2023-07-11 12:09:08

by Pu Lehui

Subject: [PATCH bpf] bpf: cpumap: Fix memory leak in cpu_map_update_elem

From: Pu Lehui <[email protected]>

Syzkaller reported a memory leak as follows:

BUG: memory leak
unreferenced object 0xff110001198ef748 (size 192):
comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
hex dump (first 32 bytes):
00 00 00 00 4a 19 00 00 80 ad e3 e4 fe ff c0 00 ....J...........
00 b2 d3 0c 01 00 11 ff 28 f5 8e 19 01 00 11 ff ........(.......
backtrace:
[<ffffffffadd28087>] __cpu_map_entry_alloc+0xf7/0xb00
[<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
[<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
[<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
[<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
[<ffffffffb029cc80>] do_syscall_64+0x30/0x40
[<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6

BUG: memory leak
unreferenced object 0xff110001198ef528 (size 192):
comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<ffffffffadd281f0>] __cpu_map_entry_alloc+0x260/0xb00
[<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
[<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
[<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
[<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
[<ffffffffb029cc80>] do_syscall_64+0x30/0x40
[<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6

BUG: memory leak
unreferenced object 0xff1100010fd93d68 (size 8):
comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
hex dump (first 8 bytes):
00 00 00 00 00 00 00 00 ........
backtrace:
[<ffffffffade5db3e>] kvmalloc_node+0x11e/0x170
[<ffffffffadd28280>] __cpu_map_entry_alloc+0x2f0/0xb00
[<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
[<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
[<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
[<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
[<ffffffffb029cc80>] do_syscall_64+0x30/0x40
[<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6

In the cpu_map_update_elem flow, when kthread_stop is called before
calling the threadfn of rcpu->kthread, since the KTHREAD_SHOULD_STOP bit
of kthread has been set by kthread_stop, the threadfn of rcpu->kthread
will never be executed, and rcpu->refcnt will never be 0, which will
lead to the allocated rcpu, rcpu->queue and rcpu->queue->queue cannot be
released.

Calling kthread_stop before executing kthread's threadfn will return
-EINTR. We can complete the release of memory resources in this state.

Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP")
Signed-off-by: Pu Lehui <[email protected]>
---
kernel/bpf/cpumap.c | 40 ++++++++++++++++++++++++----------------
1 file changed, 24 insertions(+), 16 deletions(-)

diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c
index 8a33e8747a0e..6ae02be7a48e 100644
--- a/kernel/bpf/cpumap.c
+++ b/kernel/bpf/cpumap.c
@@ -122,22 +122,6 @@ static void get_cpu_map_entry(struct bpf_cpu_map_entry *rcpu)
atomic_inc(&rcpu->refcnt);
}

-/* called from workqueue, to workaround syscall using preempt_disable */
-static void cpu_map_kthread_stop(struct work_struct *work)
-{
- struct bpf_cpu_map_entry *rcpu;
-
- rcpu = container_of(work, struct bpf_cpu_map_entry, kthread_stop_wq);
-
- /* Wait for flush in __cpu_map_entry_free(), via full RCU barrier,
- * as it waits until all in-flight call_rcu() callbacks complete.
- */
- rcu_barrier();
-
- /* kthread_stop will wake_up_process and wait for it to complete */
- kthread_stop(rcpu->kthread);
-}
-
static void __cpu_map_ring_cleanup(struct ptr_ring *ring)
{
/* The tear-down procedure should have made sure that queue is
@@ -165,6 +149,30 @@ static void put_cpu_map_entry(struct bpf_cpu_map_entry *rcpu)
}
}

+/* called from workqueue, to workaround syscall using preempt_disable */
+static void cpu_map_kthread_stop(struct work_struct *work)
+{
+ struct bpf_cpu_map_entry *rcpu;
+ int err;
+
+ rcpu = container_of(work, struct bpf_cpu_map_entry, kthread_stop_wq);
+
+ /* Wait for flush in __cpu_map_entry_free(), via full RCU barrier,
+ * as it waits until all in-flight call_rcu() callbacks complete.
+ */
+ rcu_barrier();
+
+ /* kthread_stop will wake_up_process and wait for it to complete */
+ err = kthread_stop(rcpu->kthread);
+ if (err) {
+ /* kthread_stop may be called before cpu_map_kthread_run
+ * is executed, so we need to release the memory related
+ * to rcpu.
+ */
+ put_cpu_map_entry(rcpu);
+ }
+}
+
static void cpu_map_bpf_prog_run_skb(struct bpf_cpu_map_entry *rcpu,
struct list_head *listp,
struct xdp_cpumap_stats *stats)
--
2.25.1

2023-07-11 16:22:30

by Jesper Dangaard Brouer

Subject: Re: [PATCH bpf] bpf: cpumap: Fix memory leak in cpu_map_update_elem

On 11/07/2023 13.58, Pu Lehui wrote:
> From: Pu Lehui <[email protected]>
>
> Syzkaller reported a memory leak as follows:
>
[...]>
> In the cpu_map_update_elem flow, when kthread_stop is called before
> calling the threadfn of rcpu->kthread, since the KTHREAD_SHOULD_STOP bit
> of kthread has been set by kthread_stop, the threadfn of rcpu->kthread
> will never be executed, and rcpu->refcnt will never be 0, which will
> lead to the allocated rcpu, rcpu->queue and rcpu->queue->queue cannot be
> released.
>
> Calling kthread_stop before executing kthread's threadfn will return
> -EINTR. We can complete the release of memory resources in this state.
>
> Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP")
> Signed-off-by: Pu Lehui <[email protected]>

LGTM, thanks for fixing this.

