ipfw pipe show yields "REDZONE: Buffer overflow detected..."

[Andrey V. Elsukov]

On 19.12.2017 16:17, David Wolfskill wrote:
> Similarly; "ipfw sched show" yields:
> Dec 19 13:08:54 g1-252 kernel: REDZONE: Buffer overflow detected. 16 bytes corrupted after 0xfffff80232e73748 (328 bytes allocated).
> Dec 19 13:08:54 g1-252 kernel: Allocation backtrace:
> Dec 19 13:08:54 g1-252 kernel: #0 0xffffffff80d49299 at redzone_setup+0xe9
> Dec 19 13:08:54 g1-252 kernel: #1 0xffffffff80a1175d at malloc+0x22d
> Dec 19 13:08:54 g1-252 kernel: #2 0xffffffff80c95e07 at dummynet_get+0x337
> Dec 19 13:08:54 g1-252 kernel: #3 0xffffffff80ba4102 at rip_ctloutput+0x102
> Dec 19 13:08:54 g1-252 kernel: #4 0xffffffff80ac2d9d at sogetopt+0xcd
> Dec 19 13:08:54 g1-252 kernel: #5 0xffffffff80ac756b at kern_getsockopt+0xdb
> Dec 19 13:08:54 g1-252 kernel: #6 0xffffffff80ac7462 at sys_getsockopt+0x52
> Dec 19 13:08:54 g1-252 kernel: #7 0xffffffff80e3a66a at amd64_syscall+0xa6a
> Dec 19 13:08:54 g1-252 kernel: #8 0xffffffff80e1cedb at Xfast_syscall+0xfb
> Dec 19 13:08:54 g1-252 kernel: Free backtrace:
> Dec 19 13:08:54 g1-252 kernel: #0 0xffffffff80d49604 at redzone_check+0x304
> Dec 19 13:08:54 g1-252 kernel: #1 0xffffffff80a117b6 at free+0x46
> Dec 19 13:08:54 g1-252 kernel: #2 0xffffffff80c9623d at dummynet_get+0x76d
> Dec 19 13:08:54 g1-252 kernel: #3 0xffffffff80ba4102 at rip_ctloutput+0x102
> Dec 19 13:08:54 g1-252 kernel: #4 0xffffffff80ac2d9d at sogetopt+0xcd
> Dec 19 13:08:54 g1-252 kernel: #5 0xffffffff80ac756b at kern_getsockopt+0xdb
> Dec 19 13:08:54 g1-252 kernel: #6 0xffffffff80ac7462 at sys_getsockopt+0x52
> Dec 19 13:08:54 g1-252 kernel: #7 0xffffffff80e3a66a at amd64_syscall+0xa6a
> Dec 19 13:08:54 g1-252 kernel: #8 0xffffffff80e1cedb at Xfast_syscall+0xfb
> 
> 
> Yesterday was the first full day of running the laptop with FQ-CoDel
> implemented, and I noticed that in the evening, things (such as
> echoing charcaters in an xterm) would "pause" for a second or two,
> fairly randomly.  I can't say that the above were associated with the
> pauses, but they are associated with the commands -- this is trivially
> reproducible (for me, anyway).

It would be nice if you created PR where you described steps to
reproduce this. Your kernel/modules config, commands you used to get
this result.

-- 
WBR, Andrey V. Elsukov

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 553 bytes
Desc: OpenPGP digital signature
URL: <http://lists.freebsd.org/pipermail/freebsd-ipfw/attachments/20171220/9c9754fb/attachment.sig>