User Profile

I ran a test code with ThreadSanitizer, it throws a race condition in the block of xpcconnectionseteventhandler even I set a sync queue to the connection. xpc_connection_set_event_handler(conn, ^(xpc_object_t event) { xpc_type_t type = xpc_get_type(event); if (xpc_get_type(event) == XPC_TYPE_DICTIONARY) { 		// race condition here 		// WARNING: ThreadSanitizer: data race // ... } else { if (event == XPCERRORCONNECTIONINVALID) { // Error indicates the peer has closed the connection. // Tear down any associated data structures. } else { // Error indicates that service will terminate soon. // Flush all buffers, finish all work, etc. } xpcrelease(peer); } }); https://developer.apple.com/documentation/xpc/1448786-xpc_connection_set_target_queue?language=objc: "the XPC runtime guarantees that, when the target queue is a serial queue, the event handler block will execute synchronously with respect to other blocks submitted to that same queue. When the target queue is a concurrent queue, the event handler block may run concurrently with other blocks submitted to that queue, but it will never run concurrently with other invocations of itself for the same connection" looks like there should have no race condition according to the above description, but why I encounter the race condition issue reported by ThreadSanitizer? Does the block of xpcconnectionseteventhandler need sync to avoid race condition?

Hi Experts, I knew there is LSEnvironment for defining environment variables to be set before launching App. e.g. <key>LSEnvironment</key> <dict> <key>PATH</key> <string>/Users/flori/.rvm/gems/ruby-1.9.3-p362/bin:/Users/flori/.rvm/gems/ruby-1.9.3-p362@global/bin:/Users/flori/.rvm/rubies/ruby-1.9.3-p326/bin:/Users/flori/.rvm/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:</string> </dict> How about system extension? How can I set some environment variables before running? Thanks a lot.

