linux-kernel - Re: [PATCH] USB: HCD: Fix URB giveback issue in tasklet function

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: <154b2167-47f1-4d1e-0cd7-afab25033b76@zhaoxin.com>
Date:   Thu, 21 Jul 2022 16:52:11 +0800
From:   "WeitaoWang-oc@...oxin.com" <WeitaoWang-oc@...oxin.com>
To:     Oliver Neukum <oneukum@...e.com>, <stern@...land.harvard.edu>,
        <gregkh@...uxfoundation.org>, <kishon@...com>,
        <dianders@...omium.org>, <s.shtylyov@....ru>, <mka@...omium.org>,
        <ming.lei@...onical.com>, <linux-usb@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>
CC:     <tonywwang@...oxin.com>, <weitaowang@...oxin.com>,
        <CobeChen@...oxin.com>, <TimGuo@...oxin.com>
Subject: Re: [PATCH] USB: HCD: Fix URB giveback issue in tasklet function

On 2022/7/21 16:00, Oliver Neukum wrote:
> 
> 
> On 21.07.22 08:08, Weitao Wang wrote:
>> Usb core introduce the mechanism of giveback of URB in tasklet context to
>> reduce hardware interrupt handling time. On some test situation(such as
>> FIO with 4KB block size), when tasklet callback function called to
>> giveback URB, interrupt handler add URB node to the bh->head list also.
>> If check bh->head list again after finish all URB giveback of local_list,
>> then it may introduce a "dynamic balance" between giveback URB and add URB
>> to bh->head list. This tasklet callback function may not exit for a long
>> time, which will cause other tasklet function calls to be delayed. Some
>> real-time applications(such as KB and Mouse) will see noticeable lag.
>>
> 
> Hi,
> 
> ow do you know usb_hcd_giveback_urb() will be called in time to process
> isoc URBs in time, if you leave them on the list? In fact how do
> you be sure it will be called at all? I can see no upper time limit
> on that.
> 
> 	Regards
> 		Oliver
> 
> .
If the bh->head list is not empty, patch method will raise a softirq
immediately to call work function of every tasklet. Therefore, I think
the left URB in bh->head list will be get giveback in time.

Thanks
weitao