Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   ebiederm@xmission.com (Eric W. Biederman)
To:     Alan Stern <stern@rowland.harvard.edu>
Cc:     linux-usb@vger.kernel.org,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Oliver Neukum <oneukum@suse.com>,
        <linux-kernel@vger.kernel.org>
References: <Pine.LNX.4.44L0.1905181116330.7855-100000@netrider.rowland.org>
Date:   Sat, 18 May 2019 10:25:05 -0500
In-Reply-To: <Pine.LNX.4.44L0.1905181116330.7855-100000@netrider.rowland.org>
        (Alan Stern's message of "Sat, 18 May 2019 11:20:47 -0400 (EDT)")
Message-ID: <878sv3ss32.fsf@xmission.com>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/25.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Subject: Re: [CFT][PATCH] signal/usb: Replace kill_pid_info_as_cred with kill_pid_usb_asyncio
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Alan Stern <stern@rowland.harvard.edu> writes:

> On Fri, 17 May 2019, Eric W. Biederman wrote:
>
>> Wow I got a little distracted but now I am back to this.
>> 
>> Using your test program I was able to test the basics of this.
>> 
>> I found one bug in my patch where I was missing a memset.  So I have
>> corrected that, and reorganized the patch a little bit.
>> 
>> I have not figured out how to trigger a usb disconnect so I have not
>> tested that.
>
> Heh.  Assuming the device file you tell the test program to use 
> corresponds to an actual USB device, you can trigger a disconnect by 
> literally unplugging the USB cable.  (Add a 10-second delay to the 
> program to give yourself enough time.)

I have just been running this in qemu.  But yes.  I suppose the easy
way would be to print a message asking the usb device to be unplugged
and then just wait for the signal.  I might try that.

>> The big thing I have not been able to test is running a 64bit big-endian
>> kernel with a 32bit user space.  My modified version of your test
>> program should report "Bad" without my patch, and should report "Good"
>> with it.
>> 
>> Is there any chance you can test that configuration?  I could not figure
>> out how to get a 64bit big-endian system running in qemu, and I don't
>> have the necessary hardware so I was not able to test that at all.  As
>> that is the actual bug I am still hoping someone can test it.
>
> Unfortunately, I don't have any big-endian systems either.

That probably explains why the breakage in big-endian was never noticed.
I am starting to wonder if anyone is actually doing big-endian for new
systems anymore.

Eric