> -----Original Message----- > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Xueming(Steven) Li > Sent: Thursday, December 7, 2017 11:05 PM > To: Olivier MATZ <olivier.m...@6wind.com> > Cc: dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH] lib/cmdline: init parse result memeory > > Hi Olivier, > > > -----Original Message----- > > From: Olivier MATZ [mailto:olivier.m...@6wind.com] > > Sent: Thursday, December 7, 2017 10:48 PM > > To: Xueming(Steven) Li <xuemi...@mellanox.com> > > Cc: dev@dpdk.org > > Subject: Re: [PATCH] lib/cmdline: init parse result memeory > > > > On Wed, Nov 15, 2017 at 11:54:02PM +0800, Xueming Li wrote: > > > Initialize binary result memory before parsing to avoid garbage in > > > parsing result. > > > > > > Signed-off-by: Xueming Li <xuemi...@mellanox.com> > > > --- > > > lib/librte_cmdline/cmdline_parse.c | 2 ++ > > > 1 file changed, 2 insertions(+) > > > > > > diff --git a/lib/librte_cmdline/cmdline_parse.c > > > b/lib/librte_cmdline/cmdline_parse.c > > > index 3e12ee54f..9124758f1 100644 > > > --- a/lib/librte_cmdline/cmdline_parse.c > > > +++ b/lib/librte_cmdline/cmdline_parse.c > > > @@ -267,6 +267,8 @@ cmdline_parse(struct cmdline *cl, const char * buf) > > > if (!cl || !buf) > > > return CMDLINE_PARSE_BAD_ARGS; > > > > > > + memset(tmp_result.buf, 0, sizeof(tmp_result.buf)); > > > + > > > ctx = cl->ctx; > > > > > > /* > > > > > > Did you see an issue (a bug or a crash) without the memset()? > > Or is it to avoid filling unused fields in the parsed struct? > Yes, I'm using same struct for some similar commands, have to avoid > filling unused fields. > > > > > I'm not sure if your patch is enough: cmdline_parse() calls > > match_inst() for each registered command. If a command partially > > matches (only the first tokens), the buffer is modified. So the next > > one will start with a dirty buffer. > > > > I suggest to put the memset() in match_inst() instead. Something like > this: > > > > if (resbuf != NULL) > > memset(resbuf, 0, resbuf_size); > > > > > > It will reset the buffer before using it. > It was there for performance concern, since the buffer could be tainted, > I'll upload a new version according to your suggestion, appreciate your > suggestion. Rte_flow CLI seems to be relying on modified buffer, add Adrien: testpmd> flow create 0 ingress pattern eth / ipv4 / udp / vxlan / end actions rss queues 1 2 end level 1 / mark id 0x123456 / end Caught error type 2 (flow rule (handle)): no valid action
> > > > > Olivier
