Re: [PATCH v4] USB: hub: Fix the broken detection of USB3 device in SMSC hub

From: Eugeniu Rosca <erosca@de.adit-jv.com>
To: Kai-Heng Feng <kai.heng.feng@canonical.com>
Cc: Hardik Gajjar <hgajjar@de.adit-jv.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	<stern@rowland.harvard.edu>, <thinhn@synopsys.com>,
	<Kento.A.Kobayashi@sony.com>, <atmgnd@outlook.com>,
	<linux-usb@vger.kernel.org>, <andrew_gabbasov@mentor.com>,
	<erosca@de.adit-jv.com>, <linux-renesas-soc@vger.kernel.org>,
	Eugeniu Rosca <roscaeugeniu@gmail.com>
Subject: Re: [PATCH v4] USB: hub: Fix the broken detection of USB3 device in SMSC hub
Date: Wed, 13 May 2020 23:42:07 +0200	[thread overview]
Message-ID: <20200513214207.GA24855@lxhi-065.adit-jv.com> (raw)
In-Reply-To: <73933975-6F0E-40F5-9584-D2B8F615C0F3@canonical.com>

Hi Kai-Heng,

Many thanks for reporting the issue!

On Tue, May 12, 2020 at 09:36:07PM +0800, Kai-Heng Feng wrote:
> This patch prevents my Raven Ridge xHCI from getting runtime suspend.
> 
> > On Feb 6, 2020, at 19:49, Hardik Gajjar <hgajjar@de.adit-jv.com> wrote:
> > 
> > Renesas R-Car H3ULCB + Kingfisher Infotainment Board is either not able
> > to detect the USB3.0 mass storage devices or is detecting those as
> > USB2.0 high speed devices.
> > 
> > The explanation given by Renesas is that, due to a HW issue, the XHCI
> > driver does not wake up after going to sleep on connecting a USB3.0
> > device.
> 
> Since the issue is already root caused to xHCI, sounds the workaround should be implemented in xHCI?
> 
> Functions like xhci_alloc_dev() can be a better place to instrument the workaround.

To my understanding, based on the USB Vendor ID used in this patch and
on the lsusb output [1] got on the original arm64 board, we are talking
about a hub device [2], hence drivers/usb/core/hub.c seems an
appropriate placement.

> > +#define USB_VENDOR_SMSC				0x0424

Based on the output [1], I believe the quirk could be made specific
to USB Product IDs '2134' and '5534'?

Could you please share the output of 'lsusb | grep 0424' on the machine
you experienced the regression?

Question to both USB and Renesas people. Does anybody expect SMSC hub
Product ID to vary across different Kingfisher [3] board samples?

> > static const struct usb_device_id hub_id_table[] = {
> > +    { .match_flags = USB_DEVICE_ID_MATCH_VENDOR | USB_DEVICE_ID_MATCH_INT_CLASS,
> > +      .idVendor = USB_VENDOR_SMSC,
> > +      .bInterfaceClass = USB_CLASS_HUB,
> > +      .driver_info = HUB_QUIRK_DISABLE_AUTOSUSPEND},

[1] h3ulcb-kf #> lsusb | grep 0424
    Bus 006 Device 002: ID 0424:5534 Standard Microsystems Corp. Hub
    Bus 005 Device 002: ID 0424:2134 Standard Microsystems Corp. Hub
[2] https://devicehunt.com/search/type/usb/vendor/0424/device/any
[3] https://elinux.org/R-Car/Boards/Kingfisher

-- 
Best regards,
Eugeniu Rosca