[Top][All Lists]

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

Re: [Qemu-devel] [PATCH v3 2/2] hw/vexpress.c: Add model of ARM Versatil

From: Peter Maydell
Subject: Re: [Qemu-devel] [PATCH v3 2/2] hw/vexpress.c: Add model of ARM Versatile Express board
Date: Mon, 21 Mar 2011 21:19:56 +0000

On 21 March 2011 20:41, Aurelien Jarno <address@hidden> wrote:
> On Mon, Mar 07, 2011 at 11:10:32AM +0000, Peter Maydell wrote:
>> +    /* 0x4e000000 LAN9118 Ethernet */
>> +    if (nd_table[0].vlan) {
>> +        lan9118_init(&nd_table[0], 0x4e000000, pic[15]);
>> +    }
> It basically means we silently ignore all other network cards the user
> asked. Shouldn't we but a warning/error in that case?

Well, if you ask for one network card that's the wrong kind
you get an error:
qemu-system-arm: qemu: Unsupported NIC model: ne2k_pci

but you're right that if you ask for several cards we ignore
the other ones.

I just copied this code pattern from hw/integratorcp.c ; if you
want to suggest a better code fragment I'm happy to revise
the patch.
(hw/versatilepb.c is even worse, if I'm reading the code right
it creates rtl8139 cards regardless of what model you ask for...)

Conceptually, though, it seems to me that the right place to
put "the user specified a network card/a USB device/whatever
but nothing in the board model used it" checks is in the
generic code. Otherwise even boards with no network support
at all need to have boilerplate "reject any network card
specification" code, for instance, all boards without USB need
code to reject attempts to define USB devices, and so on...

> For having the possibility to add more than one card, I guess it's
> possible to simulated the PCI controller? Well if the board has any PCI
> controller.

Sadly PCI is not supported by the A9MPx4 daughterboard.

> Otherwise the patch looks fine to me, it's very nice being able to boot
> an ARM board with more than 256MB of RAM.

PS: PBX-A9 also supports 1GB RAM :-) (but for different reasons
doesn't have PCI support and also the model is decidedly inaccurate).

-- PMM

reply via email to

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