Hi, Ferry Toth <fntoth@xxxxxxxxx> writes: >> I don't suppose you have a USB sniffer, do you? > No, I don't have one of those. >> Andy, could you see if you can reproduce the same behavior with a >> sniffer? It really seems like this either caused by the host side or >> something outside dwc3. Perhaps the PHY is going bonkers? >> >> cheers > > But got other inspiration: like you say dwc3 is behaving normally, so > problem might be elsewhere right? > > Now, to make gadget working I needed to apply the experiment from Jack Pham: > > https://github.com/edison-fw/linux/commit/da90cb63fc8f0ce18805f262d4d6b5b4d8c11d5c > > But now Jack has already sent in "[PATCH v3] usb: gadget: audio: Free > requests only after callback" > > So, removing the experiment and adding PATCH 3, this issue appears to go > away. > > And indeed this fixes a problem in u_audio.c, not in dwc3. > > Thanks all, I'm looking forward to the patch to land in 5.10-rc3/4. Awesome!!! Do you want to reply to Jack's patch with a Tested-by? -- balbi