4.12 kernel causes a minor issue with RAID

This forum is for Marvell Kirkwood devices such as the GoFlex Home/Net, PogoPlug v1/v2, SheevaPlug, and ZyXEL devices.

4.12 kernel causes a minor issue with RAID

Postby Sergeanter » Thu Jul 20, 2017 8:11 pm

Since the upgrade to 4.12 I began getting this at every startup.
Code: Select all
[   22.561215] [<c000da9c>] (show_stack) from [<c001e9c0>] (__warn+0xdc/0x104)
[   22.568177] [<c001e9c0>] (__warn) from [<c001ea98>] (warn_slowpath_null+0x20/0x28)
[   22.575794] [<c001ea98>] (warn_slowpath_null) from [<bf003df0>] (set_in_sync+0x11c/0x13c [md_mod])
[   22.584797] [<bf003df0>] (set_in_sync [md_mod]) from [<bf0106f4>] (md_check_recovery+0x288/0x624 [md_mod])
[   22.594542] [<bf0106f4>] (md_check_recovery [md_mod]) from [<bf06ed14>] (raid5d+0x1c/0x6bc [raid456])
[   22.603832] [<bf06ed14>] (raid5d [raid456]) from [<bf003afc>] (md_thread+0x144/0x164 [md_mod])
[   22.612487] [<bf003afc>] (md_thread [md_mod]) from [<c003c9a4>] (kthread+0x12c/0x13c)
[   22.620311] [<c003c9a4>] (kthread) from [<c000a550>] (ret_from_fork+0x14/0x24)
[   22.627521] ---[ end trace 91f7f45a2ca10b06 ]---
Sergeanter
 
Posts: 72
Joined: Wed Oct 02, 2013 5:14 am

Re: 4.12 kernel causes a minor issue with RAID

Postby kriztioan » Wed Jul 26, 2017 3:59 am

Same here with 4.12.3, but for raid1:

Code: Select all
[    5.431250] ------------[ cut here ]------------
[    5.437265] WARNING: CPU: 0 PID: 166 at drivers/md/md.c:2273 set_in_sync+0x11c/0x13c [md_mod]
[    5.447112] Modules linked in: raid1 md_mod
[    5.452699] CPU: 0 PID: 166 Comm: md1_raid1 Not tainted 4.12.3-1-ARCH #1
[    5.460151] Hardware name: Marvell Kirkwood (Flattened Device Tree)
[    5.466513] [<c0010560>] (unwind_backtrace) from [<c000da9c>] (show_stack+0x10/0x14)
[    5.474394] [<c000da9c>] (show_stack) from [<c001e9c0>] (__warn+0xdc/0x104)
[    5.481486] [<c001e9c0>] (__warn) from [<c001ea98>] (warn_slowpath_null+0x20/0x28)
[    5.489320] [<c001ea98>] (warn_slowpath_null) from [<bf003df0>] (set_in_sync+0x11c/0x13c [md_mod])
[    5.498668] [<bf003df0>] (set_in_sync [md_mod]) from [<bf0106f4>] (md_check_recovery+0x288/0x624 [md_mod])
[    5.508614] [<bf0106f4>] (md_check_recovery [md_mod]) from [<bf02e5a0>] (raid1d+0x1c/0x1998 [raid1])
[    5.518035] [<bf02e5a0>] (raid1d [raid1]) from [<bf003afc>] (md_thread+0x144/0x164 [md_mod])
[    5.526678] [<bf003afc>] (md_thread [md_mod]) from [<c003c9a4>] (kthread+0x12c/0x13c)
[    5.534651] [<c003c9a4>] (kthread) from [<c000a550>] (ret_from_fork+0x14/0x24)
[    5.541981] ---[ end trace 3e62c99f8360f133 ]---


Just before this message I get:

Code: Select all
[    4.904909] md: md0 stopped.
[    4.937487] md/raid1:md0: active with 2 out of 2 mirrors
[    4.944592] md0: detected capacity change from 0 to 268369920
[    4.992684] md: md1 stopped.
[    5.054064] md/raid1:md1: active with 2 out of 2 mirrors
[    5.059734] md1: detected capacity change from 0 to 8581545984
[    5.099886] md: md2 stopped.
[    5.136931] md/raid1:md2: active with 2 out of 2 mirrors
[    5.197614] md2: detected capacity change from 0 to 1991405207552


Something going on at drivers/md/md.c:2273... but beyond me.

