On Tue, Feb 3, 2015 at 8:40 AM, Eric Blake <ebl...@redhat.com> wrote: > On 01/22/2015 01:03 AM, sfel...@gmail.com wrote: >> From: Scott Feldman <sfel...@gmail.com> >> >> This is the register programming guide for the Rocker device. It's intended >> for driver writers and device writers. It covers the device's PCI space, >> the register set, DMA interface, and interrupts. >> > > In addition to typos already pointed out by Stefan, > >> + >> +Writing BASE_ADDR or SIZE will reset HEAD and TAIL to zero. HEAD cannot be >> +written passed TAIL. To do so would wrap the ring. An empty ring is when >> HEAD > > s/passed/past/
Got all the corrections, thanks guys. > >> + >> +To support forward- and backward-compatibility, descriptor and completion >> +payloads are specified in TLV format. Fields are packed with Type=field >> name, >> +Length=field length, and Value=field value. Software will ignore unknown >> fields >> +filled in by the switch. Likewise, the switch will ignore unknown fields >> +filled in by software. > > Is ignoring unknown fields always the wisest action? If the unknown > fields are supposed to have an impact according the to writer, but get > ignored by the reader, then the two can get out of sync with what they > assume the other end is doing. This is the beauty of using TLVs for the cmd structures. The driver and device can be version mismatched and still function at lowest-common set. Care must be taken when adding new options to a TLV enum list by adding new items to the end. And if an existing structure can't be extended, then a new one is created, but the old one must remain for backward-compatibility. If it's important for one side to know the level of support of the other side, then option A can be tried, and if that fails, try option B, and so on.
