLachlan 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Cheers, Dave On Oct 26, 2004, at 13:52, Jean-Edouard BABIN wrote: Hello on 26/10/04 20:55, « Jamie Wood » <eimaj79(a)hotmail.com> wrote: If I run a similar set of tests on Linux, none of the UDP packets are dropped, and sendto returns no errors. Any information about this behavior, and suggestions for possible workarounds/fixes would be greatly appreciated. Thanks, Jamie Wood This email sent to dmaclach(a)adobe.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 Have you tried looking at it with Shark to see if you can figure out where the bottleneck is? You're right, i also already setup a ppp tunnel over ssh With something like pppd pty 'ssh -x -t -e none user@server /usr/sbin/pppd passive noauth 9600' noauth 10.0.0.1:10.0.0.2 and get the same poor performance, and bad ping also :/ I have noticed that the performance of the PPP interface in Mac OS X 10.3.5 is not very good at all. I have written a simple program that brings up a PPP interface, configures it using "ifconfig", and reads incoming packets. If I use another program to repeatedly send UDP packets to my newly-created PPP interface in a tight loop, the sendto() function quickly begins to return a "No buffer space available" error. The program which is reading packets from the PPP interface only reports that about 1% of the packets even get through. In another test, I tried using iperf to measure the throughput of a ssh-pppd tunnel. The most I can get without compression enabled is 1.2 Mbits/sec over a 10 Mbit link. The same test on Linux results in numbers which are close to link speed (one of our tests resulted in ~28 Mbits/sec). One interesting thing I and my collegues have observed is that if a Linux box is used for one end of the ssh-pppd tunnel, we get about 7 Mbits/sec download speed (Linux->Mac OS X) but only about 800 Kbits/sec upload speed (Mac OS X->Linux). All of this seems to indicate that there is some sort of performance problem between the TCP stack and the PPP interface code, which is causing some sort of serious delay when the interface picks up events from the stack. -- Jean-Edouard BABIN Mail: Jeb(a)jeb.com.fr Web : www.Jeb.com.fr IRC : Jeb @ Undernet _______________________________________________ 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/ dmaclach%40adobe.com smime.p7s