/proc/mdstat reports everything OK:

Code: Select all
Personalities : [raid1]
md2 : active raid1 sda3[0] sdb3[2]
      1944731648 blocks super 1.2 [2/2] [UU]
      bitmap: 0/15 pages [0KB], 65536KB chunk

md1 : active raid1 sda2[0] sdb2[2]
      8380416 blocks super 1.2 [2/2] [UU]

md0 : active raid1 sda1[0] sdb1[1]
      262080 blocks [2/2] [UU]

unused devices: <none>
kriztioan
 
Posts: 8
Joined: Sat Apr 29, 2017 1:32 am

Re: 4.12 kernel causes a minor issue with RAID

Postby Sergeanter » Wed Jul 26, 2017 6:05 pm

Moonman is looking into this. This may or may not be related but the patch causing overflow at 8TB capacity for RAID5 has been applied to this kernel. I have rolled back to 4.11 for now.
Sergeanter
 
Posts: 72
Joined: Wed Oct 02, 2013 5:14 am

Re: 4.12 kernel causes a minor issue with RAID

Postby kriztioan » Fri Jul 28, 2017 3:12 am

Great! Using some Google-fu I found a discussion on a mailinglist archive (linux-raid;
https://www.spinics.net/lists/raid/msg58547.html) giving me the impression that the issue has to do with our systems only having a single CPU/core. My wild guess would be that some kind of a CPU-lock is tried, which wouldn't make sense on a single CPU. A patch seems to have been submitted to the main kernel-line. I suspect this will trickle through to our end at some point.
kriztioan
 
Posts: 8
Joined: Sat Apr 29, 2017 1:32 am

Re: 4.12 kernel causes a minor issue with RAID

Postby kriztioan » Wed Aug 23, 2017 3:12 am

After upgrading to kernel 4.12.8 the message is still present in the kernel logs. However, it looks like it is going to be fixed in the version 4.13-series; http://elixir.free-electrons.com/linux/v4.13-rc6/source/drivers/md/md.c#L2290 or https://github.com/torvalds/linux/blob/master/drivers/md/md.c#L2288 - a little bit more patients it seems like.
kriztioan
 
Posts: 8
Joined: Sat Apr 29, 2017 1:32 am

Re: 4.12 kernel causes a minor issue with RAID

Postby NaldoJud » Wed Aug 23, 2017 10:42 am

I'm still getting this annoying message too.
NaldoJud
 
Posts: 1
Joined: Thu Aug 17, 2017 9:57 am

Re: 4.12 kernel causes a minor issue with RAID

Postby kriztioan » Tue Sep 19, 2017 3:36 am

After upgrading to kernel 4.13.2-1 the message is gone from the kernel logs. Unfortunately now cryptodev fails to compile with the new kernel-headers:
Code: Select all
DKMS make.log for cryptodev-1.9 for kernel 4.13.2-1-ARCH (armv5tel)
Mon Sep 18 20:34:21 PDT 2017
/var/lib/dkms/cryptodev/1.9/build/ioctl.c:1127:3: error: positional initialization of field in ‘struct’ declared with ‘designated_init’ attribute [-Werror=designated-init]
  {0, },
   ^
/var/lib/dkms/cryptodev/1.9/build/ioctl.c:1127:3: note: (near initialization for ‘verbosity_ctl_dir[1]’)
/var/lib/dkms/cryptodev/1.9/build/ioctl.c:1136:3: error: positional initialization of field in ‘struct’ declared with ‘designated_init’ attribute [-Werror=designated-init]
  {0, },
   ^
/var/lib/dkms/cryptodev/1.9/build/ioctl.c:1136:3: note: (near initialization for ‘verbosity_ctl_root[1]’)
cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:303: /var/lib/dkms/cryptodev/1.9/build/ioctl.o] Error 1
make[1]: *** [Makefile:1512: _module_/var/lib/dkms/cryptodev/1.9/build] Error 2
make: *** [Makefile:27: build] Error 2
kriztioan
 
Posts: 8
Joined: Sat Apr 29, 2017 1:32 am

Re: 4.12 kernel causes a minor issue with RAID

Postby moonman » Tue Sep 19, 2017 4:24 am

Yup, a bug report has been submitted already: https://github.com/cryptodev-linux/cryp ... /issues/29

Edit: Applied a patch to the package. Should work now. Create a new thread next time please.
Pogoplug V4 | GoFlex Home | Raspberry Pi B 512 | CuBox-i4 Pro | ClearFog | BeagleBone Black | Odroid U2 | Odroid C1 | Odroid XU4
-----------------------------------------------------------------------------------------------------------------------
[armv5] Updated U-Boot | |[armv5] How to install my.pogoplug.com service | [armv5] NAND Rescue System
moonman
Developer
 
Posts: 3089
Joined: Sat Jan 15, 2011 3:36 am
Location: Calgary, Canada


Return to Marvell Kirkwood

Who is online

Users browsing this forum: No registered users and 1 guest