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:	Wed, 20 Jul 2016 21:46:11 +0900
From:	Namhyung Kim <namhyung@...nel.org>
To:	Stefan Hajnoczi <stefanha@...il.com>
Cc:	LKML <linux-kernel@...r.kernel.org>,
	Paolo Bonzini <pbonzini@...hat.com>,
	Radim Kr??m???? <rkrcmar@...hat.com>,
	"Michael S. Tsirkin" <mst@...hat.com>,
	Anthony Liguori <aliguori@...zon.com>,
	Anton Vorontsov <anton@...msg.org>,
	Colin Cross <ccross@...roid.com>,
	Kees Cook <keescook@...omium.org>,
	Tony Luck <tony.luck@...el.com>,
	Steven Rostedt <rostedt@...dmis.org>,
	Ingo Molnar <mingo@...nel.org>,
	Minchan Kim <minchan@...nel.org>, kvm@...r.kernel.org,
	qemu-devel@...gnu.org, virtualization@...ts.linux-foundation.org
Subject: Re: [PATCH 2/3] qemu: Implement virtio-pstore device

On Wed, Jul 20, 2016 at 09:29:06AM +0100, Stefan Hajnoczi wrote:
> On Mon, Jul 18, 2016 at 11:21:18PM +0900, Namhyung Kim wrote:
> > On Mon, Jul 18, 2016 at 11:03:53AM +0100, Stefan Hajnoczi wrote:
> > > On Mon, Jul 18, 2016 at 01:37:40PM +0900, Namhyung Kim wrote:
> > > > From: Namhyung Kim <namhyung@...il.com>
> > > > 
> > > > Add virtio pstore device to allow kernel log files saved on the host.
> > > > It will save the log files on the directory given by pstore device
> > > > option.
> > > > 
> > > >   $ qemu-system-x86_64 -device virtio-pstore,directory=dir-xx ...
> > > > 
> > > >   (guest) # echo c > /proc/sysrq-trigger
> > > > 
> > > >   $ ls dir-xx
> > > >   dmesg-0.enc.z  dmesg-1.enc.z
> > > > 
> > > > The log files are usually compressed using zlib.  Users can see the log
> > > > messages directly on the host or on the guest (using pstore filesystem).
> > > 
> > > The implementation is synchronous (i.e. can pause guest code execution),
> > > does not handle write errors, and does not limit the amount of data the
> > > guest can write.  This is sufficient for ad-hoc debugging and usage with
> > > trusted guests.
> > > 
> > > If you want this to be available in environments where the guest isn't
> > > trusted then there must be a limit on how much the guest can write or
> > > some kind of log rotation.
> > 
> > Right.  The synchronous IO is required by the pstore subsystem
> > implementation AFAIK (it uses a single psinfo->buf in the loop).
> 
> The pstore subsystem in Linux may be synchronous but the QEMU device
> emulation does not have to be synchronous.
> 
> Synchronous device emulation means that no other vcpu or QEMU main loop
> processing can occur while device emulation is blocked in a syscall.
> This can make the QEMU monitor unavailable for libvirt and management
> tools.  The guest can experience jitter since vcpus freeze if they
> vmexit while device emulation is blocked (it holds the QEMU global
> mutex and prevents other QEMU threads from making progress).

Thanks for your detailed explanation.  I'll try to change pstore
implementation to deal with async devices.

Thanks,
Namhyung

> 
> You could use include/io.h for asynchronous I/O (qio_channel_add_watch()).
> 
> Stefan


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