Date: Wed, 14 Dec 2022 16:42:47 -0800 From: Rick Macklem <rick.macklem@gmail.com> To: Tomoaki AOKI <junchoon@dec.sakura.ne.jp> Cc: freebsd-current@freebsd.org Subject: Re: What to do about a few lines in vfs_domount() never executed? Message-ID: <CAM5tNy5DZ0%2Bw39SmLKHcSkbh_mLWx%2BwkLN_xRuc3W__Z8UaFyg@mail.gmail.com> In-Reply-To: <CAM5tNy6gAFn5GHXycsNRXqTyWM21ALq-G70VrFLRrwnCi1EGyg@mail.gmail.com> References: <CAM5tNy7DM__tj4CQ0L_-ugo5krmEvLJgbU-WFLs47MrRBuGaNQ@mail.gmail.com> <20221215021431.d190e55ee911f5e94799f953@dec.sakura.ne.jp> <CAM5tNy6gAFn5GHXycsNRXqTyWM21ALq-G70VrFLRrwnCi1EGyg@mail.gmail.com>
index | next in thread | previous in thread | raw e-mail
[-- Attachment #1 --]
On Wed, Dec 14, 2022 at 2:04 PM Rick Macklem <rick.macklem@gmail.com> wrote:
>
>
> On Wed, Dec 14, 2022 at 9:15 AM Tomoaki AOKI <junchoon@dec.sakura.ne.jp>
> wrote:
>
>> Tracking the commits, it was originally introduced to
>> sys/kern/vfs_syscalls.c at r22521 [1][2] (Mon Feb 10 02:22:35 1997 by
>> dyson, submitted by hsu@freebsd.org) and later centralized into
>> sys/kern/vfs_mount.c at r99264 [2].
>>
>> But according to the comment above the codes, maybe it would be
>> intended to block userland programs or ports FS modules setting
>> MNT_EXPORTED.
>>
>> If I'm not mis-understanding, it can be the case when
>> *vfs.usermount sysctl is non-zero,
>> *underlying FS (to be exported) allows it, and
>> *non-root user tries to mount the FS via NFS.
>>
> It does appear that ancient code restricted doing NFS exports
> to root only.
>
It looks like the semantics change was introduced when mountd
as converted to nmount() { r158857 about 17years ago }.
The code in vfs_mount.c assumed that MNT_EXPORTED was set in the
argument passed in from mountd.c when it called nmount(), but
that was not the case.
--> As such, the check was not performed.
The check was for suser() until r164033 when it was replaced
by priv_check(td, PRIV_VFS_MOUNT_EDPORTED).
--> This does the same thing, failing if cr_uid != 0.
So, I think the snippet was supposed to enforce "only root
can set exports", but the check has not worked post r158857
because MNT_EXPORTED was never set in fsflags.
(nmount() uses the "export" option.)
So, should I set MNT_EXPORTED in fsflags when nmount()
has specified the "export" option and restore the
"must be root to export" check?
rick
ps: Thanks Tomoaki AOKI for looking at the old code and
spotting this.
> I don't think that restriction is exactly enforced now, since
> vfs_suser() allows a non-root owner to do the update (which
> would include updating exports).
> (As I noted, MNT_EXPORTED never gets set in fsflags. The variable
> is local to one of the functions in vfs_mount.c and a search shows
> it never gets set.)
>
> I suppose you could argue that priv_check(td, PRIV_VFS_MOUNT_EXPORTED)
> should check for caller being root, since that is what ancient code did.
> Or, you could argue that, if a non-root user is allowed to mount a file
> system that they should also be allowed to export it, which is what I
> think the current code allows (and has for at least a decade).
>
> Admittedly, allowing a non-root user to do a mount and also add exports
> to it could cause confusion, since the system basically assumes that
> mountd will manage all exports.
>
> Do others think adding code to check that cr_uid == 0 for
> PRIV_VFS_MOUNT_EXPORTED to be allowed makes sense?
>
> rick
>
>
>
>>
>> [1]
>>
>> https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?revision=22521&view=markup&pathrev=99264
>>
>> [2]
>>
>> https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?r1=22520&r2=22521&pathrev=99264&;
>>
>> [3]
>>
>> https://cgit.freebsd.org/src/commit/sys/kern/vfs_mount.c?id=2b4edb69f1ef62fc38b02ac22b0a3ac09e43fa77
>>
>>
>> On Tue, 13 Dec 2022 14:19:39 -0800
>> Rick Macklem <rick.macklem@gmail.com> wrote:
>>
>> > Hi,
>> >
>> > While working on getting mountd/nfsd to run in a vnet
>> > prison, I came across the following lines near the
>> > beginning of vfs_domount() in sys/kern/vfs_mount.c:
>> >
>> > if (fsflags & MNT_EXPORTED) {
>> > error = priv_check(td, PRIV_VFS_MOUNT_EXPORTED);
>> > if (error)
>> > return (error);
>> > }
>> >
>> > #1 - Since MNT_EXPORTED is never set in fsflags, this code never
>> > gets executed.
