[PATCH tty/serial 1/1] tty: serial: imx: fix sysrq lockdep issue

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



From: Fugang Duan <fugang.duan@xxxxxxx>

commit dbdda842fe96 ("printk: Add console owner and waiter logic to
load balance console writes") introduces the lockdep issue for imx
serial driver in sysrq case:
     CPU0                    CPU1
     ----                    ----
lock(&port_lock_key);
                             lock(console_owner);
                             lock(&port_lock_key);
lock(console_owner);

It should unlock port_lock_key in handle_sysrq().

lockdep splats:
======================================================
WARNING: possible circular locking dependency detected
------------------------------------------------------
swapper/0/0 is trying to acquire lock:
c141e540 (console_owner){-...}, at: console_unlock+0x180/0x5e4

but task is already holding lock:
d864e450 (&port_lock_key){-.-.}, at: imx_uart_rxint+0x18/0x294

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&port_lock_key){-.-.}:
       imx_uart_console_write+0x1b4/0x1dc
       console_unlock+0x2a4/0x5e4
       register_console+0x1ec/0x3b4
       uart_add_one_port+0x480/0x4e8
       platform_drv_probe+0x48/0x98
       really_probe+0x1d8/0x34c
       driver_probe_device+0x5c/0x168
       device_driver_attach+0x58/0x60
       __driver_attach+0x58/0xd0
       bus_for_each_dev+0x74/0xbc
       bus_add_driver+0x150/0x1dc
       driver_register+0x74/0x108
       imx_uart_init+0x20/0x40
       do_one_initcall+0x84/0x348
       kernel_init_freeable+0x11c/0x1e0
       kernel_init+0x8/0x110
       ret_from_fork+0x14/0x20
       0x0

-> #0 (console_owner){-...}:
       lock_acquire+0xcc/0x204
       console_unlock+0x208/0x5e4
       vprintk_emit+0x108/0x2c8
       vprintk_default+0x20/0x28
       printk+0x2c/0x58
       __handle_sysrq+0x178/0x1f0
       imx_uart_rxint+0x234/0x294
       imx_uart_int+0x210/0x2f0
       __handle_irq_event_percpu+0x48/0x34c
       handle_irq_event_percpu+0x2c/0x88
       handle_irq_event+0x38/0x5c
       handle_fasteoi_irq+0xc8/0x180
       generic_handle_irq+0x20/0x34
       __handle_domain_irq+0x64/0xdc
       gic_handle_irq+0x48/0x9c
       __irq_svc+0x70/0x98
       cpuidle_enter_state+0x16c/0x520
       cpuidle_enter_state+0x16c/0x520
       cpuidle_enter+0x28/0x38
       do_idle+0x1dc/0x2b0
       cpu_startup_entry+0x18/0x1c
       start_kernel+0x404/0x4cc

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&port_lock_key);
                               lock(console_owner);
                               lock(&port_lock_key);
  lock(console_owner);

 *** DEADLOCK ***

3 locks held by swapper/0/0:
 #0: d864e450 (&port_lock_key){-.-.}, at: imx_uart_rxint+0x18/0x294
 #1: c141eb88 (rcu_read_lock){....}, at: __handle_sysrq+0x0/0x1f0
 #2: c141e4c0 (console_lock){+.+.}, at: vprintk_emit+0xfc/0x2c8

