From: Ville Syrjälä ville.syrjala@linux.intel.com
[ Upstream commit 66e4f4a9cc389b277e187c115a285fad2cba5485 ]
cmos_interrupt() isn't always called from hardirq context, so we must use spin_lock_irqsave() & co.
================================ WARNING: inconsistent lock state 5.6.0-rc2-CI-CI_DRM_7981+ #1 Tainted: G U -------------------------------- inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. rtcwake/4315 [HC0[0]:SC0[0]:HE1:SE1] takes: ffffffff82635198 (rtc_lock){?...}, at: cmos_interrupt+0x18/0x100 {IN-HARDIRQ-W} state was registered at: lock_acquire+0xa7/0x1c0 _raw_spin_lock+0x2a/0x40 cmos_interrupt+0x18/0x100 rtc_handler+0x75/0xc0 acpi_ev_fixed_event_detect+0xf9/0x132 acpi_ev_sci_xrupt_handler+0xb/0x28 acpi_irq+0x13/0x30 __handle_irq_event_percpu+0x41/0x2c0 handle_irq_event_percpu+0x2b/0x70 handle_irq_event+0x2f/0x50 handle_fasteoi_irq+0x8e/0x150 do_IRQ+0x7e/0x160 ret_from_intr+0x0/0x35 mwait_idle+0x7e/0x200 do_idle+0x1bb/0x260 cpu_startup_entry+0x14/0x20 start_secondary+0x15f/0x1b0 secondary_startup_64+0xa4/0xb0 irq event stamp: 42003 hardirqs last enabled at (42003): [<ffffffff81a36567>] _raw_spin_unlock_irqrestore+0x47/0x60 hardirqs last disabled at (42002): [<ffffffff81a362ed>] _raw_spin_lock_irqsave+0xd/0x50 softirqs last enabled at (41848): [<ffffffff81e00385>] __do_softirq+0x385/0x47f softirqs last disabled at (41841): [<ffffffff810bab3a>] irq_exit+0xba/0xc0
other info that might help us debug this: Possible unsafe locking scenario:
CPU0 ---- lock(rtc_lock); <Interrupt> lock(rtc_lock);
*** DEADLOCK ***
6 locks held by rtcwake/4315: #0: ffff888175dc9408 (sb_writers#5){.+.+}, at: vfs_write+0x1a4/0x1d0 #1: ffff88817406ca80 (&of->mutex){+.+.}, at: kernfs_fop_write+0xdd/0x1b0 #2: ffff888179be85e0 (kn->count#236){.+.+}, at: kernfs_fop_write+0xe6/0x1b0 #3: ffffffff82641e00 (system_transition_mutex){+.+.}, at: pm_suspend+0xb3/0x3b0 #4: ffffffff826b3ee0 (acpi_scan_lock){+.+.}, at: acpi_suspend_begin+0x47/0x80 #5: ffff888178fc3960 (&dev->mutex){....}, at: device_resume+0x92/0x1c0
stack backtrace: CPU: 3 PID: 4315 Comm: rtcwake Tainted: G U 5.6.0-rc2-CI-CI_DRM_7981+ #1 Hardware name: Google Soraka/Soraka, BIOS MrChromebox-4.10 08/25/2019 Call Trace: dump_stack+0x71/0x9b mark_lock+0x49a/0x500 ? print_shortest_lock_dependencies+0x200/0x200 __lock_acquire+0x6d4/0x15d0 ? __lock_acquire+0x460/0x15d0 lock_acquire+0xa7/0x1c0 ? cmos_interrupt+0x18/0x100 _raw_spin_lock+0x2a/0x40 ? cmos_interrupt+0x18/0x100 cmos_interrupt+0x18/0x100 cmos_resume+0x1fd/0x290 ? __acpi_pm_set_device_wakeup+0x24/0x100 pnp_bus_resume+0x5e/0x90 ? pnp_bus_suspend+0x10/0x10 dpm_run_callback+0x64/0x280 device_resume+0xd4/0x1c0 ? dpm_watchdog_set+0x60/0x60 dpm_resume+0x106/0x410 ? dpm_resume_early+0x38c/0x3e0 dpm_resume_end+0x8/0x10 suspend_devices_and_enter+0x16f/0xbe0 ? rcu_read_lock_sched_held+0x4d/0x80 pm_suspend+0x344/0x3b0 state_store+0x78/0xe0 kernfs_fop_write+0x112/0x1b0 vfs_write+0xb9/0x1d0 ksys_write+0x9f/0xe0 do_syscall_64+0x4f/0x220 entry_SYSCALL_64_after_hwframe+0x49/0xbe RIP: 0033:0x7ff934307154 Code: 89 02 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 8d 05 b1 07 2e 00 8b 00 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 41 54 55 49 89 d4 53 48 89 f5 RSP: 002b:00007ffe2647c168 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007ff934307154 RDX: 0000000000000004 RSI: 000055de3ec4e5a0 RDI: 000000000000000a RBP: 000055de3ec4e5a0 R08: 000055de3ec4c5e0 R09: 00007ff9349f3740 R10: 000055de3ec4a010 R11: 0000000000000246 R12: 000055de3ec4c500 R13: 0000000000000004 R14: 00007ff9345df2a0 R15: 00007ff9345de760
Fixes: c6d3a278cc12 ("rtc: cmos: acknowledge ACPI driven wake alarms upon resume") Fixes: 311ee9c151ad ("rtc: cmos: allow using ACPI for RTC alarm instead of HPET") Signed-off-by: Ville Syrjälä ville.syrjala@linux.intel.com Link: https://lore.kernel.org/r/20200221144739.11746-1-ville.syrjala@linux.intel.c... Signed-off-by: Alexandre Belloni alexandre.belloni@bootlin.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/rtc/rtc-cmos.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c index b795fe4cbd2e9..fb13993fad31c 100644 --- a/drivers/rtc/rtc-cmos.c +++ b/drivers/rtc/rtc-cmos.c @@ -649,10 +649,11 @@ static struct cmos_rtc cmos_rtc;
static irqreturn_t cmos_interrupt(int irq, void *p) { + unsigned long flags; u8 irqstat; u8 rtc_control;
- spin_lock(&rtc_lock); + spin_lock_irqsave(&rtc_lock, flags);
/* When the HPET interrupt handler calls us, the interrupt * status is passed as arg1 instead of the irq number. But @@ -686,7 +687,7 @@ static irqreturn_t cmos_interrupt(int irq, void *p) hpet_mask_rtc_irq_bit(RTC_AIE); CMOS_READ(RTC_INTR_FLAGS); } - spin_unlock(&rtc_lock); + spin_unlock_irqrestore(&rtc_lock, flags);
if (is_intr(irqstat)) { rtc_update_irq(p, 1, irqstat);