November 2004 - Darwin-kernel

Re: why sometimes can't I find the panic.log file when kernel panic occurred?
by Garth Cummings 04 Nov '04

04 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi Jason, On Nov 4, 2004, at 12:57 AM, Jason Geng wrote: Normally yes. --gc __________________________________________________________________ Garth Cummings Apple Developer Technical Support gcummings(a)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 Thanks for Garth Cummings' kind help, in the email > Re: single machine debugging?/advice for noob on OS X panic analysis > Take a look at Technical Note 2063 > <http://developer.apple.com/technotes/tn2002/tn2063.html>, titled > "Understanding and Debugging Kernel Panics." I have read the article carefully, and it is quite helpful. But my question is: Sometimes after kernel panic occurred, I could find the 'panic.log' file in the directory '/Library/Logs', however, sometimes I couldn't. Why? There are a couple of reasons for this. One is described in the technote: if the debug flags are set in boot-args you won't get a panic log. But the other reason is a known bug in 10.2.x where certain long panic dumps would not be saved properly. This was fixed in Panther. The workaround is to turn on the debug flags so that you get the panic log dumped to the screen as was done prior to 10.2. My kernel version is MAC OS 10.2.8, kern.version = Darwin Kernel Version 6.8: Wed Sep 10 15:20:55 PDT 2003; root:xnu/xnu-344.49.obj~2/RELEASE_PPC I guess those unlogged kernel panics were caused by handlers of my driver dealing with DMA operation or hardware interrupt. In these situations, can darwin save the panic.log correctly? http://developer.apple.com/technicalsupport smime.p7s

1 0

why sometimes can't I find the panic.log file when kernel panic occurred?
by Jason Geng 04 Nov '04

04 Nov '04

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: opening files in the kernel
by Justin Walker 04 Nov '04

04 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Nov 3, 2004, at 19:17, Roger Seagle wrote: There is probably a better way to do what you want. Regards, Justin _______________________________________________ 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 am new to the process of kernel hacking and am working on a networking stack. But with that aside, I am wondering what is the best way to open files inside the kernel and read the contents. I started going down the route of using VOP functions, but I don't know if this is the right avenue. If someone could point me to some code or documentation on this subject, I would greatly appreciate it. There is no approved way of doing this. VOP functions are definitely the wrong way to go (too low level), and most 'user mode' kernel functions are set up to assume that the request comes from user mode. -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | It's not whether you win or lose... | It's whether *I* win or lose. *--------------------------------------*-------------------------------* This email sent to site_archiver(a)lists.apple.com

1 0

opening files in the kernel
by Roger Seagle 04 Nov '04

04 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hey guys, roger _______________________________________________ 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 am new to the process of kernel hacking and am working on a networking stack. But with that aside, I am wondering what is the best way to open files inside the kernel and read the contents. I started going down the route of using VOP functions, but I don't know if this is the right avenue. If someone could point me to some code or documentation on this subject, I would greatly appreciate it. 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 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Nov 3, 2004, at 12:10 AM, Markus Hitter wrote: Am 02.11.2004 um 21:57 schrieb David Gatwood: Sometime in the recent past, Eric R Johnson wrote: 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.… 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. What else do you want to get the potential above "zero"? Sorry, should have been more clear. I meant the potential for using such a bug to help a virus propagate. This email sent to site_archiver(a)lists.apple.com

1 0

panic - probable bug in kqueues (surfaced by Azureus p2p)
by Eric R Johnson 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com FYI - this has been filed as Bug ID# 3864002. --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.… 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. :-) This email sent to site_archiver(a)lists.apple.com