>> > --> I am asking what to do with the above code, since that
>> > changes for the patch that allows mountd to run in a vnet
>> > prison.
>> > #2 - priv_check(td, PRIV_VFS_MOUNT_EXPORTED) always returns 0
>> > because nothing in sys/kern/kern_priv.c checks
>> > PRIV_VFS_MOUNT_EXPORTED.
>> >
>> > I don't know what the original author's thinking was w.r.t. this.
>> > Setting exports already checks that the mount operation can be
>> > done by the requestor.
>> >
>> > So, what do you think should be done with the above code snippet?
>> > - Consider it cruft and delete it.
>> > - Try and figure out what PRIV_VFS_MOUNT_EXPORTED should check?
>> > - Leave it as is. After the patch that allows mountd to run in
>> > a vnet prison, MNT_EXPORTED will be set in fsflags, but the
>> > priv_check() call will just return 0. (A little overhead,
>> > but otherwise no semantics change.)
>> >
>> > rick
>>
>>
>> --
>> Tomoaki AOKI <junchoon@dec.sakura.ne.jp>
>>
>>
[-- Attachment #2 --]
<div dir="ltr"><div dir="ltr"><div class="gmail_default" style="font-family:monospace"><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Dec 14, 2022 at 2:04 PM Rick Macklem <<a href="mailto:rick.macklem@gmail.com">rick.macklem@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div dir="ltr"><div style="font-family:monospace"><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Dec 14, 2022 at 9:15 AM Tomoaki AOKI <<a href="mailto:junchoon@dec.sakura.ne.jp" target="_blank">junchoon@dec.sakura.ne.jp</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Tracking the commits, it was originally introduced to<br>
sys/kern/vfs_syscalls.c at r22521 [1][2] (Mon Feb 10 02:22:35 1997 by<br>
dyson, submitted by <a href="mailto:hsu@freebsd.org" target="_blank">hsu@freebsd.org</a>) and later centralized into<br>
sys/kern/vfs_mount.c at r99264 [2].<br>
<br>
But according to the comment above the codes, maybe it would be<br>
intended to block userland programs or ports FS modules setting <br>
MNT_EXPORTED.<br>
<br>
If I'm not mis-understanding, it can be the case when<br>
*vfs.usermount sysctl is non-zero,<br>
*underlying FS (to be exported) allows it, and<br>
*non-root user tries to mount the FS via NFS.<br></blockquote><div><span class="gmail_default" style="font-family:monospace">It does appear that ancient code restricted doing NFS exports</span></div><div><span class="gmail_default" style="font-family:monospace">to root only.</span></div></div></div></blockquote><div><span class="gmail_default" style="font-family:monospace">It looks like the semantics change was introduced when mountd</span></div><div><span class="gmail_default" style="font-family:monospace">as converted to nmount() { r158857 about 17years ago }.</span></div><div><span class="gmail_default" style="font-family:monospace">The code in vfs_mount.c assumed that MNT_EXPORTED was set in the</span></div><div><span class="gmail_default" style="font-family:monospace">argument passed in from mountd.c when it called nmount(), but</span></div><div><span class="gmail_default" style="font-family:monospace">that was not the case.</span></div><div><span class="gmail_default" style="font-family:monospace">--> As such, the check was not performed.</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">The check was for suser() until r164033 when it was replaced</span></div><div><span class="gmail_default" style="font-family:monospace">by priv_check(td, PRIV_VFS_MOUNT_EDPORTED).</span></div><div><span class="gmail_default" style="font-family:monospace">--></span> <span class="gmail_default" style="font-family:monospace">This does the same thing, failing if cr_uid != 0.</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">So, I think the snippet was supposed to enforce "only root</span></div><div><span class="gmail_default" style="font-family:monospace">can set exports", but the check has not worked post r158857</span></div><div><span class="gmail_default" style="font-family:monospace">because MNT_EXPORTED was never set in fsflags.</span></div><div><span class="gmail_default" style="font-family:monospace">(nmount() uses the "export" option.)</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">So, should I set MNT_EXPORTED in fsflags when nmount()</span></div><div><span class="gmail_default" style="font-family:monospace">has specified the "export" option and restore the</span></div><div><span class="gmail_default" style="font-family:monospace">"must be root to export" check?</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">rick</span></div><div><span class="gmail_default" style="font-family:monospace">ps: Thanks Tomoaki AOKI for looking at the old code and</span></div><div><span class="gmail_default" style="font-family:monospace"> spotting this.