On 6/21/2017 3:38 AM, Felipe Balbi wrote: > > Hi, > > John Youn <John.Youn@xxxxxxxxxxxx> writes: > > <snip> > >>>> I haven't checked, but Thinh looked and couldn't find anything >>>> relating to the macOS issue. Which is why we want more information >>>> about the problem. >>>> >>>> I can take a look at it too later today when I get a chance. >>>> >>>> When you say you fixed the issue in the vendor driver, do you mean the >>>> macOS problem (multiple clear stalls) or the regression that it caused >>>> when it was fixed (the MSC data sequence issue)? >>> >>> Hi John, >>> >>> The issue was with the host sending a CLEAR_FEATURE(ENDPOINT_HALT) when the >>> EP was not actually halted. The host does this to reset the EP data toggle / >>> sequence number to 0, and it is allowed by some USB spec. The problem was >>> that for DWC3 if you issue a Clear Stall when the EP is not halted, nothing >>> happens. I *think* the fix was to issue a Set EP Config command instead, >>> which does reset the sequence number, but I'm not 100% sure about that. >>> >> >> Hi Paul, >> >> Ok thanks, that will definitely help looking into it. >> >> Though our issue seems the opposite. Since Felipe's macOS patch >> prevents the clear stall when already halted, but *without* that, the > >> controller doesn't clear the sequence number and fails the MSC CV >> test. > > I can see this failure here. We really need a way to reset Data > toggle/seqNum without issuing Clear Stall command. I tried issuing > set_ep_config instead, but there's still something wrong going on. It > seems like g_mass_storage won't requeue a new request for the following > CBW. > > John, can you check on your list of STARS if there's anything related to > issuing Clear Halt when the endpoint is not halted? Hi Felipe, We tried Paul's workaround but it also does not work for us. We're looking into why. We're also investigating other possible workarounds and searching our list of STARs. We'll get back to you once we have some more information. Thanks, John -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
