Re: [RFC PATCH 0/5] Removing support for 32bit KVM/arm host - Russell King

From: Russell King - ARM Linux admin <linux@armlinux.org.uk>
To: Marc Zyngier <maz@kernel.org>
Cc: Anders Berg <anders.berg@lsi.com>, Arnd Bergmann <arnd@arndb.de>,
	kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	Paolo Bonzini <pbonzini@redhat.com>,
	Will Deacon <will@kernel.org>,
	kvmarm@lists.cs.columbia.edu
Subject: Re: [RFC PATCH 0/5] Removing support for 32bit KVM/arm host
Date: Mon, 10 Feb 2020 16:25:23 +0000	[thread overview]
Message-ID: <20200210162523.GF25745@shell.armlinux.org.uk> (raw)
In-Reply-To: <20200210141324.21090-1-maz@kernel.org>

On Mon, Feb 10, 2020 at 02:13:19PM +0000, Marc Zyngier wrote:
> KVM/arm was merged just over 7 years ago, and has lived a very quiet
> life so far. It mostly works if you're prepared to deal with its
> limitations, it has been a good prototype for the arm64 version,
> but it suffers a few problems:
> 
> - It is incomplete (no debug support, no PMU)
> - It hasn't followed any of the architectural evolutions
> - It has zero users (I don't count myself here)
> - It is more and more getting in the way of new arm64 developments
> 
> So here it is: unless someone screams and shows that they rely on
> KVM/arm to be maintained upsteam, I'll remove 32bit host support
> form the tree. One of the reasons that makes me confident nobody is
> using it is that I never receive *any* bug report. Yes, it is perfect.
> But if you depend on KVM/arm being available in mainline, please shout.

It is only very recently that 64-bit ARM has really started to filter
down to people like me, and people like me have setup systems which
use 32-bit VMs under 64-bit hosts (about a year ago now.)  In fact,
everything that you presently see for the *.armlinux.org.uk domain now
runs inside several 32-bit ARM VMs under a 64-bit ARM host.

It isn't perfect; I've found issues with qemu and libvirt.  One example
is the rather sub-standard RTC implementation, which means if you
use libvirt's managesave across a host reboot, the guests idea of
time-of-day freezes while it's asleep, and resumes when the guest is
reloaded - resulting in the guests time-of-day being rather wrong,
sometimes to the point that NTP gives up.  That becomes very painful
if you use kerberos authentication, where time-of-day is important.

So, just because you haven't received any bug reports doesn't mean
there aren't any users; there certainly are, there are problems,
but the problems are in places other than the kernel.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm