On Tue, 2016-03-15 at 11:25 +0100, Cédric Le Goater wrote: > On 03/15/2016 06:01 AM, Andrew Jeffery wrote: > > On Tue, 2016-03-15 at 12:34 +0800, Jeremy Kerr wrote: > > > Hi Andrew, > > > > > > > This patch series models enough of the ASPEED AST2400 ARM9 SoC[0] to > > > > boot an aspeed_defconfig Linux kernel[1][2]. Specifically, the series > > > > implements the ASPEED timer and VIC devices, integrates them into an > > > > AST2400 SoC and exposes it all through a new opbmc2400 machine. The > > > > device model patches only partially implement the hardware features of > > > > the timer and VIC, again mostly just enough to boot Linux. > > > > > > Awesome! Nice to have these patches escaping the lab :) > > > > > > In terms of naming suggestions: I think this depends on what we're > > > looking to emulate here. I see two options: > > > > > > The qemu platform becomes a "reference" for OpenPOWER bmc hardware, but > > > doesn't necessarily align with an actual machine. In that case, > > > something generic like opbmc- would make sense. > > > > > > On the other hand, if we'd like to create qemu platforms that represent > > > actual machines (eg, the OpenPOWER "palmetto" machine), then > > > -bmc would seem more appropriate. In this case, the machine > > > name would be palmetto-bmc. No need to include the SoC name in that, as > > > it's defined by the hardware implementation. > > > > > > I think the latter option may be more generally useful. > > > > Okay, agreed, I'll rework the change to use palmetto-bmc for the > > machine name. Thanks for the feedback. > > Yes. palmetto-bmc is good choice. Palmetto is a reference machine > for OpenPOWER. > > May be change also : > > + mc->desc = "OpenPOWER AST2400 BMC (ARM926EJ-S)"; > > to reflect that choice.
Will do! Thanks, Andrew
signature.asc
Description: This is a digitally signed message part