qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] virtio-rng


From: Amit Shah
Subject: Re: [Qemu-devel] virtio-rng
Date: Tue, 17 Nov 2009 16:55:54 +0530
User-agent: Mutt/1.5.19 (2009-01-05)

On (Tue) Nov 17 2009 [11:10:32], Ian Molton wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Amit Shah wrote:
> > (Any reason to take this off-list?)
> 
> None other than hitting reply rather than reply-all. CCing list once
> more :-)
> 
> >> Either way, you still need to specify the properties - they've just
> >> moved into the console driver in your patch by the look of it.
> > 
> > Yeah; there are two ways of setting properties -- from the command line
> > or from code.
> > 
> > What you're doing is something like
> > 
> > -virtiorng,a=foo,b=bar
> 
> Ah, I see why we're at cross purposes here - I thought we'd moved to
> discussing virtio-console. I've already moved to qdev based init for
> virtio-rng - 'fixing' virtio-console was where I learnt about the 'new
> way' (qdev).

I actually meant to talk about virtio-console but had the virtio-rng
example ready in mind. I thought you did the same thing for console too?
Sorry I've not really looked at the patch in detail so you can disregard
that comment.

> >> I'd prefer to do the same for virtio-rng,
> 
> Here I refer to you having (and I didnt apply the patch so I may have
> misread) dropped the virtio-pci proxy from virtio-console. Once I've got
> a tree I can apply your patch to I'll take another look :-)

It slightly works differently now: virtio is supposed to be
pci-agnostic. So what now happens is:

virtio-console hooks on to the virtio-serial-bus via -device virtconsole
virtio-serial-bus hooks on to the virtio-pci-proxy via -device
virtio-serial-pci

> >> Yes, I saw that. Would it not be better to generate the device / chardev
> >> pair dynamically though, rather than preserve the icky old array?
> > 
> > I don't understand what you mean by 'dynamically'.
> 
> Rather than have that array virtcon_hds at all, create the pairs of host
> side data and qdev properties during parsing the commandline - no need
> to store them up and iterate through the later, or to set an arbitrary
> limit on how many can be specified in that way.

qdev won't understand the old-style commandline syntax; so once
-virtioconsole is encounters, all parsing of the arguments for that
param are to be done by the code that handles -virtioconsole. The array
is maintained because multiple virtio-consoles could be spawned, upto a
max. of MAX_VIRTIO_CONSOLES. But, as it stands, MAX_VIRTIO_CONSOLES is 1
and so the array can be dropped, but a local var. would still be needed
to fetch the chardev and then init it properly using qemu_opt.

> Might be an idea for qemu to warn people that this syntax will be
> deprecated, too...

Yes, that should be done (maybe for 1-2 releases).

                Amit




reply via email to

[Prev in Thread] Current Thread [Next in Thread]