Re: kernel crash in bpf_jit on x86_64 when running nmap

[Alexei Starovoitov]

On Fri, Oct 10, 2014 at 1:44 PM, Darrick J. Wong
<darrick.wong@xxxxxxxxxx> wrote:
> Hi everyone,
>
> I was running nmap on a x86_64 qemu guest and experienced the following crash:
>
> # nmap -sS -O -vvv 192.168.122.1
> Starting Nmap 6.40 ( http://nmap.org ) at 2014-10-10 13:14 PDT
> Initiating ARP Ping Scan at 13:14
> Scanning 192.168.122.1 [1 port]
> <kaboom>
>
> dmesg output is as follows (I set net.core.bpf_jit_enable=2 the second time):
>
> [   32.376291] flen=3 proglen=82 pass=0 image=ffffffffc01ac65b
> [   32.377595] JIT code: 00000000: 55 48 89 e5 48 81 ec 28 02 00 00 48 89 9d d8 fd
> [   32.379243] JIT code: 00000010: ff ff 4c 89 ad e0 fd ff ff 4c 89 b5 e8 fd ff ff
> [   32.380984] JIT code: 00000020: 4c 89 bd f0 fd ff ff 31 c0 4d 31 ed 48 89 fb b8
> [   32.382606] JIT code: 00000030: 00 00 00 00 48 8b 9d d8 fd ff ff 4c 8b ad e0 fd
> [   32.384280] JIT code: 00000040: ff ff 4c 8b b5 e8 fd ff ff 4c 8b bd f0 fd ff ff
> [   32.385911] JIT code: 00000050: c9 c3
> [   32.386841] bpf_jit: proglen=265 != oldproglen=269

thanks for the report. The line above indicates that JIT tried to emit
4 byte longer program than during pre-final pass.
Will look at it asap.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/