Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <537635732d9cbcc42bcf7be5ed932d284b03d39f.camel@linux.ibm.com>
Subject: Re: [PATCH] ima: Calculate digest in ima_inode_hash() if not
 available
From:   Mimi Zohar <zohar@linux.ibm.com>
To:     Roberto Sassu <roberto.sassu@huawei.com>, shuah@kernel.org,
        ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org,
        kpsingh@kernel.org, Florent Revest <revest@chromium.org>
Cc:     linux-integrity@vger.kernel.org,
        linux-security-module@vger.kernel.org,
        linux-kselftest@vger.kernel.org, netdev@vger.kernel.org,
        bpf@vger.kernel.org, linux-kernel@vger.kernel.org
Date:   Sun, 13 Feb 2022 08:06:01 -0500
In-Reply-To: <20220211104828.4061334-1-roberto.sassu@huawei.com>
References: <20220211104828.4061334-1-roberto.sassu@huawei.com>
Content-Type: text/plain; charset="ISO-8859-15"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi Roberto,

On Fri, 2022-02-11 at 11:48 +0100, Roberto Sassu wrote:
> __ima_inode_hash() checks if a digest has been already calculated by
> looking for the integrity_iint_cache structure associated to the passed
> inode.
> 
> Users of ima_file_hash() and ima_inode_hash() (e.g. eBPF) might be
> interested in obtaining the information without having to setup an IMA
> policy so that the digest is always available at the time they call one of
> those functions.
> 
> Open a new file descriptor in __ima_inode_hash(), so that this function
> could invoke ima_collect_measurement() to calculate the digest if it is not
> available. Still return -EOPNOTSUPP if the calculation failed.
> 
> Instead of opening a new file descriptor, the one from ima_file_hash()
> could have been used. However, since ima_inode_hash() was created to obtain
> the digest when the file descriptor is not available, it could benefit from
> this change too. Also, the opened file descriptor might be not suitable for
> use (file descriptor opened not for reading).
> 
> This change does not cause memory usage increase, due to using a temporary
> integrity_iint_cache structure for the digest calculation, and due to
> freeing the ima_digest_data structure inside integrity_iint_cache before
> exiting from __ima_inode_hash().
> 
> Finally, update the test by removing ima_setup.sh (it is not necessary
> anymore to set an IMA policy) and by directly executing /bin/true.
> 
> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>

Although this patch doesn't directly modify either ima_file_hash() or
ima_inode_hash(),  this change affects both functions.  ima_file_hash()
was introduced to be used with eBPF.  Based on Florent's post, changing
the ima_file_hash() behavor seems fine.  Since I have no idea whether
anyone is still using ima_inode_hash(), perhaps it would be safer to
limit this behavior change to just ima_file_hash().

Please update the ima_file_hash() doc.  While touching this area, I'd
appreciate your fixing the first doc line in both ima_file_hash() and
ima_inode_hash() cases, which wraps spanning two lines.

Please split the IMA from the eBPF changes.

-- 
thanks,

Mimi