Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 23.128.96.36 as permitted sender) client-ip=23.128.96.36;
Date:   Sat, 7 Oct 2023 11:52:33 +0800
From:   Wang Jinchao <wangjinchao@xfusion.com>
To:     Steffen Klassert <steffen.klassert@secunet.com>,
        Daniel Jordan <daniel.m.jordan@oracle.com>,
        <linux-crypto@vger.kernel.org>, <linux-kernel@vger.kernel.org>
CC:     <stone.xulei@xfusion.com>
Subject: [PATCH v3] Fixes: 07928d9bfc81 ("padata: Remove broken queue
 flushing")
Message-ID: <ZSDWAcUxXcwD4YUZ@fedora>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
Precedence: bulk

In a high-load arm64 environment, the pcrypt_aead01 test in LTP can lead to
system UAF (Use-After-Free) issues. Due to the lengthy analysis of the
pcrypt_aead01 function call, I'll describe the problem scenario using a
simplified model:

Suppose there's a user of padata named `user_function` that adheres to
the padata requirement of calling `padata_free_shell` after `serial()`
has been invoked, as demonstrated in the following code:

```c
struct request {
    struct padata_priv padata;
    struct completion *done;
};

void parallel(struct padata_priv *padata) {
    do_something();
}

void serial(struct padata_priv *padata) {
    struct request *request = container_of(padata, struct request, padata);
    complete(request->done);
}

void user_function() {
    struct request request;
    DECLARE_COMPLETION(done)
    request->done = done;
    padata = &request.padata;
    padata->parallel = parallel;
    padata->serial = serial;
    padata_do_parallel();
    wait_for_completion(&done);
    padata_free_shell();
}
```

In the corresponding padata.c file, there's the following code:

```c
static void padata_serial_worker(struct work_struct *serial_work) {
    ...
    cnt = 0;

    while (!list_empty(&local_list)) {
        ...
        padata->serial(padata);
        cnt++;
    }

    local_bh_enable();

    if (refcount_sub_and_test(cnt, &pd->refcnt))
        padata_free_pd(pd);
}
```

Because of the high system load and the accumulation of unexecuted
softirq at this moment, `local_bh_enable()` in padata takes longer
to execute than usual. Subsequently, when accessing `pd->refcnt`,
`pd` has already been released by `padata_free_shell()`, resulting
in a UAF issue with `pd->refcnt`.

The fix is straightforward: add `refcount_dec_and_test` before calling
`padata_free_pd` in `padata_free_shell`.

Signed-off-by: Wang Jinchao <wangjinchao@xfusion.com>
Acked-by: Daniel Jordan <daniel.m.jordan@oracle.com>
---
V3: 
    Include Daniel's ack
V2: https://lore.kernel.org/all/ZRTLHY5A+VqIKhA2@fedora/
    To satisfy Sparse, use rcu_dereference_protected.
    Reported-by: kernel test robot <lkp@intel.com>
    Closes: https://lore.kernel.org/oe-kbuild-all/202309270829.xHgTOMKw-lkp@intel.com/

V1: https://lore.kernel.org/all/ZRE4XvOOhz4HSOgR@fedora/

 kernel/padata.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/kernel/padata.c b/kernel/padata.c
index 222d60195de6..79d04a97ded6 100644
--- a/kernel/padata.c
+++ b/kernel/padata.c
@@ -1102,12 +1102,16 @@ EXPORT_SYMBOL(padata_alloc_shell);
  */
 void padata_free_shell(struct padata_shell *ps)
 {
+	struct parallel_data *pd;
+
 	if (!ps)
 		return;
 
 	mutex_lock(&ps->pinst->lock);
 	list_del(&ps->list);
-	padata_free_pd(rcu_dereference_protected(ps->pd, 1));
+	pd = rcu_dereference_protected(ps->pd, 1);
+	if (refcount_dec_and_test(&pd->refcnt))
+		padata_free_pd(rcu_dereference_protected(ps->pd, 1));
 	mutex_unlock(&ps->pinst->lock);
 
 	kfree(ps);
-- 
2.40.0