March 2009 - Macnetworkprog

programatically modifying com.apple.eap.profiles & com.apple.eap.bindings
by John Anthony Grigutis 31 Mar '09

31 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com Hi list, Any thoughts? -- John Anthony Grigutis Systems Analyst/Programmer/Mac Specialist Apple Certified (ACSA/ACSP/ACTC) Indiana University : UITS : Support : SDD _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com I'm working on a wireless configuration assistant for our university's wireless network and just wanted to confirm something… I've found that I can configure the AirPort interface via the System Configuration framework (SCNetworkConfiguration), but I haven't found a framework to setup the EAP profiles and bindings. I can manipulate the plists through the standard NSDictionary/NSArray/etc. classes and have already written code to do it, but if there is some framework I'm overlooking that would do this better, I'd really like to know. smime.p7s

1 0

Re: how to suppres network interface magic for USB devices
by Quinn 30 Mar '09

30 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com <http://developer.apple.com/bugreporter/> /System/Library/UserEventPlugins/SCMonitor.plugin Share and Enjoy -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Relations, Developer Technical Support, Core OS/Hardware _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… At 15:10 +0200 30/3/09, Pavel Hlavnicka wrote: Is there any way to (temporarily) disable this behavior? I researched this on behalf of another developer and, AFAICT, there's no way to disable this behaviour. I encourage you to file a bug report requesting more developer-level control over this. Any clue what process/system component/driver... is responsible for this? This email sent to site_archiver(a)lists.apple.com

1 0

how to suppres network interface magic for USB devices
by Pavel Hlavnicka 30 Mar '09

30 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(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:content-transfer-encoding; bh=9sjOyhXyQKzJ/mShqn4PMTZ33cMnJVJy89MXeKV4o4E=; b=RRcTq6uNb6us8ItpPfQrQzm5nNcnJee5/uUTPaYLFZbUwgMKIfHB8+iZJOrQ6kbIda PEAsJAOdaJniqyB/k5cIy/ApUp2Idw7/wwlzFAO6MAUMrWLOKCLYOuBPOydhrCqd6Umu jh9f6KGFVADl/t845LyJ2tEEaHANCP19WlpxI= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type :content-transfer-encoding; b=kx01ggmwCU23NdECSXmpgOKO+qSZYd7KM/YaFZIWZfb/7S6e6SI0qJVRRTxfITpPjD HZwMGPafY1zkNkk+/01Wa2dWNDTMBuDx7/kErAaJFb5oc5GT2ixWrNpMQm4u40Q/SHok WXRBgdq0OhwW5uhgo7tUG6q8Zo/mlI7w/Jh5U= Nice day to everyone, I'm working on an application which handles mobile phone serial modem once it's connected using USB cable. The issue I'm dealing with right now is that the system displays notification window (trough UserNotificationCenter.app) telling the user that "New network interfaces have been detected" and challenges the user to open network preferences. Is there any way to (temporarily) disable this behavior? Any clue what process/system component/driver... is responsible for this? Many thanks Pavel H. _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Getting accurate timestamps for received datagrams
by Peter Sichel 27 Mar '09

27 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com On Mar 27, 2009, at 5:15 AM, Thomas Tempelmann wrote: and then read the time stamp as ancillary data. result = recvmsg(mSocket, &msg, flags); Kind Regards, - Peter Sichel Sustainable Softworks _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… So my big question is: Which user-level UDP APIs are as immediate as possible when it comes to delivering incoming datagrams? I think you want the kernel to time stamp your packets as they are received: result = setsockopt(mSocket, SOL_SOCKET, SO_TIMESTAMP, &value, sizeof(int)); where "msg" includes both the data buffer and a control buffer containing the kernel time stamp. I do this to measure round trip times in IPNetMonitorX. I haven't tried this with CFSocket, but I imagine it just tells you there is data available on your socket and let's you do the rest if desired. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Getting accurate timestamps for received datagrams
by ygor＠comcast.net 27 Mar '09

27 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com It sounds like you are trying to reinvent NTP I find it hard to understand why you cannot just use it. I have been in setups -- like an isolated network -- where you just designate one machine as the Time Reference and then all other machines NTP-sync on that one. ----- Original Message ----- From: Thomas Tempelmann <tempelmann(a)gmail.com> To: macnetworkprog(a)lists.apple.com Sent: Fri, 27 Mar 2009 09:15:30 +0000 (UTC) Subject: Getting accurate timestamps for received datagrams Background: I am still working on code to synchronize the clocks of two network-connected computers. Each side tries to learn the time of the peer by repeatedly sending UDP datagrams containing its current time for the other to receive. My software then uses a "smart" algorithm to compare the sender's times with its local receive times, learning the network travel time and guessing the actual time of its peer this way. (Note: I can not use system-provided times that are sync'd over NTP, as these peers will be end-user computers and I cannot mess with their time settings. Also, their firewall might block such ports. All I have control over are is single open UDP/TCP port for my app.) _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Getting accurate timestamps for received datagrams
by Thomas Tempelmann 27 Mar '09

