[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABBYNZKsxeUim=DtUR2KR2UyD4fxLzSqY452UQKkw+nK5auPpw@mail.gmail.com>
Date: Wed, 30 Oct 2024 10:26:20 -0400
From: Luiz Augusto von Dentz <luiz.dentz@...il.com>
To: Sungwoo Kim <iam@...g-woo.kim>
Cc: daveti@...due.edu, benquike@...il.com,
Marcel Holtmann <marcel@...tmann.org>, Johan Hedberg <johan.hedberg@...il.com>,
linux-bluetooth@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] Bluetooth: hci: fix null-ptr-deref in hci_read_supported_codecs
Hi Sungwoo,
On Tue, Oct 29, 2024 at 3:47 PM Sungwoo Kim <iam@...g-woo.kim> wrote:
>
> Fix __hci_cmd_sync_sk() to return not NULL for unknown opcodes.
>
> __hci_cmd_sync_sk() returns NULL if a command returns a status event.
> However, it also returns NULL where an opcode doesn't exist in the
> hci_cc table because hci_cmd_complete_evt() assumes status = skb->data[0]
> for unknown opcodes.
> This leads to null-ptr-deref in cmd_sync for HCI_OP_READ_LOCAL_CODECS as
> there is no hci_cc for HCI_OP_READ_LOCAL_CODECS, which always assumes
> status = skb->data[0].
>
> KASAN: null-ptr-deref in range [0x0000000000000070-0x0000000000000077]
> CPU: 1 PID: 2000 Comm: kworker/u9:5 Not tainted 6.9.0-ga6bcb805883c-dirty #10
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
> Workqueue: hci7 hci_power_on
> RIP: 0010:hci_read_supported_codecs+0xb9/0x870 net/bluetooth/hci_codec.c:138
> Code: 08 48 89 ef e8 b8 c1 8f fd 48 8b 75 00 e9 96 00 00 00 49 89 c6 48 ba 00 00 00 00 00 fc ff df 4c 8d 60 70 4c 89 e3 48 c1 eb 03 <0f> b6 04 13 84 c0 0f 85 82 06 00 00 41 83 3c 24 02 77 0a e8 bf 78
> RSP: 0018:ffff888120bafac8 EFLAGS: 00010212
> RAX: 0000000000000000 RBX: 000000000000000e RCX: ffff8881173f0040
> RDX: dffffc0000000000 RSI: ffffffffa58496c0 RDI: ffff88810b9ad1e4
> RBP: ffff88810b9ac000 R08: ffffffffa77882a7 R09: 1ffffffff4ef1054
> R10: dffffc0000000000 R11: fffffbfff4ef1055 R12: 0000000000000070
> R13: 0000000000000000 R14: 0000000000000000 R15: ffff88810b9ac000
> FS: 0000000000000000(0000) GS:ffff8881f6c00000(0000) knlGS:0000000000000000
> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00007f6ddaa3439e CR3: 0000000139764003 CR4: 0000000000770ef0
> PKRU: 55555554
> Call Trace:
> <TASK>
> hci_read_local_codecs_sync net/bluetooth/hci_sync.c:4546 [inline]
> hci_init_stage_sync net/bluetooth/hci_sync.c:3441 [inline]
> hci_init4_sync net/bluetooth/hci_sync.c:4706 [inline]
> hci_init_sync net/bluetooth/hci_sync.c:4742 [inline]
> hci_dev_init_sync net/bluetooth/hci_sync.c:4912 [inline]
> hci_dev_open_sync+0x19a9/0x2d30 net/bluetooth/hci_sync.c:4994
> hci_dev_do_open net/bluetooth/hci_core.c:483 [inline]
> hci_power_on+0x11e/0x560 net/bluetooth/hci_core.c:1015
> process_one_work kernel/workqueue.c:3267 [inline]
> process_scheduled_works+0x8ef/0x14f0 kernel/workqueue.c:3348
> worker_thread+0x91f/0xe50 kernel/workqueue.c:3429
> kthread+0x2cb/0x360 kernel/kthread.c:388
> ret_from_fork+0x4d/0x80 arch/x86/kernel/process.c:147
> ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
>
> Fixes: abfeea476c68 ("Bluetooth: hci_sync: Convert MGMT_OP_START_DISCOVERY")
>
> Signed-off-by: Sungwoo Kim <iam@...g-woo.kim>
> ---
> v1 -> v2: make __hci_cmd_sync_sk() not return NULL
> v2 -> v3: elaborate reasoning
>
> net/bluetooth/hci_sync.c | 18 +++++++++++-------
> 1 file changed, 11 insertions(+), 7 deletions(-)
>
> diff --git a/net/bluetooth/hci_sync.c b/net/bluetooth/hci_sync.c
> index 9482bd562..c86f4e42e 100644
> --- a/net/bluetooth/hci_sync.c
> +++ b/net/bluetooth/hci_sync.c
> @@ -206,6 +206,12 @@ struct sk_buff *__hci_cmd_sync_sk(struct hci_dev *hdev, u16 opcode, u32 plen,
> return ERR_PTR(err);
> }
>
> + /* If command return a status event skb will be set to NULL as there are
> + * no parameters.
> + */
> + if (!skb)
> + return ERR_PTR(-ENODATA);
> +
> return skb;
> }
> EXPORT_SYMBOL(__hci_cmd_sync_sk);
> @@ -255,6 +261,11 @@ int __hci_cmd_sync_status_sk(struct hci_dev *hdev, u16 opcode, u32 plen,
> u8 status;
>
> skb = __hci_cmd_sync_sk(hdev, opcode, plen, param, event, timeout, sk);
> +
> + /* If command return a status event, skb will be set to -ENODATA */
> + if (skb == ERR_PTR(-ENODATA))
> + return 0;
> +
> if (IS_ERR(skb)) {
> if (!event)
> bt_dev_err(hdev, "Opcode 0x%4.4x failed: %ld", opcode,
> @@ -262,13 +273,6 @@ int __hci_cmd_sync_status_sk(struct hci_dev *hdev, u16 opcode, u32 plen,
> return PTR_ERR(skb);
> }
>
> - /* If command return a status event skb will be set to NULL as there are
> - * no parameters, in case of failure IS_ERR(skb) would have be set to
> - * the actual error would be found with PTR_ERR(skb).
> - */
> - if (!skb)
> - return 0;
> -
> status = skb->data[0];
>
> kfree_skb(skb);
> --
> 2.43.0
There is something not quite right with this one, it seems to be
causing some errors with the likes of mgmt-tester in particular 'Read
Ext Controller Info', for some reason the index is not 0 so I suspect
something is not working as intended due to these changes.
--
Luiz Augusto von Dentz
Powered by blists - more mailing lists