[Qemu-devel] Re: [PATCH v2 14/21] qemu-kvm: Rework VCPU state writeback API

[Jan Kiszka]

Avi Kivity wrote:
> On 02/03/2010 10:53 AM, Jan Kiszka wrote:
>> This grand cleanup drops all reset and vmsave/load related
>> synchronization points in favor of four(!) generic hooks:
>>
>> - cpu_synchronize_all_states in qemu_savevm_state_complete
>>    (initial sync from kernel before vmsave)
>> - cpu_synchronize_all_post_init in qemu_loadvm_state
>>    (writeback after vmload)
>> - cpu_synchronize_all_post_init in main after machine init
>> - cpu_synchronize_all_post_reset in qemu_system_reset
>>    (writeback after system reset)
>>
>> These writeback points + the existing one of VCPU exec after
>> cpu_synchronize_state map on three levels of writeback:
>>
>> - KVM_PUT_ASYNC_STATE (during runtime, other VCPUs continue to run)
>>    
> 
> Wouldn't that be SYNC_STATE (state that is modified by the current vcpu
> only)?

It's async /wrt other VCPUs. They continue to run and may interact with
this VCPU while updating its state.

> 
>> - KVM_PUT_RESET_STATE (on synchronous system reset, all VCPUs stopped)
>> - KVM_PUT_FULL_STATE  (on init or vmload, all VCPUs stopped as well)
>>
>> This level is passed to the arch-specific VCPU state writing function
>> that will decide which concrete substates need to be written. That way,
>> no writer of load, save or reset functions that interact with in-kernel
>> KVM states will ever have to worry about synchronization again. That
>> also means that a lot of reasons for races, segfaults and deadlocks are
>> eliminated.
>>
>> cpu_synchronize_state remains untouched, just as Anthony suggested. We
>> continue to need it before reading or writing of VCPU states that are
>> also tracked by in-kernel KVM subsystems.
>>
>> Consequently, this patch removes many cpu_synchronize_state calls that
>> are now redundant, just like remaining explicit register syncs. It does
>> not touch qemu-kvm's special hooks for mpstate, vcpu_events, or tsc
>> loading. They will be cleaned up by individual patches.
>>
>>    
> 
> I'm uneasy about this.  What are the rules for putting
> cpu_synchronize_state() now?

As before for code that accesses the state during runtime: Before you
read or write some bit of it, call cpu_synchronize_state().

Only reset and save/restore handlers do not have to worry about
synchronization anymore. It makes no sense to overload them with
arch-specific KVM knowledge about what shall be written and when.

Jan

signature.asc
Description: OpenPGP digital signature