Re: [PATCH v3] arm/aspeed: Rework NIC attachment

[Markus Armbruster <armbru@redhat.com>]

Cédric Le Goater <clg@kaod.org> writes:

> On 5/27/20 3:36 PM, Markus Armbruster wrote:
>> Cédric Le Goater <clg@kaod.org> writes:
>> 
>>> The number of MACs supported by an Aspeed SoC is defined by "macs_num"
>>> under the SoC model, that is two for the AST2400 and AST2500 and four
>>> for the AST2600. The model initializes the maximum number of supported
>>> MACs but the number of realized devices is capped by the number of
>>> network device back-ends defined on the command line. This can leave
>>> unrealized devices hanging around in the QOM composition tree.
>>>
>>> Modify the machine initialization to define which MACs are attached to
>>> a network device back-end using a bit-field property "macs-mask" and
>>> let the SoC realize all network devices.
>>>
>>> The default setting of "macs-mask" is "use MAC0" only, which works for
>>> all our AST2400 and AST2500 machines. The AST2600 machines have
>>> different configurations. The AST2600 EVB machine activates MAC1, MAC2
>>> and MAC3 and the Tacoma BMC machine activates MAC2.
>> 
>> Let's be more clear on what this means, and "This is actually a device
>> modelling fix for these two machines."  Okay?
>
> Well, I guess so. It's a fix in the way we attach network backends to 
> the MACs of the machines. 

Yes, it's that too.

> On the tacoma-bmc, we had to use '-nic <foo> -nic <bar> -nic <good one>' 
> to configure the MAC2 in use by the machine. Which was dubious.

I think you had to use something like

    -nic none -nic none -nic GOOD-ONE -nic none

to get virtual hardware that matches the physical hardware, which I
understand has all four MACs on the die, but only MAC2 connected to the
outside.

In particular, the default configuration (no -nic, -nodefaults, etc.)
got you only MAC0.  With just -nodefaults, you got none at all.

> Now, a single -nic is enough.

Now you get all four MACs regardless of configuration, but only MAC2 can
be connected to a backend, e.g. with a single -nic.

The default configuration (no -nic, -nodefaults, etc.) just works: MAC2
connected to the default network backend.

With just -nodefaults, MAC2 remains unconnected.

This matches how other machines work.

>>> Inactive MACs will have no peer and QEMU may warn the user with :
>>>
>>>     qemu-system-arm: warning: nic ftgmac100.0 has no peer
>>>     qemu-system-arm: warning: nic ftgmac100.1 has no peer
>>>     qemu-system-arm: warning: nic ftgmac100.3 has no peer
>>>
>>> Signed-off-by: Cédric Le Goater <clg@kaod.org>
>>> Reviewed-by: Markus Armbruster <armbru@redhat.com>
>> 
>> Here's the "info qom-tree" change for tacoma-bmc:
>> 
>>      /machine (tacoma-bmc-machine)
>>        /peripheral (container)
>>        /peripheral-anon (container)
>>        /soc (ast2600-a1)
>>          [...]
>>          /ftgmac100[0] (ftgmac100)
>>            /ftgmac100[0] (qemu:memory-region)
>>          /ftgmac100[1] (ftgmac100)
>>     +      /ftgmac100[0] (qemu:memory-region)
>>          /ftgmac100[2] (ftgmac100)
>>     +      /ftgmac100[0] (qemu:memory-region)
>>          /ftgmac100[3] (ftgmac100)
>>     +      /ftgmac100[0] (qemu:memory-region)
>
> Yes. All are realized now.
>
>>          [...]
>>          /mii[0] (aspeed-mmi)
>>            /aspeed-mmi[0] (qemu:memory-region)
>>          /mii[1] (aspeed-mmi)
>>     +      /aspeed-mmi[0] (qemu:memory-region)
>>          /mii[2] (aspeed-mmi)
>>     +      /aspeed-mmi[0] (qemu:memory-region)
>>          /mii[3] (aspeed-mmi)
>>     +      /aspeed-mmi[0] (qemu:memory-region)
>
> Same for the MMI interfaces on AST2600.
>
>> These changes are due to realizing MAC1, MAC2, MAC3.  Looks good.
>> 
>> Here's "info qtree":
>> 
>>        dev: ftgmac100, id ""
>>          gpio-out "sysbus-irq" 1
>>          aspeed = true
>>     -    mac = "52:54:00:12:34:56"
>>     -    netdev = "hub0port0"
>>     +    mac = "52:54:00:12:34:57"
>>     +    netdev = ""
>>          mmio 000000001e660000/0000000000002000
>>        dev: ftgmac100, id ""
>>     -    aspeed = false
>>     -    mac = "00:00:00:00:00:00"
>>     +    gpio-out "sysbus-irq" 1
>>     +    aspeed = true
>>     +    mac = "52:54:00:12:34:58"
>>          netdev = ""
>>     +    mmio 000000001e680000/0000000000002000
>>        dev: ftgmac100, id ""
>>     -    aspeed = false
>>     -    mac = "00:00:00:00:00:00"
>>     -    netdev = ""
>>     +    gpio-out "sysbus-irq" 1
>>     +    aspeed = true
>>     +    mac = "52:54:00:12:34:56"
>>     +    netdev = "hub0port0"
>>     +    mmio 000000001e670000/0000000000002000
>>        dev: ftgmac100, id ""
>>     -    aspeed = false
>>     -    mac = "00:00:00:00:00:00"
>>     +    gpio-out "sysbus-irq" 1
>>     +    aspeed = true
>>     +    mac = "52:54:00:12:34:59"
>>          netdev = ""
>>     +    mmio 000000001e690000/0000000000002000
>>        [...]
>>        dev: aspeed-mmi, id ""
>>          mmio 000000001e650000/0000000000000008
>>        dev: aspeed-mmi, id ""
>>     +    mmio 000000001e650008/0000000000000008
>>        dev: aspeed-mmi, id ""
>>     +    mmio 000000001e650010/0000000000000008
>>        dev: aspeed-mmi, id ""
>>     +    mmio 000000001e650018/0000000000000008
>> 
>> Here we can see the network backend now gets connected to MAC2 instead
>> of MAC0.
>
> yes.
>
> With only one nic on the command line, the backend was attached to the 
> first (unused) MAC0 of the machine and now it is attached to the first 
> active MAC2 of the machine.
>
>> 
>> This is without any networking-related options, i.e. we get just the
>> single default network backend.
>> 
>>> ---
>>>
>>>  To be applied on top of patch :
>>>
>>>  "arm/aspeed: Compute the number of CPUs from the SoC definition" 
>>>  http://patchwork.ozlabs.org/project/qemu-devel/patch/20200519091631.1006073-1-clg@kaod.org/
>>>
>>>  Markus, do you mind taking this patch in your QOM series also ?
>> 
>> On the contrary!
>> 
>> I'll work my "info qom-tree" and "info qtree" diffs into the commit
>> message, if you don't mind.
>
> Sure. 

Today, I hope.  Thanks!