From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Serge E. Hallyn" Subject: Re: [PATCH] capabilities: Ambient capability set V1 Date: Thu, 26 Feb 2015 15:32:36 -0600 Message-ID: <20150226213236.GA19753@mail.hallyn.com> References: <20150225033247.GC29685@ubuntumail> <20150226153524.GC15182@mail.hallyn.com> <20150226193200.GA17709@mail.hallyn.com> <20150226203405.GB18926@mail.hallyn.com> <20150226211324.GB19273@mail.hallyn.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-security-module-owner@vger.kernel.org To: Christoph Lameter Cc: "Serge E. Hallyn" , Serge Hallyn , Serge Hallyn , Andy Lutomirski , Aaron Jones , Ted Ts'o , linux-security-module@vger.kernel.org, akpm@linuxfoundation.org, "Andrew G. Morgan" , Mimi Zohar , Austin S Hemmelgarn , Markku Savela , Jarkko Sakkinen , linux-kernel@vger.kernel.org, linux-api@vger.kernel.org, Michael Kerrisk , Jonathan Corbet List-Id: linux-api@vger.kernel.org On Thu, Feb 26, 2015 at 03:23:35PM -0600, Christoph Lameter wrote: > On Thu, 26 Feb 2015, Serge E. Hallyn wrote: > > > No. fE is not "not available". If you set it with setcap, you *should* > > (if i'm tinking right) find fE full when calculating the new capability > > sets, because of magic_etc. > > There is nothing in get_vfs_caps_from_disk that does this and the magic > vanishes after this function is done. get_vfs_caps_from_disk does: cpu_caps->magic_etc = magic_etc = le32_to_cpu(caps.magic_etc); then bprm_caps_from_vfs_caps does: if (caps->magic_etc & VFS_CAP_FLAGS_EFFECTIVE) *effective = true; and finally cap_bprm_set_creds does: if (effective) new->cap_effective = new->cap_permitted; else cap_clear(new->cap_effective);