[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4c9a5640-9ebf-db87-fcb0-b076b615e4e4@linuxfoundation.org>
Date: Thu, 12 Jan 2023 08:13:52 -0700
From: Shuah Khan <skhan@...uxfoundation.org>
To: Sosthène Guédon <sosthene@...rokey.com>,
Valentina Manea <valentina.manea.m@...il.com>,
Shuah Khan <shuah@...nel.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Cc: Robin Krahl <robin@...rokey.com>,
Shuah Khan <skhan@...uxfoundation.org>
Subject: Re: kernel BUG at mm/usercopy.c when using usbip
On 1/12/23 01:40, Sosthène Guédon wrote:
> Hi!
>
> I have stumbled upon a bug that is triggered reliably by using usbip.
>
> We are using usbip to test our firmware. usbip attach works, but once `opgpcard-tools` interacts with the firmware through pcscd, a kernel bug happens.
> Then usbip stops working, and `lsusb` as well as other tools interacting with usb devices hang.
> The symptoms are similar to https://bugzilla.kernel.org/show_bug.cgi?id=215487 but the kernel bug is not the same (see attached dmesg logs).
>
> The bug can be reproduced on arch (Linux archlinux 6.1.4-arch1-1 #1 SMP PREEMPT_DYNAMIC Sat, 07 Jan 2023 15:10:07 +0000 x86_64 GNU/Linux
> ) and debian (Linux nitropc 5.10.0-19-amd64 #1 SMP Debian 5.10.149-2 (2022-10-21) x86_64 GNU/Linux), though the reproduction is not minimal.
>
> To reproduce the bug on Arch Linux, with the packages rust, pcsclite, and openpgp-card-tools installed:
>
> - Compile and run the usbip runner from this PR: https://github.com/Nitrokey/nitrokey-3-firmware/pull/149 (`cd nitrokey-3-firmware/runners/usbip && cargo run --features alpha`)
> - Run `usbip attach -r localhost -b 1-1`
> - Run `pcscd`
> - Run `opgpcard status`
>
> The bug should happen.
>
Thank you for reporting the problem. I will take a look and
get back to you with possible fix. Will you be able to test
fixes?
thanks,
-- Shuah
Powered by blists - more mailing lists