</span></div><div><span class="gmail_default" style="font-family:monospace"></span></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div class="gmail_quote"><div><span class="gmail_default" style="font-family:monospace">I don't think that restriction is exactly enforced now, since</span></div><div><span class="gmail_default" style="font-family:monospace">vfs_suser() allows a non-root owner to do the update (which</span></div><div><span class="gmail_default" style="font-family:monospace">would include updating exports).</span></div><div><span class="gmail_default" style="font-family:monospace">(As I noted, MNT_EXPORTED never gets set in fsflags. The variable</span></div><div><span class="gmail_default" style="font-family:monospace"> is local to one of the functions in vfs_mount.c and a search shows</span></div><div><span class="gmail_default" style="font-family:monospace"> it never gets set.)</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">I suppose you could argue that priv_check(td, </span> <span class="gmail_default" style="font-family:monospace">PRIV_VFS_MOUNT_EXPORTED)</span></div><div><span class="gmail_default" style="font-family:monospace">should check for caller being root, since that is what ancient code did.</span></div><div><span class="gmail_default" style="font-family:monospace">Or, you could argue that, if a non-root user is allowed to mount a file</span></div><div><span class="gmail_default" style="font-family:monospace">system that they should also be allowed to export it, which is what I</span></div><div><span class="gmail_default" style="font-family:monospace">think the current code allows (and has for at least a decade).</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">Admittedly, allowing a non-root user to do a mount and also add exports</span></div><div><span class="gmail_default" style="font-family:monospace">to it could cause confusion, since the system basically assumes that</span></div><div><span class="gmail_default" style="font-family:monospace">mountd will manage all exports.</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">Do others think adding code to check that cr_uid == 0 for</span></div><div><span class="gmail_default" style="font-family:monospace">PRIV_VFS_MOUNT_EXPORTED to be allowed makes sense?</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace">rick</span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><div><span class="gmail_default" style="font-family:monospace"><br></span></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
<br>
[1]<br>
<a href="https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?revision=22521&view=markup&pathrev=99264" rel="noreferrer" target="_blank">https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?revision=22521&view=markup&pathrev=99264</a><br>;
<br>
[2]<br>
<a href="https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?r1=22520&r2=22521&pathrev=99264&" rel="noreferrer" target="_blank">https://svnweb.freebsd.org/base/head/sys/kern/vfs_syscalls.c?r1=22520&r2=22521&pathrev=99264&</a><br>;
<br>
[3]<br>
<a href="https://cgit.freebsd.org/src/commit/sys/kern/vfs_mount.c?id=2b4edb69f1ef62fc38b02ac22b0a3ac09e43fa77" rel="noreferrer" target="_blank">https://cgit.freebsd.org/src/commit/sys/kern/vfs_mount.c?id=2b4edb69f1ef62fc38b02ac22b0a3ac09e43fa77</a><br>;
<br>
<br>
On Tue, 13 Dec 2022 14:19:39 -0800<br>
Rick Macklem <<a href="mailto:rick.macklem@gmail.com" target="_blank">rick.macklem@gmail.com</a>> wrote:<br>
<br>
> Hi,<br>
> <br>
> While working on getting mountd/nfsd to run in a vnet<br>
> prison, I came across the following lines near the<br>
> beginning of vfs_domount() in sys/kern/vfs_mount.c:<br>
> <br>
> if (fsflags & MNT_EXPORTED) {<br>
> error = priv_check(td, PRIV_VFS_MOUNT_EXPORTED);<br>
> if (error)<br>
> return (error);<br>
> }<br>
> <br>
> #1 - Since MNT_EXPORTED is never set in fsflags, this code never<br>
> gets executed.<br>
> --> I am asking what to do with the above code, since that<br>
> changes for the patch that allows mountd to run in a vnet<br>
> prison.<br>
> #2 - priv_check(td, PRIV_VFS_MOUNT_EXPORTED) always returns 0<br>
> because nothing in sys/kern/kern_priv.c checks<br>
> PRIV_VFS_MOUNT_EXPORTED.<br>
> <br>
> I don't know what the original author's thinking was w.r.t. this.<br>
> Setting exports already checks that the mount operation can be<br>
> done by the requestor.<br>
> <br>
> So, what do you think should be done with the above code snippet?<br>
> - Consider it cruft and delete it.<br>
> - Try and figure out what PRIV_VFS_MOUNT_EXPORTED should check?<br>
> - Leave it as is. After the patch that allows mountd to run in<br>
> a vnet prison, MNT_EXPORTED will be set in fsflags, but the<br>
> priv_check() call will just return 0. (A little overhead,<br>
> but otherwise no semantics change.)<br>
> <br>
> rick<br>
<br>
<br>
-- <br>
Tomoaki AOKI <<a href="mailto:junchoon@dec.sakura.ne.jp" target="_blank">junchoon@dec.sakura.ne.jp</a>><br>
<br>
</blockquote></div></div>
</blockquote></div></div>
help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAM5tNy5DZ0%2Bw39SmLKHcSkbh_mLWx%2BwkLN_xRuc3W__Z8UaFyg>