Allow building building the driver with COMPILE_TEST.

Link: https://lore.kernel.org/r/20200306131629.18837-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306074404.58909-4-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The datasheet states that 4052 is the maximum value for year. However, the
mpc5121 read_time and set_time function abuse the target time register
instead of using the broken down time so it is limited to 2106.

Link: https://lore.kernel.org/r/20200306074404.58909-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This simplifies the path for the rtc_ops selection.

Link: https://lore.kernel.org/r/20200306074404.58909-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Use devm managed function to simplify probe and remove.

Link: https://lore.kernel.org/r/20200306074404.58909-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

rtc_time64_to_tm never generates an invalid rtc_tm, stop validating it.

Link: https://lore.kernel.org/r/20200306073758.58050-4-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306073758.58050-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The pm8xxx are 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306073758.58050-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This allows further improvement of the driver. Also remove the unnecessary
error string as the core will already display error messages.

Link: https://lore.kernel.org/r/20200306073758.58050-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit version of rtc_tm to time conversion.

Link: https://lore.kernel.org/r/20200306073459.57197-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The ds1305 is a BCD rtc valid from 2000 to 2099.

Link: https://lore.kernel.org/r/20200306073459.57197-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306073404.56921-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The ds1374 is a 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306073404.56921-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The RTC IRQ is requested before the struct rtc_device is allocated,
this may lead to a NULL pointer dereference in the IRQ handler.

To fix this issue, allocating the rtc_device struct before requesting
the RTC IRQ using devm_rtc_allocate_device, and use rtc_register_device
to register the RTC device.

Link: https://lore.kernel.org/r/20200306073404.56921-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306015703.42101-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The CPCAP rtc is a 14bit day counter plus a 17bit seconds counter.

Note that this failed on Nov 10 2014 so it is very likely this driver as
never been used since.

Link: https://lore.kernel.org/r/20200306015703.42101-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This allows further improvement of the driver.

Link: https://lore.kernel.org/r/20200306015703.42101-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306010146.39762-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The SA1100 RTC is a 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306010146.39762-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Both RTC IRQs are requested before the struct rtc_device is allocated,
this may lead to a NULL pointer dereference in the IRQ handler.

To fix this issue, allocating the rtc_device struct before requesting
the IRQs using devm_rtc_allocate_device, and use rtc_register_device
to register the RTC device.

Link: https://lore.kernel.org/r/20200306010146.39762-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306005958.39203-4-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The Alchemy counter0 is a 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306005958.39203-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Simplify the driver by removing the goto label as it only does return ret.

Link: https://lore.kernel.org/r/20200306005958.39203-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This allows further improvement of the driver.

Link: https://lore.kernel.org/r/20200306005958.39203-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit version of rtc_tm time conversion.

Link: https://lore.kernel.org/r/20200306005910.38939-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The starfire RTC is a 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306005910.38939-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion to avoid the y2106 issue.

Link: https://lore.kernel.org/r/20200306005730.38268-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The core will never pass an invalid alarm to .set_alarm, it is not
necessary to check for its validity.

Link: https://lore.kernel.org/r/20200306005730.38268-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This RTC is a 32bit seconds counter.

Link: https://lore.kernel.org/r/20200306005730.38268-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

convertfromdays and convert2days never return errors, stop handling non
existent errors.

Link: https://lore.kernel.org/r/20200305215022.32533-4-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Let the core handle offsetting and windowing the RTC range.

This rtc has hours, minutes, seconds and 16bit days. As the driver has the
RTC epoch set to year 2000, this means that the end of the range is 2^16
days minus one second later. This is Sun Jun  6 23:59:59 UTC 2179. This
is better than the currently set year 2099.

Link: https://lore.kernel.org/r/20200305215022.32533-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

The code handling invalid alarms meaning the alarm is in the next 24 hours
is not necessary since commit f8245c26 ("rtc: remove "RTC_ALM_SET mode"
bugs") which actually predates this driver. Since then, .set_alarm is never
called with an invalid alarm and this handling is not necessary in the
driver.

Link: https://lore.kernel.org/r/20200305215022.32533-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This allows further improvement of the driver. Also remove the unnecessary
error string as the core will already display error messages.

Link: https://lore.kernel.org/r/20200305215022.32533-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Call the 64bit versions of rtc_tm time conversion to avoid the y2106 issue.

Link: https://lore.kernel.org/r/20200305160452.27808-3-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This RTC is a 32bit counter running at 16Hz. This overflows every eight
years and a half. However, the driver uses the SW_VALUE register to store
the overflow, extending the counter to 64bit as long as the update happens
before the overflow.

Link: https://lore.kernel.org/r/20200305160452.27808-2-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

This allows further improvement of the driver. Also remove the unnecessary
error string as the core will already display error messages.

Link: https://lore.kernel.org/r/20200305160452.27808-1-alexandre.belloni@bootlin.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

i.MX SNVS RTC should depend on ARCH_MXC or COMPILE_TEST.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Link: https://lore.kernel.org/r/1583136785-4973-1-git-send-email-Anson.Huang@nxp.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

use kobj_to_dev instead of open-conding it
Signed-off-by: suguosong <suguosong@xiaomi.com>
Link: https://lore.kernel.org/r/20200225021923.8570-1-guosongsu@gmail.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

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: c6d3a278 ("rtc: cmos: acknowledge ACPI driven wake alarms upon resume")
Fixes: 311ee9c1 ("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.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

There is nothing in use from of_device.h, remove it.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Link: https://lore.kernel.org/r/1581823666-16944-1-git-send-email-Anson.Huang@nxp.comSigned-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>