Process: sysextd [633] Path: /System/Library/Frameworks/SystemExtensions.framework/Versions/A/Helpers/sysextd Identifier: sysextd Version: 79.40.4 Code Type: X86-64 (Native) Parent Process: launchd [1] Responsible: sysextd [633] User ID: 0 Date/Time: 2020-11-21 19:06:59.948 +0800 OS Version: macOS 11.0.1 (20B50) Report Version: 12 Anonymous UUID: E7E0F5B9-898B-B40D-21A3-AF796FC8960A Time Awake Since Boot: 120 seconds System Integrity Protection: disabled Crashed Thread: 1 Dispatch queue: sysextd.extensionmanager Exception Type: EXCBADINSTRUCTION (SIGILL) Exception Codes: 0x0000000000000001, 0x0000000000000000 Exception Note: EXCCORPSENOTIFY Termination Signal: Illegal instruction: 4 Termination Reason: Namespace SIGNAL, Code 0x4 Terminating Process: exc handler [633] Application Specific Information: dyld3 mode Fatal error: 'try!' expression unexpectedly raised an error: sysextd.ExtensionFinishUninstallingAction.actionErrors.outstandingreferences: file sysextd/extensionmanager.swift, line 422 Thread 0:: Dispatch queue: com.apple.main-thread 0 libsystemkernel.dylib 0x00007fff203185ce _ulockwait + 10 1 libdispatch.dylib 0x00007fff201a0f98 dlockwait + 44 2 libdispatch.dylib 0x00007fff201a122c dispatchgroupwaitslow + 49 3 libdispatch.dylib 0x00007fff201a3b37 dispatchblockwait + 212 4 sysextd 0x0000000104ff8c3f 0x104ff0000 + 35903 5 sysextd 0x0000000104ffb95f 0x104ff0000 + 47455 6 sysextd 0x0000000104ffb199 0x104ff0000 + 45465 7 libdyld.dylib 0x00007fff20366631 start + 1 Thread 1 Crashed:: Dispatch queue: sysextd.extensionmanager 0 libswiftCore.dylib 0x00007fff2c72f124 assertionFailure(::file:line:flags:) + 532 1 libswiftCore.dylib 0x00007fff2c784629 swiftunexpectedError + 585 2 sysextd 0x0000000105009168 0x104ff0000 + 102760 3 sysextd 0x000000010501047f 0x104ff0000 + 132223 4 sysextd 0x0000000105012018 0x104ff0000 + 139288 5 sysextd 0x0000000104ff8c86 0x104ff0000 + 35974 6 libdispatch.dylib 0x00007fff201ad1d5 dispatchblockasyncinvoke2 + 83 7 libdispatch.dylib 0x00007fff201a07c7 dispatchclientcallout + 8 8 libdispatch.dylib 0x00007fff201a65fe dispatchlaneserialdrain + 606 9 libdispatch.dylib 0x00007fff201a70cb dispatchlaneinvoke + 375 10 libdispatch.dylib 0x00007fff201b0c5d dispatchworkloopworkerthread + 819 11 libsystempthread.dylib 0x00007fff20348499 pthreadwqthread + 314 12 libsystempthread.dylib 0x00007fff20347467 startwqthread + 15 Thread 1 crashed with X86 Thread State (64-bit): rax: 0x0000000200000003 rbx: 0x0000000105066360 rcx: 0xfffffffe00000000 rdx: 0x0000000000000003 rdi: 0x00007fd5ddd06f08 rsi: 0x00000000000bcfa0 rbp: 0x00007000082ed2f0 rsp: 0x00007000082ed2a0 r8: 0x0000000000000110 r9: 0x000000000000001b r10: 0x00007fd5ddc00000 r11: 0x0000000000000000 r12: 0x0000000000000000 r13: 0x000000000000000b r14: 0x0000000000000001 r15: 0x00007fff2cab8944 rip: 0x00007fff2c72f124 rfl: 0x0000000000010246 cr2: 0x00007fff80068470 Logical CPU: 0 Error Code: 0x00000000 Trap Number: 6 Thread 1 instruction stream: 41 0f b6 c7 48 89 03 49-c7 46 10 01 00 00 00 41 A...H..I.F.....A bf 01 00 00 00 48 8b 7d-d0 e9 19 01 00 00 48 89 .....H.}......H. 4d c0 4c 89 65 c8 4c 89-e7 41 89 f4 e8 5b 8a 34 M.L.e.L..A...[.4 00 41 0f b6 cc 48 8d 7d-c0 48 89 de 48 8b 55 10 .A...H.}.H..H.U. 4d 89 f8 4d 89 e9 45 31-e4 41 56 ff 75 20 e8 29 M..M..E1.AV.u .) 18 00 00 48 83 c4 10 48-8b 7d c8 e8 cc 8a 34 00 ...H...H.}....4. [0f]0b 44 89 f9 c1 e9 06-44 89 f8 83 e0 3f c1 e0 ..D.....D....?.. <== 08 41 81 ff 00 08 00 00-73 0d 09 c8 ba c1 81 00 .A......s....... 00 48 8b 7d d0 eb 38 83-e1 3f 09 c8 44 89 f9 c1 .H.}..8..?..D... e9 0c c1 e0 08 41 81 ff-ff ff 00 00 48 8b 7d d0 .....A......H.}. 77 09 09 c8 ba e1 81 81-00 eb 14 83 e1 3f 09 c8 w............?.. 41 c1 ef 12 c1 e0 08 44-09 f8 ba f1 81 81 81 8d A......D........ Thread 1 last branch register state not available. Binary Images: 0x104ff0000 - 0x105073fff sysextd (79.40.4) <4ED43E22-4E63-3CA2-9354-AB354CE49D62> /System/Library/Frameworks/SystemExtensions.framework/Versions/A/Helpers/sysextd

Do I need to release the block which passed to the dispatchbarrierasync when no ARC ? I noticed "The barrier block to submit to the target dispatch queue. This block is copied and retained until it finishes executing, at which point it is released." in dispatchbarrierasync.dispatch_block_t work = dispatch_block_create(0, ^{ //... }); if (work) { dispatchbarrierasync(dispatchQueue, work); auto res = dispatchblockwait(work, timeout); if (res) { // timeout, then cancel, I should release work here? dispatchblockcancel(work); } Blockrelease(work); // do I need to release work when no ARC? the dispatchbarrierasync would release it if it's executed? } } Thanks!

Hi Experts, If multiple NEFilterDataProvider/NEFilterPacketProvider(s) are on the machine and all are active at the same time. which provider will receive the TCP/UDP segment first? If one of network extension drops the flow or packet, then there is no chance for others to parse it? which is instantiated first or instantiated last if no earlyboot? Thanks a lot!

