sigsegv test failed with Ubuntu T-3.13
Cypresslin opened this issue · comments
Hello Colin,
with recent SRU test for Trusty 3.13, we have found that sigsegv test is failing with Ubuntu Trusty 3.13.0-195, no matter it's a baremetal or VM. And it can be reproduced with 3.13.0-170-generic as well.
Test log:
Free memory: 7209 MB
Memory used: 6488 MB
Using cgroup version 1
/home/ubuntu/autotest/client/tests/ubuntu_stress_smoke_test/ubuntu_stress_single_smoke_test.sh: line 32: [: too many arguments
Machine Configuration
Physical Pages: 2020977
Pages available: 1845513
Page Size: 4096
Zswap enabled: N
Free memory:
total used free shared buffers cached
Mem: 8083908 701980 7381928 444 23280 473656
-/+ buffers/cache: 205044 7878864
Swap: 5242872 880 5241992
Number of CPUs: 4
Number of CPUs Online: 4
Maximum bogo ops: 3000
sigsegv STARTING
sigsegv RETURNED 2
sigsegv FAILED
stress-ng: debug: [18923] invoked with './stress-ng -v -t 5 --sigsegv 4 --sigsegv-ops 3000 --ignite-cpu --syslog --verbose --verify --oomable' by user 0 'root'
stress-ng: debug: [18923] stress-ng 0.17.04 g955e7a758572
stress-ng: debug: [18923] system: Linux durin 3.13.0-195-generic #246-Ubuntu SMP Mon Jan 15 15:10:06 UTC 2024 x86_64, gcc 4.8.4, glibc 2.19
stress-ng: debug: [18923] RAM total: 7.7G, RAM free: 7.0G, swap free: 5.0G
stress-ng: debug: [18923] temporary file path: '/home/ubuntu/autotest/client/tmp/ubuntu_stress_smoke_test/src/stress-ng', filesystem type: ext2 (226160480 blocks available)
stress-ng: debug: [18923] CPUs have 3 idle states: C1, C2, POLL
stress-ng: debug: [18923] 4 processors online, 4 processors configured
stress-ng: info: [18923] setting to a 5 secs run per stressor
stress-ng: debug: [18923] CPU data cache: L1: 32K, L2: 256K, L3: 8192K
stress-ng: debug: [18923] cache allocate: shared cache buffer size: 8192K
stress-ng: info: [18923] dispatching hogs: 4 sigsegv
stress-ng: debug: [18923] starting stressors
stress-ng: debug: [18923] 4 stressors started
stress-ng: debug: [18924] sigsegv: [18924] started (instance 0 on CPU 3)
stress-ng: debug: [18925] sigsegv: [18925] started (instance 1 on CPU 0)
stress-ng: debug: [18926] sigsegv: [18926] started (instance 2 on CPU 0)
stress-ng: debug: [18927] sigsegv: [18927] started (instance 3 on CPU 0)
stress-ng: fail: [18924] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18924] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18925] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18925] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18925] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18925] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18924] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18925] sigsegv: expecting fault address 0x8, got 0x10 instead
info: 5 failures reached, aborting stress process
stress-ng: fail: [18924] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: debug: [18925] sigsegv: [18925] exited (instance 1 on CPU 1)
stress-ng: fail: [18924] sigsegv: expecting fault address 0x8, got 0x10 instead
info: 5 failures reached, aborting stress process
stress-ng: debug: [18924] sigsegv: [18924] exited (instance 0 on CPU 3)
stress-ng: fail: [18926] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: error: [18923] sigsegv: [18924] terminated with an error, exit status=2 (stressor failed)
stress-ng: debug: [18923] sigsegv: [18924] terminated (stressor failed)
stress-ng: error: [18923] sigsegv: [18925] terminated with an error, exit status=2 (stressor failed)
stress-ng: fail: [18926] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18926] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18926] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: debug: [18923] sigsegv: [18925] terminated (stressor failed)
stress-ng: fail: [18926] sigsegv: expecting fault address 0x8, got 0x10 instead
info: 5 failures reached, aborting stress process
stress-ng: fail: [18927] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: debug: [18926] sigsegv: [18926] exited (instance 2 on CPU 2)
stress-ng: fail: [18927] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18927] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: error: [18923] sigsegv: [18926] terminated with an error, exit status=2 (stressor failed)
stress-ng: debug: [18923] sigsegv: [18926] terminated (stressor failed)
stress-ng: fail: [18927] sigsegv: expecting fault address 0x8, got 0x10 instead
stress-ng: fail: [18927] sigsegv: expecting fault address 0x8, got 0x10 instead
info: 5 failures reached, aborting stress process
stress-ng: debug: [18927] sigsegv: [18927] exited (instance 3 on CPU 1)
stress-ng: error: [18923] sigsegv: [18927] terminated with an error, exit status=2 (stressor failed)
stress-ng: debug: [18923] sigsegv: [18927] terminated (stressor failed)
stress-ng: debug: [18923] metrics-check: all stressor metrics validated and sane
stress-ng: info: [18923] skipped: 0
stress-ng: info: [18923] passed: 0
stress-ng: info: [18923] failed: 4: sigsegv (4)
stress-ng: info: [18923] metrics untrustworthy: 0
stress-ng: info: [18923] unsuccessful run completed in 0.01 secs
Bisect shows commit e9ddb1c is the first bad commit.
Sorry for not catching this early on, we seldom test Trusty 3.13 in recent years.
Good catch, thanks for the bisect information, I'll try and see what's happing.
I've fixed this and tested on 3.13.0-170-generic. Do you mind giving this a check to see if it fully addresses the problem you are seeing?
Fix verified with both 3.13.0-170 and -195, thank you!