[Bug 253616] Panic when concurrently trying to start/stop auditing

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=253616

            Bug ID: 253616
           Summary: Panic when concurrently trying to start/stop auditing
           Product: Base System
           Version: CURRENT
          Hardware: Any
                OS: Any
            Status: New
          Severity: Affects Only Me
          Priority: ---
         Component: kern
          Assignee: bugs at FreeBSD.org
          Reporter: arichardson at FreeBSD.org

This can be reproduced by removing the is_exclusive tag from the audit tests
after https://reviews.freebsd.org/D28451 and running them in parallel. 

root at freebsd-amd64:/usr/tests/sys/audit # kyua -v parallelism=16 test
administrative:adjtime_failure  ->  passed  [0.153s]
administrative:adjtime_success  ->  passed  [0.149s]
administrative:auditctl_failure  ->  failed:
/local/scratch/alr48/cheri/freebsd/tests/sys/audit/utils.c:300: 0 !=
audit_quick_start(): Failed to start fake auditd: File exists  [0.148s]
administrative:auditctl_success  ->  passed  [0.145s]
administrative:auditon_default_failure  ->  passed  [0.057s]
administrative:auditon_default_success  ->  failed:
/local/scratch/alr48/cheri/freebsd/tests/sys/audit/utils.c:300: 0 !=
audit_quick_start(): Failed to start fake auditd: File exists  [0.143s]
panic: audit_syscall_enter: td->td_ar != NULL
cpuid = 0
time = 1612265481
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe0093bb59d0
vpanic() at vpanic+0x188/frame 0xfffffe0093bb5a20
panic() at panic+0x43/frame 0xfffffe0093bb5a80
audit_syscall_enter() at audit_syscall_enter+0x1b7/frame 0xfffffe0093bb5ac0
amd64_syscall() at amd64_syscall+0x702/frame 0xfffffe0093bb5bf0
fast_syscall_common() at fast_syscall_common+0xf8/frame 0xfffffe0093bb5bf0
--- syscall (3, FreeBSD ELF64, sys_read), rip = 0x8003e556a, rsp =
0x7fffffffca48, rbp = 0x7fffffffca60 ---
KDB: enter: panic
[ thread pid 809 tid 100093 ]
Stopped at      kdb_enter+0x37: movq    $0,0x10a084e(%rip)

-- 
You are receiving this mail because:
You are the assignee for the bug.