[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5e51f9a5-ba76-a42d-fc2b-9255f8544859@tycho.nsa.gov>
Date: Tue, 19 Nov 2019 13:59:40 -0500
From: Stephen Smalley <sds@...ho.nsa.gov>
To: Will Deacon <will@...nel.org>, selinux@...r.kernel.org
Cc: linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH 1/2] selinux: Don't call avc_compute_av() from RCU
path walk
On 11/19/19 1:40 PM, Will Deacon wrote:
> 'avc_compute_av()' can block, so we carefully exit the RCU read-side
> critical section before calling it in 'avc_has_perm_noaudit()'.
> Unfortunately, if we're calling from the VFS layer on the RCU path walk
> via 'selinux_inode_permission()' then we're still actually in an RCU
> read-side critical section and must not block.
avc_compute_av() should never block AFAIK. The blocking concern was with
slow_avc_audit(), and even that appears dubious to me. That seems to be
more about misuse of d_find_alias in dump_common_audit_data() than anything.
>
> 'avc_denied()' already handles this by simply returning success and
> postponing the auditing until we're called again on the slowpath, so
> follow the same approach here and return early if the node lookup fails
> on the RCU walk path.
>
> Signed-off-by: Will Deacon <will@...nel.org>
> ---
> security/selinux/avc.c | 9 ++++++---
> 1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/security/selinux/avc.c b/security/selinux/avc.c
> index ecd3829996aa..9c183c899e92 100644
> --- a/security/selinux/avc.c
> +++ b/security/selinux/avc.c
> @@ -1159,16 +1159,19 @@ inline int avc_has_perm_noaudit(struct selinux_state *state,
> rcu_read_lock();
>
> node = avc_lookup(state->avc, ssid, tsid, tclass);
> - if (unlikely(!node))
> + if (unlikely(!node)) {
> + if (flags & AVC_NONBLOCKING)
> + goto out;
> node = avc_compute_av(state, ssid, tsid, tclass, avd, &xp_node);
> - else
> + } else {
> memcpy(avd, &node->ae.avd, sizeof(*avd));
> + }
>
> denied = requested & ~(avd->allowed);
> if (unlikely(denied))
> rc = avc_denied(state, ssid, tsid, tclass, requested, 0, 0,
> flags, avd);
> -
> +out:
> rcu_read_unlock();
> return rc;
> }
>
Powered by blists - more mailing lists