RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 12:28:12 +00:00
Code Issues Activity Actions

Kernel: Leak a ref() on the new Process ASAP in sys$fork()

Browse source 
This fixes an issue where failing the fork due to OOM or other error,
we'd end up destroying the Process too early. By the time we got to
WaitBlockerSet::finalize(), it was long gone.
This commit is contained in:
Andreas Kling 2022-08-15 00:53:28 +02:00
parent ec4b83326b
commit ae8f1c7dc8
1 changed files with 4 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

7
Kernel/Syscalls/fork.cpp
View file

@ -29,6 +29,10 @@ ErrorOr<FlatPtr> Process::sys$fork(RegisterState& regs)
auto child_name = TRY(m_name->try_clone());
auto child = TRY(Process::try_create(child_first_thread, move(child_name), uid(), gid(), pid(), m_is_kernel_process, current_directory(), m_executable, m_tty, this));
// NOTE: All user processes have a leaked ref on them. It's balanced by Thread::WaitBlockerSet::finalize().
child->ref();
TRY(m_unveil_data.with([&](auto& parent_unveil_data) -> ErrorOr<void> {
return child->m_unveil_data.with([&](auto& child_unveil_data) -> ErrorOr<void> {
child_unveil_data.state = parent_unveil_data.state;
@ -143,9 +147,6 @@ ErrorOr<FlatPtr> Process::sys$fork(RegisterState& regs)
auto child_pid = child->pid().value();
// NOTE: All user processes have a leaked ref on them. It's balanced by Thread::WaitBlockerSet::finalize().
(void)child.leak_ref();
return child_pid;
}