linux-kernel - Re: [PATCH] usb: gadget: f_mass_storage: Disable eps during disconnect

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] [day] [month] [year] [list]

Date:   Tue, 26 Oct 2021 18:18:37 -0700
From:   Wesley Cheng <quic_wcheng@...cinc.com>
To:     Alan Stern <stern@...land.harvard.edu>
CC:     <balbi@...nel.org>, <gregkh@...uxfoundation.org>,
        <linux-usb@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        Wesley Cheng <wcheng@...eaurora.org>
Subject: Re: [PATCH] usb: gadget: f_mass_storage: Disable eps during
 disconnect

Hi Alan,

On 10/26/2021 1:57 PM, Alan Stern wrote:
> On Tue, Oct 26, 2021 at 12:40:04PM -0700, Wesley Cheng wrote:
>> Hi Alan,
>>
>> On 10/26/2021 7:41 AM, Alan Stern wrote:
>>> On Mon, Oct 25, 2021 at 05:44:56PM -0700, Wesley Cheng wrote:
>>>> From: Wesley Cheng <wcheng@...eaurora.org>
>>>>
>>>> When receiving a disconnect event from the UDC, the mass storage
>>>> function driver currently runs the handle_exception() routine
>>>> asynchronously.  For UDCs that support runtime PM, there is a
>>>> possibility the UDC is already suspended by the time the
>>>> do_set_interface() is executed.  This can lead to HW register access
>>>> while the UDC is already suspended.
>>>>
>>>> Signed-off-by: Wesley Cheng <wcheng@...eaurora.org>
>>>> ---
>>>>  drivers/usb/gadget/function/f_mass_storage.c | 10 ++++++++++
>>>>  1 file changed, 10 insertions(+)
>>>>
>>>> diff --git a/drivers/usb/gadget/function/f_mass_storage.c b/drivers/usb/gadget/function/f_mass_storage.c
>>>> index 3cabf7692ee1..752439690fda 100644
>>>> --- a/drivers/usb/gadget/function/f_mass_storage.c
>>>> +++ b/drivers/usb/gadget/function/f_mass_storage.c
>>>> @@ -2342,6 +2342,16 @@ static void fsg_disable(struct usb_function *f)
>>>>  {
>>>>  	struct fsg_dev *fsg = fsg_from_func(f);
>>>>  
>>>> +	/* Disable the endpoints */
>>>> +	if (fsg->bulk_in_enabled) {
>>>> +		usb_ep_disable(fsg->bulk_in);
>>>
>>> According to the kerneldoc, this routine must be called in process 
>>> context.
>>>
>>>> +		fsg->bulk_in_enabled = 0;
>>>> +	}
>>>> +	if (fsg->bulk_out_enabled) {
>>>> +		usb_ep_disable(fsg->bulk_out);
>>>> +		fsg->bulk_out_enabled = 0;
>>>> +	}
>>>> +
>>>>  	__raise_exception(fsg->common, FSG_STATE_CONFIG_CHANGE, NULL);
>>>>  }
>>>
>>> Looks like you'll have to find a different way to avoid the problem.  
>>> For example, if an exception is pending then you might prevent the 
>>> gadget from going into runtime suspend until the exception has been 
>>> handled.
>>>
>> Thanks for the suggestion.  I noticed that a lot of other FDs currently
>> call usb_ep_disable() in the disable/disconnect path as well.  Actually,
>> f_mass_storage seems to be the only one that doesn't do so.  Maybe we
>> should change the kerneldoc :)
> 
> You're right; I don't see any real reason why the usb_ep_{en,dis}able 
> routines can't run in an atomic context.  So if you to make this change 
> to f_mass_storage, you should first submit a patch changing the 
> kerneldoc.
> 
Sure, will do that and include it in the next rev.

Thanks
Wesley Cheng