On Fri, 2018-05-18 at 07:02 +0000, Phil Edworthy wrote:
> Hi Greg,
>
> On 18 May 2018 07:56 Greg Kroah-Hartman wrote:
> > On Fri, May 18, 2018 at 10:20:13AM +0800, Shawn Lin wrote:
> > > Hi,
> > >
> > > With the following patch appled,
> > >
> > > commit d76c74387e1c978b6c5524a146ab0f3f72206f98
> > > Author: Phil Edworthy <phil.edworthy@xxxxxxxxxxx>
> > > Date: Thu May 10 15:13:56 2018 +0100
> > >
> > > serial: 8250_dw: Fix runtime PM handling
> > >
> > > When using kgdb, you get an abort when accessing the UART
> > > registers.
> > > This is because the driver has already entered runtime PM and
> > > so turned
> > > off the bus clock needed to access the registers.
> > >
> > > To fix this, set the capability indicating Runtime PM is
> > > active
> > > while idle.
> > >
> > > Signed-off-by: Phil Edworthy <phil.edworthy@xxxxxxxxxxx>
> > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > >
> > >
> > > leading to the following splat:
> > >
> > >
> > > [ 30.419892] Serial: 8250/16550 driver, 4 ports, IRQ sharing
> > > enabled
> > > [ 30.682642] ff180000.serial: ttyS0 at MMIO 0xff180000 (irq =
> > > 38,
> > > base_baud = 1500000) is a 16550A
> > > [ 30.770040] ff1a0000.serial: ttyS2 at MMIO 0xff1a0000 (irq =
> > > 39,
> > > base_baud = 1500000) is a 16550A
> > > [ 30.780849] console [ttyS2] enabled
> > > [ 30.780849] console [ttyS2] enabled
> > > [ 30.785492] BUG: sleeping function called from invalid context
> > > at
> > > drivers/base/power/runtime.c:1034
> > > [ 30.785492] BUG: sleeping function called from invalid context
> > > at
> > > drivers/base/power/runtime.c:1034
> > > [ 30.785554] in_atomic(): 1, irqs_disabled(): 128, pid: 1, name:
> > > swapper/0
> > > [ 30.785554] in_atomic(): 1, irqs_disabled(): 128, pid: 1, name:
> > > swapper/0
> > > [ 30.785596] 7 locks held by swapper/0/1:
> > > [ 30.785596] 7 locks held by swapper/0/1:
> > > [ 30.785630] #0: (____ptrval____) (&dev->mutex){....}, at:
> > > __driver_attach+0x7c/0xf0
> > > [ 30.785630] #0: (____ptrval____) (&dev->mutex){....}, at:
> > > __driver_attach+0x7c/0xf0
> > > [ 30.785839] #1: (____ptrval____) (&dev->mutex){....}, at:
> > > __driver_attach+0x8c/0xf0
> > > [ 30.785839] #1: (____ptrval____) (&dev->mutex){....}, at:
> > > __driver_attach+0x8c/0xf0
> > > [ 30.786042] #2: (____ptrval____) (serial_mutex){+.+.}, at:
> > > serial8250_register_8250_port+0x48/0x708
> > > [ 30.786042] #2: (____ptrval____) (serial_mutex){+.+.}, at:
> > > serial8250_register_8250_port+0x48/0x708
> > > [ 30.786253] #3: (____ptrval____) (port_mutex){+.+.}, at:
> > > uart_add_one_port+0xe0/0x848
> > > [ 30.786253] #3: (____ptrval____) (port_mutex){+.+.}, at:
> > > uart_add_one_port+0xe0/0x848
> > > [ 30.786457] #4: (____ptrval____) (&port->mutex){+.+.}, at:
> > > uart_add_one_port+0xf4/0x848
> > > [ 30.786457] #4: (____ptrval____) (&port->mutex){+.+.}, at:
> > > uart_add_one_port+0xf4/0x848
