楼主 #1 2018-08-17 12:48:28 分享评论
- Lusien4
- 会员
- 注册时间： 2017-12-29
- 已发帖子： 7
- 积分： 7
在全志平台跑android7.1，时钟误差很大的问题
大家中午好，我把我现在的问题描述下：
一、经过几次测量，在锁屏情况下，实际时间过去27分钟，板子上时间过去33分钟；实际时间过去1小时16分，板子上时间过去1小时45分钟。在关机情况下，时间同样快了很多。在一直亮屏情况下，时钟就是比较准确的。
二、晶振部分电路如下：负载电容实际用的型号是CAP CER;NPO;12pF;+/-5%;50V;0402
用示波器测量如下图，测量频率为32.7689KHz，平板亮屏和锁屏均无很大变化。
放大后
三、关于驱动方面，我还未修改任何地方，系统默认使用的驱动是，/lichee/linux-3.10/drivers/rtc/rtc-sunxi.c
四、截取一段从锁屏到亮屏，再到锁屏的日志
CPU1: Booted secondary processor
[ 369.870645] CPU1: update cpu_power 1024
[ 369.886974] CPU Budget hotplug: cluster0 min:0 max:4
[ 369.892677] CPU Budget:Try to up cpu 2, cluster0 online 2, max 4
[ 369.901787] CPU2: Booted secondary processor
[ 369.901787] CPU2: update cpu_power 1024
[ 369.917055] CPU Budget:Try to up cpu 3, cluster0 online 2, max 4
[ 369.924409] CPU3: Booted secondary processor
[ 369.924409] CPU3: update cpu_power 1024
[ 369.942342] PM: suspend entry 2010-01-02 03:52:38.762305228 UTC
[ 369.948817] [pm]valid
[ 369.951472] Notice: sun9i & sun8iw5 & sun50i not need support normal standby, change to super standby.
[ 369.962203] PM: Syncing filesystems ... done.
[ 369.968739] PM: Preparing system for mem sleep
[ 369.974163] Freezing user space processes ... (elapsed 0.001 seconds) done.
[ 369.983610] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[ 369.993273] PM: Entering mem sleep
[ 369.996985] [pm]3 state begin
[ 370.000405] Suspending console(s) (use no_console_suspend to debug)
[ 370.008381] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 03:55:03
[ 370.008504] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 03:55:03
[ 370.008515] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 03:55:03
[ 370.008523] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 03:55:03
[ 370.008530] sunxi-rtc 1f00000.rtc: set alarm seconds:4260 enable:1
[ 370.009321] [BT_LPM] bluesleep_suspend: bluesleep_suspend
[ 370.009384] Mali: Disable gpu power successfully.
[ 370.009401] sunxi_dramfreq_suspend:1441
[ 370.009485] sunxi_keyboard_suspend apb1_keyadc mclk handle is invalid!
[ 370.009533] [VFE]vfe suspend
[ 370.009571] enter tr_suspend
[ 370.009572] exit tr_suspend
[ 370.009594] disp_suspend
[ 370.009603] hwc_suspend
[ 370.590019] LCD_power_off++++++++
[ 370.590264] [DISP] disp_sys_pin_set_state,line:422:pinctrl_lookup_state for allwinner,sunxi-lcd0 fail
[ 370.590286] [DISP] disp_sys_gpio_set_value,line:381:OSAL_GPIO_DevWRITE_ONEPIN_DATA, hdl is NULL
[ 371.350185] [HDMI]hdmi suspend
[ 371.350195] disp_suspend finish
[ 371.350232] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 0Hz bm OD pm OFF vdd 0 width 1 timing LEGACY(SDR12) dt B
[ 371.351468] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 0Hz bm OD pm OFF vdd 0 width 1 timing LEGACY(SDR12) dt B
[ 371.353955] [ohci0-controller]: is disable, can not suspend
[ 371.353974] [ehci0-controller]: is disable, can not suspend
[ 371.353985] udc is disable, need not enter to suspend
[ 371.354044] sw_uart_suspend()1463 - uart1 suspend
[ 371.354083] sw_uart_suspend()1463 - uart0 suspend
[ 371.354118] [cedar] standby suspend
[ 371.354169] MBUS_PMU 1c62000.mbus-controller: suspend okay
[ 371.354293] PM: suspend of devices complete after 1346.268 msecs
[ 371.354298] PM: suspend devices took 1.350 seconds
[ 371.354301] [pm]prepare
[ 371.355017] PM: late suspend of devices complete after 0.720 msecs
[ 371.355665] PM: noirq suspend of devices complete after 0.654 msecs
[ 371.355668] [pm]prepare_late
[ 371.357861] Disabling non-boot CPUs ...
