Revert "usb: gadget: u_ether: synchronize with transmit when stopping queue"
authorLi RongQing <roy.qing.li@gmail.com>
Fri, 25 Jul 2014 06:22:40 +0000 (14:22 +0800)
committerFelipe Balbi <balbi@ti.com>
Tue, 19 Aug 2014 14:51:19 +0000 (09:51 -0500)
This reverts commit a9232076374334ca2bc2a448dfde96d38a54349a.

It introduced a dead lock, and did not fix anything.

it made netif_tx_lock() be called in IRQ context, but in softirq context,
the same lock is locked without disabling IRQ. In fact, the commit a923207637
did not fix anything, since netif_stop_queue did not free the any resource

[   10.154920] =================================
[   10.156026] [ INFO: inconsistent lock state ]
[   10.156026] 3.16.0-rc5+ #13 Not tainted
[   10.156026] ---------------------------------
[   10.156026] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage.
[   10.156026] swapper/1/0 [HC0[0]:SC1[5]:HE1:SE0] takes:
[   10.156026]  (_xmit_ETHER){?.-...}, at: [<80948b6a>] sch_direct_xmit+0x7a/0x250
[   10.156026] {IN-HARDIRQ-W} state was registered at:
[   10.156026]   [<804811f0>] __lock_acquire+0x800/0x17a0
[   10.156026]   [<804828ba>] lock_acquire+0x6a/0xf0
[   10.156026]   [<809ed477>] _raw_spin_lock+0x27/0x40
[   10.156026]   [<8088d508>] gether_disconnect+0x68/0x280
[   10.156026]   [<8088e777>] eem_set_alt+0x37/0xc0
[   10.156026]   [<808847ce>] composite_setup+0x30e/0x1240
[   10.156026]   [<8088b8ae>] pch_udc_isr+0xa6e/0xf50
[   10.156026]   [<8048abe8>] handle_irq_event_percpu+0x38/0x1e0
[   10.156026]   [<8048adc1>] handle_irq_event+0x31/0x50
[   10.156026]   [<8048d94b>] handle_fasteoi_irq+0x6b/0x140
[   10.156026]   [<804040a5>] handle_irq+0x65/0x80
[   10.156026]   [<80403cfc>] do_IRQ+0x3c/0xc0
[   10.156026]   [<809ee6ae>] common_interrupt+0x2e/0x34
[   10.156026]   [<804668c5>] finish_task_switch+0x65/0xd0
[   10.156026]   [<809e89df>] __schedule+0x20f/0x7d0
[   10.156026]   [<809e94aa>] schedule_preempt_disabled+0x2a/0x70
[   10.156026]   [<8047bf03>] cpu_startup_entry+0x143/0x410
[   10.156026]   [<809e2e61>] rest_init+0xa1/0xb0
[   10.156026]   [<80ce2a3b>] start_kernel+0x336/0x33b
[   10.156026]   [<80ce22ab>] i386_start_kernel+0x79/0x7d
[   10.156026] irq event stamp: 52070
[   10.156026] hardirqs last  enabled at (52070): [<809375de>] neigh_resolve_output+0xee/0x2a0
[   10.156026] hardirqs last disabled at (52069): [<809375a8>] neigh_resolve_output+0xb8/0x2a0
[   10.156026] softirqs last  enabled at (52020): [<8044401f>] _local_bh_enable+0x1f/0x50
[   10.156026] softirqs last disabled at (52021): [<80404036>] do_softirq_own_stack+0x26/0x30
[   10.156026]
[   10.156026] other info that might help us debug this:
[   10.156026]  Possible unsafe locking scenario:
[   10.156026]
[   10.156026]        CPU0
[   10.156026]        ----
[   10.156026]   lock(_xmit_ETHER);
[   10.156026]   <Interrupt>
[   10.156026]     lock(_xmit_ETHER);
[   10.156026]
[   10.156026]  *** DEADLOCK ***
[   10.156026]
[   10.156026] 4 locks held by swapper/1/0:
[   10.156026]  #0:  (((&idev->mc_ifc_timer))){+.-...}, at: [<8044b100>] call_timer_fn+0x0/0x190
[   10.156026]  #1:  (rcu_read_lock){......}, at: [<a0577c40>] mld_sendpack+0x0/0x590 [ipv6]
[   10.156026]  #2:  (rcu_read_lock_bh){......}, at: [<a055680c>] ip6_finish_output2+0x4c/0x7f0 [ipv6]
