Date: Fri, 29 Apr 2016 08:15:40 -0700
From: Greg KH <gregkh@linuxfoundation.org>
To: "Richard W.M. Jones" <rjones@redhat.com>
Cc: "Matwey V. Kornilov" <matwey@sai.msu.ru>, Jiri Slaby <jslaby@suse.com>,
        Peter Hurley <peter@hurleysoftware.com>,
        andriy.shevchenko@linux.intel.com, phillip.raffeck@fau.de,
        anton.wuerfel@fau.de, yamada.masahiro@socionext.com,
        valentinrothberg@gmail.com, linux-serial@vger.kernel.org,
        linux-kernel <linux-kernel@vger.kernel.org>, ddutile@redhat.com
Subject: Re: [PATCH] 8250: Hypervisors always export working 16550A UARTs.
Message-ID: <20160429151540.GA16895@kroah.com>
References: <1461881913-23967-1-git-send-email-rjones@redhat.com>
 <CAJs94EYUs2Gx9z9YmK2htXnOQSE4EobNdbnDCF6W3AebsXmQ+Q@mail.gmail.com>
 <20160429074114.GC3826@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20160429074114.GC3826@redhat.com>
User-Agent: Mutt/1.6.0 (2016-04-01)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1661
Lines: 35

On Fri, Apr 29, 2016 at 08:41:14AM +0100, Richard W.M. Jones wrote:
> On Fri, Apr 29, 2016 at 10:01:08AM +0300, Matwey V. Kornilov wrote:
> > 2016-04-29 1:18 GMT+03:00 Richard W.M. Jones <rjones@redhat.com>:
> > > [This is an opinionated patch, mainly for discussion.]
> > >
> > > I'm trying to reduce the time taken in the kernel in initcalls, with
> > > my aim being to reduce the current ~700ms spent in initcalls before
> > > userspace, down to something like 100ms.  All times on my Broadwell-U
> > > laptop, under virtualization.  The purpose of this is to be able to
> > > launch VMs around containers with minimal overhead, like Intel Clear
> > > Containers, but using standard distro kernels and qemu.
> > >
> > > Currently the kernel spends 25ms inspecting the UART that we passed to
> > > it from qemu to find out whether it's an 8250/16550/16550A perhaps
> > > with a non-working FIFO or other quirks.  Well, it isn't -- it's a
> > > working emulated 16550A, with a FIFO and no quirks, and if it isn't,
> > > we should fix qemu.
> > >
> > > So the patch detects if we're running virtualized (perhaps it should
> > > only check for qemu/KVM?) and if so, shortcuts the tests.
> > 
> > Does anybody know, whether it is possible to pass through real
> > hardware serial port to a guest? It seems to be as simple as to pass
> > through an interrupt and memory IO ports.
> 
> In theory it seems like something VFIO could do.  Passing something as
> low performance as a serial port through would seem to make little
> sense though.

Multi-port serial PCI cards are not "low performance" and might be
passed through directly to a guest.

thanks,

greg k-h