Re: [PATCH v2] staging: rtl8723bs: remove possible deadlock when disconnect
From: Hans de Goede
Date: Fri Sep 10 2021 - 17:07:09 EST
Hi,
On 9/2/21 11:35 AM, Fabio Aiuto wrote:
> when turning off a connection, lockdep complains with the
> following warning (a modprobe has been done but the same
> happens with a disconnection from NetworkManager,
> it's enough to trigger a cfg80211_disconnect call):
>
> [ 682.855867] ======================================================
> [ 682.855877] WARNING: possible circular locking dependency detected
> [ 682.855887] 5.14.0-rc6+ #16 Tainted: G C OE
> [ 682.855898] ------------------------------------------------------
> [ 682.855906] modprobe/1770 is trying to acquire lock:
> [ 682.855916] ffffb6d000332b00 (&pxmitpriv->lock){+.-.}-{2:2},
> at: rtw_free_stainfo+0x52/0x4a0 [r8723bs]
> [ 682.856073]
> but task is already holding lock:
> [ 682.856081] ffffb6d0003336a8 (&pstapriv->sta_hash_lock){+.-.}-{2:2},
> at: rtw_free_assoc_resources+0x48/0x110 [r8723bs]
> [ 682.856207]
> which lock already depends on the new lock.
>
> [ 682.856215]
> the existing dependency chain (in reverse order) is:
> [ 682.856223]
> -> #1 (&pstapriv->sta_hash_lock){+.-.}-{2:2}:
> [ 682.856247] _raw_spin_lock_bh+0x34/0x40
> [ 682.856265] rtw_get_stainfo+0x9a/0x110 [r8723bs]
> [ 682.856389] rtw_xmit_classifier+0x27/0x130 [r8723bs]
> [ 682.856515] rtw_xmitframe_enqueue+0xa/0x20 [r8723bs]
> [ 682.856642] rtl8723bs_hal_xmit+0x3b/0xb0 [r8723bs]
> [ 682.856752] rtw_xmit+0x4ef/0x890 [r8723bs]
> [ 682.856879] _rtw_xmit_entry+0xba/0x350 [r8723bs]
> [ 682.856981] dev_hard_start_xmit+0xee/0x320
> [ 682.856999] sch_direct_xmit+0x8c/0x330
> [ 682.857014] __dev_queue_xmit+0xba5/0xf00
> [ 682.857030] packet_sendmsg+0x981/0x1b80
> [ 682.857047] sock_sendmsg+0x5b/0x60
> [ 682.857060] __sys_sendto+0xf1/0x160
> [ 682.857073] __x64_sys_sendto+0x24/0x30
> [ 682.857087] do_syscall_64+0x3a/0x80
> [ 682.857102] entry_SYSCALL_64_after_hwframe+0x44/0xae
> [ 682.857117]
> -> #0 (&pxmitpriv->lock){+.-.}-{2:2}:
> [ 682.857142] __lock_acquire+0xfd9/0x1b50
> [ 682.857158] lock_acquire+0xb4/0x2c0
> [ 682.857172] _raw_spin_lock_bh+0x34/0x40
> [ 682.857185] rtw_free_stainfo+0x52/0x4a0 [r8723bs]
> [ 682.857308] rtw_free_assoc_resources+0x53/0x110 [r8723bs]
> [ 682.857415] cfg80211_rtw_disconnect+0x4b/0x70 [r8723bs]
> [ 682.857522] cfg80211_disconnect+0x12e/0x2f0 [cfg80211]
> [ 682.857759] cfg80211_leave+0x2b/0x40 [cfg80211]
> [ 682.857961] cfg80211_netdev_notifier_call+0xa9/0x560 [cfg80211]
> [ 682.858163] raw_notifier_call_chain+0x41/0x50
> [ 682.858180] __dev_close_many+0x62/0x100
> [ 682.858195] dev_close_many+0x7d/0x120
> [ 682.858209] unregister_netdevice_many+0x416/0x680
> [ 682.858225] unregister_netdevice_queue+0xab/0xf0
> [ 682.858240] unregister_netdev+0x18/0x20
> [ 682.858255] rtw_unregister_netdevs+0x28/0x40 [r8723bs]
> [ 682.858360] rtw_dev_remove+0x24/0xd0 [r8723bs]
> [ 682.858463] sdio_bus_remove+0x31/0xd0 [mmc_core]
> [ 682.858532] device_release_driver_internal+0xf7/0x1d0
> [ 682.858550] driver_detach+0x47/0x90
> [ 682.858564] bus_remove_driver+0x77/0xd0
