[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <939ee428816825b5b28641d6e09b5e75b4172917.camel@linux.ibm.com>
Date: Wed, 18 Nov 2020 13:28:38 -0500
From: Mimi Zohar <zohar@...ux.ibm.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Christoph Hellwig <hch@...radead.org>,
Roberto Sassu <roberto.sassu@...wei.com>,
"linux-integrity@...r.kernel.org" <linux-integrity@...r.kernel.org>,
"linux-security-module@...r.kernel.org"
<linux-security-module@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Silviu Vlasceanu <Silviu.Vlasceanu@...wei.com>,
"stable@...r.kernel.org" <stable@...r.kernel.org>,
"viro@...iv.linux.org.uk" <viro@...iv.linux.org.uk>,
"linux-fsdevel@...r.kernel.org" <linux-fsdevel@...r.kernel.org>
Subject: Re: [RESEND][PATCH] ima: Set and clear FMODE_CAN_READ in
ima_calc_file_hash()
On Tue, 2020-11-17 at 15:36 -0800, Linus Torvalds wrote:
> Another alternative is to change the policy and say "any write-only
> open gets turned into a read-write open".
>
> But it needs to be done at *OPEN* time, not randomly afterwards by
> just lying to the 'struct file'.
The ima_file_check hook is at open, but it is immediately after
vfs_open(). Only after the file is opened can we determine if the
file is in policy. If the file was originally opened without read
permission, a new file instance (dentry_open) with read permission is
opened. Would limiting opening a new file instance with read
permission to just the ima_file_check hook be acceptable?
thanks,
Mimi
Powered by blists - more mailing lists