Kernel: Don't acquire Mutex for hostname() before scheduling is enabled

There's no reason to acquire the mutex for this resource before APs are booted, before scheduling is enabled, and before interrupts are enabled.
Author: https://github.com/ADKaster Commit: https://github.com/SerenityOS/serenity/commit/4f2520674ca Pull-request: https://github.com/SerenityOS/serenity/pull/9580 Reviewed-by: https://github.com/alimpfard Reviewed-by: https://github.com/awesomekling Reviewed-by: https://github.com/bgianfo
2025-04-21 03:55:24 +00:00 · 2021-08-22 22:03:16 -06:00 · 2021-08-22 22:03:16 -06:00 · 4f2520674c · 2024-07-18 05:08:37 +09:00
commit 4f2520674c
parent 54161bf5b4
1 changed files with 3 additions and 3 deletions
--- a/Kernel/Process.cpp
+++ b/Kernel/Process.cpp
@ -76,9 +76,9 @@ UNMAP_AFTER_INIT void Process::initialize()

    next_pid.store(0, AK::MemoryOrder::memory_order_release);

-    hostname().with_exclusive([&](auto& name) {
-        name = "courage";
-    });
+    // Note: This is called before scheduling is initialized, and before APs are booted.
+    //       So we can "safely" bypass the lock here.
+    reinterpret_cast<String&>(hostname()) = "courage";

    create_signal_trampoline();
 }