> [ 682.858579] rtw_drv_halt+0xc/0x678 [r8723bs]
> [ 682.858685] __x64_sys_delete_module+0x13f/0x250
> [ 682.858699] do_syscall_64+0x3a/0x80
> [ 682.858715] entry_SYSCALL_64_after_hwframe+0x44/0xae
> [ 682.858729]
> other info that might help us debug this:
>
> [ 682.858737] Possible unsafe locking scenario:
>
> [ 682.858744] CPU0 CPU1
> [ 682.858751] ---- ----
> [ 682.858758] lock(&pstapriv->sta_hash_lock);
> [ 682.858772] lock(&pxmitpriv->lock);
> [ 682.858786] lock(&pstapriv->sta_hash_lock);
> [ 682.858799] lock(&pxmitpriv->lock);
> [ 682.858812]
> *** DEADLOCK ***
>
> [ 682.858820] 5 locks held by modprobe/1770:
> [ 682.858831] #0: ffff8d870697d980 (&dev->mutex){....}-{3:3},
> at: device_release_driver_internal+0x1a/0x1d0
> [ 682.858869] #1: ffffffffbdbbf1c8 (rtnl_mutex){+.+.}-{3:3},
> at: unregister_netdev+0xe/0x20
> [ 682.858906] #2: ffff8d87054ee5e8 (&rdev->wiphy.mtx){+.+.}-{3:3},
> at: cfg80211_netdev_notifier_call+0x9e/0x560 [cfg80211]
> [ 682.859131] #3: ffff8d870f2bc8f0 (&wdev->mtx){+.+.}-{3:3},
> at: cfg80211_leave+0x20/0x40 [cfg80211]
> [ 682.859354] #4: ffffb6d0003336a8 (&pstapriv->sta_hash_lock){+.-.}-{2:2},
> at: rtw_free_assoc_resources+0x48/0x110 [r8723bs]
> [ 682.859482]
> stack backtrace:
> [ 682.859491] CPU: 1 PID: 1770 Comm: modprobe Tainted: G
> C OE 5.14.0-rc6+ #16
> [ 682.859507] Hardware name: LENOVO 80NR/Madrid, BIOS DACN25WW 08/20/2015
> [ 682.859517] Call Trace:
> [ 682.859531] dump_stack_lvl+0x56/0x6f
> [ 682.859551] check_noncircular+0xdb/0xf0
> [ 682.859579] __lock_acquire+0xfd9/0x1b50
> [ 682.859606] lock_acquire+0xb4/0x2c0
> [ 682.859623] ? rtw_free_stainfo+0x52/0x4a0 [r8723bs]
> [ 682.859752] ? mark_held_locks+0x48/0x70
> [ 682.859769] ? rtw_free_stainfo+0x4a/0x4a0 [r8723bs]
> [ 682.859898] _raw_spin_lock_bh+0x34/0x40
> [ 682.859914] ? rtw_free_stainfo+0x52/0x4a0 [r8723bs]
> [ 682.860039] rtw_free_stainfo+0x52/0x4a0 [r8723bs]
> [ 682.860171] rtw_free_assoc_resources+0x53/0x110 [r8723bs]
> [ 682.860286] cfg80211_rtw_disconnect+0x4b/0x70 [r8723bs]
> [ 682.860397] cfg80211_disconnect+0x12e/0x2f0 [cfg80211]
> [ 682.860629] cfg80211_leave+0x2b/0x40 [cfg80211]
> [ 682.860836] cfg80211_netdev_notifier_call+0xa9/0x560 [cfg80211]
> [ 682.861048] ? __lock_acquire+0x4dc/0x1b50
> [ 682.861070] ? lock_is_held_type+0xa8/0x110
> [ 682.861089] ? lock_is_held_type+0xa8/0x110
> [ 682.861104] ? find_held_lock+0x2d/0x90
> [ 682.861120] ? packet_notifier+0x173/0x300
> [ 682.861141] ? lock_release+0xb3/0x250
> [ 682.861160] ? packet_notifier+0x192/0x300
> [ 682.861184] raw_notifier_call_chain+0x41/0x50
> [ 682.861205] __dev_close_many+0x62/0x100
> [ 682.861224] dev_close_many+0x7d/0x120
> [ 682.861245] unregister_netdevice_many+0x416/0x680
> [ 682.861264] ? find_held_lock+0x2d/0x90
> [ 682.861284] unregister_netdevice_queue+0xab/0xf0
> [ 682.861306] unregister_netdev+0x18/0x20
> [ 682.861325] rtw_unregister_netdevs+0x28/0x40 [r8723bs]
> [ 682.861434] rtw_dev_remove+0x24/0xd0 [r8723bs]
> [ 682.861542] sdio_bus_remove+0x31/0xd0 [mmc_core]
> [ 682.861615] device_release_driver_internal+0xf7/0x1d0
> [ 682.861637] driver_detach+0x47/0x90