[ 371.359184] CPU1: shutdown
[ 371.370093] psci: Retrying again to check for CPU kill
[ 371.370108] psci: CPU1 killed.
[ 371.371775] CPU2: shutdown
[ 371.390037] psci: Retrying again to check for CPU kill
[ 371.390054] psci: CPU2 killed.
[ 371.391492] CPU3: shutdown
[ 371.410037] psci: Retrying again to check for CPU kill
[ 371.410054] psci: CPU3 killed.
[ 371.410542] [pm]enter state 3
[ 371.410848] info: can not find np for sys_pwr_dm_para.
[ 371.410853] after customized: sys_mask config = 0x4a55.
[ 371.410866] scence_lock: super_standby
[ 371.410871]
[ 371.410875] dynamic config wakeup_src: 0x 800000
[ 371.410879] wakeup_gpio_map 0x 20
[ 371.410882] wakeup_gpio_group 0x 0
[ 371.410885] extended_standby id = 0x 8
[ 371.410888] extended_standby pmu_id = 0x 0
[ 371.410891] extended_standby soc_id = 0x 0
[ 371.410894] extended_standby pwr dep as follow:
[ 371.410896] pwr dm state as follow:
[ 371.410899] pwr dm state = 0x c244.
[ 371.410903] pwr dm sys mask = 0x 4a55.
[ 371.410905] pwr on = 0x4244.
[ 371.410908] pwr off = 0xfffff7ee.
[ 371.410916] count 0 io config: addr 0x1c208b4, value_mask 0x f0f0ff, value 0x 707077.
[ 371.410921] count 1 io config: addr 0x1c208b4, value_mask 0x f0f00, value 0x 70700.
[ 371.420010] [pm]platform wakeup, standby wakesource is:0x100000
[ 371.420010] Suspended for 0.000 seconds
[ 371.420010] Enabling non-boot CPUs ...
[ 371.420010] CPU1: Booted secondary processor
[ 371.420010] CPU1: update cpu_power 1024
[ 371.420010] CPU1 is up
[ 371.420010] CPU2: Booted secondary processor
[ 371.420010] CPU2: update cpu_power 1024
[ 371.420010] CPU2 is up
[ 371.420010] CPU3: Booted secondary processor
[ 371.420010] CPU3: update cpu_power 1024
[ 371.420010] CPU3 is up
[ 371.420010] [pm]aw_pm_wake
[ 371.420010] PM: noirq resume of devices complete after 0.658 msecs
[ 371.420010] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 05:06:04
[ 371.420010] PM: early resume of devices complete after 0.676 msecs
[ 371.420010] [pm]platform wakeup finish
[ 371.420010] [ehci0-controller]: is disable, can not resume
[ 371.420010] [ohci0-controller]: is disable, can not resume
[ 371.420010] MBUS_PMU 1c62000.mbus-controller: resume okay
[ 371.420010] [cedar] standby resume
[ 371.420010] sw_uart_resume()1505 - uart0 resume. DLH: 0, DLL: 13.
[ 371.420010] sw_uart_resume()1505 - uart1 resume. DLH: 0, DLL: 0.
[ 371.420010] udc is disable, need not enter to resume
[ 371.420992] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 0Hz bm PP pm UP vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.440045] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.460039] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.460088] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm OD pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.460128] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm OD pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.462533] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm OD pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.481295] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 22 width 1 timing LEGACY(SDR12) dt B
[ 371.482606] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 22 width 1 timing MMC-HS(SDR20) dt B
[ 371.482645] mmc0: Check switch ready sunxi_mmc_check_timing_switch_done
[ 371.482972] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 50000000Hz bm PP pm ON vdd 22 width 1 timing MMC-HS(SDR20) dt B
[ 371.483137] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 50000000Hz bm PP pm ON vdd 22 width 8 timing MMC-HS(SDR20) dt B
[ 371.483280] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 50000000Hz bm PP pm ON vdd 22 width 8 timing UHS-DDR50 dt B
[ 371.483332] sunxi-mmc 1c11000.sdmmc: sdc set ios: clk 50000000Hz bm PP pm ON vdd 22 width 8 timing UHS-DDR50 dt B
[ 371.483453] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 0Hz bm PP pm UP vdd 21 width 1 timing LEGACY(SDR12) dt B
[ 371.483528] sunxi-mmc 1c10000.sdmmc: no vqmmc,Check if there is regulator
[ 371.500037] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 21 width 1 timing LEGACY(SDR12) dt B
[ 371.520027] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 21 width 1 timing LEGACY(SDR12) dt B
[ 371.520876] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 52, RTO !!
[ 371.521697] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 52, RTO !!
