intr-msg / hurdsig looks broken, is my analysis correct?

Samuel Thibault samuel.thibault@gnu.org
Wed Mar 1 09:24:21 GMT 2023


Sergey Bugaev, le mer. 01 mars 2023 12:21:17 +0300, a ecrit:
> * eip < _hurd_intr_rpc_msg_about_to => thread is still to check
> ss->cancel, so just set ss->cancel = 1;
> * _hurd_intr_rpc_msg_about_to <= eip < _hurd_intr_rpc_msg_setup_done
> => thread is still to check eax, so set eax = MACH_SEND_INTERRUPTED;
> * _hurd_intr_rpc_msg_setup_done <= eip => the thread will not check
> anything, but we can make it skip over the trap safely, so do that.

Yup!

Samuel


More information about the Libc-alpha mailing list