Hi Jianfeng, On Tue, 18 Apr 2017 19:15:51 +0800, "Tan, Jianfeng" <jianfeng....@intel.com> wrote: > On 4/18/2017 4:50 PM, Olivier MATZ wrote: > > Hi Jianfeng, > > > > On Fri, 14 Apr 2017 13:33:49 +0800, "Tan, Jianfeng" > > <jianfeng....@intel.com> wrote: > >> Hi Olivier, > >> > >> If I understand it correctly, this patch is to shift log level setting > >> earlier. But we did not remove the one in eal_parse_common_option(). So > >> we can see this parameter will be analyzed twice. Does it make sense to > >> remove analysis of log level in eal_parse_common_option()? > >> > > The patch does not change the way the log level is parsed: it was > > already parsed twice, because we want to know the log level as soon > > as possible. > > Oh, yes, it's not introduced in this patch. How do you think if we move > log parser out from eal_parse_common_option to avoid parsing log twice?
Not sure it's so easy to do. The second pass can ignore the --log-level argument, but it has to be recognized, else we would have an error. I think it's ok like this. > > > > > But the patch introduces a bug, as seen by Ferruh: the default log > > level is not set properly when no --log-level parameter is passed. > > > > Before this one, there is another one on usage of regexec(), which I put > it here: http://dpdk.org/ml/archives/dev/2017-April/064031.html. Yes, I've seen it, I'll send a patch for this today. Thanks. Olivier
