`_remote_debugging`: incorrect error for `process_vm_readv()` w/ `EPERM` when detecting the cookie

[maurycy]

Bug report

Bug description:

Note Failed to find the PyRuntime section in process 175853 on Linux platform here:

2026-05-09T21:00:48.747402752+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % sudo sysctl -w kernel.yama.ptrace_scope=1
kernel.yama.ptrace_scope = 1
2026-05-09T21:00:49.153189782+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % ./python -c "import time; time.sleep(60)" &
[1] 175853
2026-05-09T21:00:51.504230609+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % ./python -m profiling.sampling attach $!
Failed to find the PyRuntime section in process 175853 on Linux platform
[1] 2026-05-09T21:00:54.515838389+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % sudo sysctl -w kernel.yama.ptrace_scope=0
kernel.yama.ptrace_scope = 0
2026-05-09T21:00:57.313995182+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % ./python -c "import time; time.sleep(60)" &
[2] 175951
2026-05-09T21:00:59.679641859+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % ./python -m profiling.sampling attach $!
[1] - done ./python -c "import time; time.sleep(60)"
Captured 57,731 samples in 57.73 seconds
Sample rate: 1,000.00 samples/sec
Error rate: 0.00
[2] + done ./python -c "import time; time.sleep(60)"
Profile Stats:
 nsamples sample% tottime (s) cumul% cumtime (s) filename:lineno(function)
 57729/57729 100.0 57.729 100.0 57.729 <string>:1(<module>)

Legend:
 nsamples: Direct/Cumulative samples (direct executing / on call stack)
 sample%: Percentage of total samples this function was directly executing
 tottime: Estimated total time spent directly in this function
 cumul%: Percentage of total samples when this function was on the call stack
 cumtime: Estimated cumulative time (including time in called functions)
 filename:lineno(function): Function location and name

Summary of Interesting Functions:

Functions with Highest Direct/Cumulative Ratio (Hot Spots):
 1.000 direct/cumulative ratio, 100.0% direct samples: <string>:(<module>)

Functions with Highest Call Frequency (Indirect Calls):

Functions with Highest Call Magnification (Cumulative/Direct):
2026-05-09T21:01:59.697134414+0000 ubuntu@f4-metal-medium-lon-1 /home/ubuntu/cpython (main 7241f27) % 

The wrapper handles PermissionError well:

cpython/Lib/profiling/sampling/__main__.py

Lines 63 to 67 in 57ef219

	if __name__ == '__main__':
	try:
	main()
	except PermissionError:
	handle_permission_error()

cpython/Lib/profiling/sampling/__main__.py

Lines 51 to 61 in 57ef219

	def handle_permission_error():
	"""Handle PermissionError by displaying appropriate error message."""
	if sys.platform == "darwin":
	print(MACOS_PERMISSION_ERROR, file=sys.stderr)
	elif sys.platform.startswith("linux"):
	print(LINUX_PERMISSION_ERROR, file=sys.stderr)
	elif sys.platform.startswith("win"):
	print(WINDOWS_PERMISSION_ERROR, file=sys.stderr)
	else:
	print(GENERIC_PERMISSION_ERROR, file=sys.stderr)
	sys.exit(1)

cpython/Lib/profiling/sampling/__main__.py

Lines 17 to 36 in 57ef219

	LINUX_PERMISSION_ERROR = """
	🔒 Tachyon was unable to access process memory. This could be because tachyon
	has insufficient privileges (the required capability is CAP_SYS_PTRACE).
	Unprivileged processes cannot trace processes that they cannot send signals
	to or those running set-user-ID/set-group-ID programs, for security reasons.
	If your uid matches the uid of the target process you want to analyze, you
	can do one of the following to get 'ptrace' scope permissions:
	* If you are running inside a Docker container, you need to make sure you
	start the container using the '--cap-add=SYS_PTRACE' or '--privileged'
	command line arguments. Notice that this may not be enough if you are not
	running as 'root' inside the Docker container as you may need to disable
	hardening (see next points).
	* Try running again with elevated permissions by running 'sudo -E !!'.
	* You can disable kernel hardening for the current session temporarily (until
	a reboot happens) by running 'echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope'.
	"""

but it's swallowed as RuntimeError:

cpython/Python/remote_debug.h

Lines 954 to 965 in 57ef219

	#elif defined(__linux__) && HAVE_PROCESS_VM_READV
	// On Linux, search for 'python' in executable or DLL
	address = search_linux_map_for_section(handle, "PyRuntime", "python",
	_Py_RemoteDebug_ValidatePyRuntimeCookie);
	if (address == 0) {
	// Error out: 'python' substring covers both executable and DLL
	PyObject *exc = PyErr_GetRaisedException();
	PyErr_Format(PyExc_RuntimeError,
	"Failed to find the PyRuntime section in process %d on Linux platform",
	handle->pid);
	_PyErr_ChainExceptions1(exc);
	}

The error bubbles out of

cpython/Python/remote_debug.h

Lines 164 to 176 in 57ef219

	static int
	_Py_RemoteDebug_ValidatePyRuntimeCookie(proc_handle_t *handle, uintptr_t address)
	{
	if (address == 0) {
	return 0;
	}
	char buf[sizeof(_Py_Debug_Cookie) - 1];
	if (_Py_RemoteDebug_ReadRemoteMemory(handle, address, sizeof(buf), buf) != 0) {
	PyErr_Clear();
	return 0;
	}
	return memcmp(buf, _Py_Debug_Cookie, sizeof(buf)) == 0;
	}

hence such a cookie panic

CPython versions tested on:

CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-149619: Fix _remote_debugging permissions error on Linux #150012
• [3.15] gh-149619: Fix _remote_debugging permissions error on Linux (GH-150012) #150339
• gh-149619: Harden _remote_debugging error paths #150349
• [3.15] gh-149619: Harden _remote_debugging error paths (GH-150349) #150434
• [3.15] gh-149619: Harden _remote_debugging error paths (GH-150349) #150435
• [3.14] gh-149619: Fix _remote_debugging permissions error on Linux (GH-150012) #154825
• [3.14] gh-149619: Harden _remote_debugging error paths (GH-150349) #154844

[maurycy]

cc @pablogsal

[pablogsal]

@ivonastojanovic can you take a look?