[ 371.521712] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 21 width 1 timing LEGACY(SDR12) dt B
[ 371.524104] sunxi-mmc 1c10000.sdmmc: sdc set ios: clk 400000Hz bm PP pm ON vdd 21 width 1 timing LEGACY(SDR12) dt B
[ 371.525967] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 5, RTO !!
[ 371.526787] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 5, RTO !!
[ 371.527607] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 5, RTO !!
[ 371.528427] sunxi-mmc 1c10000.sdmmc: smc 2 p1 err, cmd 5, RTO !!
[ 371.528439] mmc2: error -110 during resume (card was removed?)
[ 371.630149] [HDMI]hdmi resume
[ 371.640490] [DISP] lcd_clk_config,line:813:disp 0, clk: pll(474000000),clk(474000000),dclk(79000000) dsi_rate(79000000)
[ 371.640490] clk real:pll(475200000),clk(475200000),dclk(118800000) dsi_rate(148500000)
[ 371.641010] +++++++++LCD_open_flow
[ 371.641605] disp_resume finish
[ 371.641620] tr_resume
[ 371.641622] exit tr_resume
[ 371.641653] [VFE]vfe resume
[ 371.641713] sunxi_keyboard_resume apb1_keyadc mclk handle is invalid!
[ 371.641782] sunxi_dramfreq_resume:1477
[ 371.641792] Mali: Enable gpu power successfully.
[ 371.641854] [BT_LPM] bluesleep_resume: bluesleep_resume
[ 371.661024] sunxi-rtc 1f00000.rtc: Read hardware RTC time 2010-01-02 05:06:04
[ 371.662056] PM: resume of devices complete after 243.121 msecs
[ 371.662787] PM: resume devices took 0.250 seconds
[ 371.687105] [DISP] disp_sys_gpio_set_value,line:381:OSAL_GPIO_DevWRITE_ONEPIN_DATA, hdl is NULL
[ 371.713540] [DISP] disp_sys_pin_set_state,line:422:pinctrl_lookup_state for allwinner,sunxi-lcd0 fail
[ 371.731453] LCD_panel_init
[ 371.731454] inet_dsi_panel init
[ 372.496461] [pm]aw_pm_end!
[ 372.496463] [pm]aw_pm_end wake_lock timeout: 2000 ms
[ 372.496469] PM: Finishing wakeup.
[ 372.498601] healthd: battery l=89 v=4081 t=30.0 h=2 st=3 c=-26 chg=
[ 372.499160] healthd: battery l=89 v=4081 t=30.0 h=2 st=3 c=-26 chg=
[ 372.499629] healthd: battery l=89 v=4081 t=30.0 h=2 st=3 c=-26 chg=
[ 372.501199] Restarting tasks ... done.
[ 372.539042] disp_runtime_idle
[ 372.540305] PM: suspend exit 2010-01-02 05:03:40.707760146 UTC
[ 373.841791] CPU3: shutdown
[ 373.844810] psci: CPU3 killed.
[ 373.871755] CPU2: shutdown
[ 373.874766] psci: CPU2 killed.
[ 374.071539] CPU1: shutdown
[ 374.074553] psci: CPU1 killed.
[ 374.490148] PM: suspend entry 2010-01-02 05:03:42.657593313 UTC
[ 374.496682] [pm]valid
[ 374.499187] Notice: sun9i & sun8iw5 & sun50i not need support normal standby, change to super standby.
[ 374.513059] PM: Syncing filesystems ... done.
[ 374.519439] PM: Preparing system for mem sleep
[ 374.525202] Freezing user space processes ... (elapsed 0.002 seconds) done.
[ 374.534988] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[ 374.545039] PM: Entering mem sleep
[ 374.548769] [pm]3 state begin
[ 374.552214] Suspending console(s) (use no_console_suspend to debug)五、还有一个其他问题，板子上跑了一个app，通过socketio与服务器保持连接，若板子长时间休眠后，就会导致socket断开，因此在app中加入了通过广播唤醒平板的机制，心跳为40秒，由于rtc时钟不准，实际唤醒时间为30秒左右，但在app跑了一天左右会导致平板死机，死机日志如下
[10745.140141] PM: suspend entry 2017-12-01 05:57:31.785462936 UTC
[10745.146950] [pm]valid
[10745.149460] Notice: sun9i & sun8iw5 & sun50i not need support normal standby, change to super standby.
[10745.160489] PM: Syncing filesystems ... done.
[10745.167265] PM: Preparing system for mem sleep
[10745.172633] Freezing user space processes ... (elapsed 0.001 seconds) done.
