snowball/igloo-kernel.git/drivers/tty, branch integration-linux-ux500

Merge Linaro Android branch ' jstultz/linaro-android-3.4

2012-06-04T11:46:24+00:00

' into integration-android-ux500 Signed-off-by: Philippe Langlais

tty: serial: amba-pl011: fix cancel_delayed_work() usage

2012-05-22T09:01:37+00:00

PL011 calls del_timer_sync() (via cancel_delayed_work()) from atomic context. Fix it. ====================================================== [ INFO: HARDIRQ-safe -> HARDIRQ-unsafe lock order detected ] 3.0.8+ #1188 ------------------------------------------------------ kworker/u:0/5 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire: (&(&uap->clk_off_work)->timer){+.-...}, at: [] del_timer_sync+0x0/0xc8 and this task is already holding: (&port_lock_key){-.-...}, at: [] pl011_clock_on+0x80/0xc4 which would create a new lock dependency: (&port_lock_key){-.-...} -> (&(&uap->clk_off_work)->timer){+.-...} but this new dependency connects a HARDIRQ-irq-safe lock: (&port_lock_key){-.-...} ... which became HARDIRQ-irq-safe at: [] mark_irqflags+0x68/0x170 [] __lock_acquire+0x598/0x788 [] lock_acquire+0x10c/0x130 [] _raw_spin_lock_irqsave+0x5c/0x98 [] pl011_int+0x10/0x128 [] handle_irq_event_percpu+0x98/0x1f4 [] handle_irq_event+0x3c/0x5c [] handle_fasteoi_irq+0xc8/0x100 [] generic_handle_irq+0x28/0x30 [] asm_do_IRQ+0x78/0xb8 [] __irq_svc+0x50/0xf4 [] enter_sleep+0x550/0x608 [] cpuidle_idle_call+0x1c4/0x33c [] cpu_idle+0xa4/0x10c [] start_kernel+0x228/0x27c [<00008040>] 0x8040 to a HARDIRQ-irq-unsafe lock: (&(&uap->clk_off_work)->timer){+.-...} ... which became HARDIRQ-irq-unsafe at: ... [] mark_irqflags+0xec/0x170 [] __lock_acquire+0x598/0x788 [] lock_acquire+0x10c/0x130 [] call_timer_fn+0x68/0x1e4 [] run_timer_softirq+0x1c8/0x224 [] __do_softirq+0x110/0x23c [] irq_exit+0x5c/0xc0 [] do_local_timer+0x50/0x80 [] __irq_svc+0x50/0xf4 [] __mutex_unlock_slowpath+0x150/0x17c [] sysfs_addrm_finish+0x10/0x74 [] sysfs_add_file_mode+0x70/0xb0 [] create_files+0x7c/0xb4 [] internal_create_group+0xb8/0x118 [] device_add_groups+0x1c/0x70 [] device_add_attrs+0x50/0xa8 [] device_add+0x110/0x2b4 [] hci_register_sysfs+0x54/0x114 [] hci_register_dev+0x378/0x4d8 [] register_bluetooth+0xb4/0x124 [] probe_common+0x7c/0xcc [] btcg2900_evt_probe+0x38/0x6c [] platform_drv_probe+0x18/0x1c [] really_probe+0x98/0x148 [] driver_probe_device+0x48/0x60 [] bus_for_each_drv+0x44/0x74 [] device_attach+0x78/0xa4 [] bus_probe_device+0x24/0x40 [] device_add+0x174/0x2b4 [] platform_device_add+0x104/0x160 [] mfd_add_device+0x194/0x1d4 [] mfd_add_devices+0x70/0xbc [] work_power_off_chip+0x11c/0x1c8 [] process_one_work+0x2ac/0x4c8 [] worker_thread+0x144/0x234 [] kthread+0x80/0x88 [] kernel_thread_exit+0x0/0x8 other info that might help us debug this: Possible interrupt unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&(&uap->clk_off_work)->timer); local_irq_disable(); lock(&port_lock_key); lock(&(&uap->clk_off_work)->timer); lock(&port_lock_key); *** DEADLOCK *** Change-Id: I55d565d484baaee2dc76f258a5066aacbb6e6dee Signed-off-by: Rabin Vincent ST-Ericsson ID: 413918 ST-Ericsson Linux next: NA ST-Ericsson FOSS-OUT ID: Trivial Change-Id: I1a45effb0afeb4344f03ca500a5b0767bd760ec3 Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/47174 Reviewed-by: QABUILD Reviewed-by: Shreshtha Kumar SAHU Reviewed-by: Rabin VINCENT Tested-by: Rabin VINCENT

tty: serial: amba-pl011: use pm_runtime_irq_safe()

