linux-kernel - Re: [PATCH] security: Fix ret values doc for security_inode_init

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAHC9VhSbya3Q-VM8v43qkQDWCuYWFqQ801j9_HfdwWJ9RLzkjw@mail.gmail.com>
Date:   Tue, 25 Jul 2023 14:38:26 -0400
From:   Paul Moore <paul@...l-moore.com>
To:     Roberto Sassu <roberto.sassu@...weicloud.com>
Cc:     jmorris@...ei.org, serge@...lyn.com,
        linux-security-module@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Roberto Sassu <roberto.sassu@...wei.com>
Subject: Re: [PATCH] security: Fix ret values doc for security_inode_init_security()

On Tue, Jul 25, 2023 at 3:02 AM Roberto Sassu
<roberto.sassu@...weicloud.com> wrote:
> On Mon, 2023-07-24 at 17:54 -0400, Paul Moore wrote:
> > On Mon, Jul 24, 2023 at 10:52 AM Roberto Sassu
> > <roberto.sassu@...weicloud.com> wrote:
> > >
> > > From: Roberto Sassu <roberto.sassu@...wei.com>
> > >
> > > Commit 6bcdfd2cac55 ("security: Allow all LSMs to provide xattrs for
> > > inode_init_security hook") unified the !initxattrs and initxattrs cases. By
> > > doing that, security_inode_init_security() cannot return -EOPNOTSUPP
> > > anymore, as it is always replaced with zero at the end of the function.
> > >
> > > Also, mentioning -ENOMEM as the only possible error is not correct. For
> > > example, evm_inode_init_security() could return -ENOKEY.
> > >
> > > Fix these issues in the documentation of security_inode_init_security().
> > >
> > > Fixes: 6bcdfd2cac55 ("security: Allow all LSMs to provide xattrs for inode_init_security hook")
> > > Signed-off-by: Roberto Sassu <roberto.sassu@...wei.com>
> > > ---
> > >  security/security.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/security/security.c b/security/security.c
> > > index cfdd0cbbcb9..5aa9cb91f0f 100644
> > > --- a/security/security.c
> > > +++ b/security/security.c
> > > @@ -1604,8 +1604,8 @@ EXPORT_SYMBOL(security_dentry_create_files_as);
> > >   * a security attribute on this particular inode, then it should return
> > >   * -EOPNOTSUPP to skip this processing.
> > >   *
> > > - * Return: Returns 0 on success, -EOPNOTSUPP if no security attribute is
> > > - * needed, or -ENOMEM on memory allocation failure.
> > > + * Return: Returns 0 on success or on -EOPNOTSUPP error, a negative value other
> > > + *         than -EOPNOTSUPP otherwise.
> >
> > How about "Returns 0 if the LSM successfully initialized all of the
> > inode security attributes that are required, negative values
> > otherwise."?  The caller doesn't need to worry about the individual
> > LSMs returning -EOPNOTSUPP in the case of no security attributes, and
> > if they really care, they are likely reading the description above (or
> > the code) which explains it in much better detail.
>
> Maybe this could be better:
>
> Return 0 if security attributes initialization is successful or not
> necessary, a negative value otherwise.

Well, I'm trying to avoid differentiating between the non-zero, but
successful attribute initialization and the zero attribute case; from
a caller's perspective it doesn't matter (and why we don't
differentiate between the two with different error codes).  If the
distinction between the two states is important from a caller's
perspective, there should be different return codes.

-- 
paul-moore.com