linux-kernel - Re: [PATCH] usb: dwc2: host: fix isoc urb actual length

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]

Date:   Mon, 6 Nov 2017 18:08:05 +0800
From:   wlf <wulf@...k-chips.com>
To:     Minas Harutyunyan <Minas.Harutyunyan@...opsys.com>,
        William Wu <william.wu@...k-chips.com>,
        "John.Youn@...opsys.com" <John.Youn@...opsys.com>,
        "balbi@...nel.org" <balbi@...nel.org>,
        "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>
Cc:     "heiko@...ech.de" <heiko@...ech.de>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
        "linux-rockchip@...ts.infradead.org" 
        <linux-rockchip@...ts.infradead.org>,
        "frank.wang@...k-chips.com" <frank.wang@...k-chips.com>,
        "huangtao@...k-chips.com" <huangtao@...k-chips.com>,
        "dianders@...gle.com" <dianders@...gle.com>,
        "daniel.meng@...k-chips.com" <daniel.meng@...k-chips.com>,
        "fml@...k-chips.com" <fml@...k-chips.com>
Subject: Re: [PATCH] usb: dwc2: host: fix isoc urb actual length

Hi Minas,

在 2017年11月06日 17:28, Minas Harutyunyan 写道:
> Hi,
>
> On 11/6/2017 12:46 PM, William Wu wrote:
>> The actual_length in dwc2_hcd_urb structure is used
>> to indicate the total data length transferred so far,
>> but in dwc2_update_isoc_urb_state(), it just updates
>> the actual_length of isoc frame, and don't update the
>> urb actual_length at the same time, this will cause
>> device drivers working error which depend on the urb
>> actual_length.
>>
>> we can easily find this issue if use an USB camera，
>> the userspace use libusb to get USB data from kernel
>> via devio driver.In usb devio driver, processcompl()
>> function will process urb complete and copy data to
>> userspace depending on urb actual_length.
>>
>> Let's update the urb actual_length if the isoc frame
>> is valid.
>>
>> Signed-off-by: William Wu <william.wu@...k-chips.com>
>> ---
>>    drivers/usb/dwc2/hcd_intr.c | 2 ++
>>    1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/usb/dwc2/hcd_intr.c b/drivers/usb/dwc2/hcd_intr.c
>> index 28a8210..01b1e13 100644
>> --- a/drivers/usb/dwc2/hcd_intr.c
>> +++ b/drivers/usb/dwc2/hcd_intr.c
>> @@ -580,6 +580,7 @@ static enum dwc2_halt_status dwc2_update_isoc_urb_state(
>>    		frame_desc->status = 0;
>>    		frame_desc->actual_length = dwc2_get_actual_xfer_length(hsotg,
>>    					chan, chnum, qtd, halt_status, NULL);
>> +		urb->actual_length += frame_desc->actual_length;
>>    		break;
>>    	case DWC2_HC_XFER_FRAME_OVERRUN:
>>    		urb->error_count++;
>> @@ -599,6 +600,7 @@ static enum dwc2_halt_status dwc2_update_isoc_urb_state(
>>    		frame_desc->status = -EPROTO;
>>    		frame_desc->actual_length = dwc2_get_actual_xfer_length(hsotg,
>>    					chan, chnum, qtd, halt_status, NULL);
>> +		urb->actual_length += frame_desc->actual_length;
>>    
>>    		/* Skip whole frame */
>>    		if (chan->qh->do_split &&
>>
> According urb description in usb.h urb->actual_length used for non-iso
> transfers:
>
> "@actual_length: This is read in non-iso completion functions, and ...
>
>    * ISO transfer status is reported in the status and actual_length fields
>    * of the iso_frame_desc array, ...."
Yes,  urb->actual_length is used for non-iso transfers.  And for ISO 
transfer,  the most
usb class drivers can only use iso frame status and actual_length to 
handle usb data,
like: kernel v4.4 drivers/media/usb/uvc/uvc_video.c

But the usb devio driver really need the urb actual_length in the 
processcompl() if
handle ISO transfer, just as I mentioned in the commit message.

And I also found the same issue on raspberrypi board:
https://github.com/raspberrypi/linux/issues/903

So do you think we need to fix the usb devio driver, rather than fix dwc2?
I think maybe we can use urb actual length for ISO transfer, it seems that
don't cause any side-effect.

>
> Thanks,
> Minas
>
>
>
>
>
>

-- 
吴良峰 William.Wu
福建省福州市铜盘路软件大道89号软件园A区21号楼
No.21 Building, A District, No.89,software Boulevard Fuzhou,Fujian, PRC
手机: 13685012275
座机: 0591-83991906-8520
邮件:wulf@...k-chips.com