27 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(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=0hfY8D+n7B6O0RT3RN7XJJS2r+mzUUkvTAC3Nf3BRn0=; b=LmFaU++f40ej9BUrTOnaK6ncR/FYurSOGC9Q1MVJu4MRcWZQ8pGtXj9u2PgO9tBZAy 2G/yE1HXXwab8UUN7wkblw6myPWIRHQTHrpfkwti4SqwBIJujuEuTd0bfpbHsob0N8db Kmibj1eQLk+Vl+toqJ1QM+RBtD8fCSThHGewQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=kk+mE9gJ46VIoD0t9LdHRAgdqkLGW3GPcG5m0Vh8wggrA7/Ad37eofseILy+2LY/n1 3USVjRvjGuwjUQxFJ630EW5a9jTiQwRKogAwtRykxaJoJAbvjelY9y3/wKo8k1tqMFJP oh2GJPVhcZBm0Ohly/gs9lntCIk/IDuTa1zgY= _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

GSSAPI on x86_64 dying?
by Edward Thomson 24 Mar '09

24 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com This is 10.5.6 (9G55). I'm hoping someone can shed some light on this? This is the expected behavior - working correctly on i386: % gcc test.c -o test -g -lgssapi_krb5 -arch i386 % file test test: Mach-O executable i386 % ./test Beginning authentication for http @ HOST.DOMAIN.COM Received auth token of length 2571 test.c: #include <stdio.h> #include <stdlib.h> #include <string.h> #include <gssapi.h> #include <krb5.h> int main(int argc, char **argv) { char *target = "http @ HOST.DOMAIN.COM"; char *output = NULL; int outputlen = 0; gss_OID mechanism_oid = &auth_oid_negotiate; gss_ctx_id_t gsscontext = NULL; OM_uint32 statusMajor, statusMinor; gss_buffer_desc target_buffer = GSS_C_EMPTY_BUFFER; gss_buffer_desc output_token = GSS_C_EMPTY_BUFFER; gss_buffer_t input_token_ptr = GSS_C_NO_BUFFER; gss_name_t server; printf("Beginning authentication for %s\n", target); target_buffer.value = (target != NULL) ? (void *) target : ""; target_buffer.length = (target != NULL) ? strlen(target) + 1 : 0; statusMajor = gss_init_sec_context( &statusMinor, GSS_C_NO_CREDENTIAL, &gsscontext, server, mechanism_oid, GSS_C_DELEG_FLAG | GSS_C_MUTUAL_FLAG, GSS_C_INDEFINITE, GSS_C_NO_CHANNEL_BINDINGS, input_token_ptr, NULL, &output_token, NULL, NULL); printf("Received auth token of length %d\n", output_token.length); return; } Sincerely, Edward Thomson <ethomson(a)teamprise.com> Teamprise _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… I'm having difficulties using GSSAPI on x86_64. When I build a simple program using GSSAPI to do SPNEGO, it crashes on the x86_64 architecture. (It works fine on Mac OS for both x86 and ppc architectures, as well as the non-Mac architectures we support.) Sample code is included below. (Note that principal names have spaces added to hopefully avoid email address mangling. If that fails, they are http at HOST.) This is the behavior on x86_64: % gcc test.c -o test -g -lgssapi_krb5 -arch x86_64 % file test test: Mach-O 64-bit executable x86_64 % ./test Beginning authentication for http @ HOST.DOMAIN.COM 2009-03-24 09:28:55.428 test[594:10b] *** NSInvocation: warning: object 0x100079040 of class 'ReplicaFile' does not implement methodSignatureForSelector: -- trouble ahead 2009-03-24 09:28:55.429 test[594:10b] *** NSInvocation: warning: object 0x100079040 of class 'ReplicaFile' does not implement doesNotRecognizeSelector: -- abort Trace/BPT trap Stack trace: Program received signal SIGTRAP, Trace/breakpoint trap. 0x00007fff804e4c15 in ___forwarding___ () (gdb) bt #0 0x00007fff804e4c15 in ___forwarding___ () #1 0x00007fff804e4dd8 in __forwarding_prep_0___ () #2 0x0000000100019e81 in apply_ldap_hints () #3 0x000000010001a53d in get_real_ip_for_loopback () #4 0x00007fff83472b65 in krb5int_locate_server () #5 0x00007fff83474545 in krb5_sendto_kdc () #6 0x00007fff8346aa11 in krb5_send_tgs () #7 0x00007fff8345ba46 in krb5_get_cred_via_tkt () #8 0x00007fff8345a71e in krb5_fwd_tgt_creds () #9 0x00007fff8347b5a9 in make_gss_checksum () #10 0x00007fff83461f65 in krb5_mk_req_extended () #11 0x00007fff834168da in krb5_gss_init_sec_context () #12 0x00007fff83415f67 in gss_init_sec_context () #13 0x00007fff834881bb in spnego_gss_init_sec_context () #14 0x00007fff83415f67 in gss_init_sec_context () #15 0x0000000100000e32 in main (argc=1, argv=0x7fff5fbff470) at test.c: 37 static gss_OID_desc auth_oid_negotiate = { 6, (void *) "\x2b \x06\x01\x05\x05\x02" }; statusMajor = gss_import_name(&statusMinor, &target_buffer, GSS_C_NT_HOSTBASED_SERVICE, &server); if (GSS_ERROR(statusMajor)) { fprintf(stderr, "could not locate principal: %x.%x (%s)", statusMajor, statusMinor, error_message(statusMinor)); return; } if (GSS_ERROR(statusMajor)) { fprintf(stderr, "negotiate failure: %x.%x (%s)\n", statusMajor, statusMinor, error_message(statusMinor)); return; } Please let me know if I can provide more information. Thanks for your time. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Can CFReadStream use proxy set in the system configuration automatically?
by Jens Alfke 19 Mar '09