2012-05-22T09:01:35+00:00

PL011 should use pm_runtime_irq_safe() since it calls the pm_runtime routines with interrupts disabled. Found via lockdep: ================================= [ INFO: inconsistent lock state ] 3.0.8+ #1187 --------------------------------- inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. kworker/u:0/5 [HC0[0]:SC0[0]:HE1:SE1] takes: (&port_lock_key){?.....}, at: [] pl011_clock_on+0xc/0xc4 {IN-HARDIRQ-W} state was registered at: [] mark_irqflags+0x68/0x170 [] __lock_acquire+0x598/0x788 [] lock_acquire+0x10c/0x130 [] _raw_spin_lock_irqsave+0x5c/0x98 [] pl011_int+0x10/0x128 [] handle_irq_event_percpu+0x98/0x1f4 [] handle_irq_event+0x3c/0x5c [] handle_fasteoi_irq+0xc8/0x100 [] generic_handle_irq+0x28/0x30 [] asm_do_IRQ+0x78/0xb8 [] __irq_svc+0x50/0xf4 [] _raw_spin_unlock_irq+0x28/0x54 [] uart_carrier_raised+0x3c/0x44 [] tty_port_carrier_raised+0x1c/0x20 [] tty_port_block_til_ready+0x220/0x2cc [] tty_open+0x1f0/0x32c [] chrdev_open+0x20c/0x238 [] __dentry_open+0x220/0x338 [] nameidata_to_filp+0x50/0x5c [] do_last+0x43c/0x4c4 [] path_openat+0xb8/0x37c [] do_filp_open+0x30/0x7c [] do_sys_open+0xd8/0x170 [] ret_fast_syscall+0x0/0x3c irq event stamp: 60637 hardirqs last enabled at (60636): [] mutex_lock_nested+0x364/0x3c8 hardirqs last disabled at (60637): [] _raw_spin_lock_irqsave+0x1c/0x98 softirqs last enabled at (60509): [] irq_exit+0x5c/0xc0 softirqs last disabled at (60490): [] irq_exit+0x5c/0xc0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&port_lock_key); lock(&port_lock_key); *** DEADLOCK *** 7 locks held by kworker/u:0/5: #0: ((hdev->name)){.+.+.+}, at: [] process_one_work+0x1cc/0x4c8 #1: ((&hdev->power_on)){+.+.+.}, at: [] process_one_work+0x1cc/0x4c8 #2: (&hdev->req_lock){+.+.+.}, at: [] hci_dev_open+0x28/0x334 #3: (&main_info->man_mutex#2){+.+.+.}, at: [] cg2900_open+0xa8/0x49c #4: (&(uart_info->sleep_state_lock)){+.+.+.}, at: [] uart_set_chip_power+0x64/0x228 #5: (&tty->termios_mutex){+.+...}, at: [] cg2900_hci_uart_set_baudrate+0x34/0x8c #6: (&port_lock_key){?.....}, at: [] pl011_clock_on+0xc/0xc4 stack backtrace: [] (unwind_backtrace+0x0/0xec) from [] (print_usage_bug+0x184/0x1c4) [] (print_usage_bug+0x184/0x1c4) from [] (mark_lock_irq+0xbc/0x280) [] (mark_lock_irq+0xbc/0x280) from [] (mark_lock+0x2c8/0x3f8) [] (mark_lock+0x2c8/0x3f8) from [] (mark_held_locks+0x5c/0x80) [] (mark_held_locks+0x5c/0x80) from [] (trace_hardirqs_on_caller+0x138/0x18c) [] (trace_hardirqs_on_caller+0x138/0x18c) from [] (_raw_spin_unlock_irq+0x24/0x54) [] (_raw_spin_unlock_irq+0x24/0x54) from [] (rpm_callback+0x3c/0x5c) [] (rpm_callback+0x3c/0x5c) from [] (rpm_resume+0x2ec/0x3c0) [] (rpm_resume+0x2ec/0x3c0) from [] (__pm_runtime_resume+0x48/0x60) [] (__pm_runtime_resume+0x48/0x60) from [] (pl011_clock_on+0x34/0xc4) [] (pl011_clock_on+0x34/0xc4) from [] (pl011_set_termios+0x28/0x308) [] (pl011_set_termios+0x28/0x308) from [] (uart_change_speed+0x90/0x94) [] (uart_change_speed+0x90/0x94) from [] (uart_set_termios+0x68/0x1b4) [] (uart_set_termios+0x68/0x1b4) from [] (cg2900_hci_uart_set_baudrate+0x78/0x8c) [] (cg2900_hci_uart_set_baudrate+0x78/0x8c) from [] (uart_set_chip_power+0x100/0x228) [] (uart_set_chip_power+0x100/0x228) from [] (cg2900_open+0x228/0x49c) [] (cg2900_open+0x228/0x49c) from [] (btcg2900_open+0x84/0x188) [] (btcg2900_open+0x84/0x188) from [] (hci_dev_open+0x94/0x334) [] (hci_dev_open+0x94/0x334) from [] (hci_power_on+0x14/0x6c) [] (hci_power_on+0x14/0x6c) from [] (process_one_work+0x2ac/0x4c8) [] (process_one_work+0x2ac/0x4c8) from [] (worker_thread+0x144/0x234) [] (worker_thread+0x144/0x234) from [] (kthread+0x80/0x88) [] (kthread+0x80/0x88) from [] (kernel_thread_exit+0x0/0x8) Change-Id: I8b05184e348656a0d422b5b68267ee03246c6563 Signed-off-by: Rabin Vincent ST-Ericsson ID: 413918 ST-Ericsson Linux next: NA ST-Ericsson FOSS-OUT ID: Trivial Change-Id: Id5a17f56dfa1b1b8c469237be5b9c236ad212137 Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/47173 Reviewed-by: QABUILD Tested-by: Rabin VINCENT Reviewed-by: Jonas ABERG Reviewed-by: Rabin VINCENT