1 0

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

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Andrew Gallatin writes: > > Markus Hitter writes: > > 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. > > > > You mention "default install" -- is it even possible to get a mac to > autoreboot after a panic w/o rebuilding a kernel from source? I > thought there was a while(1) at the end of panic()... Hardware watch-dog (which is not part of "default installation") can do this. > > Thanks, > > Drew Nikita. _______________________________________________ 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
by Andrew Gallatin 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Markus Hitter writes: > 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. > You mention "default install" -- is it even possible to get a mac to autoreboot after a panic w/o rebuilding a kernel from source? I thought there was a while(1) at the end of panic()... Thanks, Drew _______________________________________________ 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

New Partition Scheme Driver (not a filter scheme)
by Developer 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Importance: Normal Hi, In addition to the Filter Scheme, I would like to check a possibility to develop a New Partition Scheme Driver (not a filter scheme). Where can I get the Sample Source Code and Documentation to develop a New Partition Scheme Driver (not a filter scheme)? Thanks! D _______________________________________________ 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

Sample Filter Scheme (read doesn't work)
by Developer 03 Nov '04

03 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Importance: Normal Hi, I'm trying to implement Filter Scheme using "Sample Filter Scheme" from the Apple's documentation. My driver is successfully loaded, creating a Volume for the partition I created with hdiutil. At this stage I want just to process a content of a buffer in my driver's read()/write() methods. In both methods I allocate a temporary char *myBuff of IOMemoryDescriptor* buffer->getLength() size, pass it to byffer->readBytes() and then send myBuff to IOLog() method that prints it to /var/log/system.log. It works for getProvider()->write() but it shows also the data which is not a content of a file. It DOESN'T work at all for getProvider()->read(). QUESTION 1: Why getProvider()->read() doesn't bring the file data at all? ------------------------------------------------------------------------- It seems that getProvider()->read() doesn't work, since calling buffer->readBytes() in read() before (for debug only) and after getProvider()->read() gives the same junk in myBuff. What is wrong with my read() method? QUESTION 2: How can I know what data is read/written with getProvider()->read/write() ? ---------------------------------------------------------------------------- ----------- IOLog() shows that read/write() receives some strange data which is not a file content. How can I know what data contains buffer? Following is the source code of my read() and write() methods. Any help will be appreciated. Thanks in advance! // - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - void com_MyTutorial_driver_MyFilterscheme::read(IOService* /* client */, UInt64 byteStart, IOMemoryDescriptor* buffer, IOStorageCompletion completion) { IOByteCount readBytes; IOByteCount bufLength = buffer->getLength(); char *myBuff = new char[bufLength]; readBytes = buffer->readBytes(0, myBuff, bufLength); IOLog("MyFilterscheme::read BEFORE*** readBytes = %lu byteStart = %lu bufLength = %lu buffer = %s ***\n", (long)readBytes, (long)byteStart, (long)bufLength, myBuff); getProvider()->read(this, byteStart, buffer, completion); readBytes = buffer->readBytes(0, myBuff, bufLength); for (IOByteCount i = 0; i < bufLength - 1; i++) { if (myBuff [i] == 't') if (myBuff [i+1] == 'e') IOLog("MyFilterscheme::read ***FOUND*** buffer @ %lu = %s ***\n", (long)i, myBuff + i); } IOLog("MyFilterscheme::read AFTER *** readBytes = %lu byteStart = %lu bufLength = %lu buffer = %s ***\n", (long)readBytes, (long)byteStart, (long)bufLength, myBuff); delete[] myBuff; } // - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - void com_MyTutorial_driver_MyFilterscheme::write(IOService* /* client */, UInt64 byteStart, IOMemoryDescriptor* buffer, IOStorageCompletion completion) { IOByteCount readBytes; IOByteCount bufLength = buffer->getLength(); char *myBuff = new char[bufLength]; readBytes = buffer->readBytes(0, myBuff, bufLength); IOLog("MyFilterscheme::write readBytes = %lu byteStart = %lu bufLength = %lu buffer = %s ***\n", (long)readBytes, (long)byteStart, (long)bufLength, myBuff); getProvider()->write(this, byteStart, buffer, completion); delete[] myBuff; } _______________________________________________ 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