On Thu, Apr 11, 2024 at 5:52 AM Thinh Nguyen <Thinh.Nguyen@xxxxxxxxxxxx> wrote: > > > On Wed, Apr 10, 2024, Manan Aurora wrote: > > Hi Thinh, > > > > I'm working on a patch to bring EBC support back, I had a doubt > > regarding some of the required corrections though (inlined) > > > > Please take a look and advise, I'll proceed accordingly. > > > > > > > On Fri, Feb 9, 2024 at 6:55 AM Thinh Nguyen <Thinh.Nguyen@xxxxxxxxxxxx> wrote: > > > > > > This reverts commit 398aa9a7e77cf23c2a6f882ddd3dcd96f21771dc. > > > > > > The update to the gadget API to support EBC feature is incomplete. It's > > > missing at least the following: > > > * New usage documentation > > I will address this > > > * Gadget capability check > > > * Condition for the user to check how many and which endpoints can be > > > used as "fifo_mode" > > > The easiest option seems to be to add a new function that lets users > > specifically request > > fifo_mode endpoints eg: usb_fifo_mode_ep_autoconfig_ss > > This function will cover ensuring that the device supports > > fifo_endpoints and returning a suitable > > endpoint (if available) and NULL otherwise. This can be indicated by > > adding a new bit to > > the existing ep_caps structure. > > Does this seem like an acceptable solution? > > That sounds fine to me. For the naming, perhaps just name it as EBC for > External Buffer Control and provide proper description in documentation. > "fifo_mode" may not be clear. > > Maybe usb_ep_autoconfig_ss_with_ebc()? > > How are you planning to implement the above function? Are you going to > apply the DWC_usb3x specific requirements directly or implement > gadget_ops->match_ep() to properly return the right endpoint base on the > endpoint desc? As you're aware, EBC is only applicable to non-streams > bulk only. Also it doesn't apply to full-speed. The issue with implementing match_ep is that the API doesn't let us specify that an EBC endpoint is desired. What about adding a new function to usb_gadget_ops? Then we could apply IP-specific restrictions in one place. Speed can be enforced when we attempt to configure the EP (config_ep_by_speed etc) > > > > > > * Description of how it can affect completed request (e.g. dwc3 won't > > > update TRB on completion -- ie. how it can affect request's actual > > > length report) > > I will remove the NO_WB bit for the EBC endpoint and leave it up to > > the user to enable/disable this > > Thanks, > Thinh
