linux-kernel - dwc2: calls disconnect instead of reset on port reset

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Sun, 20 Aug 2023 01:06:38 +0200
From:   Andrey Konovalov <andreyknvl@...il.com>
To:     Minas Harutyunyan <hminas@...opsys.com>
Cc:     Felipe Balbi <balbi@...nel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        USB list <linux-usb@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Alan Stern <stern@...land.harvard.edu>
Subject: dwc2: calls disconnect instead of reset on port reset

Hi Minas,

I have been testing Raw Gadget with the dwc2 UDC driver and stumbled
upon an issue related to how dwc2 behaves during USB port reset.

The problem is that dwc2 calls the gadget driver's ->disconnect()
callback when the port to which the device is connected gets reset by
the host (via SET_FEATURE/PORT_RESET). My expectation was that the
->reset() callback should be called instead.

I grepped the source code of dwc2, and it looks like dwc2 never calls
the ->reset() callback of the gadget driver, not via
usb_gadget_udc_reset() nor via call_gadget(..., reset).

I was wondering if it would be possible to add proper reset handling to dwc2?

This would make it possible to write UDC-agnostic gadgets with Raw
Gadget without adding special handing for this dwc2 behavior.

Thank you!