1 0

Re: Our FS make Finder hangs on
by yf-263 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3) Gecko/20040922 Hi, gurus, Thanks for your notice/glance :) /yftty yf-263 wrote: Hi, Quinn, Following is output of all bt: (gdb) thread apply all bt Quinn wrote: At 11:37 +0800 22/10/04, yf-263 wrote: (gdb) thread apply all bt _______________________________________________ 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/yfyoufeng%40263.net This email sent to yfyoufeng(a)263.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.… We have getten it fixed. The problem existed from that we didn't tell the node type during _readdir( ) Thread 2 (process 1137 thread 0x5b33): #0 0x90293508 in getdirentries_exhaustive(int, unsigned long*) () #1 0x9028a6d8 in DCBlock_BuildContents(int, DCBlock*) () #2 0x9029aa60 in DCBlock_Create(VolumeInfo const*, unsigned long, DCBlock**) () #3 0x90290550 in GetIndexedDirectoryItem(VolumeInfo*, unsigned long, unsigned long, unsigned long, char*, unsigned char*) () #4 0x9028c3cc in POSIXMount::getattrsbulk(void*, unsigned long, unsigned long*, unsigned long, unsigned long, FSAttributeInfo*, unsigned char*) () #5 0x9028d310 in PBGetCatalogInfoBulkSync () #6 0x902a2104 in FSGetCatalogInfoBulk () #7 0x9091d494 in THFSPlusIterator::Next(THFSPlusRef&) () #8 0x9092b6dc in THFSPlusIterator::First(THFSPlusRef&) () #9 0x9091d40c in THFSPlusIterator::Next(THFSPlusRef&) () #10 0x9091c4d0 in THFSPlusSynchronizer::Next(THFSPlusStore&) () #11 0x9091f724 in THFSPlusStore::CreateChildren(TChildrenList&, TNodeTask*) () #12 0x9092b610 in TNode::CreateChildren(TNode::StInitialCreate const&, TNodeTask*) () #13 0x90923964 in TNodeOpenSyncTask::OpenSyncTaskProc(void*) () #14 0x902c6da8 in PrivateMPEntryPoint () #15 0x900246e8 in _pthread_body () Thread 1 (process 1137 thread 0x40b): #0 0x90014528 in semaphore_wait_trap () #1 0x900021f0 in pthread_mutex_lock () #2 0x90281798 in TSLockMutex () #3 0x90284584 in IsForkedDirectoryItem(VolumeInfo*, unsigned long, char const*, unsigned char*) () #4 0x90288efc in POSIXMount::getattrscore(bool, char const*, char const*, POSIXNode*, unsigned long, FSAttributeInfo*, unsigned long, unsigned char*) () #5 0x90298584 in POSIXMount::_getattrs(unsigned long, char const*, unsigned long, unsigned long, FSAttributeInfo*, unsigned long, unsigned char*) () #6 0x902888c8 in FSMount::getattrs(unsigned long, char const*, unsigned long, unsigned long, FSAttributeInfo*, unsigned long, unsigned char*) () #7 0x9028ba90 in GetFSRefAttributes(FSMount*, FSRefPrivate const*, unsigned long, FSAttributeInfo*, unsigned long, char*) () #8 0x9028bb3c in PBGetCatalogInfoSync () #9 0x902955d8 in FSGetCatalogInfo () #10 0x902a62e0 in AL_fillInternalAliasRecord () #11 0x902a7e88 in AL_fillAlias () #12 0x902c0c90 in AL_newAliasCommon () #13 0x902cf45c in FSNewAlias () #14 0x0005200c in ?? () #15 0x00051dd0 in ?? () #16 0x000251b4 in ?? () #17 0x0002a898 in ?? () #18 0x0003f100 in ?? () #19 0x0003eae8 in ?? () #20 0x0003e92c in ?? () #21 0x0002c36c in ?? () #22 0x000439d0 in ?? () #23 0x00073dc8 in ?? () #24 0x0005c71c in ?? () #25 0x0002734c in ?? () #26 0x0009da18 in ?? () #27 0x0005c128 in ?? () #28 0x0002fe3c in ?? () #29 0x0005ba60 in ?? () #30 0x0005b8b4 in ?? () #31 0x0005b874 in ?? () #32 0x0005b75c in ?? () #33 0x0005b100 in ?? () #34 0x0000f954 in ?? () #35 0x00021990 in ?? () #36 0x0000a078 in ?? () #37 0x0000ca54 in ?? () #38 0x927d2330 in DispatchEventToHandlers () #39 0x927d25a4 in SendEventToEventTargetInternal () #40 0x927e4a34 in SendEventToEventTarget () #41 0x927f31b4 in HandleMouseEventForWindow(OpaqueWindowPtr*, OpaqueEventRef*, unsigned short) () #42 0x927e8ad0 in HandleMouseEvent(OpaqueEventRef*) () #43 0x927e2fd4 in ToolboxEventDispatcherHandler(OpaqueEventHandlerCallRef*, OpaqueEventRef*, void*) () #44 0x927d23ec in DispatchEventToHandlers () #45 0x927d25a4 in SendEventToEventTargetInternal () #46 0x927e4a34 in SendEventToEventTarget () #47 0x927e892c in ToolboxEventDispatcher () #48 0x92805c5c in HLTBEventDispatcher () #49 0x927fbaf8 in RunApplicationEventLoop () #50 0x00035798 in ?? () #51 0x00034090 in ?? () #52 0x0003313c in ?? () #53 0x00032fb0 in ?? () (gdb) Recently we ported a Linux FS to Darwin, and it can works in CLE. But on openning with Finder, it always hangs with following info: Can anyone give us some hints about it here ? This backtrace isn't very useful because it only shows one thread, and that thread is blocked waiting for a mutex that arbitrates access to Core Services File Manager's AppleDouble code. Can you post the results of: S+E This email sent to site_archiver(a)lists.apple.com

