> -----Original Message----- > From: Ilya Maximets [mailto:i.maxim...@samsung.com] > Sent: Friday, November 09, 2018 12:43 PM > To: Ophir Munk <ophi...@mellanox.com>; Wenzhuo Lu > <wenzhuo...@intel.com>; Jingjing Wu <jingjing...@intel.com>; Bernard > Iremonger <bernard.iremon...@intel.com>; dev@dpdk.org; Adrien > Mazarguil <adrien.mazarg...@6wind.com> > Cc: Asaf Penso <as...@mellanox.com>; Shahaf Shuler > <shah...@mellanox.com>; Thomas Monjalon <tho...@monjalon.net>; > Olga Shern <ol...@mellanox.com>; sta...@dpdk.org; Ferruh Yigit > <ferruh.yi...@intel.com> > Subject: Re: [v2] app/testpmd: revert setting default RSS > > On 09.11.2018 12:38, Ophir Munk wrote: > > This reverts the patch that enables default RSS action by setting > > key=NULL and key_len=0. > > In current testpmd implementation a key pointer must exist if > > key_len!=0. For example, the following flow rule will cause a > > segmentation fault: > > flow create 0 <pattern> actions rss queues 0 1 end key_len 40 / end > > Maybe it's better to check that 'key_len' and 'key' passed both or none?
I agree. However I don't see this option easily added to current testpmd flow implementation. Adrien - how would you recommend adding this check? Please note that currently if no key and no key_len are specified - testpmd still assign a dummy string. > > BTW, is there any profit from the 'key_len' argument for testpmd? > Can we just always use the size of the passed 'key' and drop the configurable > from the user interface? > If you just specify key without key_len then the key length is calculated implicitly from the key itself. So this is an already implemented feature. You can still use key_len (with different values) maybe for special case handling / debugging in the PMD. > Best regards, Ilya Maximets.