Re: [PATCH v7 0/5] IMA: restrict the accepted digest algorithms for the security.ima xattr

From: Steve Grubb <sgrubb@redhat.com>
To: "linux-integrity@vger.kernel.org"
	<linux-integrity@vger.kernel.org>,
	linux-audit@redhat.com, Mimi Zohar <zohar@linux.ibm.com>
Cc: "dmitry.kasatkin@gmail.com" <dmitry.kasatkin@gmail.com>,
	BARVAUX Didier <Didier.BARVAUX@viveris.fr>,
	THOBY Simon <Simon.THOBY@viveris.fr>
Subject: Re: [PATCH v7 0/5] IMA: restrict the accepted digest algorithms for the security.ima xattr
Date: Wed, 11 Aug 2021 19:53:15 -0400	[thread overview]
Message-ID: <2072391.irdbgypaU6@x2> (raw)
In-Reply-To: <f06d6b278bf228d21bf3ec01da8173e6518557f9.camel@linux.ibm.com>

Hello,

On Wednesday, August 11, 2021 3:40:51 PM EDT Mimi Zohar wrote:
> On Wed, 2021-08-11 at 11:40 +0000, THOBY Simon wrote:
> Other than the two questions on " IMA: add a policy option to restrict
> xattr hash algorithms on appraisal" patch, the patch set is looking
> good.
> 
> thanks,
> 
> Mimi
> 
> > Here is also a short description of the new audit messages, but I can
> > send it in a followup mail if that is not the proper place:
> > 
> > When writing the xattr with an algorithm not built in the kernel (here
> > the
> > kernel was built with CONFIG_CRYPTO_MD5 unset), e.g. with
> > 
> > "evmctl ima_hash -a md5 /usr/bin/strace":
> > 	audit(1628066120.418:121): pid=1344 uid=0 auid=0 ses=1
> > 	subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 op=set_data
> > 	cause=unavailable-hash-algorithm comm="evmctl" name="/usr/bin/strace"
> > 	dev="dm-0" ino=2632657 res=0 errno=0> 

Is this audit event accurate? I seem to be seeing name=value=value. I'm 
hoping this is a copy/paste/mail client issue.

-Steve

> > With the same command and the policy rule
> > 
> > "appraise func=SETXATTR_CHECK appraise_algos=sha256,sha384,sha512", we 
get:
> > 	audit(1628066210.141:127): pid=1362 uid=0 auid=0 ses=1
> > 	subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 op=set_data
> > 	cause=denied-hash-algorithm comm="evmctl" name="/usr/bin/strace"
> > 	dev="dm-0" ino=2632657 res=0 errno=0> 
> > Note that the cause is now 'denied-hash-algorithm' instead of
> > 'unavailable-hash-algorithm'. We get that audit message for any algorithm
> > outside of sha256/384/512 (including algorithms not compiled in the
> > kernel
> > like MD5). In a sense, 'denied-hash-algorithm' takes predecence over
> > 'unavailable-hash-algorithm'.
> > 
> > When appraising files, e.g. trying to execute a file whose xattr was
> > hashed with sha1 while the policy rule
> > 
> > "appraise func=BPRM_CHECK fowner=0 appraise_algos=sha256" is enabled:
> > 	audit(1628066349.230:130): pid=1369 uid=0 auid=0 ses=1
> > 	subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023
> > 	op=collect_data cause=denied-hash-algorithm comm="bash"
> > 	name="/usr/bin/strace" dev="dm-0" ino=2632657 res=0 errno=0> 
> > This series is based on the following repo/branch:
> >  repo:
> >  https://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity.g
> >  it branch: next-integrity-testing
> >  commit e37be5343ae2b9419aea1442b07e5d2428b437b4 ("Merge branch
> >  'ima-buffer-measurement-changes-v4' into next-integrity")
> --
> Linux-audit mailing list
> Linux-audit@redhat.com
> https://listman.redhat.com/mailman/listinfo/linux-audit

--
Linux-audit mailing list
Linux-audit@redhat.com
https://listman.redhat.com/mailman/listinfo/linux-audit