There have been occasional questions about remote debugging. Unfortunately I see a couple of them which were reasonable basic questions got no answer at all. I will try this again. A couple questions had some mention of lldb server. In some cases clearly the people asking the questions already knew a bit about remote debugging. Yet I can find no documentation about it.Does XCode or other tools that will work with XCode support debugging over the internet? If so what are the reqirements re XCode version and macOS version, whether the client and server can have unmatched versions, etc.?My immediate need (without extra steps) is to debug a process running on Sierra (no XCode installed yet) from a development system running XCode 7.3.1 under El Capitan.The need for remote debugging is to be able to diagnose crashes that don't happen on my development systems. There have also been questions concerning crash logs, but crash logs are of limited use unless the crash can be opened into a debugger with full ability to examine the stack and local variables.Thanks in advance!
LLDB
RSS for tagThe LLDB command-line debugger provides underlying debugging services for development on all Apple platforms.
Posts under LLDB tag
35 Posts
Sort by:
Post
Replies
Boosts
Views
Activity
I am using Xcode 13 on a Macbook Pro (Intel/2019) with macOS Big Sur 11.6
I am building and launching the app on a connecting iPhone X running iOS 15.0, I am also running a UI test suite based on XCUITest while I do that.
The app takes extremely long to launch (2+ minutes), then it sits on the splash screen for another 2-3 minutes. Meanwhile, Xcode shows the alert attached:
I have 2 questions:
Is this a widespread Xcode 13 / iOS 15 issue?
Can someone please explain what the message in the alert means "LLDB is likely reading from device memory to resolve symbols"? Is there anything I can do to reduce or avoid this wait?
HI devs, help me please, i want to debug Big Sur kernel on inter-based macbook from Monterey on m1, i have installed KDK_11.6.4_20G417.kdk in Monterey system on m1 macmini, then launch lldb, have created target and got this message : WARNING! Python version 3 is not supported for xnu lldbmacros.
(lldb) target create /Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel
warning: 'kernel' contains a debug script. To run this script in this debug session:
command script import "/Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel.dSYM/Contents/Resources/Python/kernel.py"
To run all discovered debug scripts in this session:
settings set target.load-script-from-symbol-file true
Current executable set to '/Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel' (x86_64).
(lldb) settings set target.load-script-from-symbol-file true
##############################
WARNING! Python version 3 is not supported for xnu lldbmacros.
Please restart your debugging session with the following workaround
defaults write com.apple.dt.lldb DefaultPythonVersion 2
##############################
Loading kernel debugging from /Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel.dSYM/Contents/Resources/Python/kernel.py
LLDB version lldb-1300.0.42.3
Swift version 5.5.2-dev
settings set target.process.python-os-plugin-path "/Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel.dSYM/Contents/Resources/Python/lldbmacros/core/operating_system.py"
settings set target.trap-handler-names hndl_allintrs hndl_alltraps trap_from_kernel hndl_double_fault hndl_machine_check _fleh_prefabt _ExceptionVectorsBase _ExceptionVectorsTable _fleh_undef _fleh_dataabt _fleh_irq _fleh_decirq _fleh_fiq_generic _fleh_dec
command script import "/Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel.dSYM/Contents/Resources/Python/lldbmacros/xnu.py"
error: module importing failed: Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/Library/Developer/KDKs/KDK_11.6.4_20G417.kdk/System/Library/Kernels/kernel.dSYM/Contents/Resources/Python/lldbmacros/xnu.py", line 123
print "Execution interrupted by user"
^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print("Execution interrupted by user")?
settings set target.process.optimization-warnings false
How can i solve this problem? lldb linked with python 3, but kdk uses python 2, also command line tools version 12.5.1 which uses python 2 i can not install on monterey too.
I don't know when this started but it's basically brought my development to a standstill. I know it worked fine in macOS 13.2.1, though I didn't try Xcode 14.3 RC 2 in 13.2.1.
Basically, Xcode 14.3 RC 2 (14E222b) and also Xcode 14.2 (14C18) (which had been stable up till now) are both unable to launch the app in debug mode. As soon as the launched/debugged app finally starts, the following is printed to Xcode's console:
Message from debugger: The LLDB RPC server has crashed. You may need to manually terminate your process. The crash log is located in ~/Library/Logs/DiagnosticReports and has a prefix 'lldb-rpc-server'. Please file a bug and attach the most recent crash log.
The crash log is always the same. The full log is at https://gist.github.com/NSGod/5e194900bd6f5a9ac9328b936a26037e.
