On 08/09/2013 06:55 PM, Christoffer Dall wrote:
> On Fri, Aug 09, 2013 at 05:03:04PM +0200, Andre Przywara wrote:
>> (for GIT URL and Changelog see below)
>>
>> ARM CPUs with the virtualization extension have a new mode called
>> HYP mode, which allows hypervisors to safely control and monitor
>> guests. The current hypervisor implementations (KVM and Xen)
>> require the kernel to be entered in that HYP mode.
>>
>> This patch series introduces a configuration variable
>> CONFIG_ARMV7_VIRT which enables code to switch all cores into HYP
>> mode. This is done automatically during execution of the bootm
>> command.
>>
>> The process of switching into HYP mode requires the CPU to be in
>> secure state initially when entering u-boot, it will then setup some
>> register and switch to non-secure state. This requires the GIC to be
>> programmed properly first. Explanations about the details are in the
>> commit messages of the respective patches.
>>
>> The patches are structured like this:
>> 1/8: prepare header file
>> 2/8: add monitor handler (assembly)
>> 3/8: add per CPU non-secure switch routine (assembly)
>> 4/8: add system wide non-secure setup (C)
>> 5/8: trigger non-secure switch during bootm command
>> 6/8: add generic SMP functionality
>> 7/8: add HYP mode switching
>> 8/8: board specific code for ARM Versatile Express TC2
>>
>> Since up to patch 6/8 this code works on non-virtualization capable
>> CPUs also and there has been a request, there is now a second
>> configuration variable CONFIG_ARMV7_NONSEC, which omits the final
>> HYP mode switch and just goes into non-secure SVC state.
>> You can specify either (or none) of them, the code cares about
>> the dependency.
>>
>> The code aims to be as generic as possible, though currently it has
>> only been tested on the Versatile Express TC-2 board. The last patch
>> thus enables the feature for that board and should serve as an
>> example for supporting other boards.
>>
>> For convenience there is a GIT tree which you can pull these patches
>> from ("hypmode_v4" branch):
>> git://git.linaro.org/people/aprzywara/u-boot.git
>>
>> Changes RFC..v1
>> * not a dedicated command anymore, code run by bootm & friends
>> * protecting code with #ifdefs to avoid unnecessary inclusion and
>> accidental crashing (when accessing restricted registers)
>> * moving prototypes to header file to meet checkpatch recommendation
>> * adding comment as proposed by Christoffer
>>
>> Changes v1..v2
>> mostly style and code layout changes
>> * restructure assembly code to live in a new file and not start.S
>> * split smp, nonsec_init and hyp_init to be separate functions
>> * used named constants from common header files
>> * split C function to be more readable
>> * extend comments to be more precise and elaborate
>> * add provision to override GIC base address (needed for Arndale?)
>> * add configuration variable to enable VExpress specific SMP code
>> * use writel/readl for MMIO GIC accesses
>> * remove superfluous isb instructions
>> * more minor fixes
>>
>> Changes v2..v3
>> * fix clobbering of GICC address actually spoiling the stack
>> * do CNTFRQ setup in assembly per core (and not only once per SoC)
>> * moving the new code files into arch/arm/cpu/armv7
>> * add config variable for doing non-secure switch only
>> * use actual HYP and secure instructions mnemonics instead of
>> the encoded byte sequence. This requires more recent compilers.
>> * make the identification of the CPU core more robust and saner
>> * use enum for error codes and rename them
>> * lots of smaller layout and style fixes
>>
>> Changes v3..v4
>> * mask reserved bits in CBAR register
>> * move the VExpress board specific SMP code into the board directory
>> * embed error reporting in the respective functions and getting
>> rid of the error code enum at all (by popular demand ;-)
>> * minor style fixes
>>
>> Please review and comment!
>>
> Only had those very few nits sent separately, which can be fixed later
> if we want. Otherwise looks good.
>
> Reviewed-by: Christoffer Dall <christoffer.dall@xxxxxxxxxx>
Thanks, Christoffer.
Are there more concerns or comments about this?
Is there anything left I can do get these patches merged?
Albert, Tom,
do you need more ACKs or Reviewed-bys?
Kim, do you have time and expertise to look at these?
Thanks,
Andre.
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
