Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <b1d1c907-406a-57bd-daf5-4366f8714cad@google.com>
Date:   Tue, 30 May 2023 14:50:52 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.11.0
Subject: Re: [PATCH v4 1/2] usb: gadget: udc: core: Invoke usb_gadget_connect
 only when started
To:     Badhri Jagan Sridharan <badhri@google.com>
Cc:     Francesco Dolcini <francesco@dolcini.it>,
        Alistair <alistair@alistair23.me>, gregkh@linuxfoundation.org,
        stern@rowland.harvard.edu, colin.i.king@gmail.com,
        xuetao09@huawei.com, quic_eserrao@quicinc.com,
        water.zhangjiantao@huawei.com, peter.chen@freescale.com,
        balbi@ti.com, linux-usb@vger.kernel.org,
        linux-kernel@vger.kernel.org, stable@vger.kernel.org,
        Linux regressions mailing list <regressions@lists.linux.dev>
References: <20230407030741.3163220-1-badhri@google.com>
 <0cf8c588b701d7cf25ffe1a9217b81716e6a5c51.camel@alistair23.me>
 <1ac16f0a-3cca-40ca-c444-82719f85a24c@leemhuis.info>
 <ZGStr1oZvmJ0XzSu@francesco-nb.int.toradex.com>
 <53213ba8-be58-0f97-8a06-3c9380bdff53@leemhuis.info>
 <CAPTae5Lz=efNKhCnoy6Hp78HFqCLqxtWdZQ7zSDx2omeYf2KDw@mail.gmail.com>
Content-Language: en-US
From:   Avichal Rakesh <arakesh@google.com>
In-Reply-To: <CAPTae5Lz=efNKhCnoy6Hp78HFqCLqxtWdZQ7zSDx2omeYf2KDw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 5/17/23 10:45, Badhri Jagan Sridharan wrote:
> Keeping the thread updated. I sent out
> https://www.spinics.net/lists/kernel/msg4792009.html few hours earlier
> and have requested help from Francesco, Alistair and others who
> reported the issue.
> Discussing with Alan stern on the feedback he had left.
> 

Tangential to the original issues: it looks like patch 2/2 also
breaks gadget functions that sets `bind_deactivated` to true.

When usb_gadget_connect is called while the gadget is deactivated from
functions binding, it sets gadget->connected to true, but does not call
the pullup function. Later, when the gadget function calls
usb_gadget_activate which in turn calls usb_gadget_connect, the pullup
function isn't called because gadget->connected is true.

- Avi