There seems to be an infinite loop in thread 6:
Thread 6 Crashed:: <lldb.process.internal-state(pid=30877)>
0 LLDB 0x12030cfb4 ParseTrieEntries(....) + 20
1 LLDB 0x12030d653 ParseTrieEntries(....) + 1715
After I initially encountered this issue in Xcode 14.3 RC 2 (14E222b) with my business project, I tried creating a new Cocoa app using Xibs and Objective-C and still encountered the problem in that project. I've cleaned out the build folders, etc. but that made no difference. I opened both my business project and a new project in Xcode 14.2 (14C18) but saw no difference (it still crashes). I also tried in a new user account using both versions of Xcode but they both crash.
Could having Python 3.10 and 3.11 installed through MacPorts anything to do with it (I noticed Python 3.9 in the binary images, but I'm not sure).
Any ideas?
I’m trying to use the allocated variables given after p self ($R1 for instance) but after resuming the process and then pausing it again, I cannot access to $R1.
(lldb) p self
(Signals.MainContainerViewController) $R1 = 0x000000012ba19dc0 {
...
(lldb) po $R1
<Signals.MainContainerViewController: 0x12ba19dc0>
(lldb) c
Process 88944 resuming
(lldb) expression -O -l swift -- $R1
warning: Module "/usr/lib/system/libsystem_kernel.dylib" uses triple "arm64-apple-macosx13.4.0", which is not compatible with the target triple "arm64-apple-ios16.0.0-simulator". Enabling per-module Swift scratch context.
error: expression failed to parse:
error: <EXPR>:3:1: error: cannot find '$R1' in scope
$R1
^~~
(lldb)
Does someone have an idea?
macOS: 13.4.1 (22F82)
Xcode Version 14.3.1
How are debug maps saved in Mach-O files? I'm trying to essentially recreate the output of dsymutil -dump-debug-map.
We've got hundreds of crashes in our SwiftUI app which we think are "silent" crashes as there are no complaints from clients and yet - it happens on the main thread in the foreground so I'm not completely sure.
The annoying thing is that we have no idea by the stack trace what is causing this issue, I feel helpless as this is causing some very loud noise through management and honestly - myself who wants to have this noise cleared.
this particular crash is the highest impact (one of a few different weird crashes in our app without clear stack trace)
0 SwiftUI 0x895d90 OUTLINED_FUNCTION_2 + 836
1 SwiftUI 0x895da8 OUTLINED_FUNCTION_2 + 860
2 SwiftUI 0x1329880 OUTLINED_FUNCTION_2 + 424
3 SwiftUI 0x6806c OUTLINED_FUNCTION_441 + 584
4 SwiftUI 0x481b0 OUTLINED_FUNCTION_194 + 544
5 UIKitCore 0x1b7194 -[UIViewController removeChildViewController:notifyDidMove:] + 128
6 UIKitCore 0x77d6e8 -[UINavigationController removeChildViewController:notifyDidMove:] + 80
7 UIKitCore 0x205224 -[UIViewController dealloc] + 768
8 UIKitCore 0x1036c -[UINavigationController viewDidDisappear:] + 372
9 UIKitCore 0xd9c4 -[UIViewController _setViewAppearState:isAnimating:] + 1012
10 UIKitCore 0x46e61c -[UIViewController __viewDidDisappear:] + 136
11 UIKitCore 0x7ec024 __64-[UIViewController viewDidMoveToWindow:shouldAppearOrDisappear:]_block_invoke_3 + 44
12 UIKitCore 0x1a3f24 -[UIViewController _executeAfterAppearanceBlock] + 84
13 UIKitCore 0x1a3e68 -[_UIAfterCACommitBlock run] + 72
14 UIKitCore 0x1a3d9c -[_UIAfterCACommitQueue flush] + 176
15 UIKitCore 0x1a3ca8 _runAfterCACommitDeferredBlocks + 496
16 UIKitCore 0x3f530 _cleanUpAfterCAFlushAndRunDeferredBlocks + 108
17 CoreFoundation 0x43564 __CFRUNLOOP_IS_CALLING_OUT_TO_A_BLOCK__ + 28
18 CoreFoundation 0xabd9c __CFRunLoopDoBlocks + 368
19 CoreFoundation 0x7bbbc __CFRunLoopRun + 856
20 CoreFoundation 0x80ed4 CFRunLoopRunSpecific + 612
21 GraphicsServices 0x1368 GSEventRunModal + 164
22 UIKitCore 0x3a23d0 -[UIApplication _run] + 888
23 UIKitCore 0x3a2034 UIApplicationMain + 340
24 SwiftUI 0x1d1014 OUTLINED_FUNCTION_895 + 2420
25 SwiftUI 0x13216c block_copy_helper.1 + 388
26 SwiftUI 0x11b4bc OUTLINED_FUNCTION_901 + 2868
Number 27 will be our app's Main function and that it - no other trace of our apps code.
