Thread (46 messages) flat view 46 messages, 8 authors, 2013-10-03

Re: [PATCH 3/3] KVM: PPC: Book3S: Add support for hwrng found on some powernv systems

From: Gleb Natapov <hidden>
Date: 2013-10-03 06:29:18
Also in: kvm, lkml

On Thu, Oct 03, 2013 at 08:07:22AM +1000, Benjamin Herrenschmidt wrote:
On Wed, 2013-10-02 at 13:02 +0300, Gleb Natapov wrote:
quoted
Yes, I alluded to it in my email to Paul and Paolo asked also. How this
interface is disabled? Also hwrnd is MMIO in a host why guest needs to
use hypercall instead of emulating the device (in kernel or somewhere
else?). Another things is that on a host hwrnd is protected from
direct userspace access by virtue of been a device, but guest code (event
kernel mode) is userspace as far as hosts security model goes, so by
implementing this hypercall in a way that directly access hwrnd you
expose hwrnd to a userspace unconditionally. Why is this a good idea? 
BTW. Is this always going to be like this ?
If something questionable will be noticed explanation will be required.
It is like that for all arches and all parts of kernel.
Every *single* architectural or design decision we make for our
architecture has to be justified 30 times over, every piece of code bike
shedded to oblivion for month, etc... ?
This is simply not true, most powerpc patches go in without any comments.
Do we always have to finally get to some kind of agreement on design, go
to the 6 month bike-shedding phase, just to have somebody else come up
and start re-questioning the whole original design (without any
understanding of our specific constraints of course) ?
Do you really think that nobody here understands that exit to userspace
is slow?
You guys are the most horrendous community I have ever got to work with.
It's simply impossible to get anything done in any reasonable time
frame .

At this stage, it would have taken us an order of magnitude less time to
simply rewrite an entire hypervisor from scratch.
Of course, it is always much easier to ignore other people input and do
everything your way. Why listen to people who deal with migration issues
for many years if you can commit the patch and forget about it until
migration fails, but who cares, you got there in an order of magnitude
less time and this is what counts.
This is sad.
Agree.

--
			Gleb.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help