linux-kernel - Re: [PATCH v2][RESEND] usb: gadget: core: Fix ->udc_set

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

Message-ID: <bcaa7d8a-c179-0e5b-e91c-2e04d6f7cb2a@ti.com>
Date:   Wed, 1 Nov 2017 17:11:18 +0200
From:   Roger Quadros <rogerq@...com>
To:     Alan Stern <stern@...land.harvard.edu>
CC:     <balbi@...nel.org>, <Dylan.Howey@...nantco.com>,
        <linux-usb@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        "# v4 . 13+" <stable@...r.kernel.org>
Subject: Re: [PATCH v2][RESEND] usb: gadget: core: Fix ->udc_set_speed() speed
 handling

Hi Alan,

On 01/11/17 16:24, Alan Stern wrote:
> On Tue, 31 Oct 2017, Roger Quadros wrote:
> 
>> Currently UDC core calls ->udc_set_speed() with the speed parameter
>> containing the maximum speed supported by the gadget function
>> driver. This might very well be more than that supported by the
>> UDC controller driver.
>>
>> Select the lesser of the 2 speeds so both UDC and gadget function
>> driver are operating within limits.
>>
>> This fixes PHY Erratic errors and 2 second enumeration delay on
>> TI's AM437x platforms.
>>
>> Fixes: 6099eca796ae ("usb: gadget: core: introduce ->udc_set_speed() method")
>> Cc: <stable@...r.kernel.org> # v4.13+
>> Reported-by: Dylan Howey <Dylan.Howey@...nantco.com>
>> Signed-off-by: Roger Quadros <rogerq@...com>
>> ---
>> v2:
>> - Move speed limiting code to UDC core.
>>
>>  drivers/usb/gadget/udc/core.c | 8 ++++++--
>>  1 file changed, 6 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
>> index d41d07a..def1b05 100644
>> --- a/drivers/usb/gadget/udc/core.c
>> +++ b/drivers/usb/gadget/udc/core.c
>> @@ -1080,8 +1080,12 @@ static inline void usb_gadget_udc_stop(struct usb_udc *udc)
>>  static inline void usb_gadget_udc_set_speed(struct usb_udc *udc,
>>  					    enum usb_device_speed speed)
>>  {
>> -	if (udc->gadget->ops->udc_set_speed)
>> -		udc->gadget->ops->udc_set_speed(udc->gadget, speed);
>> +	if (udc->gadget->ops->udc_set_speed) {
>> +		enum usb_device_speed s;
>> +
>> +		s = min(speed, udc->gadget->max_speed);
>> +		udc->gadget->ops->udc_set_speed(udc->gadget, s);
>> +	}
>>  }
>>  
>>  /**
> 
> Roger, would you like to send in a follow-up patch that removes the now 
> unnecessary code from dummy_udc_set_speed() in dummy_hcd.c?
> 

Sure. I have a question though.

In dummy_udc_set_speed() we should just set gadget.speed to whatever is passed in speed.
This will be the maximum supported speed by the gadget function driver.

However, is it possible that the gadget will be connected at a lower speed than that?

I see gadget.speed being updated in dummy_hub_control() GetPortStatus: case, but not
in case of USB_SPEED_HIGH and USB_SPEED_LOW.

-- 
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki