February 2006 - Darwin-dev - lists.apple.com

Reading CTS from a USB serial adaptor
by Mark Gilbert 11 Feb '06

11 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Folks. Any suggestions would be welcomed. Thanks Mark Gilbert -- mark(a)gallery.co.uk Tel: +44 208 340 5677 fax: +44 870 055 7790 http://www.gallery.co.uk New Product ! - MetaCorder Location Sound Field Recorder http://www.metacorder.info _______________________________________________ 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 have a special USB serial adaptor using the FTDI RS232 chip. The device has a sensor attached to the CTS line of the chip, and I need to be able to sense the status of the CTS line from my app. Ideally I would be able to setup a blocking call to only return when the status changes. I use Posix to open and configure the port, then I have tried ioctl using TIOCMGET to read the CTS status, but TIOCMGET always returns the same thing (102 decimal) regardless of the value on the CTS line. This denotes CTS is high. Another developer using the same box on Windows is able to read the CTS no problem, so I know the device is OK. Also, I see that in Linux there is a call TIOCMIWAIT which appears to do the blocking sense, but I can't find this in any Mac headers. This would be desirable but not essential. Could anyone suggest anything else to try ? I am not having any luck with the TIOCMGET which should be working according to the Apple docs I have found ('WorkingwithSerial.pdf') This email sent to site_archiver(a)lists.apple.com

1 0

Re: Any plan to have something similar to this for Darwin?
by AgentM 10 Feb '06

10 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -M |-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|- AgentM agentm(a)themactionfaction.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… I use http://fxr.watson.org to surf Darwin sources. It's not as fancy, but it works great and has excellent linking between sources- a bunch of other OS sources are indexed there, too. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Any plan to have something similar to this for Darwin?
by Andrew Gallatin 10 Feb '06

10 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Stephane Sudre writes: > http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/inet/tcp/ > tcp.c > > CVSWeb is good but IMHO this kind of feature is better when you want to > browse source code. It strikes me as a symbol of the difference in spirit of the two projects. I don't know if you are on any of the Open Solaris mailing lists, but, unlike Darwin, Open Solaris is really open to community feedback. Apple could really learn a lot from how Sun deal with Open Solaris.. Sun engineers are subscribed Open Solaris mailing lists and are actually allowed to discuss (*gasp*) future development directions openly, and they actively solicit feedback from the community. They make snapshots of the future version of Solaris available for free download. Contrast this with Apple, who charges $500 for the ADC membership required to be able to download future versions of Darwin, and will not let you talk about them in a public forum. There are interesting discussions going on now on their networking discuss list which are a good example of the openness I'm talking about. See http://www.opensolaris.org/jive/forum.jspa?forumID=3 Drew _______________________________________________ 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

Any plan to have something similar to this for Darwin?
by Stephane Sudre 10 Feb '06

10 Feb '06

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… http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/inet/tcp/ tcp.c CVSWeb is good but IMHO this kind of feature is better when you want to browse source code. This email sent to site_archiver(a)lists.apple.com

1 0

Re: question about interposing routines in libSystem.B.dylib
by Kevin Van Vechten 09 Feb '06

09 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com See the darwintrace library for an example: - Kevin On Feb 9, 2006, at 4:50 PM, Will Stockwell wrote: Hi, I'm using the following simple code for a dry run: // *** BEGIN CODE *** #include <stdio.h> #include <dlfcn.h> void *malloc(size_t size) { static void * (*func)(); printf ("malloc(%d) called\n", size); if(!func) func = (void *(*)()) dlsym(RTLD_NEXT, "malloc"); return(func(size)); } // *** END CODE *** And building the dylib as follows: $ gcc -dynamiclib -o interpose.dylib -fPIC interpose.c Your help is greatly appreicated, _______________________________________________ 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… There are some interesting constraints, such as when you override one symbol in a .o, you need to override them all. The syscalls you're interested should be relatively easy to intercept for a couple of reasons--first, each is in their own .o, and second, you don't need to use RTLD_NEXT, you can simply invoke the syscall directly. Of course invoking syscalls directly isn't recommended for a production system (since that ABI is not guaranteed), so use with caution. <http://cvs.opendarwin.org/cgi-bin/cvsweb.cgi/proj/darwinbuild/ darwintrace/darwintrace.c?rev=1.13&content-type=text/x-cvsweb-markup> I'm attempting to develop an interposition library to catch a few system calls on the application side (i.e., interpose the system call's system library analogue). It seems, however, whenever I attempt to build such a library that interposes, for example, malloc, I have a dependency on libSystem.B.dylib. This is the library in which the malloc system library routine is stored. otool also tells me dlsym (needed to do interposition) is contained in libSystem.B.dylib. According to Apple documentation, it is not possible to interpose dependent libraries and it further seems I can't do interposition without this particular library. Of course, this turns up the following messages about duplicate malloc symbols: ld: warning multiple definitions of symbol _malloc /var/tmp//cczzOOvT.o definition of _malloc in section (__TEXT,__text) /usr/lib/gcc/powerpc-apple -darwin8/4.0.0/../../../libSystem.dylib(malloc.So) definition of _malloc And attemptnig to have it loaded via DYLD_INSERT_LIBRARIES results in a seg fault (not a suprise). My sense is that it's just plain silly that it's not possible to do something like this. Any ideas out there? It's definitely possible to interpose malloc, open, etc. on other systems. Does anyone know how I can achieve this under darwin? Googling around hasn't turned up much. And, if it matters, the calls I'm most interested in interposing are open, close, read, write, connect, accept, pipe, and fork. Will _______________________________________________ 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/kevin% 40opendarwin.org This email sent to kevin(a)opendarwin.org This email sent to site_archiver(a)lists.apple.com