Hi Experts,I can add WaitForDebugger to file in /Library/LaunchDaemon/*.plist to debug launch daemons from scratch with LLDB.Is it possible to debug a system extension from scratch with LLDB? Thanks a lot.

Hi Experts,I created a simple deadlock program with dispatch concurrent queue but no label:dispatch_queue_t queueA = dispatch_queue_create(NULL, DISPATCH_QUEUE_CONCURRENT);dispatch_async(queueA, ^(){ NSLog(@"3:%@", [NSThread currentThread]); dispatch_barrier_sync(queueA, ^(){ NSLog(@"4:%@", [NSThread currentThread]); }); });the output of sample shows the dispatch queue is serial, not concurrent.2733 Thread_&lt;multiple&gt; DispatchQueue_25: (null) (serial) 2733 start_wqthread (in libsystem_pthread.dylib) + 13 [0x1003bce01] 2733 _pthread_wqthread (in libsystem_pthread.dylib) + 583 [0x1003bd0b7] 2733 _dispatch_worker_thread2 (in libdispatch.dylib) + 125 [0x1003572da] 2733 _dispatch_root_queue_drain (in libdispatch.dylib) + 334 [0x100356824] 2733 _dispatch_async_redirect_invoke (in libdispatch.dylib) + 769 [0x1003463b4] 2733 _dispatch_continuation_pop (in libdispatch.dylib) + 563 [0x1003472be] 2733 _dispatch_client_callout (in libdispatch.dylib) + 8 [0x100343f1b] 2733 _dispatch_call_block_and_release (in libdispatch.dylib) + 12 [0x100342e7c] 2733 __main_block_invoke_2.14 (in learn_oc) + 125 [0x10000176d] main.mm:249 2733 _dispatch_sync_f_slow (in libdispatch.dylib) + 223 [0x10035346c] 2733 __DISPATCH_WAIT_FOR_QUEUE__ (in libdispatch.dylib) + 301 [0x100353c90] 2733 _dispatch_thread_event_wait_slow (in libdispatch.dylib) + 33 [0x1003448ec] 2733 _dispatch_ulock_wait (in libdispatch.dylib) + 47 [0x10034480d] 2733 __ulock_wait (in libsystem_kernel.dylib) + 10 [0x7fff7e3e19de]But if I create the same concurrent queue with the label, the result is correct:dispatch_queue_t queueA = dispatch_queue_create("test", DISPATCH_QUEUE_CONCURRENT);2891 Thread_34199 DispatchQueue_25: test (concurrent) + 2891 start_wqthread (in libsystem_pthread.dylib) + 13 [0x1003bce01] + 2891 _pthread_wqthread (in libsystem_pthread.dylib) + 583 [0x1003bd0b7] + 2891 _dispatch_worker_thread2 (in libdispatch.dylib) + 125 [0x1003572da] + 2891 _dispatch_root_queue_drain (in libdispatch.dylib) + 334 [0x100356824] + 2891 _dispatch_async_redirect_invoke (in libdispatch.dylib) + 769 [0x1003463b4] + 2891 _dispatch_continuation_pop (in libdispatch.dylib) + 563 [0x1003472be] + 2891 _dispatch_client_callout (in libdispatch.dylib) + 8 [0x100343f1b] + 2891 _dispatch_call_block_and_release (in libdispatch.dylib) + 12 [0x100342e7c] + 2891 __main_block_invoke_2.15 (in learn_oc) + 125 [0x10000176d] main.mm:249 + 2891 _dispatch_sync_f_slow (in libdispatch.dylib) + 223 [0x10035346c] + 2891 __DISPATCH_WAIT_FOR_QUEUE__ (in libdispatch.dylib) + 301 [0x100353c90] + 2891 _dispatch_thread_event_wait_slow (in libdispatch.dylib) + 33 [0x1003448ec] + 2891 _dispatch_ulock_wait (in libdispatch.dylib) + 47 [0x10034480d] + 2891 __ulock_wait (in libsystem_kernel.dylib) + 10 [0x7fff7e3e19de]It seems like if there is a label in the queue, that means you want to uniquely identify it in debugging tools such as Instruments, sample, stackshots, and crash reports, or else the system somehow would change the queue?

