Bug 368923

Summary: WARNING: unhandled arm64-linux syscall: 268 (setns)
Product: [Developer tools] valgrind Reporter: Marcin Juszkiewicz <marcin>
Component: generalAssignee: Julian Seward <jseward>
Status: RESOLVED DUPLICATE    
Severity: normal CC: bjb, jreiser, mark, michal.privoznik
Priority: NOR    
Version: 3.11.0   
Target Milestone: ---   
Platform: Other   
OS: Linux   
See Also: https://bugs.kde.org/show_bug.cgi?id=343099
Latest Commit: Version Fixed In:
Sentry Crash Report:
Attachments: 0001-Add-support-for-setns-syscall.patch

Description Marcin Juszkiewicz 2016-09-16 20:40:05 UTC 
<<<test_start>>>
tag=setns01_valgrind_memory_leak_check stime=1473974802
cmdline=" valgrind -q --leak-check=full --trace-children=yes setns01"
contacts=""
analysis=exit
<<<test_output>>>
--26151-- WARNING: unhandled arm64-linux syscall: 268
--26151-- You may be able to write your own handler.
--26151-- Read the file README_MISSING_SYSCALL_OR_IOCTL.
--26151-- Nevertheless we consider this a bug.  Please report
--26151-- it at http://valgrind.org/support/bug_reports.html.
setns01     1  TCONF  :  setns01.c:209: syscall(268) __NR_setns not supported on your arch
setns01     2  TCONF  :  setns01.c:209: Remaining cases not appropriate for configuration
<<<execution_status>>>
initiation_status="ok"
duration=1 termination_type=exited termination_id=32 corefile=no
cutime=61 cstime=2
<<<test_end>>>
<<<test_start>>>
tag=setns01_valgrind_thread_concurrency_check stime=1473974803
cmdline=" valgrind -q --tool=helgrind --trace-children=yes setns01"
contacts=""
analysis=exit
<<<test_output>>>
--26158-- WARNING: unhandled arm64-linux syscall: 268
--26158-- You may be able to write your own handler.
--26158-- Read the file README_MISSING_SYSCALL_OR_IOCTL.
--26158-- Nevertheless we consider this a bug.  Please report
--26158-- it at http://valgrind.org/support/bug_reports.html.
setns01     1  TCONF  :  setns01.c:209: syscall(268) __NR_setns not supported on your arch
setns01     2  TCONF  :  setns01.c:209: Remaining cases not appropriate for configuration
<<<execution_status>>>
initiation_status="ok"
duration=0 termination_type=exited termination_id=32 corefile=no
cutime=35 cstime=3
<<<test_end>>>


Reproducible: Always
Comment 1 Marcin Juszkiewicz 2016-09-19 06:58:18 UTC 
https://bugs.kde.org/show_bug.cgi?id=369031 - same for amd64
Comment 2 Marcin Juszkiewicz 2018-04-30 08:07:37 UTC 
Can we close it? IIRC it was merged into valgrind.
Comment 3 Brenda J. Butler 2018-11-21 03:11:14 UTC 
This bug is similar to 343099 and 369031.

But this bug is reporting on arm64 arch, while the others have to do with amd64.  So, not marking this as duplicate of 343099.  Is there another way to link this bug with those ones besides writing a comment?
Comment 4 John Reiser 2019-05-06 03:43:05 UTC 
Still happens in Valgrind-3.15.0 on arm64, notably in /usr/bin/perf.

# gdb perf
GNU gdb (GDB) Fedora 8.2.91.20190424-24.fc30
(gdb) b setns
Breakpoint 1 at 0x95ce4
(gdb) run report >bar
Starting program: /usr/bin/perf report >bar
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".

Breakpoint 1, 0x0000ffffbee68844 in setns () from /lib64/libc.so.6
(gdb) bt
#0  0x0000ffffbee68844 in setns () from /lib64/libc.so.6
#1  0x0000aaaaaac112c4 in nsinfo.mountns_enter ()
#2  0x0000aaaaaabe90b8 in ?? ()
#3  0x0000aaaaaabebfd8 in dso.type ()
#4  0x0000aaaaaac357fc in machine.findnew_vdso ()
#5  0x0000aaaaaac084dc in map.new ()
#6  0x0000aaaaaac065e8 in machine.process_mmap2_event ()
#7  0x0000aaaaaac0cd2c in ?? ()
#8  0x0000aaaaaac0d684 in ?? ()
#9  0x0000aaaaaac0ec90 in perf_session.process_events ()
#10 0x0000aaaaaab579cc in cmd_report ()
#11 0x0000aaaaaabb8fe4 in ?? ()
#12 0x0000aaaaaab42130 in main ()
(gdb) x/5i $pc-4
   0xffffbee68840 <setns>:	mov	x8, #0x10c                 	// #268
=> 0xffffbee68844 <setns+4>:	svc	#0x0
   0xffffbee68848 <setns+8>:	cmn	x0, #0xfff
   0xffffbee6884c <setns+12>:	b.cs	0xffffbee68854 <setns+20>  // b.hs, b.nlast
   0xffffbee68850 <setns+16>:	ret

strace of perf shows:
=====
12549 openat(AT_FDCWD, "/proc/self/ns/mnt", O_RDONLY) = 10
12549 openat(AT_FDCWD, "/proc/424/ns/mnt", O_RDONLY) = 11
12549 setns(11, CLONE_NEWNS)
=====
and similarly for many other /proc/PID.
Comment 6 Michal Prívozník 2019-12-03 09:02:26 UTC 
Created attachment 124292 [details]
0001-Add-support-for-setns-syscall.patch

Based on IRC conversation, I'm attaching the patch I've sent to the list.
Comment 7 Mark Wielaard 2020-04-28 14:20:13 UTC 
*** This bug has been marked as a duplicate of bug 343099 ***