On Tue, Feb 19, 2019 at 06:42:08PM +0000, Julien Grall wrote: > On 1/22/19 5:50 AM, Will Deacon wrote: > > On Thu, Dec 20, 2018 at 03:21:17PM +0000, Julien Grall wrote: > > > At the moment, a user is only able to specify the amount of RAM used by the > > > guest. Where the RAM will live is left to the software and hardcoded. > > > > > > It could be useful for testing purpose to move the RAM in different place. > > > This series adds the possibility for the user to specify multiple RAM region. > > > > > > The option -m/--mem is extended to specify the address using the following > > > format: <size>@<addr>. The option needs to be repeated as many times as the > > > number of RAM region in the guest layout. > > > > > > For instance, if you want 512MB at 3GB and 512MB 4GB it would look like: > > > -m 512@0xc0000000 -m 512@0x100000000 > > > > > > Note that the memory layout is not yet fully configurable by the user, so the > > > MMIO region is still living below 2GB. This means RAM cannot live in the > > > region 0-2GB. This could be changed in the future. > > > > Although the implementation mostly looks ok (I've commented on some things I > > spotted), I'm not completely sold on the need for this. What sort of testing > > do you have in mind that requires multiple memory banks to be specified at > > runtime? I'm keen to keep kvmtool as simple and lean as we can, whilst > > supporting enough functionality to be useful as a prototyping tool for the > > latest KVM features. I'm just struggling to see how support for multiple > > memory banks really fits in with that, and whether or not we can't just make > > it easier to configure this at compile time instead. > This series turns kvmtool to a useful tool for exercising booting a kernel > with different memory layout. You can imagine use it when touching the > memory code (i.e 52-bit...) or even as a testsuite. > > For this use case, the compile time option would make things a bit more > difficult. Hence the runtime option. Ok, fair enough. I spoke to Julien Thierry for a second opinion on this, and perhaps a way forward would be to restrict ourselves to a single memory bank, but allow it's base address to be provided on the command line? I think that covers your use-cases above, and should avoid a fair amount of complication. What do you reckon? Will
