From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Thu, 14 Mar 2019 20:45:07 +0100 From: aszlig Subject: Re: Failure to execute file on overlayfs during switch_root/chroot Message-ID: <20190314194507.GA27112@dnyarri> References: <20181207121027.GA5996@dnyarri> <20190202172914.GA17406@dnyarri> <20190203101340.GA6934@dnyarri> <20190314010928.GA748@dnyarri> <20190314103757.GA4654@dnyarri> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="7AUc2qLy4jB3hD7Z" Content-Disposition: inline In-Reply-To: To: Amir Goldstein Cc: Miklos Szeredi , overlayfs , Graham Christensen , Samuel Dionne-Riel , michael bishop List-ID: --7AUc2qLy4jB3hD7Z Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Mar 14, 2019 at 09:38:05PM +0200, Amir Goldstein wrote: > OTOH, ovl_path_open() from ovl_copy_up_data() does open lower > file with O_NOATIME, so I am a bit surprised you did not see the > same issue there. Good point actually, I did forget about that. The reason we did not run into this issue is because Nix store paths are *never* changed (only new ones are added) and the store is mounted read-only (new paths are crated using a mou= nt namespace where the store path is writable). So yes, actually the O_NOATIME should also be removed from ovl_copy_up_data= (). a! --=20 aszlig Universal dilettante --7AUc2qLy4jB3hD7Z Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iHUEABYIAB0WIQTdUmvHdn26KBbAleVoQInOZ+u2kQUCXIqvQQAKCRBoQInOZ+u2 kZzSAP4sH05kJZEOmBFTTWtXeo9l5ytgjH5Ko7WTofkT8omsHwEAt/9q93kNCH/O cVjq8S8WaE7bbQArjXMUoKinuTm07A4= =QD4F -----END PGP SIGNATURE----- --7AUc2qLy4jB3hD7Z--