> [ 682.861656] bus_remove_driver+0x77/0xd0
> [ 682.861674] rtw_drv_halt+0xc/0x678 [r8723bs]
> [ 682.861782] __x64_sys_delete_module+0x13f/0x250
> [ 682.861801] ? lockdep_hardirqs_on_prepare+0xf3/0x170
> [ 682.861817] ? syscall_enter_from_user_mode+0x20/0x70
> [ 682.861836] do_syscall_64+0x3a/0x80
> [ 682.861855] entry_SYSCALL_64_after_hwframe+0x44/0xae
> [ 682.861873] RIP: 0033:0x7f6dbe85400b
> [ 682.861890] Code: 73 01 c3 48 8b 0d 6d 1e 0c 00 f7 d8 64 89
> 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa
> b8 b0 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 3d
> 1e 0c 00 f7 d8 64 89 01 48
> [ 682.861906] RSP: 002b:00007ffe7a82f538 EFLAGS: 00000206 ORIG_RAX: 00000000000000b0
> [ 682.861923] RAX: ffffffffffffffda RBX: 000055a64693bd20 RCX: 00007f6dbe85400b
> [ 682.861935] RDX: 0000000000000000 RSI: 0000000000000800 RDI: 000055a64693bd88
> [ 682.861946] RBP: 000055a64693bd20 R08: 0000000000000000 R09: 0000000000000000
> [ 682.861957] R10: 00007f6dbe8c7ac0 R11: 0000000000000206 R12: 000055a64693bd88
> [ 682.861967] R13: 0000000000000000 R14: 000055a64693bd88 R15: 00007ffe7a831848
>
> This happens because when we enqueue a frame for
> transmission we do it under xmit_priv lock, then calling
> rtw_get_stainfo (needed for enqueuing) takes sta_hash_lock
> and this leads to the following lock dependency:
>
> xmit_priv->lock -> sta_hash_lock
>
> Turning off a connection will bring to call
> rtw_free_assoc_resources which will set up
> the inverse dependency:
>
> sta_hash_lock -> xmit_priv_lock
>
> This could lead to a deadlock as lockdep complains.
>
> Fix it by removing the xmit_priv->lock around
> rtw_xmitframe_enqueue call inside rtl8723bs_hal_xmit
> and put it in a smaller critical section inside
> rtw_xmit_classifier, the only place where
> xmit_priv data are actually accessed.
>
> Replace spin_{lock,unlock}_bh(pxmitpriv->lock)
> in other tx paths leading to rtw_xmitframe_enqueue
> call with spin_{lock,unlock}_bh(psta->sleep_q.lock)
> - it's not clear why accessing a sleep_q was protected
> by a spinlock on xmitpriv->lock.
>
> This way is avoided the same faulty lock nesting
> order.
>
> CC: Larry Finger <Larry.Finger@xxxxxxxxxxxx>
> Tested-on: Lenovo Ideapad MiiX 300-10IBY
> Signed-off-by: Fabio Aiuto <fabioaiuto83@xxxxxxxxx>
Thank you for your work on this. Overall this looks good.
I have one remark, since now you are relying on the
sta->sleep_q.lock to protect the sleep_q data, you also
need to update the sleep_q accesses in rtw_free_stainfo()
specifically you need to add a spin_{lock,unlock}_bh(psta->sleep_q.lock)
around these lines:
rtw_free_xmitframe_queue(pxmitpriv, &psta->sleep_q);
psta->sleepq_len = 0;
Note there also is a:
spin_lock_bh(&pxmitpriv->lock);
Just above this which needs to be pushed down to below the
block which takes the psta->sleep_q.lock, so that the entire
thing ends up looking like this:
spin_lock_bh(&psta->sleep_q.lock);
rtw_free_xmitframe_queue(pxmitpriv, &psta->sleep_q);
psta->sleepq_len = 0;
spin_unlock_bh(&psta->sleep_q.lock);
spin_lock_bh(&pxmitpriv->lock);
Other then that this patch looks good, thanks.