> > > [ 30.786661] #5: (____ptrval____) (console_lock){+.+.}, at:
> > > vprintk_emit+0x284/0x4a0
> > > [ 30.786661] #5: (____ptrval____) (console_lock){+.+.}, at:
> > > vprintk_emit+0x284/0x4a0
> > > [ 30.786864] #6: (____ptrval____) (console_owner){....}, at:
> > > console_unlock+0x238/0x690
> > > [ 30.786864] #6: (____ptrval____) (console_owner){....}, at:
> > > console_unlock+0x238/0x690
> > > [ 30.787077] irq event stamp: 577424
> > > [ 30.787077] irq event stamp: 577424
> > > [ 30.787136] hardirqs last enabled at (577423):
> > > [<ffff20000818988c>]
> > > __down_trylock_console_sem+0x6c/0xb0
> > > [ 30.787136] hardirqs last enabled at (577423):
> > > [<ffff20000818988c>]
> > > __down_trylock_console_sem+0x6c/0xb0
> > > [ 30.787197] hardirqs last disabled at (577424):
> > > [<ffff20000818a050>]
> > > console_unlock+0x148/0x690
> > > [ 30.787197] hardirqs last disabled at (577424):
> > > [<ffff20000818a050>]
> > > console_unlock+0x148/0x690
> > > [ 30.787254] softirqs last enabled at (577412):
> > > [<ffff20000808215c>]
> > > __do_softirq+0x2d4/0x3b4
> > > [ 30.787254] softirqs last enabled at (577412):
> > > [<ffff20000808215c>]
> > > __do_softirq+0x2d4/0x3b4
> > > [ 30.787309] softirqs last disabled at (577403):
> > > [<ffff2000080fe3f8>]
> > > irq_exit+0x150/0x160
> > > [ 30.787309] softirqs last disabled at (577403):
> > > [<ffff2000080fe3f8>]
> > > irq_exit+0x150/0x160
> > > [ 30.787367] Preemption disabled at:[<ffff20000818bcb8>]
> > > vprintk_emit+0x280/0x4a0
> > > [ 30.787367] Preemption disabled at:[<ffff20000818bcb8>]
> > > vprintk_emit+0x280/0x4a0
> > > [ 30.787457] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
> > > 4.17.0-rc5-next-20180517-00019-gae3b4aa-dirty #358
> > > [ 30.787457] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
> > > 4.17.0-rc5-next-20180517-00019-gae3b4aa-dirty #358
> > > [ 30.787502] Hardware name: Excavator-RK3399 Board (DT)
> > > [ 30.787502] Hardware name: Excavator-RK3399 Board (DT)
> > > [ 30.787538] Call trace:
> > > [ 30.787538] Call trace:
> > > [ 30.787579] dump_backtrace+0x0/0x238
> > > [ 30.787579] dump_backtrace+0x0/0x238
> > > [ 30.787618] show_stack+0x14/0x20
> > > [ 30.787618] show_stack+0x14/0x20
> > > [ 30.787656] dump_stack+0xb0/0xec
> > > [ 30.787656] dump_stack+0xb0/0xec
> > > [ 30.787695] ___might_sleep+0x254/0x2a8
> > > [ 30.787695] ___might_sleep+0x254/0x2a8
> > > [ 30.787734] __might_sleep+0x74/0xe0
> > > [ 30.787734] __might_sleep+0x74/0xe0
> > > [ 30.787777] __pm_runtime_resume+0xb0/0xb8
> > > [ 30.787777] __pm_runtime_resume+0xb0/0xb8
> > > [ 30.787820] serial8250_rpm_get.part.2+0x24/0x30
> > > [ 30.787820] serial8250_rpm_get.part.2+0x24/0x30
> > > [ 30.787863] serial8250_console_write+0x3a8/0x3d8
> > > [ 30.787863] serial8250_console_write+0x3a8/0x3d8
> > > [ 30.787909] univ8250_console_write+0x44/0x58
> > > [ 30.787909] univ8250_console_write+0x44/0x58