Acked-by: Jesper Dangaard Brouer <[email protected]>

2023-07-12 03:13:14

by Hou Tao

Subject: Re: [PATCH bpf] bpf: cpumap: Fix memory leak in cpu_map_update_elem

On 7/11/2023 7:58 PM, Pu Lehui wrote:
> From: Pu Lehui <[email protected]>
>
> Syzkaller reported a memory leak as follows:
>
> BUG: memory leak
> unreferenced object 0xff110001198ef748 (size 192):
> comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> hex dump (first 32 bytes):
> 00 00 00 00 4a 19 00 00 80 ad e3 e4 fe ff c0 00 ....J...........
> 00 b2 d3 0c 01 00 11 ff 28 f5 8e 19 01 00 11 ff ........(.......
> backtrace:
> [<ffffffffadd28087>] __cpu_map_entry_alloc+0xf7/0xb00
> [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
>
> BUG: memory leak
> unreferenced object 0xff110001198ef528 (size 192):
> comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> hex dump (first 32 bytes):
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> backtrace:
> [<ffffffffadd281f0>] __cpu_map_entry_alloc+0x260/0xb00
> [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
>
> BUG: memory leak
> unreferenced object 0xff1100010fd93d68 (size 8):
> comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> hex dump (first 8 bytes):
> 00 00 00 00 00 00 00 00 ........
> backtrace:
> [<ffffffffade5db3e>] kvmalloc_node+0x11e/0x170
> [<ffffffffadd28280>] __cpu_map_entry_alloc+0x2f0/0xb00
> [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
>
> In the cpu_map_update_elem flow, when kthread_stop is called before
> calling the threadfn of rcpu->kthread, since the KTHREAD_SHOULD_STOP bit
> of kthread has been set by kthread_stop, the threadfn of rcpu->kthread
> will never be executed, and rcpu->refcnt will never be 0, which will
> lead to the allocated rcpu, rcpu->queue and rcpu->queue->queue cannot be
> released.
>
> Calling kthread_stop before executing kthread's threadfn will return
> -EINTR. We can complete the release of memory resources in this state.
>
> Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP")
> Signed-off-by: Pu Lehui <[email protected]>

Acked-by: Hou Tao <[email protected]>

2023-07-12 03:32:27

by Alexei Starovoitov

Subject: Re: [PATCH bpf] bpf: cpumap: Fix memory leak in cpu_map_update_elem

On Tue, Jul 11, 2023 at 7:12 PM Hou Tao <[email protected]> wrote:
>
>
>
> On 7/11/2023 7:58 PM, Pu Lehui wrote:
> > From: Pu Lehui <[email protected]>
> >
> > Syzkaller reported a memory leak as follows:
> >
> > BUG: memory leak
> > unreferenced object 0xff110001198ef748 (size 192):
> > comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> > hex dump (first 32 bytes):
> > 00 00 00 00 4a 19 00 00 80 ad e3 e4 fe ff c0 00 ....J...........
> > 00 b2 d3 0c 01 00 11 ff 28 f5 8e 19 01 00 11 ff ........(.......
> > backtrace:
> > [<ffffffffadd28087>] __cpu_map_entry_alloc+0xf7/0xb00
> > [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> > [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> > [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> > [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> > [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> > [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
> >
> > BUG: memory leak
> > unreferenced object 0xff110001198ef528 (size 192):
> > comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> > hex dump (first 32 bytes):
> > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > backtrace:
> > [<ffffffffadd281f0>] __cpu_map_entry_alloc+0x260/0xb00
> > [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> > [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> > [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> > [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> > [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> > [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
> >
> > BUG: memory leak
> > unreferenced object 0xff1100010fd93d68 (size 8):
> > comm "syz-executor.3", pid 17672, jiffies 4298118891 (age 9.906s)
> > hex dump (first 8 bytes):
> > 00 00 00 00 00 00 00 00 ........
> > backtrace:
> > [<ffffffffade5db3e>] kvmalloc_node+0x11e/0x170
> > [<ffffffffadd28280>] __cpu_map_entry_alloc+0x2f0/0xb00
> > [<ffffffffadd28d8e>] cpu_map_update_elem+0x2fe/0x3d0
> > [<ffffffffadc6d0fd>] bpf_map_update_value.isra.0+0x2bd/0x520
> > [<ffffffffadc7349b>] map_update_elem+0x4cb/0x720
> > [<ffffffffadc7d983>] __se_sys_bpf+0x8c3/0xb90
> > [<ffffffffb029cc80>] do_syscall_64+0x30/0x40
> > [<ffffffffb0400099>] entry_SYSCALL_64_after_hwframe+0x61/0xc6
> >
> > In the cpu_map_update_elem flow, when kthread_stop is called before
> > calling the threadfn of rcpu->kthread, since the KTHREAD_SHOULD_STOP bit
> > of kthread has been set by kthread_stop, the threadfn of rcpu->kthread
> > will never be executed, and rcpu->refcnt will never be 0, which will
> > lead to the allocated rcpu, rcpu->queue and rcpu->queue->queue cannot be
> > released.
> >
> > Calling kthread_stop before executing kthread's threadfn will return
> > -EINTR. We can complete the release of memory resources in this state.
> >
> > Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP")
> > Signed-off-by: Pu Lehui <[email protected]>
>
> Acked-by: Hou Tao <[email protected]>
>

Applied. Thanks