Regards,
Hans
> ---
> Changes in v2:
> - remove an unused local variable to suppress
> gcc warning intoduced by the patch
>
> drivers/staging/rtl8723bs/core/rtw_mlme_ext.c | 7 ++-----
> drivers/staging/rtl8723bs/core/rtw_recv.c | 10 +++-------
> drivers/staging/rtl8723bs/core/rtw_xmit.c | 13 +++++++------
> drivers/staging/rtl8723bs/hal/rtl8723bs_xmit.c | 2 --
> 4 files changed, 12 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/staging/rtl8723bs/core/rtw_mlme_ext.c b/drivers/staging/rtl8723bs/core/rtw_mlme_ext.c
> index 375d2a742dd2..a1ae16ec69eb 100644
> --- a/drivers/staging/rtl8723bs/core/rtw_mlme_ext.c
> +++ b/drivers/staging/rtl8723bs/core/rtw_mlme_ext.c
> @@ -5919,7 +5919,6 @@ u8 chk_bmc_sleepq_hdl(struct adapter *padapter, unsigned char *pbuf)
> struct sta_info *psta_bmc;
> struct list_head *xmitframe_plist, *xmitframe_phead, *tmp;
> struct xmit_frame *pxmitframe = NULL;
> - struct xmit_priv *pxmitpriv = &padapter->xmitpriv;
> struct sta_priv *pstapriv = &padapter->stapriv;
>
> /* for BC/MC Frames */
> @@ -5930,8 +5929,7 @@ u8 chk_bmc_sleepq_hdl(struct adapter *padapter, unsigned char *pbuf)
> if ((pstapriv->tim_bitmap&BIT(0)) && (psta_bmc->sleepq_len > 0)) {
> msleep(10);/* 10ms, ATIM(HIQ) Windows */
>
> - /* spin_lock_bh(&psta_bmc->sleep_q.lock); */
> - spin_lock_bh(&pxmitpriv->lock);
> + spin_lock_bh(&psta_bmc->sleep_q.lock);
>
> xmitframe_phead = get_list_head(&psta_bmc->sleep_q);
> list_for_each_safe(xmitframe_plist, tmp, xmitframe_phead) {
> @@ -5954,8 +5952,7 @@ u8 chk_bmc_sleepq_hdl(struct adapter *padapter, unsigned char *pbuf)
> rtw_hal_xmitframe_enqueue(padapter, pxmitframe);
> }
>
> - /* spin_unlock_bh(&psta_bmc->sleep_q.lock); */
> - spin_unlock_bh(&pxmitpriv->lock);
> + spin_unlock_bh(&psta_bmc->sleep_q.lock);
>
> /* check hi queue and bmc_sleepq */
> rtw_chk_hi_queue_cmd(padapter);
> diff --git a/drivers/staging/rtl8723bs/core/rtw_recv.c b/drivers/staging/rtl8723bs/core/rtw_recv.c
> index 105fe0e3482a..41bfca549c64 100644
> --- a/drivers/staging/rtl8723bs/core/rtw_recv.c
> +++ b/drivers/staging/rtl8723bs/core/rtw_recv.c
> @@ -957,10 +957,8 @@ static signed int validate_recv_ctrl_frame(struct adapter *padapter, union recv_
> if ((psta->state&WIFI_SLEEP_STATE) && (pstapriv->sta_dz_bitmap&BIT(psta->aid))) {
> struct list_head *xmitframe_plist, *xmitframe_phead;
> struct xmit_frame *pxmitframe = NULL;
> - struct xmit_priv *pxmitpriv = &padapter->xmitpriv;
>
> - /* spin_lock_bh(&psta->sleep_q.lock); */
> - spin_lock_bh(&pxmitpriv->lock);
> + spin_lock_bh(&psta->sleep_q.lock);
>
> xmitframe_phead = get_list_head(&psta->sleep_q);
> xmitframe_plist = get_next(xmitframe_phead);
> @@ -991,12 +989,10 @@ static signed int validate_recv_ctrl_frame(struct adapter *padapter, union recv_
> update_beacon(padapter, WLAN_EID_TIM, NULL, true);
> }
>
> - /* spin_unlock_bh(&psta->sleep_q.lock); */
> - spin_unlock_bh(&pxmitpriv->lock);
> + spin_unlock_bh(&psta->sleep_q.lock);
>
> } else {
> - /* spin_unlock_bh(&psta->sleep_q.lock); */
> - spin_unlock_bh(&pxmitpriv->lock);
> + spin_unlock_bh(&psta->sleep_q.lock);
>
> if (pstapriv->tim_bitmap&BIT(psta->aid)) {
> if (psta->sleepq_len == 0) {
> diff --git a/drivers/staging/rtl8723bs/core/rtw_xmit.c b/drivers/staging/rtl8723bs/core/rtw_xmit.c