serial/amba-pl011: Return early in deadlock work-around if there is no tty

2012-05-22T09:01:17+00:00

The first amba port does not always have an attached tty, which will result in a crash if the work-around is activated. This seems to be the case for example on Snowball which has no modem hw, which seems to be the main use-case for this work-around.

ux500: pl011: remove hardcoding of uart instance

2012-05-22T09:01:12+00:00

amba-pl011 mechanism detects the hardware register deadlock situation, and triggers a workaround tasklet. Tasklet implementation should execute workaround for the instance of the uart on which deadlock is detected. This patch provides the uart line number on whcich the deadlock is detected to the tasklet. ST-Ericsson ID: 355971 ST-Ericsson Linux next: NA ST-Ericsson FOSS-OUT ID: Trivial Change-Id: Ie8fbea42dc20cae1b604b555815935bcb5a70241 Signed-off-by: Shreshtha Kumar Sahu Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/38415 Reviewed-by: Srinidhi KASAGAR

drivers: serial: amba-pl011: Introduce runtime pm

2012-05-22T09:01:11+00:00

ST-Ericsson Linux next: - ST-Ericsson ID: 373766 ST-Ericsson FOSS-OUT ID: Trivial Change-Id: Ie2a7719ca602abf3fbb3e4cef05adc4f7ab07035 Signed-off-by: Jonas Aaberg Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/37021 Reviewed-by: QABUILD Reviewed-by: Ulf HANSSON

tty: serial: amba-pl011: Release clocks and regulators after probe

2012-05-22T09:01:10+00:00

These will be enabled once in use. ST-Ericsson Linux next: - ST-Ericsson ID: 367599 ST-Ericsson FOSS-OUT ID: Trivial Change-Id: I1cfe9b00262d78a3326ca01af85268e918170735 Signed-off-by: Jonas Aaberg Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/35397 Reviewed-by: QABUILD

serial/amba-pl011: prevent console port from holding regulator

2012-05-22T09:01:08+00:00

Register the console in a console_initcall() to ensure that pl011_console_setup() gets called early and thus the uart_console() macro usage in pl011_power_startup() returns the correct value the first time it is called, so that the vAPE regulator is not taken for the console port. Note that this problem is seen only when the console port is line 0, since even without this patch pl011_console_setup() gets called (and thus the correct console->index is set) after the first port is added. ST-Ericsson ID: 368522 ST-Ericsson Linux next: NA ST-Ericsson FOSS-OUT ID: Trivial Change-Id: I7f4ab80e8d8be014bfdbe695ea45007212c9c4ea Signed-off-by: Rabin Vincent Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/34411 Reviewed-by: Srinidhi KASAGAR Reviewed-by: Shreshtha Kumar SAHU

serial: amba-pl011: delay driver initialization

2012-05-22T09:00:57+00:00

During pl011 probe() On u8500 platform, amba bus driver tries to access regulator driver and fails as the regulator driver is also at arch_init level. Moving serial driver to the next level but still as early as possible. Change-Id: I29d23067df9465b3c25f4010dca8af253fe6d070 Signed-off-by: Mian Yousaf Kaukab Reviewed-on: http://gerrit.lud.stericsson.com/gerrit/32121 Reviewed-by: Jonas ABERG Tested-by: Jonas ABERG

UART:pl011: migrate pl011 support to 3.0

2012-05-22T09:00:51+00:00

pl011 changes from the following commits have been manually merged. ef45647, d57a627, 588c0bb, 9b4ecf9, 913afa0, 13aa593, e7d5dcb, 5115d25, 19fcf44 ST-Ericsson ID: NA ST-Ericsson Linux next: NA ST-Ericsson FOSS-OUT ID: NA Change-Id: I1759c825add127867cf6edc688b860e82743ce6f Signed-off-by: Rajanikanth H.V