> > > [ 30.787949] console_unlock+0x56c/0x690
> > > [ 30.787949] console_unlock+0x56c/0x690
> > > [ 30.787988] vprintk_emit+0x28c/0x4a0
> > > [ 30.787988] vprintk_emit+0x28c/0x4a0
> > > [ 30.788027] vprintk_default+0x34/0x40
> > > [ 30.788027] vprintk_default+0x34/0x40
> > > [ 30.788067] vprintk_func+0x1e0/0x2b8
> > > [ 30.788067] vprintk_func+0x1e0/0x2b8
> > > [ 30.788105] printk+0xac/0xd8
> > > [ 30.788105] printk+0xac/0xd8
> > > [ 30.788146] register_console+0x2d0/0x508
> > > [ 30.788146] register_console+0x2d0/0x508
> > > [ 30.788187] uart_add_one_port+0x834/0x848
> > > [ 30.788187] uart_add_one_port+0x834/0x848
> > > [ 30.788233] serial8250_register_8250_port+0x514/0x708
> > > [ 30.788233] serial8250_register_8250_port+0x514/0x708
> > > [ 30.788271] dw8250_probe+0x6bc/0x8f0
> > > [ 30.788271] dw8250_probe+0x6bc/0x8f0
> > > [ 30.788312] platform_drv_probe+0x6c/0xd0
> > > [ 30.788312] platform_drv_probe+0x6c/0xd0
> > > [ 30.788353] driver_probe_device+0x2e0/0x3f8
> > > [ 30.788353] driver_probe_device+0x2e0/0x3f8
> > > [ 30.788393] __driver_attach+0xec/0xf0
> > > [ 30.788393] __driver_attach+0xec/0xf0
> > > [ 30.788433] bus_for_each_dev+0xdc/0x150
> > > [ 30.788433] bus_for_each_dev+0xdc/0x150
> > > [ 30.788472] driver_attach+0x30/0x40
> > > [ 30.788472] driver_attach+0x30/0x40
> > > [ 30.788515] bus_add_driver+0x1c0/0x318
> > > [ 30.788515] bus_add_driver+0x1c0/0x318
> > > [ 30.788556] driver_register+0xb4/0x198
> > > [ 30.788556] driver_register+0xb4/0x198
> > > [ 30.788602] __platform_driver_register+0x7c/0x88
> > > [ 30.788602] __platform_driver_register+0x7c/0x88
> > > [ 30.788648] dw8250_platform_driver_init+0x18/0x20
> > > [ 30.788648] dw8250_platform_driver_init+0x18/0x20
> > > [ 30.788688] do_one_initcall+0xac/0x210
> > > [ 30.788688] do_one_initcall+0xac/0x210
> > > [ 30.788729] kernel_init_freeable+0x294/0x33c
> > > [ 30.788729] kernel_init_freeable+0x294/0x33c
> > > [ 30.788768] kernel_init+0x10/0x118
> > > [ 30.788768] kernel_init+0x10/0x118
> > > [ 30.788808] ret_from_fork+0x10/0x18
> > > [ 30.788808] ret_from_fork+0x10/0x18
> > > [ 30.857151] bootconsole [uart8250] disabled
> > > [ 30.857151] bootconsole [uart8250] disabled
> >
> >
> > Ick. That's crazy. And reverting this patch makes things work
> > again?
> > If so, I'll go revert this, thanks for letting us know.
> >
> > Phil, any ideas?
>
> Not off the top of my head. When I get some time I'll have another
> look
> at the dw8250 driver.
>
Tony, Sebastian, JFYI.
This is exact example why we have to disable runtime PM for kernel
console.
Phil, I suppose the series [1] fixes an issue.
[1]: https://www.spinics.net/lists/linux-serial/msg30335.html
--
Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Intel Finland Oy
--
To unsubscribe from this list: send the line "unsubscribe linux-serial" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