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Tainted: G        W
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[<c0112e30>] (unwind_backtrace) from [<c010cdc0>] (show_stack+0x10/0x14)
[<c010cdc0>] (show_stack) from [<c0d2a73c>] (dump_stack+0xe0/0x114)
[<c0d2a73c>] (dump_stack) from [<c018f744>] (check_noncircular+0x138/0x1ec)
[<c018f744>] (check_noncircular) from [<c0191730>] (__lock_acquire+0xe00/0x2418)
[<c0191730>] (__lock_acquire) from [<c01935c0>] (lock_acquire+0xcc/0x204)
[<c01935c0>] (lock_acquire) from [<c019e41c>] (console_unlock+0x208/0x5e4)
[<c019e41c>] (console_unlock) from [<c019fd38>] (vprintk_emit+0x108/0x2c8)
[<c019fd38>] (vprintk_emit) from [<c019ff18>] (vprintk_default+0x20/0x28)
[<c019ff18>] (vprintk_default) from [<c01a04e0>] (printk+0x2c/0x58)
[<c01a04e0>] (printk) from [<c061e990>] (__handle_sysrq+0x178/0x1f0)
[<c061e990>] (__handle_sysrq) from [<c0635cb0>] (imx_uart_rxint+0x234/0x294)
[<c0635cb0>] (imx_uart_rxint) from [<c06376a8>] (imx_uart_int+0x210/0x2f0)
[<c06376a8>] (imx_uart_int) from [<c01a1cdc>] (__handle_irq_event_percpu+0x48/0x34c)
[<c01a1cdc>] (__handle_irq_event_percpu) from [<c01a200c>] (handle_irq_event_percpu+0x2c/0x88)
[<c01a200c>] (handle_irq_event_percpu) from [<c01a20a0>] (handle_irq_event+0x38/0x5c)
[<c01a20a0>] (handle_irq_event) from [<c01a6930>] (handle_fasteoi_irq+0xc8/0x180)
[<c01a6930>] (handle_fasteoi_irq) from [<c01a0e10>] (generic_handle_irq+0x20/0x34)
[<c01a0e10>] (generic_handle_irq) from [<c01a1418>] (__handle_domain_irq+0x64/0xdc)
[<c01a1418>] (__handle_domain_irq) from [<c054bcd0>] (gic_handle_irq+0x48/0x9c)
[<c054bcd0>] (gic_handle_irq) from [<c0101a70>] (__irq_svc+0x70/0x98)
Exception stack(0xc1401ed8 to 0xc1401f20)
1ec0:                                                       00000001 00000006
1ee0: 00000000 c140c600 00000001 5b0881e8 c140f934 db6eadb8 c1504668 0000098b
1f00: 0000098b 5c28eb3d 00000000 c1401f28 c0194520 c0938fcc 20070013 ffffffff
[<c0101a70>] (__irq_svc) from [<c0938fcc>] (cpuidle_enter_state+0x16c/0x520)
[<c0938fcc>] (cpuidle_enter_state) from [<c09393bc>] (cpuidle_enter+0x28/0x38)
[<c09393bc>] (cpuidle_enter) from [<c016f6a4>] (do_idle+0x1dc/0x2b0)
[<c016f6a4>] (do_idle) from [<c016fb2c>] (cpu_startup_entry+0x18/0x1c)
[<c016fb2c>] (cpu_startup_entry) from [<c1300e20>] (start_kernel+0x404/0x4cc)

Signed-off-by: Fugang Duan <fugang.duan@xxxxxxx>
---
 drivers/tty/serial/imx.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/tty/serial/imx.c b/drivers/tty/serial/imx.c
index a9e20e6..a665439 100644
--- a/drivers/tty/serial/imx.c
+++ b/drivers/tty/serial/imx.c
@@ -736,6 +736,7 @@ static irqreturn_t imx_uart_rxint(int irq, void *dev_id)
 
 	while (imx_uart_readl(sport, USR2) & USR2_RDR) {
 		u32 usr2;
+		int sysrq;
 
 		flg = TTY_NORMAL;
 		sport->port.icount.rx++;
@@ -749,7 +750,11 @@ static irqreturn_t imx_uart_rxint(int irq, void *dev_id)
 				continue;
 		}
 
-		if (uart_handle_sysrq_char(&sport->port, (unsigned char)rx))
+		spin_unlock(&sport->port.lock);
+		sysrq = uart_handle_sysrq_char(&sport->port, (unsigned char)rx);
+		spin_lock(&sport->port.lock);
+
+		if (sysrq)
 			continue;
 
 		if (unlikely(rx & URXD_ERR)) {
-- 
2.7.4





[Index of Archives]     [Kernel Newbies]     [Security]     [Netfilter]     [Bugtraq]     [Linux PPP]     [Linux FS]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Samba]     [Video 4 Linux]     [Linmodem]     [Device Mapper]     [Linux Kernel for ARM]

  Powered by Linux