April 2009 - Darwin-dev - lists.apple.com

Re: Force a crash on SIGABRT
by James Walker 14 Apr '09

14 Apr '09

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com User-agent: Thunderbird 2.0.0.21 (Macintosh/20090302) Just an untested idea... try kill(getpid(), SIGILL);? On Apr 13, 2009, at 3:14 PM, James Walker wrote: static void AbortHandler( int /* inSignal */ ) { long* badAddr = 0; *badAddr = 7; } -- James W. Walker, Innoventive Software LLC <http://www.frameforge3d.com/> _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Kevin Van Vechten wrote: When my app receives SIGABRT, I want to force some kind of crash that will generate a crash report. I tried installing a SIGABRT signal handler like this: On Tiger, it partly works: There is a crash log, but the app keeps running in a hung state. On Leopard, there's just a hang, no crash report. It turns out that on Leopard, I don't need the handler at all, just calling abort() causes the "unexpectedly quit" dialog to appear, and the app to go away. On Tiger, your suggestion is no better or worse than what I had before... the dialog appears, but the app hangs around. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Force a crash on SIGABRT
by Kevin Van Vechten 14 Apr '09

14 Apr '09

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Just an untested idea... try kill(getpid(), SIGILL);? On Apr 13, 2009, at 3:14 PM, James Walker wrote: static void AbortHandler( int /* inSignal */ ) { long* badAddr = 0; *badAddr = 7; } _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… When my app receives SIGABRT, I want to force some kind of crash that will generate a crash report. I tried installing a SIGABRT signal handler like this: On Tiger, it partly works: There is a crash log, but the app keeps running in a hung state. On Leopard, there's just a hang, no crash report. This email sent to site_archiver(a)lists.apple.com

1 0

Force a crash on SIGABRT
by James Walker 13 Apr '09

13 Apr '09

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com User-agent: Thunderbird 2.0.0.21 (Macintosh/20090302) static void AbortHandler( int /* inSignal */ ) { long* badAddr = 0; *badAddr = 7; } -- James W. Walker, Innoventive Software LLC <http://www.frameforge3d.com/> _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… When my app receives SIGABRT, I want to force some kind of crash that will generate a crash report. I tried installing a SIGABRT signal handler like this: On Tiger, it partly works: There is a crash log, but the app keeps running in a hung state. On Leopard, there's just a hang, no crash report. This email sent to site_archiver(a)lists.apple.com

1 0

Force a crash on SIGABRT?
by James Walker 13 Apr '09

13 Apr '09

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com User-agent: Thunderbird 2.0.0.21 (Macintosh/20090302) static void AbortHandler( int /* inSignal */ ) { long* badAddr = 0; *badAddr = 7; } -- James W. Walker, Innoventive Software LLC <http://www.frameforge3d.com/> _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… When my app gets SIGABRT, I'd like to get an access violation or something that produces a crash log with a stack crawl. I tried saying signal( SIGABRT, AbortHandler ), where AbortHandler is defined as On Tiger, it sort of works... the crash report window comes up, though the app remains running in a hung state. On Leopard, it just hangs. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Run daemon from installer as root
by Rakesh Singhal 13 Apr '09

13 Apr '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:date:message-id:subject :from:to:content-type; bh=i2KC5/X+UOwfF5CBiapt2URb8pfHUxKP+DHe4q1m9+g=; b=P43qq4CvBfLU7qjTRTKXxqkcI4bPaA1if+RXjpGF8ESM7UErXZI0Bf5bfUPF8EmV+G g+nq5AIqpuI/iqlu+qnC5v6iu1Ljr0aCzylJfXPZPub/L/YGye2V10Vlapw/zBxSbAn1 vCwhqAY4hUhyMcox/x8CLCzztrSsPHKcBwwSQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=EiCYH2lyhkpo4R/ymBwJTXRtGDWTwFRg1/2SaQvkI8D1d27Vf5eSs903wZwiLtUGMg PWO0ggY7inC1xq1Jx4oW3ykDtyiH4Mmx5KgHUPkhUDp4MNzH9djcjxlr/YnCYeBOtzK9 wG58ud3qeXWkgFiM1P/7TnfplU89T6AK78GY8= _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: kext symbol exports?
by Ryan McGann 12 Apr '09

