Bug#813764: linux-source-3.16: "Dazed and confused, but trying to continue" on X10SDV-TLN4F while using perf top
I can reproduce this bug on Intel Broadwell processors by running "perf top" and another cpu intensive process like "stress -c 8". Within 5 minutes, the kernel reports an unexpected NMI: [2005170.748842] Uhhuh. NMI received for unknown reason 01 on CPU 70. [2005170.748882] Do you have a strange power saving mode enabled? [2005170.748900] Dazed and confused, but trying to continue -- Hardware CPU: Intel Xeon E5-2695 v4 RAM: 512GB DDR4 2400MHz System: Supermicro B10DRT -- Operating System Debian Release: 8.7 Architecture: amd64 (x86_64) Kernel: Linux 3.16.0-4-amd64 Init: systemd
Bug#813764: linux-source-3.16: "Dazed and confused, but trying to continue" on X10SDV-TLN4F while using perf top
Control: reassign -1 src:linux 3.16.7-ckt20-1+deb8u3 Control: tag -1 unreproducible moreinfo On Thu, 2016-02-04 at 23:25 -0500, Rich Ercolani wrote: > Package: linux-source-3.16 > Version: 3.16.0-4 > Severity: normal > > Dear Maintainer, > > I was going about my business, using perf top to see what I was spending a > bunch of my time in on this system, when suddenly, I got this written to > console: > > [2425302.546957] Uhhuh. NMI received for unknown reason 11 on CPU 1. > [2425302.547625] Do you have a strange power saving mode enabled? > [2425302.548291] Dazed and confused, but trying to continue > > I've been running this system for several months at this point, and have not > seen this issue at any point prior. > > I can't easily reproduce it again, so it may have been something other than > "perf top" that resulted in the behavior, but that's the only unusual thing > I can think of. [...] perf uses the hardware performance monitoring units (PMUs) which do trigger NMIs periodically. Obviously the kernel should expect and recognise these as related to performance monitoring, but there could be a bug in that code. I'm tagging this report as needing more information as we won't be able to make any progress unless you can find a way to reproduce the bug. Ben. -- Ben Hutchings It is a miracle that curiosity survives formal education. - Albert Einstein signature.asc Description: This is a digitally signed message part
Bug#813764: linux-source-3.16: "Dazed and confused, but trying to continue" on X10SDV-TLN4F while using perf top
Package: linux-source-3.16 Version: 3.16.0-4 Severity: normal Dear Maintainer, I was going about my business, using perf top to see what I was spending a bunch of my time in on this system, when suddenly, I got this written to console: [2425302.546957] Uhhuh. NMI received for unknown reason 11 on CPU 1. [2425302.547625] Do you have a strange power saving mode enabled? [2425302.548291] Dazed and confused, but trying to continue I've been running this system for several months at this point, and have not seen this issue at any point prior. I can't easily reproduce it again, so it may have been something other than "perf top" that resulted in the behavior, but that's the only unusual thing I can think of. Supermicro X10SDV-TLN4F, which is a Xeon D-1540 board. -- System Information: Debian Release: 8.2 APT prefers stable APT policy: (1000, 'stable'), (900, 'testing'), (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 3.16.0-4-amd64 (SMP w/16 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system)