Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
From:   John Stultz <john.stultz@linaro.org>
Date:   Tue, 7 May 2019 19:18:36 -0700
Message-ID: <CALAqxLUMRaNxwTUi9QS7-Cy-Ve4+vteBm8-jW4yzZg_QTJVChA@mail.gmail.com>
Subject: [REGRESSION] usb: gadget: f_fs: Allow scatter-gather buffers
To:     Felipe Balbi <balbi@kernel.org>,
        Andrzej Pietrasiewicz <andrzejtp2010@gmail.com>
Cc:     "Yang, Fei" <fei.yang@intel.com>,
        Bjorn Andersson <bjorn.andersson@linaro.org>,
        Chen Yu <chenyu56@huawei.com>,
        lkml <linux-kernel@vger.kernel.org>,
        Linux USB List <linux-usb@vger.kernel.org>,
        Amit Pundir <amit.pundir@linaro.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Since commit 772a7a724f69 ("usb: gadget: f_fs: Allow scatter-gather
buffers"), I've been seeing trouble with adb transfers in Android on
HiKey960, HiKey and now Dragonboard 845c.

Sometimes things crash, but often the transfers just stop w/o any
obvious error messages.

Initially I thought it was an issue with the HiKey960 dwc3 usb patches
being upstreamed, and was using the following hack workaround:
  https://git.linaro.org/people/john.stultz/android-dev.git/commit/?h=dev/hikey960-5.1&id=dcdadaaec9db7a7b78ea9b838dd1453359a2f388

Then dwc2 added sg support, and I ended up having to revert it to get
by on HiKey:
  https://git.linaro.org/people/john.stultz/android-dev.git/commit/?h=dev/hikey-5.1&id=6e91b4c7bd1e94bdd835263403c53e85a677b848

(See thread here: https://lkml.org/lkml/2019/3/8/765)

And now I've reproduced the same issue (with the same dwc3 workaround)
on the already upstream code for Dragonboard 845c.

Fei Yang has also reached out and mentioned he was seeing similar
problems with the f_fs sg support.

Andrzej: Do you have any ideas or suggestions on this? I'm happy to
test or run any debug patches, if it would help narrow the issue down.

If not, should we consider reverting the f_fs sg support?

thanks
-john