1 0

Re: single machine debugging?/advice for noob on OS X panic
by David Gatwood 02 Nov '04

02 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Sometime in the recent past, Eric R Johnson wrote: Subject: Re: single machine debugging?/advice for noob on OS X panic analysis To: darwin-kernel(a)lists.apple.com Message-ID: <1FDEE7BB-2CFA-11D9-ABB3-000A95DC3D44(a)mac.com> Content-Type: text/plain; charset=US-ASCII; format=flowed Andrew, Rohan, Garth, Markus, Thank you for sharing your insights, pointers, etc. I'll need some time to struggle up the learning curve w/OS X, but it promises to be an interesting challenge. Given my available free time, it'll probably be a while before I get back to the list with any findings or follow-up questions. Backtrace explained here: 0x00000006 junk 0x0020B1F4 <kevent+476>: lwz r1,0(r1) 0x002452B40 <unix_syscall+580>: lwz r0,164(r30) 0x00094200 <.L_call_server_syscall_exception> 0xFF8BB9DF junk The link register (which tells where the current stack frame should return) contains: 0x0020B83C <kqueue_scan+784>: cmpwi cr7,r3,0 The program counter contains; 0x0020B7B8 <kqueue_scan+652>: lwz r9,20(r31) Okay. Now that I've reiterated that.... David _______________________________________________ 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 probably missed my reply because I forgot to change the subject line. Whoops. Here's the relevant bit: The address accessed was 0x14. It's a null pointer dereference (or, more accurately, a dereference to some field in a pointer to a struct where the struct pointer itself was NULL, thus the field was at some very small value). So if these (or similar) results are reproducible on more than one machine, that suggests a bug somewhere in the kqueues part of the xnu kernel itself. Please file a bug. :-) Markus, to your point regarding the user-land Java app crashing the kernel... I couldn't agree more. There's been some bluster on the SF thread about writing a virus when they figure out the issue, but I think that's idle talk born of frustration. The potential is pretty ugly, though. The potential should be pretty much zero. It's a null pointer dereference, not a buffer overflow problem. A buffer overflow would, more often than not, cause a jump to an invalid location, resulting in a 0x400 (instruction access). The only way I can think of that this could turn into anything more than a crash would be if the null pointer dereference was a fluke side-effect of a buffer overflow. While there's the most distant possibility that a buffer overflow caused the null dereference, given that the stack backtrace doesn't look like Swiss cheese, I'd say the odds are pretty slim.... To put it more in perspective, the odds of this being helpful to a virus (unless the goal of the virus were to cause a kernel panic) are probably roughly on the same order as the odds of Ralph Nader winning today's election by a landslide. :-) Just my $0.02. As always, my opinions rarely (if ever) reflect the views of my employer. :-) 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 R Johnson 02 Nov '04

02 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Andrew, Rohan, Garth, Markus, Thanks again for the help. --EJ On Nov 02, 2004, at 12:08 AM, Markus Hitter wrote: Am 01.11.2004 um 19:53 schrieb Eric R Johnson: Am 26.08.2004 um 06:31 schrieb Joe Groff: HTH, Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ 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.… Thank you for sharing your insights, pointers, etc. I'll need some time to struggle up the learning curve w/OS X, but it promises to be an interesting challenge. Given my available free time, it'll probably be a while before I get back to the list with any findings or follow-up questions. Markus, to your point regarding the user-land Java app crashing the kernel... I couldn't agree more. There's been some bluster on the SF thread about writing a virus when they figure out the issue, but I think that's idle talk born of frustration. The potential is pretty ugly, though. As far as filing a bug, unfortunately, the Azureus developers and users don't yet have a reproducible test case. It seems to be a frequently occuring intermittent problem with no clear causal event or condition. Anyone with a dual cpu mac and an extended time to let the app run (while downloading and seeding torrents) will probably see a panic, though. I'll see if I can push to get users to submit information to get a rough idea of MTBF or TDR trends. It's pretty hard to get good data out of your average user, though. If anyone wants to talk about this off-list, feel free to contact me directly. (It seems odd to me that a Java app could panic the kernel. But ... A user-land app never ever should be able to cause a kernel panic. Think of a java app embedded into a web page, causing your machine to halt ... If you have a reproducible case, please file it at <http://bugreporter.apple.com/> Unfortunately, my wife has only allowed me one Mac (so far)... One guy on this list obviously was successful in setting up the remote environment on a x86 box: After making some modifications to the Darwin gdb codebase, I was able to compile apple's gdb on a Debian 3.0 system, run it, and connect to the OS X 10.3.4 kernel running on my iBook. Here's a document describing how I did it: http://pknet.com/~joe/apple-cross-gdb.html And here's a patch for gdb-292 to make it compile and run on a Debian 3.0 system, and hopefully other Unix-like systems as well: http://pknet.com/~joe/apple-cross-gdb-292.patch This email sent to site_archiver(a)lists.apple.com