[   10.156026]  #3:  (rcu_read_lock_bh){......}, at: [<8092e510>] __dev_queue_xmit+0x0/0x5f0
[   10.156026]
[   10.156026] stack backtrace:
[   10.156026] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.16.0-rc5+ #13
[   10.156026]  811dbb10 00000000 9e919d10 809e6785 9e8b8000 9e919d3c 809e561e 80b95511
[   10.156026]  80b9545a 80b9543d 80b95450 80b95441 80b957e4 9e8b84e0 00000002 8047f7b0
[   10.156026]  9e919d5c 8048043b 00000002 00000000 9e8b8000 00000001 00000004 9e8b8000
[   10.156026] Call Trace:
[   10.156026]  [<809e6785>] dump_stack+0x48/0x69
[   10.156026]  [<809e561e>] print_usage_bug+0x18f/0x19c
[   10.156026]  [<8047f7b0>] ? print_shortest_lock_dependencies+0x170/0x170
[   10.156026]  [<8048043b>] mark_lock+0x53b/0x5f0
[   10.156026]  [<804810cf>] __lock_acquire+0x6df/0x17a0
[   10.156026]  [<804828ba>] lock_acquire+0x6a/0xf0
[   10.156026]  [<80948b6a>] ? sch_direct_xmit+0x7a/0x250
[   10.156026]  [<809ed477>] _raw_spin_lock+0x27/0x40
[   10.156026]  [<80948b6a>] ? sch_direct_xmit+0x7a/0x250
[   10.156026]  [<80948b6a>] sch_direct_xmit+0x7a/0x250
[   10.156026]  [<8092e6bf>] __dev_queue_xmit+0x1af/0x5f0
[   10.156026]  [<80947fc0>] ? ether_setup+0x80/0x80
[   10.156026]  [<8092eb0f>] dev_queue_xmit+0xf/0x20
[   10.156026]  [<8093764c>] neigh_resolve_output+0x15c/0x2a0
[   10.156026]  [<a0556927>] ip6_finish_output2+0x167/0x7f0 [ipv6]
[   10.156026]  [<a0559b05>] ip6_finish_output+0x85/0x1c0 [ipv6]
[   10.156026]  [<a0559cb7>] ip6_output+0x77/0x240 [ipv6]
[   10.156026]  [<a0578163>] mld_sendpack+0x523/0x590 [ipv6]
[   10.156026]  [<80480501>] ? mark_held_locks+0x11/0x90
[   10.156026]  [<a057947d>] mld_ifc_timer_expire+0x15d/0x280 [ipv6]
[   10.156026]  [<8044b168>] call_timer_fn+0x68/0x190
[   10.156026]  [<a0579320>] ? igmp6_group_added+0x150/0x150 [ipv6]
[   10.156026]  [<8044b3fa>] run_timer_softirq+0x16a/0x240
[   10.156026]  [<a0579320>] ? igmp6_group_added+0x150/0x150 [ipv6]
[   10.156026]  [<80444984>] __do_softirq+0xd4/0x2f0
[   10.156026]  [<804448b0>] ? tasklet_action+0x100/0x100
[   10.156026]  [<80404036>] do_softirq_own_stack+0x26/0x30
[   10.156026]  <IRQ>  [<80444d05>] irq_exit+0x65/0x70
[   10.156026]  [<8042d758>] smp_apic_timer_interrupt+0x38/0x50
[   10.156026]  [<809ee91f>] apic_timer_interrupt+0x2f/0x34
[   10.156026]  [<8048007b>] ? mark_lock+0x17b/0x5f0
[   10.156026]  [<8040a912>] ? default_idle+0x22/0xf0
[   10.156026]  [<8040b13e>] arch_cpu_idle+0xe/0x10
[   10.156026]  [<8047bfc6>] cpu_startup_entry+0x206/0x410
[   10.156026]  [<8042bfbd>] start_secondary+0x19d/0x1e0

Acked-by: Tony Lindgren <tony@atomide.com>
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Jeff Westfahl <jeff.westfahl@ni.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: <linux-usb@vger.kernel.org>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
drivers/usb/gadget/function/u_ether.c

index d50adda..6e6f876 100644 (file)
@@ -1127,10 +1127,7 @@ void gether_disconnect(struct gether *link)
 
        DBG(dev, "%s\n", __func__);
 
-       netif_tx_lock(dev->net);
        netif_stop_queue(dev->net);
-       netif_tx_unlock(dev->net);
-
        netif_carrier_off(dev->net);
 
        /* disable endpoints, forcing (synchronous) completion