Firebase is saying this happens 100% on iOS 16 only and always on the foreground.
How can I get to the bottom of this? how can I debug such a crash?
xcrun devicectl has support for installing and launching apps but does it support launching a debugserver? I've tried xcrun devicectl device process launch /usr/libexec/debugserver to no avail.
Is there something I'm missing or perhaps another tool that would let me start a debugserver from command line?
Hello, I am facing an issue when debugging cpp code through the XCode after updating to XCode 15.0.
During debugging, in a variables watch view, it is shown for the collections (let's take std::vector as an example) that its size is 0 even if it is not. If I add it as an expression to watch, it is fine. Also if I print it in lldb console, it is also fine. Any thoughts on this?
Hello,
My purpose is to understand how macOS works.
Here is what i've done: I have wrote a c program on a M1 CPU with this lines:
printf("Before breakpoint\n");
asm volatile("brk #0");
printf("After breakpoint\n");
When i run this program with lldb, a breakpoint is hit on the second line. So i suppose lldb is writing a "brk #0" instruction when we put a breakpoint manually.
I can't continue to next line with lldb "c" command. PC stays on the brk instruction. I need to manually set PC to next instruction in lldb.
Now, what i want to do is to create my own debugger. (I want to understand how lldb works).
I have managed to ptrace the target program and i was able to catch an event with waitpid when "brk #0" is hit. But i don't know how i can increase PC value and continue execution because i can't do this on Silicon CPU:
ptrace(PTRACE_GETREGS, child_pid, NULL, &regs);
ptrace(PTRACE_SETREGS, child_pid, NULL, &regs);
kill(child_pid, SIGCONT);
So my question is: How does lldb managed to change ARM64 registers of a remote process ?
Thanks
Hello,
I have tried to create a thread with thread_create_running API.
It works but i would like to suspend this thread. I can call thread_suspend, but my thread has already been start before i call this API.
Is there a way to create a thread without running it automaticaly.
Thanks
Hello.
I am currently developing a C++ application for x86_64 using XCode, and I am experiencing the same problem as described in the following link.
https://discourse.llvm.org/t/lldb-doesnt-stop-on-sigabrt-assert-with-x86-executable-on-apple-arm/69749/1
Due to this bug, when the application crashes, it is difficult to tell which source code caused the crash, and this is reducing our development efficiency.
How can this be fixed?
Thank you in advance for your help.
As I was watching WWDC19 Session: "LLDB: beyond "po" " I got confused about "p" command. In the session there was a following example to show how "p" command works:
// Struct for demonstration purposes
struct Trip {
var name: String
var destinations: [String]
let cruise = Trip (
name: "Mediterranean Cruise"
destinations: ["Sorrento", "Capri", "Taormina"])
Using "p" to get info about cruise instance:
(lldb) p cruise
(Travel.Trip) $R0 = {
name = "Mediterranean Cruise"
destinations = 3 values {
[0] = "Sorrento"
[1] = "Capri"
[2] = "Taormina"
}
}
I was following along and wrote the same code. But the output from LLDB turned out to be different:
(lldb) p cruise
(LLDB_Apple_Session_FollowAlong.Trip) {
name = "Mediterranean Cruise"
destinations = 3 values {
[0] = "Sorrento"
[1] = "Capri"
[2] = "Taormina"
}
}
As you can see LLDB didn't create a global variable R0 which I could later use in my debugging session and that seemed strange to me.
Then the presenter said the following:
"p" is just an alias for the "expression" command.
So, I tried to use the "expr" command to see if they're actually the same and they turned out to be different commands. The output I got from "expr" was the one I expected from "p":
(lldb) expr cruise
(LLDB_Apple_Session_FollowAlong.Trip) $R0 = {
name = "Mediterranean Cruise"
destinations = 3 values {
[0] = "Sorrento"
[1] = "Capri"
[2] = "Taormina"
}
}
Finally, my question is:
Am I wrong somewhere or did something change in LLDB regarding "p" and "expr" commands and if so, where could I get more information about the changes?
