dnie...@gmail.com, le ven. 15 août 2025 02:57:28 +0100, a ecrit:
> I have, however, some trouble running the test case using setrlimit
> under rpctrace:
>
>
> 27<--40(pid707)->io_write_request ("setrlimit failed: (os/kern) invalid
> task\n" -1)setrlimit failed: (os/kern) invalid task
> = 0 41
I'm not completely surprised since there are probably confusions about
which task one is talking about: your program, or rpctrace.
> The new setrlimit call uses the __mach_task_self () call to get the
> task port.
That should be fine.
> It's also inside a HURD_CRITICAL_BEGIN/HURD_CRITICAL_END, could that
> be a problem?
No, rpctrace does not really care if the process is inside a critical
section.
I'm thinking that probably the __vm_set_size_limit call, since it is
directed to the host port, does not actually go through rpctrace. Do
you actually see the RPC call? If it doesn't, probably the task port
that the process has is the wrapper port in rpctrace, and not the kernel
task port, and thus the kernel does not like it.
Samuel
