[v4] padata: Fix refcnt handling in padata_free_shell()

  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() {
    DECLARE_COMPLETION(done)
    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`.

Fixes: 07928d9bfc81 ("padata: Remove broken queue flushing")

Signed-off-by: WangJinchao <wangjinchao@xfusion.com>
Acked-by: Daniel Jordan <daniel.m.jordan@oracle.com>
---
V4:
    Included Daniel's ack
    Included Herbert's ack
V3: https://lore.kernel.org/all/ZSDWAcUxXcwD4YUZ@fedora/
    Included Daniel's ack
    introduced wrong patch 
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 | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
  

WangJinchao <wangjinchao@xfusion.com> wrote:
> 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() {
>    DECLARE_COMPLETION(done)
>    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`.
> 
> Fixes: 07928d9bfc81 ("padata: Remove broken queue flushing")
> 
> Signed-off-by: WangJinchao <wangjinchao@xfusion.com>
> Acked-by: Daniel Jordan <daniel.m.jordan@oracle.com>
> ---
> V4:
>    Included Daniel's ack
>    Included Herbert's ack
> V3: https://lore.kernel.org/all/ZSDWAcUxXcwD4YUZ@fedora/
>    Included Daniel's ack
>    introduced wrong patch 
> 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 | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/padata.c b/kernel/padata.c
> index 222d60195de6..73108ac75f03 100644
> --- a/kernel/padata.c
> +++ b/kernel/padata.c
> @@ -1102,12 +1102,15 @@ 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(pd);
>        mutex_unlock(&ps->pinst->lock);
> 
>        kfree(ps);

Daniel, could you please reconfirm that you're still with v4?

Thanks,
  

On Fri, Oct 20, 2023 at 01:03:59PM +0800, Herbert Xu wrote:
> WangJinchao <wangjinchao@xfusion.com> wrote:
> > diff --git a/kernel/padata.c b/kernel/padata.c
> > index 222d60195de6..73108ac75f03 100644
> > --- a/kernel/padata.c
> > +++ b/kernel/padata.c
> > @@ -1102,12 +1102,15 @@ EXPORT_SYMBOL(padata_alloc_shell);
> >  */
> > void padata_free_shell(struct padata_shell *ps)
> > {
> > +       struct parallel_data *pd;
> >        if (!ps)
> >                return;

The \n after the declaration disappeared in this version but would be
nice for consistency with the file.  Maybe when applying, no need for
another post.

> Daniel, could you please reconfirm that you're still with v4?

Yes, regardless of above,
Acked-by: Daniel Jordan <daniel.m.jordan@oracle.com>
  

WangJinchao <wangjinchao@xfusion.com> wrote:
> 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() {
>    DECLARE_COMPLETION(done)
>    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`.
> 
> Fixes: 07928d9bfc81 ("padata: Remove broken queue flushing")
> 
> Signed-off-by: WangJinchao <wangjinchao@xfusion.com>
> Acked-by: Daniel Jordan <daniel.m.jordan@oracle.com>
> ---
> V4:
>    Included Daniel's ack
>    Included Herbert's ack
> V3: https://lore.kernel.org/all/ZSDWAcUxXcwD4YUZ@fedora/
>    Included Daniel's ack
>    introduced wrong patch 
> 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 | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)

Patch applied.  Thanks.