12 Apr '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Is there any advantage to limiting the export symbol list for a kext? Ryan _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I worry that the kernel is a single namespace with a high potential for symbol collision. Does using a symbol export list help to mitigate the potential for symbol collision? I would assume it has the same positive/negative affects it does for userspace—limits namespace collision, reduces binary size, reduces time spent in runtime linker, etc. The speed impact may be less/more for kernel extensions because it's a totally different linker; I haven't done any testing to that affect. But for userspace applications it definitely helps. Beware though, I have a lot of problems stripping my IOKit kernel extensions. Things might be better on 10.5, but on 10.4 stripping our kernel extension would cause some pretty weird problems—from the small like kextd crashing randomly, to the very big, like the computer panicing early on in the boot cycle. I eventually got it working, but not 100% correct. Theoretically one should only have to leave in the C+ + vtable and class information for the IOKit classes, plus implementation of the classes themselves, but that turned out to be wrong. I had to leave in some random global variables (because ld complained about relocation entries, which I never fully understood) and some random C functions contained in a static library we link in (because the computer would just panic if I stripped them out, which has me totally baffled). I still recommend stripping, just be careful. This email sent to site_archiver(a)lists.apple.com

1 0

Re: kext symbol exports?
by Michael Smith 11 Apr '09

11 Apr '09

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

1 0

kext symbol exports?
by Duane Murphy 11 Apr '09

11 Apr '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Is there any advantage to limiting the export symbol list for a kext? I worry that the kernel is a single namespace with a high potential for symbol collision. Does using a symbol export list help to mitigate the potential for symbol collision? ...Duane _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOBufferMemoryDescriptor::getVirtualSegment Deprecated and Missing?
by Duane Murphy 11 Apr '09

11 Apr '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com --- At Sat, 11 Apr 2009 07:13:16 -0400, Chas Williams (CONTRACTOR) wrote: >In message <20090410204924.1368388445 >@mini.internal.murphyslogic.com>,"Duane Murphy" writes: >>We are updating our drivers and moving them to 10.5 from 10.4. I am >>getting a link/load error when loading the kext. >> >>Undefined Symbols: >>IOBufferMemoryDescriptor::getVirtualSegment > >sure you dont want to use getBytesNoCopy()? getVirtualSegment() is >basically the same except it optionally returned the remaining length >of the segment. Well, that's the interesting part; I'm _not_ calling getVirtualSegment(); not using it at all. Turns out I was subclassing IOBufferMemoryDescriptor. I haven't looked at the implementation yet (I should do that), but I suspect that the way it is setup, getVirtualSegment() is not part of the vtable for the class, causing the vtable to not be able to be filled for a subclass. Somehow, this manages to work for using IOBufferMemoryDescriptor, but not for subclassing. I've changed to sublcassing IOMemoryDescriptor, the root class of memory management. The subclass keeps an IOBufferMemoryDescriptor as a member and delegates the functions to that. I'm still having trouble with this, so I suspect that I need to override and delegate more functions. ...Duane _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOBufferMemoryDescriptor::getVirtualSegment Deprecated and Missing?
by Chas Williams (CONTRACTOR) 11 Apr '09

11 Apr '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com In message <20090410204924.1368388445(a)mini.internal.murphyslogic.com>,"Duane Murphy" writes: >We are updating our drivers and moving them to 10.5 from 10.4. I am >getting a link/load error when loading the kext. > >Undefined Symbols: >IOBufferMemoryDescriptor::getVirtualSegment sure you dont want to use getBytesNoCopy()? getVirtualSegment() is basically the same except it optionally returned the remaining length of the segment. you could subclass IOBufferMemoryDescriptor and just implement this yourself if you really needed this functionality. you should know the length you allocated. /* DEPRECATED */ void * IOBufferMemoryDescriptor::getVirtualSegment(IOByteCount offset, /* DEPRECATED */ IOByteCount * lengthOfSegment) { void * bytes = getBytesNoCopy(offset, 0); if (bytes && lengthOfSegment) *lengthOfSegment = _length - offset; return bytes; } _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0