Re: [PATCH 2/2] init: ensure that /dev/null is (nearly) always available in initramfs
Askar Safin <[email protected]> Tue, 24 Feb 2026 00:18:48 +0300
| Newsgroups | org.kernel.vger.initramfs,dev.linux.lists.patches,org.kernel.vger.linux-fsdevel,org.kernel.vger.linux-kbuild,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <CAPnZJGDDFrOObdxa70nSS0bqXQ0zqRPnKWLRFUBCj-FtnkpPtg@mail.gmail.com> |
On Mon, Feb 23, 2026 at 4:15=E2=80=AFAM Rob Landley <[email protected]> wrote= : > Are you saying it's reoccurred, or that you plan to run a 5 year old > userspace with a current kernel? I was not aware about this. Okay, I agree that /dev/null is not needed. > (Although I can't say I've regression tested an init task statically > linked against bionic launching itself in an empty chroot recently > either. I tested current Android NDK LTS. I found that if stdin/stdout/stderr already opened, then bionic binary will not crash if /dev/null is absent. It will not crash even if PID is not 1 (so chroot should work). But I didn't test what happens if stdin/stdout/stderr are closed. So, assuming you already somehow got /dev/console, then /dev/null is not needed. According to code, bionic binary should work even if stdin/stdout/stderr ar= e closed, assuming that PID is 1. -- Askar Safin