Skip to content
Snippets Groups Projects
New look: Off

[samsung-kylessopen] Phone reboots every 10 minutes

    • Closed Imported Issue created by Administrator @root

    The phone reboots consistently after 10 minutes of running. As per the wiki, I looked at the kernel config and did find a certain watchdog that was enabled. There are the steps I did.

    zap
init
kconfig edit samsung-kylessopen
	CONFIG_MSM_RPCSERVER_WATCHDOG=y ; to
	CONFIG_MSM_RPCSERVER_WATCHDOG=n
build --force linux-samsung-kylessopen
install --no-fde
flasher flash_kernel

    The steps above caused no change. Did I do this right or could it be something else?

    Edited
    To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information

    Child items ...

    • Activity

    • Administrator changed the description · Imported

      changed the description

      By Ghost User on 2019-05-03T09:11:04

    • Administrator
      Administrator @root · Imported
      Author Owner

      Just to make sure, the steps you did don't include flashing the actual system, you're just flashing the kernel. Is this right or did you just leave that out of this issue?

      By Bart Ribbers on 2019-05-03T19:28:00

    • Administrator
      Administrator @root · Imported
      Author Owner

      You're right. I assumed since I only changed the kernel, I'd only flash the kernel. I'm going to flash the rootfs now and get back to you.

      By Ghost User on 2019-05-03T20:14:52

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      I've flashed both the kernel and rootfs with the same results.

      By Ghost User on 2019-05-03T22:04:17

    • Administrator
      Administrator @root · Imported
      Author Owner

      Just to make sure that it's not only the watchdog, have you checked the last_kmsg?

      By Danct12 on 2019-05-04T04:20:14

    • Administrator
      Administrator @root · Imported
      Author Owner

      /proc/last_kmsg doesn't exist. I did however just capture an output of dmesg ~26ish seconds before a crash. It may not say much but it could give us a lead. https://pastebin.com/cphAM30p

      By Ghost User on 2019-05-04T06:27:11

    • Administrator
      Administrator @root · Imported
      Author Owner

      Another dmesg output taken about 8ish seconds before crash via bash trickery. https://pastebin.com/EmRSSFG9

      By Ghost User on 2019-05-04T06:56:33

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      Is there a /dev/watchdog or similar ("wd", "wd0"...?) in /dev?

      If so, you could set up "watchdog-kick" to continuously kick it, see device-nokia-n9 for example.

      By Oliver Smith on 2019-05-05T08:45:26

    • Administrator
      Administrator @root · Imported
      Author Owner

      I couldn't find such a thing. Here's a directory listing of /dev/ https://pastebin.com/AyndwmkZ.

      By Ghost User on 2019-05-05T22:08:29

    • Administrator
      Administrator @root · Imported
      Author Owner

      And here's a directory tree listing of /dev https://pastebin.com/KMj3tyi1.

      By Ghost User on 2019-05-05T22:31:29

    • Administrator
      Administrator @root · Imported
      Author Owner

      Thanks for posting these, I don't have an idea then, besides digging in the kernel source for watchdog related code, maybe commenting out stuff or adding debug statements.

      (btw, I would prefer if you could attach the logs as files, instead of linking to pastebin. pastebin wants tor users to type in captchas, which is annoying.)

      By Oliver Smith on 2019-05-06T21:51:33

    • Administrator
      Administrator @root · Imported
      Author Owner

      @ollieparanoid While searching through the wiki pages of similar devices, I stumbled upon the Samsung Galaxy Mini 2 (samsung-s6500d). If you check it out, they have similar issue with the reboots as well (the devices are very similar). Sadly, my device is partially damaged, I'm busy with life and I'm broke. If you can see anything you recognize or could give any pointers, it'd be greatly appreciated.

      By Ghost User on 2019-06-04T08:57:54

    • Administrator
      Administrator @root · Imported
      Author Owner

      Maybe @Halamix2 has a tip for you, he is the one who ported the device. But it seems like it is stuck there as well.

      TBH I would suggest to sort your life out first, so you can come back with recharged batteries and fresh motivation and ideas if you want :smiley:

      By Oliver Smith on 2019-06-04T20:15:51

    • Administrator
      Administrator @root · Imported
      Author Owner

      I have a weird obsession with this project. I would have definitely given up if I were doing anything else by now. I'm definitely taking it easy, so my contributions will be spaced out a lot more and lesser in volume, but I bet our kernels both have that file with the hardcoded message, and if I ran diff on them, there might be a parameter or something for the timeout. It's just a thought. I don't have the time (or patience) to do kernel stuff right now, but I'll get to it eventually if someone doesn't beat me to it or my device doesn't become completely unusable.

      By Ghost User on 2019-06-05T11:58:23

    • Administrator
      Administrator @root · Imported
      Author Owner

      In my case (s6500d) I couldn't find watchdog anywhere either. I was able to run on pmos (dowbnstream) kernel longer than 3 minutes: pmos kernel + cyanogenmod initramfs allowed me to do this (but also disallowed me to sue ssh, see anything on screen except backlight etc., basically brick that doesn't reboot every 3 minutes). I wasn't able to determine what magic is done in CM initramfs for the phone to not reboot

      By Piotr Halama on 2019-06-05T19:42:18

    • Administrator
      Administrator @root · Imported
      Author Owner

      Does PostmarketOS boot past the CM initramfs?

