Re: nVMX x86: Re-name check_vmentry_pre[post]reqs and re-organize check_vmentry_prereqs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Nov 13, 2018 at 01:12:03AM -0500, Krish Sadhukhan wrote:
> According to Intel SDM vol 3C, VMCS data is of six different types. But the
> single function check_vmentry_prereqs currently holds all the different types
> of checks related to those six types of VMCS data. This patch series
> re-organizes the checks in that function into six different sub-functions
> based on the VMCS data organization. The re-organized code is easier for
> readability, enhancement and maintenance.

I like the idea, but I think it'd be better to follow the SDM's section
for the actual checks (26.1 - 26.3 in my copy) rather than the section
that defines the VMCS fields.  Many of the consistency checks involve
fields of multiple types, e.g. a lot of the checks on host/guest state
interact with control settings.  Following the SDM means we don't have
to make judgment calls on where to put such checks.  This also avoids
the issue of identifying which checks do/don't result in VMExit as the
the SDM explicitly calls out in "Checking and Loading Guest State" that
the subsequent checks cause VMExit.  And cross-checking our code with
SDM should be even easier if we follow the SDM's ordering.



[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux