[dpdk-users] Is this a bug?//Re: pktgen segment fault when running save cmd

[Wiles, Keith]

Regards,
Keith

> On Sep 14, 2016, at 1:34 AM, Caianning <caianning at huawei.com> wrote:
> 
> In pktgen_config_ports():
> 
> /* Start up the ports and display the port Link status */
>         for (pid = 0; pid < pktgen.nb_ports; pid++) {
>                   …..
>                   pktgen_rnd_bits_init(&pktgen.info[pid].rnd_bitfields);
>         }
> 
> And in pktgen_script_save()
> for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
>                   info = &pktgen.info[i];
>                   pkt = &info->seq_pkt[SINGLE_PKT];
>                   range = &info->range;
> 
>                   if (info->tx_burst == 0)
>                            continue;
> 
> 
> (gdb) p info->tx_burst
> $12 = 0
> (gdb) p pktgen.info[31].tx_burst
> $13 = 0
> (gdb) p pktgen.info[30].tx_burst
> $14 = 0
> (gdb) p i
> $15 = 32
> 
> Is this a bug?
> Thx
> 
> 发件人: Caianning
> 发送时间: 2016年9月14日 14:18
> 收件人: users at dpdk.org
> 抄送: Houzhipeng; 'Hu, Xuekun'
> 主题: pktgen segment fault when running save cmd
> 
> Hi
>         I got a segment fault when saving pktgen configurations:

Please test out Pktgen version 3.0.14 for this fix.

> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000000043db17 in pktgen_script_save (path=path at entry=0x7ffc1c1dced0 "/home/vmse/a/pktgen-nsh.cfg") at /home/vmse/a/pktgen-v3.0.09/app/pktgen-cmds.c:420
> 420             if (info->rnd_bitfields->active_specs) {
> (gdb) p info
> $2 = (port_info_t *) 0x8dfe08 <pktgen+280872>
> (gdb) p info->rnd_bitfields
> $3 = (struct rnd_bits_s *) 0x0
> (gdb) p sizeof(*info)
> $5 = 90322
> 
> Version of dpdk is 16.07, and version of pktgen is 3.0.09.
> the cmdline is:
> pktgen -c 0x1ff -n 3 -- -P -m "[1-2:3-4].0,[5-6:7-8].1"
> 
> 
> 
> Thx