1 0

Re: compiling kernel
by Felix Kronlage 02 Nov '04

02 Nov '04

site_archiver(a)lists.apple.com Delivered-To: Darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Nov 2, 2004, at 07:20, Michael Hogsett wrote: Where is a good starting point for reading documentation about kernel compilation on Darwin 7.5 / MacOSX 10.3.5? <url: http://www.opendarwin.org/en/articles/building_xnu/> I searched around for sources to download from Apple but could not locate kernel sources for Darwin 7.5.1 in an obvious place. Where do I locate the sources? <url: http://darwinsource.opendarwin.org/> or from <url: http://darwinsource.opendarwin.org/> that all should give you a good start. - -fkr - -- 076E 1E87 3E05 1C7F B1A0 8A48 0D31 9BD3 D9AC 74D0 http://fkr.hazardous.org/ | http://opendarwin.org/~fkr/ | FKR-RIPE -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin) iD8DBQFBhz3yDTGb09msdNARAmVVAJ0YE5dh6dL1uoMFvAfqtBz1p6NbkwCfcgQ/ mslCsBfZlI4GKYUkXttd32s= =tHMu -----END PGP SIGNATURE----- _______________________________________________ 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 Markus Hitter 02 Nov '04

02 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Am 01.11.2004 um 19:53 schrieb Eric R Johnson: Am 26.08.2004 um 06:31 schrieb Joe Groff: HTH, Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ 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.… (It seems odd to me that a Java app could panic the kernel. But ... A user-land app never ever should be able to cause a kernel panic. Think of a java app embedded into a web page, causing your machine to halt ... If you have a reproducible case, please file it at <http://bugreporter.apple.com/> Unfortunately, my wife has only allowed me one Mac (so far)... One guy on this list obviously was successful in setting up the remote environment on a x86 box: After making some modifications to the Darwin gdb codebase, I was able to compile apple's gdb on a Debian 3.0 system, run it, and connect to the OS X 10.3.4 kernel running on my iBook. Here's a document describing how I did it: http://pknet.com/~joe/apple-cross-gdb.html And here's a patch for gdb-292 to make it compile and run on a Debian 3.0 system, and hopefully other Unix-like systems as well: http://pknet.com/~joe/apple-cross-gdb-292.patch This email sent to site_archiver(a)lists.apple.com

1 0

compiling kernel
by Michael Hogsett 02 Nov '04

02 Nov '04

site_archiver(a)lists.apple.com Delivered-To: Darwin-kernel(a)lists.apple.com User-agent: Mozilla Thunderbird 0.8 (Macintosh/20040913) Where is a good starting point for reading documentation about kernel compilation on Darwin 7.5 / MacOSX 10.3.5? I am coming primarily from a FreeBSD/x86 environment (with lots of Linux/x86, and some FreeBSD/Alpha and FreeBSD/Sparc64 thrown in for good measure). I searched around for sources to download from Apple but could not locate kernel sources for Darwin 7.5.1 in an obvious place. Where do I locate the sources? Is there a source checkout, a la cvsup on FreeBSD, for keeping in sync with a stable, patched release? Is there anything better than a Mac UI on top of a BSD system with GNU tools on PPC hardware? Thanks in advance, - Mike Hogsett _______________________________________________ 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