Hi Experts,I created a simple deadlock program with dispatch concurrent queue but no label:dispatch_queue_t queueA = dispatch_queue_create(NULL, DISPATCH_QUEUE_CONCURRENT);dispatch_async(queueA, ^(){ NSLog(@"3:%@", [NSThread currentThread]); dispatch_barrier_sync(queueA, ^(){ NSLog(@"4:%@", [NSThread currentThread]); }); });the output of sample shows the dispatch queue is serial, not concurrent.2733 Thread_&lt;multiple&gt; DispatchQueue_25: (null) (serial) 2733 start_wqthread (in libsystem_pthread.dylib) + 13 [0x1003bce01] 2733 _pthread_wqthread (in libsystem_pthread.dylib) + 583 [0x1003bd0b7] 2733 _dispatch_worker_thread2 (in libdispatch.dylib) + 125 [0x1003572da] 2733 _dispatch_root_queue_drain (in libdispatch.dylib) + 334 [0x100356824] 2733 _dispatch_async_redirect_invoke (in libdispatch.dylib) + 769 [0x1003463b4] 2733 _dispatch_continuation_pop (in libdispatch.dylib) + 563 [0x1003472be] 2733 _dispatch_client_callout (in libdispatch.dylib) + 8 [0x100343f1b] 2733 _dispatch_call_block_and_release (in libdispatch.dylib) + 12 [0x100342e7c] 2733 __main_block_invoke_2.14 (in learn_oc) + 125 [0x10000176d] main.mm:249 2733 _dispatch_sync_f_slow (in libdispatch.dylib) + 223 [0x10035346c] 2733 __DISPATCH_WAIT_FOR_QUEUE__ (in libdispatch.dylib) + 301 [0x100353c90] 2733 _dispatch_thread_event_wait_slow (in libdispatch.dylib) + 33 [0x1003448ec] 2733 _dispatch_ulock_wait (in libdispatch.dylib) + 47 [0x10034480d] 2733 __ulock_wait (in libsystem_kernel.dylib) + 10 [0x7fff7e3e19de]But if I create the same concurrent queue with the label, the result is correct:dispatch_queue_t queueA = dispatch_queue_create("test", DISPATCH_QUEUE_CONCURRENT);2891 Thread_34199 DispatchQueue_25: test (concurrent) + 2891 start_wqthread (in libsystem_pthread.dylib) + 13 [0x1003bce01] + 2891 _pthread_wqthread (in libsystem_pthread.dylib) + 583 [0x1003bd0b7] + 2891 _dispatch_worker_thread2 (in libdispatch.dylib) + 125 [0x1003572da] + 2891 _dispatch_root_queue_drain (in libdispatch.dylib) + 334 [0x100356824] + 2891 _dispatch_async_redirect_invoke (in libdispatch.dylib) + 769 [0x1003463b4] + 2891 _dispatch_continuation_pop (in libdispatch.dylib) + 563 [0x1003472be] + 2891 _dispatch_client_callout (in libdispatch.dylib) + 8 [0x100343f1b] + 2891 _dispatch_call_block_and_release (in libdispatch.dylib) + 12 [0x100342e7c] + 2891 __main_block_invoke_2.15 (in learn_oc) + 125 [0x10000176d] main.mm:249 + 2891 _dispatch_sync_f_slow (in libdispatch.dylib) + 223 [0x10035346c] + 2891 __DISPATCH_WAIT_FOR_QUEUE__ (in libdispatch.dylib) + 301 [0x100353c90] + 2891 _dispatch_thread_event_wait_slow (in libdispatch.dylib) + 33 [0x1003448ec] + 2891 _dispatch_ulock_wait (in libdispatch.dylib) + 47 [0x10034480d] + 2891 __ulock_wait (in libsystem_kernel.dylib) + 10 [0x7fff7e3e19de]It seems like if there is a label in the queue, that means you want to uniquely identify it in debugging tools such as Instruments, sample, stackshots, and crash reports, or else the system somehow would change the queue?

Hi Experts,Instruments 11.1 has the option to choose to profile a LunachDaemon by selecting the relevant launchd plist file. Then when you start to record it shows an alert with:Waiting for /Library/LaunchDaemons/com.mydaemon.plistPlease take appropriate action to initiate the launch of '/Library/LaunchDaemons/com.mydaemon.plist.'With a Cancel button.I noticed the daemon has been restarted but Instruments does nothing, it's still waiting...Any ideas? Thanks for your help!