[Devel] nsproxy c/r bug?

Serge E. Hallyn serue at us.ibm.com
Tue May 26 11:04:19 PDT 2009

On a ckpt-v15-dev kernel, if I do:

	git clone git://git.sr71.net/~hallyn/cr_tests.git
	cd cr_tests; make; make install
	mkdir /cgroup
	mount -t cgroup -o freezer cgroup /cgroup
	mkdir /cgroup/1
	cd userns
	sh run_userns.sh
	cd ../fileio
	sh runtests.sh

I get something like:

Unable to handle kernel pointer dereference at virtual kernel address fffffffd132b4000
Oops: 0038 [#4] SMP 
CPU: 0 Tainted: G      D    2.6.30-rc3-00080-g2b1009c #261
Process ns_exec (pid: 3842, task: 000000001fdf9d50, ksp: 0000000012d83d90)
Krnl PSW : 0404100180000000 00000000000ae266 (kmem_cache_alloc+0xa6/0x104)
           R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:1 PM:0 EA:3
Krnl GPRS: 000003ff00000000 0000000000000000 00000000004bf308 000000000000063b
           00000000000ae20e 0000000000422088 0000000000dc5098 0000000000000030
           000000000005f0ea 000000001fc19200 00000000000000d0 0000000012d83d58
           0700000000000400 00000000004075a8 fffffffd132b4090 0000000012d83d58
Krnl Code: 00000000000ae256: a7f4000e		brc	15,ae272
           00000000000ae25a: e31060140016	llgf	%r1,20(%r6)
           00000000000ae260: eb110003000d	sllg	%r1,%r1,3
          >00000000000ae266: e321e0000004	lg	%r2,0(%r1,%r14)
           00000000000ae26c: e32060000024	stg	%r2,0(%r6)
           00000000000ae272: e3c0b0a00024	stg	%r12,160(%r11)
           00000000000ae278: 8000b0a0		ssm	160(%r11)
           00000000000ae27c: a7a18000		tmll	%r10,32768
Call Trace:
([<00000000000ae20e>] kmem_cache_alloc+0x4e/0x104)
 [<000000000005f0ea>] create_new_namespaces+0x4e/0x204
 [<000000000005f4fa>] unshare_nsproxy_namespaces+0x82/0xf8
 [<000000000003c222>] SyS_unshare+0x11a/0x2f4
 [<0000000000026b94>] sysc_do_restart+0x20/0x26
 [<0000004d7db60ce6>] 0x4d7db60ce6
INFO: lockdep is turned off.
Last Breaking-Event-Address:
 [<00000000000ae238>] kmem_cache_alloc+0x78/0x104
 <4>---[ end trace 3165f2ef06617eea ]---

I have not figured what what is going on.  Is there an nsproxy
get/put imbalance that could cause this?

Containers mailing list
Containers at lists.linux-foundation.org

More information about the Devel mailing list