Re: [kvm PATCH] KVM: nVMX - enable VMX preemption timer migration

[Sean Christopherson <sean.j.christopherson@intel.com>]

Please read through Thomas' diatribe^W comments on top-posting and HTML
email[*].  I appreciate that it can be frustrating to configure and get
used to, especially in a corporate environment, but it really does help
get your code upstreamed.

https://people.kernel.org/tglx/notes-about-netiquette

On Thu, Apr 23, 2020 at 11:48:34AM -0700, Makarand Sonare wrote:
> Thanks for the feedback!
> *>>Any plans to enhance the vmx_set_nested_state_test.c to verify this
> works as intended?*
> I am working on writing the test as part of state_test.c as that test
> already has the nested state save/restore logic.
> 
> *>>Build tested only, but {get,put}_user() compiles just fine, as
> requested.*
> The impacted functions are using copy_to_user/copy_from_user already so I
> used the same for consistency.

That's because they're copying complex structures that don't fit in a
single memory access.

> I will send a v3 PATCH after incorporating rest of the feedback.

In the future, please give folks a chance to digest and respond before
spinning another version, especially when there is disagreement over which
direction to take.  E.g. the above {get,put}_user() thing is easy to sort
out with a few back-and-forth mails.  Sending v3 prematurely (and v2 for
that matter) means that closing that discussion requires tracking down the
new version and providing the necessary context.  Which, as an aside, is
why trimming mails and bottom-posting is helpful as it allows the scope of
the conversation to be narrowed to the remaining opens.