Re: vexpress hdlcd driver: error: kvm run failed function not implemented

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

 



On Thu, 2013-08-08 at 19:29 +0100, Peter Maydell wrote:
> On 8 August 2013 14:36, Jon Medhurst (Tixy) <tixy@xxxxxxxxxx> wrote:
> > Shows how long it is since I last plugged in the TC1 CoreTile.
> >
> > TC2 has 2GB, so this probably crept in with some merge or cut'n'paste
> > foobar.
> >
> > I'll fix the patch in my tree, and dust off that TC1 :-)
> 
> Do you think the TC1 device tree/kernel support is likely to
> bitrot in future (compared with the TC2 one)?

It's probably _more_ likely to bitrot simply because people don't really
use TC1.

>  I originally
> made QEMU model the TC1 because the TC1 is A15x2, and QEMU
> can't model a heterogenous cluster setup like the TC2's
> A15x2+A7x3. However if TC1 is going to effectively die
> as a supported platform the other possibility would be
> to make vexpress-a15 be "a TC2 with the A7 cores
> permanently powered off/held in reset".

I believe that is also what people wanting real hardware for A15 (and
A7) development are being recommended to do, i.e. use TC2 and disable
the cluster they don't want.

> I'd rather not change if I can avoid it though because
> it'll just confuse everybody more and/or require me to
> keep the current TC1 model as well as adding TC2 ;-)

Well, I don't think there many (any) devices which are TC1 specific, so
I would hope that any regression on TC1 would be simple failures to
update TC1 device-trees correctly.

> On the HDLCD: QEMU doesn't currently model this; we could
> in theory add a model, but I'd need to add support in
> QEMU's display-output infrastructure for selecting
> which display to output based on guest control (ie
> the model equivalent of the h/w's DVI multiplexer)
> so as long as the kernel has continued to be able to
> just fall back to the PL111 when it can't find the
> HDCLD it hasn't seemed necessary.

Is that what happens? I thought that having both devices in device-tree
didn't work, so one had to manually disable one of them in device-tree.
I'm also not sure if the motherboard CLCD display actually works in the
Linaro kernel tree, may do though. (You would never want to use it on
real hardware as the performance impact is very bad.)

-- 
Tixy

_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm




[Index of Archives]     [Linux KVM]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux