Yes, that's correct, glad I could be helpful.
On Wed, Jul 13, 2016 at 04:32:14PM -0500, chris annonch wrote: > Okay to clarify and for the archives, > > original VLAN_ID field = 0x1001 > > oxm_mask = 0x1002 > oxm_value = 0x1002 > > result = 0x1003 > > where in 0x*1*00 the *1* is the OFPVID_PRESENT bit > > Thank you for your help! > > > On Wed, Jul 13, 2016 at 4:02 PM, Ben Pfaff <b...@ovn.org> wrote: > > > On Wed, Jul 13, 2016 at 09:16:20AM -0500, chris annonch wrote: > > > Hello all, > > > > > > On page 99 the documentation states that if oxm_hasmask is set to 1 that > > > when setting a field with oxm_value, oxm_mask is used. > > > > > > The documentation further states that for each 1 bit in oxm_mask, the > > > corresponding bit is written in, and for each 0 in oxm_mask the header > > > field is preserved. > > > > > > Does this mean that the if if the vlan field is 01 0000 0000 0000 > > > > > > oxm_mask = 00 0000 0000 0001 > > > > > > oxm_value = 00 0000 0000 0001 > > > > > > that the resulting vlan after processing should be: 01 0000 0000 0001 ? > > > > That would be true for any other field that is modifiable, but the VLAN > > field is special, see this sentence on the same page: > > > > In an OXM_OF_VLAN_VID set-field action, the OFPVID_PRESENT bit must > > be a 1-bit in oxm_value and in oxm_mask. > > > > Thus, a conformant switch will reject the set_field that you describe. > > > > > Or does the original vlan get clobbered? > > > > No. > > > > > If anyone has experience with this I would really appreciate your input. > > > > _______________________________________________ openflow-discuss mailing list openflow-discuss@lists.stanford.edu https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