19 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com On Mar 19, 2009, at 3:57 AM, Quinn wrote: Four, sir! —Jens _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… To get the system proxy settings, you should use one of three different techniques: 4. Use NSURLConnection instead — it does automatically use the system proxy settings (and is a heck of a lot easier to use than CFStream, too.) This email sent to site_archiver(a)lists.apple.com

1 0

Re: Can CFReadStream use proxy set in the system configuration automatically?
by Quinn 19 Mar '09

19 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com Share and Enjoy -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Relations, Developer Technical Support, Core OS/Hardware _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… At 13:27 +0800 19/3/09, µÀéuº· wrote: Some one tell me that when using CFReadStream to query http, I don't have to care the proxy set in system configuration. Because CFReadStream can use the proxy set in system configuration. But I can't find the related description in the document. Can anyone tell me the fact? CFStream will not automatically pick up the proxy settings. If you want to use the system-defined proxy settings, you have to get the settings from the system (more on this below) and apply them using various stream properties (for example, kCFStreamPropertySOCKSProxy and kCFStreamPropertyHTTPProxy). To get the system proxy settings, you should use one of three different techniques: o In Mac OS X 10.5 and later, or iPhone OS, the best approach is to use CFNetworkCopyProxiesForURL. o If you don't have a URL handy, then on iPhone OS you can call CFNetworkCopySystemProxySettings to get the raw proxy settings. o On Mac OS X, prior to 10.5, or if you don't have a URL handy, you can use SCDynamicStoreCopyProxies. This email sent to site_archiver(a)lists.apple.com

1 0

Socket recv problems in real-time context : lock or priority inversion issues?
by Stéphane Letz 19 Mar '09

19 Mar '09

site_archiver(a)lists.apple.com Delivered-To: macnetworkprog(a)lists.apple.com Hi, 0.3% 28.9% mach_kernel recvfrom_nocancel 0.3% 27.5% mach_kernel socketpair 4.0% 24.4% mach_kernel soreceive 0.3% 10.6% mach_kernel sbwait 0.0% 10.0% mach_kernel msleep 0.0% 10.0% mach_kernel uiomove 0.0% 9.7% mach_kernel lck_mtx_sleep_deadline 0.0% 7.7% mach_kernel thread_block 0.0% 7.7% mach_kernel thread_block_reason 7.7% 7.7% mach_kernel ml_set_interrupts_enabled 0.0% 1.4% mach_kernel assert_wait_deadline 0.0% 1.4% mach_kernel wait_queues 1.4% 1.4% mach_kernel ml_set_interrupts_enabled 0.6% 0.6% mach_kernel lck_mtx_lock 0.3% 0.3% mach_kernel thread_block 0.3% 0.3% mach_kernel _mutex_assert Thanks. Stephane Letz _______________________________________________ Do not post admin requests to the list. They will be ignored. Macnetworkprog mailing list (Macnetworkprog(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macnetworkprog/site_archiver%40lists… We are using UDP sockets to transmit audio in real-time between 2 machines. Basically the "master" machine is synchronized on the audio interrupt and send/recv audio packets on the "slave" machine. The "slave" machine is synchronized on incoming packets (that is waiting for an incoming packet on the socket to start it's audio cycle). The "slave" machine reception thread is set-up in real-time mode (using thread_policy_set /THREAD_TIME_CONSTRAINT_POLICY kind of API). It works quite reliably until we start to load the slave machine (like starting a big compilation). In this case the RT slave thread does not "wake up" the the right time, and we have quite big latencies (up to 50 millisecond latency measured on a MacPro Intel 4 cores 10.5.6 when the audio cycle is supposed to be 5 millisecond). More precisely it seems the socket recv operation blocks a lot of time. We configure the socket with SO_SNDBUF/SO_RCVBUF and possibly SO_RCVTIMEO kind of options. The same code works without problem on Linux or Solaris. Using Shark "system trace", we were able to measure abnormal latencies on the recv function (20 ms in a given test), more precisely in the "recvfrom_nocancel" one. Using the Shark "Time profile" we see the following call stack for "recvfrom_nocancel" in the kernel: So it may be a lock or priority inversion of some functions used in kernel mode? Is this a known problem for sockets on OSX? It this trace enough to do a bug report? This email sent to site_archiver(a)lists.apple.com

1 0