> index 3eb6db2f26bb..878264f46980 100644
> --- a/drivers/staging/rtl8723bs/core/rtw_xmit.c
> +++ b/drivers/staging/rtl8723bs/core/rtw_xmit.c
> @@ -1797,6 +1797,7 @@ s32 rtw_xmit_classifier(struct adapter *padapter, struct xmit_frame *pxmitframe)
> struct sta_info *psta;
> struct tx_servq *ptxservq;
> struct pkt_attrib *pattrib = &pxmitframe->attrib;
> + struct xmit_priv *xmit_priv = &padapter->xmitpriv;
> struct hw_xmit *phwxmits = padapter->xmitpriv.hwxmits;
> signed int res = _SUCCESS;
>
> @@ -1814,12 +1815,14 @@ s32 rtw_xmit_classifier(struct adapter *padapter, struct xmit_frame *pxmitframe)
>
> ptxservq = rtw_get_sta_pending(padapter, psta, pattrib->priority, (u8 *)(&ac_index));
>
> + spin_lock_bh(&xmit_priv->lock);
> if (list_empty(&ptxservq->tx_pending))
> list_add_tail(&ptxservq->tx_pending, get_list_head(phwxmits[ac_index].sta_queue));
>
> list_add_tail(&pxmitframe->list, get_list_head(&ptxservq->sta_pending));
> ptxservq->qcnt++;
> phwxmits[ac_index].accnt++;
> + spin_unlock_bh(&xmit_priv->lock);
>
> exit:
>
> @@ -2202,11 +2205,10 @@ void wakeup_sta_to_xmit(struct adapter *padapter, struct sta_info *psta)
> struct list_head *xmitframe_plist, *xmitframe_phead, *tmp;
> struct xmit_frame *pxmitframe = NULL;
> struct sta_priv *pstapriv = &padapter->stapriv;
> - struct xmit_priv *pxmitpriv = &padapter->xmitpriv;
>
> psta_bmc = rtw_get_bcmc_stainfo(padapter);
>
> - spin_lock_bh(&pxmitpriv->lock);
> + spin_lock_bh(&psta->sleep_q.lock);
>
> xmitframe_phead = get_list_head(&psta->sleep_q);
> list_for_each_safe(xmitframe_plist, tmp, xmitframe_phead) {
> @@ -2307,7 +2309,7 @@ void wakeup_sta_to_xmit(struct adapter *padapter, struct sta_info *psta)
>
> _exit:
>
> - spin_unlock_bh(&pxmitpriv->lock);
> + spin_unlock_bh(&psta->sleep_q.lock);
>
> if (update_mask)
> update_beacon(padapter, WLAN_EID_TIM, NULL, true);
> @@ -2319,9 +2321,8 @@ void xmit_delivery_enabled_frames(struct adapter *padapter, struct sta_info *pst
> struct list_head *xmitframe_plist, *xmitframe_phead, *tmp;
> struct xmit_frame *pxmitframe = NULL;
> struct sta_priv *pstapriv = &padapter->stapriv;
> - struct xmit_priv *pxmitpriv = &padapter->xmitpriv;
>
> - spin_lock_bh(&pxmitpriv->lock);
> + spin_lock_bh(&psta->sleep_q.lock);
>
> xmitframe_phead = get_list_head(&psta->sleep_q);
> list_for_each_safe(xmitframe_plist, tmp, xmitframe_phead) {
> @@ -2374,7 +2375,7 @@ void xmit_delivery_enabled_frames(struct adapter *padapter, struct sta_info *pst
> }
> }
>
> - spin_unlock_bh(&pxmitpriv->lock);
> + spin_unlock_bh(&psta->sleep_q.lock);
> }
>
> void enqueue_pending_xmitbuf(struct xmit_priv *pxmitpriv, struct xmit_buf *pxmitbuf)
> diff --git a/drivers/staging/rtl8723bs/hal/rtl8723bs_xmit.c b/drivers/staging/rtl8723bs/hal/rtl8723bs_xmit.c
> index 156d6aba18ca..5f5c4719b586 100644
> --- a/drivers/staging/rtl8723bs/hal/rtl8723bs_xmit.c
> +++ b/drivers/staging/rtl8723bs/hal/rtl8723bs_xmit.c
> @@ -507,9 +507,7 @@ s32 rtl8723bs_hal_xmit(
> rtw_issue_addbareq_cmd(padapter, pxmitframe);
> }
>
> - spin_lock_bh(&pxmitpriv->lock);
> err = rtw_xmitframe_enqueue(padapter, pxmitframe);
> - spin_unlock_bh(&pxmitpriv->lock);
> if (err != _SUCCESS) {
> rtw_free_xmitframe(pxmitpriv, pxmitframe);
>
>