I am building simple command-line programs in C. When I run LLDB on the command line on any such executable , even though I set a breakpoint in main, LLDB always runs to termination and reports an error. Below is an example:
~ % clang -Wall -g prog.c -o prog aviram@csaamdS278stma ~ % lldb prog (lldb) target create "prog" Current executable set to '/Users/aviram/prog' (x86_64). (lldb) breakpoint set --name main Breakpoint 1: where = progmain + 15 at prog.c:73:21, address = 0x0000000100003ebf
(lldb) run
Process 99926 launched: '/Users/aviram/prog' (x86_64)
Process 99926 exited with status = -1 (0xffffffff) lost connection
(lldb)
`
Note that it says "lost connection" even though I am _ remote debugging. This happens with every executable I build._
Details: 2019 MacBook Pro (x86_64, Sonoma); Apple Clang 12.0.0 (downloaded with Xcode 15.0.1).
Why is this happening and what can I do to debug my program (for real) using LLDB from the command line? Thanks.
In an old document from Apple(Debugging An Authorization Plug-In With Xcode), a method was taught to debug plugins on the device remotely, but I can't find this feature in the new Xcode version. Does it exist In the new Xcodes?
https://developer.apple.com/library/archive/technotes/tn2108/_index.html#//apple_ref/doc/uid/DTS40007965-CH1-SECCONFIGXCODE
Yesterday, my code ran just fine under the previous Xcode version. Today, some print() statements seem to come with extra lines. I know it sounds stupid, but my code did not change in the meantime. It doesn't appear to come from anything I control, almost like some Apple code emits an extra line feed somewhere. It's just a Swift Mac App I built to make my digital art; otherwise, nothing else is incorrect, just these odd lines.
It's not as simple as just making a test case with a few print("***") statements, it seems to require other code to run in between calls to print. Most of my app is using CoreGraphics. It has no UI.
It's like when you see spurious Apple debugging info in the console sometimes, but it's only a blank line this time. It's not a big issue, just annoying.
When po an NSObject variable in lldb in Xcode 15.1,an error occurs,this error description is "error: Internal error [IRForTarget]: Couldn't change a static reference to an Objective-C selector to a dynamic reference"
Has anyone encountered this situation?
How can i fix this?
I am trying to debug a kernel panic in our kext. I can attach to the target Mac over ethernet if I:
cause an NMI using
add an IOPanic call to my kext and cause it to be executed
use Dtrace to invoke a panic
However if I reproduce the kernel panic which I am investigating, the Mac just restarts.
How can I make the Mac wait for me to attach with lldb rather than restarting?
My target configuration is:
Mac is 2021 M1 Pro 14" MacBook Pro
macOS 14.2 (23C64)
Network: Apple Thunderbolt 3 <-> Thunderbolt 3 adapter + Apple Thunderbolt 2 to ethernet adapters
Boot-args = "debug=0x44 wdt=-1 kdp_match_name=en8" (I have also tried debug=0x104C0C)
Hello, is there any way that I can automatically fold items in the console?
I do log all API requests and responses, printed in a nice json to console. It works well, but having a lot of requests makes the console very big, and it kills readability, I have noticed that I can manually fold json printed in console with CMD+OPTION+Left shortcut.
I am asking if that's possible to fold everything, or just display first lines of the console.
So I can see more entries easier, and if I need I can unfold, or disable Line Wrap, to see details.
I am attaching the photo of how code folds works there. It's all good, but I need to do it one by one, which is bad.
If I could select everything and auto-fold, or It has the line wrap disabled.
I
I'm writing a C/C++ command line program which, at some point, calls IOHIDManagerOpen. I've added both my program executable and lldb as permitted for input monitoring (as far as I remember, my program was added after showing up a permission prompt, I've added lldb manually later, trying to resolve the problem).
My problem is that when I run my program from within lldb in Terminal, the call to IOHIDManagerOpen returns kIOReturnNotPermitted. When I run my program directly in the terminal session (without lldb), this call returns kIOReturnSuccess. Such behaviour means it will be impractical to use lldb for any debugging of this program.
What can be done to make lldb session behave the same way, the normal execution works?
I'm on:
23.2.0 Darwin Kernel Version 23.2.0: Wed Nov 15 21:55:06 PST 2023; root:xnu-10002.61.3~2/RELEASE_ARM64_T6020 arm64
and:
lldb-1500.0.200.58
Apple Swift version 5.9.2 (swiftlang-5.9.2.2.56 clang-1500.1.0.2.5)