segfault when running busybox dmesg

[Leif Lindholm]

Linaro 12.12.

When running (busybox) dmesg on the foundation model, it first prints the kernel log, and then it is terminated on a segfault:
---
[63368.251146] dmesg[1474]: unhandled level 1 translation fault (11) at 0x100415faf, esr 0x92000045
[63368.251219] pgd = ffffffc87fe05000
[63368.251275] [100415faf] *pgd=0000000000000000
[63368.251347]
[63368.251399] Pid: 1474, comm: dmesg
[63368.251475] CPU: 0 Not tainted (3.6.0-1-linaro-vexpress64 #1~ci+121213151217)
[63368.251545] PC is at 0x401d1c
[63368.251602] LR is at 0x401e74
[63368.251669] pc : [<0000000000401d1c>] lr : [<0000000000401e74>] pstate: 20000000
[63368.251736] sp : 0000007ff16e6410
[63368.251791] x29: 0000007ff16e6410 x28: 0000007ff16e6528
[63368.251908] x27: 0000000000404df0 x26: 0000007f92a5f710
[63368.252024] x25: 0000000000417f38 x24: 0000000000415fb0
[63368.252140] x23: 0000000000415f38 x22: 0000000000415fc4
[63368.252256] x21: 0000000000415000 x20: 0000000000415fc4
[63368.252372] x19: 0000007ff16e64f0 x18: 0000007ff16e6130
[63368.252488] x17: 0000007f929b5560 x16: 0000000000000000
[63368.252604] x15: 0000007f92a69028 x14: 0000000000002710
[63368.252720] x13: 0000007f9290e3e8 x12: ffffffffffffffff
[63368.252836] x11: 0000000000000005 x10: 0000000000000000
[63368.252950] x9 : 0000000000000000 x8 : 000000000000003f
[63368.253066] x7 : ffffffffffffffc4 x6 : fffffffffffffdfe
[63368.253182] x5 : 000000000000003b x4 : 0000000000415000
[63368.253296] x3 : 0000000000000000 x2 : 0000007f92a5f6f0
[63368.253412] x1 : 0000000100415f37 x0 : 00000000ffffffff
[63368.253520]
---

[Marcin Juszkiewicz (hrw)]

Works fine for me with OE kernel.

[Leif Lindholm (leif-lindholm)]

That was with the (12.12) OE kernel.

However, this behaviour is gone with build #132 filesystem/kernel, which uses glibc 2.17. Feel free to close.

[Dan Aloni (alonid)]

This is very similar to what I'm getting with the newer rootfs of vexpress64-openembedded_sdk-armv8_20130127-242, combined with the kernel: Linux genericarmv8 3.8.0-1-linaro-vexpress64 #1ubuntu1~ci+130127041142 SMP Sun Jan 27 04:15:58 UTC 2013 aarch64 GNU/Linux.

When I login with sshd:

[ 1494.369684] sshd[861]: unhandled level 2 translation fault (11) at 0x00000000, esr 0x92000006
[ 1494.369763] pgd = ffffffc87e450000
[ 1494.369826] [00000000] *pgd=00000008fe7d7003, *pmd=0000000000000000

[ 1494.369994] Pid: 861, comm: sshd
[ 1494.370079] CPU: 0 Not tainted (3.8.0-1-linaro-vexpress64 #1ubuntu1~ci+130127041142)
[ 1494.370160] PC is at 0x7fb7cf345c
[ 1494.370224] LR is at 0x40bf90
[ 1494.370299] pc : [<0000007fb7cf345c>] lr : [<000000000040bf90>] pstate: a0000000
[ 1494.370373] sp : 0000007ffffff430
[ 1494.370434] x29: 0000007ffffff430 x28: 000000000044b000
[ 1494.370554] x27: 000000000046bf14 x26: 0000000000448928
[ 1494.370673] x25: 0000007ffffff5c0 x24: 000000000047e248
[ 1494.370791] x23: 0000000000000006 x22: 0000000000000001
[ 1494.370909] x21: 0000000000000001 x20: 000000000047e660
[ 1494.371028] x19: 000000000044b468 x18: 0000000000000018
[ 1494.371146] x17: 0000007fb7cf3458 x16: 000000000046a9f8
[ 1494.371264] x15: 0000000000000fc0 x14: 00000000000000c0
[ 1494.371383] x13: 0000007fb7dc2e50 x12: 0000000000000000
[ 1494.371500] x11: 0000000000000040 x10: 00000000fffffff8
[ 1494.371618] x9 : 0000000000000000 x8 : 0000000000000000
[ 1494.371735] x7 : 0000000000000fff x6 : 0000000000000000
[ 1494.371853] x5 : 0000000002800010 x4 : 0000000042800520
[ 1494.371972] x3 : 4280052002800010 x2 : 0000000000000000
[ 1494.372090] x1 : 000000000044b468 x0 : 0000000000000000

A bit of analysis shows the place in libc where it faults:

7fb7c79000-7fb7da4000 r-xp 00000000 fe:02 487 /lib/libc-2.17.so
7fb7da4000-7fb7db3000 ---p 0012b000 fe:02 487 /lib/libc-2.17.so
7fb7db3000-7fb7db7000 r-xp 0012a000 fe:02 487 /lib/libc-2.17.so
7fb7db7000-7fb7db9000 rwxp 0012e000 fe:02 487 /lib/libc-2.17.so

000000000007a458 <strcmp>:
   7a458: d2800002 movz x2, #0x0
 ** 7a45c: 38626803 ldrb w3, [x0,x2] **
   7a460: 38626824 ldrb w4, [x1,x2]
   7a464: 340000c3 cbz w3, 7a47c <strcmp+0x24>
   7a468: 6b04007f cmp w3, w4
   7a46c: 91000442 add x2, x2, #0x1
   7a470: 54ffff60 b.eq 7a45c <strcmp+0x4>
   7a474: 4b040060 sub w0, w3, w4
   7a478: d65f03c0 ret
   7a47c: 4b0403e0 neg w0, w4
   7a480: d65f03c0 ret

Caller inside sshd (according to LR):
  40bf84: 9400e1b4 bl 444654 <setlogin+0x200>
  40bf88: aa1303e1 mov x1, x19
  40bf8c: 97ffedf9 bl 407770 <strcmp@plt>
  40bf90: 6b1f001f cmp w0, wzr
  40bf94: 1a9f17e0 cset w0, eq

Any takers?