[10745.182188] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[10745.192159] PM: Entering mem sleep
[10745.195870] [pm]3 state begin
[10745.199105] Suspending console(s) (use no_console_suspend to debug)
[SCP ERROR] :allocate message [0] frame is invalid
[SCP WARING] :allocate message failed
[SCP ERROR] :allocate message [0] frame is invalid
[SCP WARING] :allocate message failed
[SCP ERROR] :allocate message [0] frame is invalid
[SCP WARING] :allocate message failed
[SCP ERROR] :allocate message [0] frame is invalid
[SCP WARING] :allocate message failed
[SCP ERROR] :allocate message [0] frame is invalid
[SCP WARING] :allocate message failed最后两句话一直循环打印。另外通过增加app中心跳时长，或者降低板子sys_config.fex中dram_clk频率，均可使死机情况有所缓解，就是说板子能运行的时间更长了，但最终还是会出现上面的死机现象
大神们，这个rtc和死机问题困扰我很久了，我分别该从哪个地方去排查或修改，不胜感激
离线
#2 2018-08-17 13:40:49 分享评论
- 超级萌新
- 会员
- 注册时间： 2018-05-04
- 已发帖子： 408
- 积分： 407
Re: 在全志平台跑android7.1，时钟误差很大的问题
RTC的时间你是如何读的？用什么命令？
离线
楼主 #4 2018-08-17 15:02:02 分享评论
- Lusien4
- 会员
- 注册时间： 2017-12-29
- 已发帖子： 7
- 积分： 7
Re: 在全志平台跑android7.1，时钟误差很大的问题
晕哥 说：
1. 你试下不关屏，只用pwm调背光会不会影响.
2. 或者把屏幕拆了，看下有什么影响.
晕哥好：1.我目前这块平板没有调背光的功能，所以不能测试这一条
2.我最开始的数据就是把屏幕拆了之后测的，刚刚又测试了一遍，实际过了13分钟，板子上过去了18分钟，板子上的时间是通过date命令来看的
离线
#5 2018-08-17 15:14:56 分享评论
- 晕哥
- 管理员
- 注册时间： 2017-09-06
- 已发帖子： 9,342
- 积分： 9202
Re: 在全志平台跑android7.1，时钟误差很大的问题
date 命令看到的不是 RTC 时间, 而是系统的时间(经过RTC初始化之后的CPU时间).
date 命令不会去读 RTC 外设的.
如果我没记错的话，是用 hwclock 读RTC时间
hwclock --show
写RTC时间:
date
hwclock --set
上面的时间不准，
很有可能是因为关屏或者其他低功耗行为，
系统降频了，
导致时间出错了,
应该是软件有哪些地方没有重新换算。
离线
楼主 #6 2018-08-17 16:04:04 分享评论
- Lusien4
- 会员
- 注册时间： 2017-12-29
- 已发帖子： 7
- 积分： 7
Re: 在全志平台跑android7.1，时钟误差很大的问题
晕哥 说：
date 命令看到的不是 RTC 时间, 而是系统的时间(经过RTC初始化之后的CPU时间).
date 命令不会去读 RTC 外设的.
如果我没记错的话，是用 hwclock 读RTC时间
hwclock --show
写RTC时间:
date
hwclock --set上面的时间不准，
很有可能是因为关屏或者其他低功耗行为，
系统降频了，
导致时间出错了,
应该是软件有哪些地方没有重新换算。
好的，我刚用这个命令测了下，也是在锁屏情况下，实际走了16分钟，板子走了22分钟
我这个板子上没有外接RTC芯片。从示波器上看，是不是可以说明不是硬件问题了，我该从哪些方面去查软件方面的问题呢
离线
#7 2018-08-17 16:11:19 分享评论
- 晕哥
- 管理员
- 注册时间： 2017-09-06
- 已发帖子： 9,342
- 积分： 9202
Re: 在全志平台跑android7.1，时钟误差很大的问题
这样看来 hwclock 也不对？
hwclock是一个相对独立的IP,
与系统并没有什么关系.
这样试一试，
你找到驱动里面读 RTC 寄存器的代码，
在那里面加入调试语句,
把寄存器的内容输出来。
或者更简单一点,
找到 RTC 的寄存器地址，
直接用 devmem 读出来，
对照datasheet,
换算成时间.
devmem 参考链接:
https://whycan.cn/t_1430.html
https://whycan.cn/t_463.html
离线
楼主 #8 2018-08-17 18:26:04 分享评论
- Lusien4
- 会员
- 注册时间： 2017-12-29
- 已发帖子： 7
- 积分： 7
Re: 在全志平台跑android7.1，时钟误差很大的问题
好的，谢谢晕哥，rtc驱动这一块，我再去熟悉一下。
请问晕哥你有遇到过我上面说的第五点的问题吗，有没有一些查找原因的建议，或者在socket上有没有其他保持长连接的方法等。
离线
感谢为中文互联网持续输出优质内容的各位老铁们。
QQ: 516333132, 微信(wechat): whycan_cn (哇酷网/挖坑网/填坑网) service@whycan.cn