RE: Commit 'sched/cpuacct: Split usage accounting into user_usage and sys_usage' causing crashes

[Zhao Lei]

Hi, Guenter,

> From: Guenter Roeck [mailto:linux@xxxxxxxxxxxx]
> Sent: Monday, April 11, 2016 12:15 AM
> To: Dongsheng Yang <yangds.fnst@xxxxxxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx; Zhao Lei <zhaolei@xxxxxxxxxxxxxx>; Peter
> Zijlstra <peterz@xxxxxxxxxxxxx>; Tejun Heo <tj@xxxxxxxxxx>; Ingo Molnar
> <mingo@xxxxxxxxxx>; linux-next@xxxxxxxxxxxxxxx
> Subject: Commit 'sched/cpuacct: Split usage accounting into user_usage and
> sys_usage' causing crashes
> 
> Hi,
> 
> your commit 'sched/cpuacct: Split usage accounting into user_usage and
> sys_usage' in -next causes crashes with several of my qemu tests (mac99,
> q3beige). Reverting the patch fixes the problem.
> 
> Details and complete logs are available at
> 	http://kerneltests.org/builders/qemu-ppc-next
> The scripts are published at
> 	https://github.com/groeck/linux-build-test/tree/master/rootfs/ppc
> 
> Crash log and bisect log are attached.
> 
Thanks for report.

Seems it is same problem discussed in
https://lkml.org/lkml/2016/4/4/312
and fixed by:
https://lkml.org/lkml/2016/4/6/306

Thanks
Zhaolei

> Guenter
> 
> ---
> Crash log:
> 
> [ ... ]
> ftrace: allocating 20986 entries in 62 pages
> Unable to handle kernel paging request for data at address 0x00000084
> Faulting instruction address: 0xc0067188
> Oops: Kernel access of bad area, sig: 11 [#1]
> PREEMPT PowerMac
> Modules linked in:
> CPU: 0 PID: 2 Comm: kthreadd Not tainted
> 4.6.0-rc2-next-20160408-yocto-standard #1
> task: c78d0e00 ti: c78e8000 task.ti: c78e8000
> NIP: c0067188 LR: c005c0d0 CTR: c005cf84
> REGS: c78e9da0 TRAP: 0300   Not tainted
> (4.6.0-rc2-next-20160408-yocto-standard)
> MSR: 00001032 <ME,IR,DR,RI>  CR: 82040002  XER: 00000000
> DAR: 00000084 DSISR: 40000000
> GPR00: c005c0d0 c78e9e50 c78d0e00 c78d0e00 00000000 00000000 000a25f0
> 0037ecae
> GPR08: ffffffff 00000000 00000000 0027ecae 22000008 00000000 c004decc
> 00000000
> GPR16: 00000000 00000000 00000000 00000000 00000000 00000000
> 00000000 00000000
> GPR24: 00000001 c78d0e00 c0918580 c78d0e30 00000000 c78d0e00
> 00000000 000a25f0
> NIP [c0067188] cpuacct_charge+0x2c/0x90
> LR [c005c0d0] update_curr+0x110/0x250
> Call Trace:
> [c78e9e70] [c005c0d0] update_curr+0x110/0x250
> [c78e9ea0] [c005d048] dequeue_task_fair+0xc4/0x27c
> [c78e9ed0] [c06c67d4] __schedule+0x338/0x770
> [c78e9f00] [c06c6c54] schedule+0x48/0xb8
> [c78e9f10] [c004e070] kthreadd+0x1a4/0x1dc
> [c78e9f40] [c0010540] ret_from_kernel_thread+0x5c/0x64
> Instruction dump:
> 4bffffd0 7c0802a6 90010004 60000000 9421ffe0 7c0802a6 bf810010 90010024
> 81230cb4 7c7d1b78 7cdf3378 7cbe2b78 <83890084> 4800fbfd 813d047c
> 6b9c4000
> ---[ end trace dc8fa200cb88537f ]---
> 
> 
> ---
> bisect:
> 
> # bad: [6e5921edae25005c5c25a79bb2ff5335650c1b66] Add linux-next specific
> files
> # for 20160408
> # good: [9735a22799b9214d17d3c231fe377fc852f042e9] Linux 4.6-rc2
> git bisect start 'HEAD' 'v4.6-rc2'
> # good: [e61201ac50d9599aef64e8885d6bfa5b35d574d7] Merge
> remote-tracking branch
> # 'mac80211-next/master'
> git bisect good e61201ac50d9599aef64e8885d6bfa5b35d574d7
> # bad: [5c4015574ec4d63b7d3a9efbd8b77f90bcac88be] Merge
> remote-tracking branch
> # 'usb-chipidea-next/ci-for-usb-next'
> git bisect bad 5c4015574ec4d63b7d3a9efbd8b77f90bcac88be
> # good: [35fda07ad91c8d46253f6dd262800a747adcede9] Merge
> remote-tracking branch
> # 'sound-asoc/for-next'
> git bisect good 35fda07ad91c8d46253f6dd262800a747adcede9
> # good: [a2089aa1dcd0a33b8eb67d8db0e07f859b986e04] Merge
> remote-tracking branch
> # 'spi/for-next'
> git bisect good a2089aa1dcd0a33b8eb67d8db0e07f859b986e04
> # bad: [129f0d75057f9e78ad70e3e01f6c9edbc7de553c] Merge
> remote-tracking branch
> # 'irqchip/irqchip/for-next'
> git bisect bad 129f0d75057f9e78ad70e3e01f6c9edbc7de553c
> # bad: [9338637e6e2256ac26ac75391dc1349fb492d7c6] Merge branch
> 'x86/asm'
> git bisect bad 9338637e6e2256ac26ac75391dc1349fb492d7c6
> # good: [66d219014a4ee47ad4ca2b9db5fe6547353e2a56] perf/x86/intel/pt:
> Move
> # transaction start/stop to PMU start/stop callbacks
> git bisect good 66d219014a4ee47ad4ca2b9db5fe6547353e2a56
> # good: [5367d2a3279cd7aa40535efa526fe936ec24528d] Merge branch
> 'mm/urgent'
> git bisect good 5367d2a3279cd7aa40535efa526fe936ec24528d
> # bad: [2b8c41daba327c633228169e8bd8ec067ab443f8] sched/fair: Initiate a
> new
> # task's util avg to a bounded value
> git bisect bad 2b8c41daba327c633228169e8bd8ec067ab443f8
> # bad: [d02c071183e1c01a76811c878c8a52322201f81f] sched/fair: Reset
> # nr_balance_failed after active balancing
> git bisect bad d02c071183e1c01a76811c878c8a52322201f81f
> # bad: [d740037fac7052e49450f6fa1454f1144a103b55] sched/cpuacct: Split
> usage
> # accounting into user_usage and sys_usage
> git bisect bad d740037fac7052e49450f6fa1454f1144a103b55
> # good: [5ca3726af7f66a8cc71ce4414cfeb86deb784491] sched/cpuacct: Show
> all
> # possible CPUs in cpuacct output
> git bisect good 5ca3726af7f66a8cc71ce4414cfeb86deb784491
> # first bad commit: [d740037fac7052e49450f6fa1454f1144a103b55]
> sched/cpuacct:
> # Split usage accounting into user_usage and sys_usage
>