[OpenBIOS] [PATCHv2 0/3] Enable bus master support for rtl8139 network cards on Apple PPC machines

[Cédric Le Goater]

On 08/07/2016 03:52 PM, Mark Cave-Ayland wrote:
> On 07/08/16 11:05, Cédric Le Goater wrote:
> 
>> Hello Mark,
>>
>> The recent changes in openbios and qemu really improved the support 
>> of 10.4.11. I can now logon on the raw disk I have and the system 
>> looks pretty good and stable. All these issues are gone :
>>
>> 	https://www.coreboot.org/pipermail/openbios/2016-June/009436.html
>>
>> I still need :
>>
>> 	-prom-env boot-device=hd:3,\System\Library\CoreServices\BootX
>>
>> to boot, but I suppose this is due to the state of the disk which also 
>> has a macosx 9 on it.
> 
> Right. I think I asked the question before if we should simply add this
> to the search patch, e.g. "hd:,\System\Library\CoreServices\BootX"
> before "hd:,\\:tbxi" if an installed BootX should always take priority?

yes. That works. sorry for the late answer.

>> However, the network is not detected. This is curious as the command 
>> 'ioreg' does list a rtl8139 with this patchset, see below, and I suppose 
>> the driver is correct  : 
>>
>>    28    1 0x55c000   0x1f000    0x1e000    com.apple.iokit.IONetworkingFamily (1.5.0) <6 5 4 3 2>
>>    29    0 0x57b000   0x5000     0x4000     com.apple.driver.AppleRTL8139Ethernet (1.1.0) <28 16 5 4 3 2>
>>
>>  
>> Is there anything I could do to learn a bit more why the adapter is not
>> showing in the system ?  I did not find anything in the osx logs.
> 
> Following on from Howard's testing earlier, maybe something in the
> ppc-for-2.8 branch breaks this? If you try running QEMU git master with
> an OpenBIOS from
> https://github.com/mcayland/openbios/commits/ppc-busmaster and -cpu 750
> does networking work?

That's better. I was using the mainline openbios with your set of patches
for busmaster but I was missing Ben's fixes.

So, with your ppc-busmaster branch, macosx 10.4.11 finds the adapter. 
However, dhcp does not work. Investigation in progress.
 
> If so that suggest something in the ppc-for-2.8 queue breaks it in which
> case if you try and bisect that would be very helpful :)

There are no regression with David's branch. 

Cheers,

C.

> 
> ATB,
> 
> Mark.
>