Message-ID: <6340bd43-c96e-9702-e00a-b426f05b0271@kernel.dk>
Date:   Fri, 10 Feb 2023 09:58:55 -0700
From:   Jens Axboe <axboe@...nel.dk>
To:     Paul Moore <paul@...l-moore.com>
Cc:     Steve Grubb <sgrubb@...hat.com>,
        Richard Guy Briggs <rgb@...hat.com>,
        LKML <linux-kernel@...r.kernel.org>, io-uring@...r.kernel.org,
        Eric Paris <eparis@...isplace.org>,
        Christian Brauner <brauner@...nel.org>,
        Stefan Roesch <shr@...com>
Subject: Re: [PATCH v2] io_uring,audit: don't log IORING_OP_MADVISE

On 2/10/23 9:52?AM, Paul Moore wrote:
> On Fri, Feb 10, 2023 at 11:00 AM Jens Axboe <axboe@...nel.dk> wrote:
>> On 2/10/23 8:39?AM, Paul Moore wrote:
>>> On Thu, Feb 9, 2023 at 7:15 PM Jens Axboe <axboe@...nel.dk> wrote:
>>>> On 2/9/23 3:54?PM, Steve Grubb wrote:
>>>>> On Thursday, February 9, 2023 5:37:22 PM EST Paul Moore wrote:
>>>>>> On Thu, Feb 9, 2023 at 4:53 PM Richard Guy Briggs <rgb@...hat.com> wrote:
>>>>>>> On 2023-02-01 16:18, Paul Moore wrote:
>>>>>>>> On Wed, Feb 1, 2023 at 3:34 PM Richard Guy Briggs <rgb@...hat.com>
>>>>> wrote:
>>>>>>>>> fadvise and madvise both provide hints for caching or access pattern
>>>>>>>>> for file and memory respectively.  Skip them.
>>>>>>>>
>>>>>>>> You forgot to update the first sentence in the commit description :/
>>>>>>>
>>>>>>> I didn't forget.  I updated that sentence to reflect the fact that the
>>>>>>> two should be treated similarly rather than differently.
>>>>>>
>>>>>> Ooookay.  Can we at least agree that the commit description should be
>>>>>> rephrased to make it clear that the patch only adjusts madvise?  Right
>>>>>> now I read the commit description and it sounds like you are adjusting
>>>>>> the behavior for both fadvise and madvise in this patch, which is not
>>>>>> true.
>>>>>>
>>>>>>>> I'm still looking for some type of statement that you've done some
>>>>>>>> homework on the IORING_OP_MADVISE case to ensure that it doesn't end
>>>>>>>> up calling into the LSM, see my previous emails on this.  I need more
>>>>>>>> than "Steve told me to do this".
>>>>>>>>
>>>>>>>> I basically just want to see that some care and thought has gone into
>>>>>>>> this patch to verify it is correct and good.
>>>>>>>
>>>>>>> Steve suggested I look into a number of iouring ops.  I looked at the
>>>>>>> description code and agreed that it wasn't necessary to audit madvise.
>>>>>>> The rationale for fadvise was detemined to have been conflated with
>>>>>>> fallocate and subsequently dropped.  Steve also suggested a number of
>>>>>>> others and after investigation I decided that their current state was
>>>>>>> correct.  *getxattr you've advised against, so it was dropped.  It
>>>>>>> appears fewer modifications were necessary than originally suspected.
>>>>>>
>>>>>> My concern is that three of the four changes you initially proposed
>>>>>> were rejected, which gives me pause about the fourth.  You mention
>>>>>> that based on your reading of madvise's description you feel auditing
>>>>>> isn't necessary - and you may be right - but based on our experience
>>>>>> so far with this patchset I would like to hear that you have properly
>>>>>> investigated all of the madvise code paths, and I would like that in
>>>>>> the commit description.
>>>>>
>>>>> I think you're being unnecessarily hard on this. Yes, the commit message
>>>>> might be touched up. But madvise is advisory in nature. It is not security
>>>>> relevant. And a grep through the security directory doesn't turn up any
>>>>> hooks.
>>>>
>>>> Agree, it's getting a bit anal... FWIW, patch looks fine to me.
>>>
>>> Call it whatever you want, but the details are often important at this
>>> level of code, and when I see a patch author pushing back on verifying
>>> that their patch is correct it makes me very skeptical.
>>
>> Maybe it isn't intended, but the replies have generally had a pretty
>> condescending tone to them. That's not the best way to engage folks, and
>> may very well be why people just kind of give up on it. Nobody likes
>> debating one-liners forever, particularly not if it isn't inviting.
> 
> I appreciate that you are coming from a different space, but I stand
> by my comments.  Of course you are welcome to your own opinion, but I
> would encourage you to spend some time reading the audit mail archives
> going back a few years before you make comments like the above ... or
> not, that's your call; I recognize it is usually easier to criticize.

I'm just saying how it was received on my end, you can take that as
constructive feedback or ignore it. I don't need to read the archives
for that as it is not related to anything but this thread, it was not
meant to reflect a general concern outside of this thread.

> On a quasi related note to the list/archives: unfortunately there was
> continued resistance to opening up the linux-audit list so I've setup
> audit@...r for upstream audit kernel work moving forward.  The list
> address in MAINTAINERS will get updated during the next merge window
> so hopefully some of the problems you had in the beginning of this
> discussion will be better in the future.

OK good, I keep forgetting to delete it from the replies and get annoyed
at the spam I get back... Thanks for fixing that going forward.

>>> I really would have preferred that you held off from merging this
>>> until this was resolved and ACK'd ... oh well.
>>
>> It's still top of tree. If you want to ack it, let me know and I'll add
>> it. If you want to nak it, give me something concrete to work off of.
> 
> I can't in good conscience ACK it without some comment from Richard
> that he has traced the code paths; this shouldn't be surprising at
> this point.  I'm not going to NACK it or post a revert, I would have
> done that already if I felt that was appropriate.  Right now this
> patch is in a gray area for me in that I suspect it is good, but I
> can't ACK it without some comment that it has been properly
> researched.

Richard, can you do the due diligence here? Steve did say:

"But madvise is advisory in nature. It is not security relevant. And a
grep through the security directory doesn't turn up any hooks."

Seems to me if we're not currently auditing madvise outside of io_uring,
then why would we do it here?

-- 
Jens Axboe

Powered by blists - more mailing lists