      By Ghost User on 2019-06-06T15:07:31

    • Administrator
      Administrator @root · Imported
      Author Owner

      No, it tries to boot Android which I didn't had on the phone. I could try this combination: boot.img(pmos kernel+MC initramfs) + rootfs(CyanogenMod) to see if the kernel is okay. (IF Android will boot with that kernel of course)

      By Piotr Halama on 2019-06-07T12:46:05

    • Administrator
      Administrator @root · Imported
      Author Owner

      The pmos kernel should be almost exactly the same as the Android one, depending on where it's from. Do give it a try and see if the issue goes away, I'm curious.

      By Ghost User on 2019-06-08T05:14:58

    • Administrator
      Administrator @root · Imported
      Author Owner

      samsung-s6500d: as it turns out the config.gz in the final CyanogenMod11 .zip differs from defconfig in git repo, I try to make diff file and find what it could be (if this really is problem in the kernel) @lofenyy , can you check your config and compare if any of the options with "<" in front differs from your pmos kernel (besides CONFIG_CMDLINE_*, these were set to force cmdline without overwriting by bootloader)?

      < CONFIG_HAVE_HW_BRKPT_RESERVED_RW_ACCESS=y
< CONFIG_IOSCHED_BFQ=y
< CONFIG_MACH_QRD_SKUD_PRIME=y
< CONFIG_MACH_MSM8625Q_EVBD=y
< CONFIG_MACH_MSM8625Q_SKUD=y
< CONFIG_MSM7X27A_OVERCLOCK=y
< CONFIG_MSM_PBUS=y
< CONFIG_ARCH_HAS_BARRIERS=y
> CONFIG_BOUNCE=y
< CONFIG_CMDLINE="init=/sbin/init root=/dev/ram rw initrd=0x11000000,16M console=ttyDCC0 mem=88M"
< CONFIG_HAVE_AOUT=y
< CONFIG_ANDROID_PARANOID_NETWORK=y
< CONFIG_NETWORK_SECMARK=y
< CONFIG_NF_CONNTRACK_SECMARK=y
< CONFIG_NETFILTER_XT_TARGET_CONNSECMARK=y
< CONFIG_NETFILTER_XT_TARGET_SECMARK=y
< CONFIG_IP_NF_SECURITY=y
> CONFIG_SYNC=y
> CONFIG_SW_SYNC=y
> CONFIG_SW_SYNC_USER=y
< CONFIG_EXFAT_FS=m
< CONFIG_EXFAT_DEFAULT_CODEPAGE=437
< CONFIG_EXFAT_DEFAULT_IOCHARSET="utf8"
< CONFIG_SECURITY=y
< CONFIG_SECURITY_NETWORK=y
< CONFIG_LSM_MMAP_MIN_ADDR=4096
< CONFIG_SECURITY_SELINUX=y
< CONFIG_SECURITY_SELINUX_DEVELOP=y
< CONFIG_SECURITY_SELINUX_AVC_STATS=y
< CONFIG_SECURITY_SELINUX_CHECKREQPROT_VALUE=1
< CONFIG_DEFAULT_SECURITY_SELINUX=y
< CONFIG_DEFAULT_SECURITY="selinux"
> CONFIG_DEFAULT_SECURITY_DAC=y
> CONFIG_DEFAULT_SECURITY=""

      UPDATE: List of things that changed between Android .zip and git repo for samsung-s6500d AND things that are different between samsung-s6500d Android.zip and pmos samsung-kylessopen:

      < CONFIG_HAVE_HW_BRKPT_RESERVED_RW_ACCESS=y
< CONFIG_MACH_QRD_SKUD_PRIME=y
< CONFIG_MACH_MSM8625Q_EVBD=y
< CONFIG_MACH_MSM8625Q_SKUD=y
< CONFIG_ARCH_HAS_BARRIERS=y
---
< CONFIG_NF_CONNTRACK_SECMARK=y
< CONFIG_NETFILTER_XT_TARGET_CONNSECMARK=y
< CONFIG_NETFILTER_XT_TARGET_SECMARK=y
< CONFIG_IP_NF_SECURITY=y

      lg-p700, which is 3rd of of 3 booting devices with MSMS7227a SoC has CONFIG_ARCH_HAS_BARRIERS=y ( netfilters/2nd part of diff also are set to "y", but I don't think this causes problems) , rest of the options is unset, like in the rest of pmos kernels for this SoC, I'm going to change that in s6500d kernel and see it this helps nope, this does not help

      By Piotr Halama on 2019-08-31T18:30:47

      Edited by Administrator
    • Administrator mentioned in issue #348 (closed) · Imported

      mentioned in issue #348 (closed)

      By Piotr Halama on 2019-09-01T00:04:17

    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    • Loading
    Please register or sign in to reply