1 0

question about interposing routines in libSystem.B.dylib
by Will Stockwell 09 Feb '06

09 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type; b=HtehMOFmgVg0S340GbVNcnT81ir9qChw8P27tsJglu4mOKpqWEkIuOkmBOYVTssd090cRblPwsv1/c10HMqTA9dasL7qmfkQImN/65CfPpA0QqGsUGwvlPj4anaihg50VfURCUjT9555apmKhbYe9qzftOKyUmOQa67sIl7cp1E= _______________________________________________ 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: Gloabal object initalisation on MAC-OSX 10.3.3
by Steve Checkoway 07 Feb '06

07 Feb '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Feb 7, 2006, at 9:23 PM, Gaurav Kumar wrote: Steve, The problem was with translation unit.My global variable was not initialized because of dependency factor. Actually the order of initialization was a problem. class A{ public: A(){ } fun(){ } }; A& initfun() { static A a; return a; } In call to class A functions use this statement: That looks good to me. You could also add a global variable A& myA = initfun(); - Steve _______________________________________________ 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 I have solved the problem by using Singleton object with following code snippet initfun().fun(); And then just use myA anywhere that would be accessed via a global initializer. smime.p7s

1 0

Gloabal object initalisation on MAC-OSX 10.3.3
by Gaurav Kumar 07 Feb '06

07 Feb '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Organization: Steve, The problem was with translation unit.My global variable was not initialized because of dependency factor. Actually the order of initialization was a problem. I have solved the problem by using Singleton object with following code snippet class A{ public: A(){ } fun(){ } }; A& initfun() { static A a; return a; } In call to class A functions use this statement: initfun().fun(); This will solve the order of initialization of global variables. Gaurav _______________________________________________ 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: Global and Static Variable initialisation Problem on MAC-OSX 10.3.3 & GCC 3.3 ........
by Steve Checkoway 06 Feb '06

06 Feb '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Feb 6, 2006, at 10:20 AM, Cliff Russell wrote: Gaurav, - Steve _______________________________________________ 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 Generally getting global and static variables to work correctly is relatively straight forward. Can you post some code snippets showing the general organization of the classes and object you are having trouble with, and perhaps someone on the list can help. The only things that I can think of that could cause the problems your seeing are, either a misunderstanding of how global and static variables work. Or, you in your code you have one or more global or static variables that depend on other global/static variables being initialized. Some compilers allow you to control the order that globals and statics are initialized, I don't know if gcc is one of those compilers. Typically, the safest route is to write your code in such a way that the order that global and statics are initialized isn't important. C (and C++) offer no guarantees that about the order that global variables are initialized other than constant init will happen before nontrivial ctors (for c++). Relying on the order is a mistake and your code _will_ break. smime.p7s

1 0

Re: Global and Static Variable initialisation Problem on MAC-OSX 10.3.3 & GCC 3.3 ........
by Cliff Russell 06 Feb '06

06 Feb '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Gaurav, On 6-Feb-06, at 5:58 AM, Gaurav Kumar wrote: hi all I am using GCC 3.3 for compiling my process on MAC-OS X 10.3.3. I am constructing a global static variable i.e object of a class in another class file. I have included that class header. The error is Abort Trap after the object construction. Anywhere in my project if i make global object of that class, the execution don't go beyond that. That class object depends upon some other global variables. If we explicitly initialize those variables then also it throws error. The global variable calls its constructor properly. But the pointer of the global variable works when we initialize it in main i.e Local pointer. Thanks Gaurav _______________________________________________ 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… Generally getting global and static variables to work correctly is relatively straight forward. Can you post some code snippets showing the general organization of the classes and object you are having trouble with, and perhaps someone on the list can help. The only things that I can think of that could cause the problems your seeing are, either a misunderstanding of how global and static variables work. Or, you in your code you have one or more global or static variables that depend on other global/static variables being initialized. Some compilers allow you to control the order that globals and statics are initialized, I don't know if gcc is one of those compilers. Typically, the safest route is to write your code in such a way that the order that global and statics are initialized isn't important. Anyway, you can probably get some more specific help if you provide a few more details. Can you suggest what is real problem in initializing global variable or the in which i should proceed to solve the bug. if you require any more information do mail me. _______________________________________________ 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/cliff_russell% 40atimi.com This email sent to cliff_russell(a)atimi.com This email sent to site_archiver(a)lists.apple.com

1 0