[PATCH v7 1/4] arm: add ARMv7 HYP API to flush VM TLBs without address param

[Christoffer Dall]

On Mon, Jun 09, 2014 at 10:06:29AM -0700, Mario Smarduch wrote:
> On 06/08/2014 05:05 AM, Christoffer Dall wrote:
> > On Tue, Jun 03, 2014 at 04:19:24PM -0700, Mario Smarduch wrote:
> >> Patch adds HYP interface for global VM TLB invalidation without address
> >> parameter. Added ARM version of kvm_flush_remote_tlbs(), made the generic
> >> implementation a weak symbol.
> >>
> >> Signed-off-by: Mario Smarduch <m.smarduch at samsung.com>
> >> ---
> >>  arch/arm/include/asm/kvm_asm.h |    1 +
> >>  arch/arm/kvm/interrupts.S      |   11 +++++++++++
> >>  arch/arm/kvm/mmu.c             |   14 ++++++++++++++
> >>  virt/kvm/kvm_main.c            |    2 +-
> >>  4 files changed, 27 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/arch/arm/include/asm/kvm_asm.h b/arch/arm/include/asm/kvm_asm.h
> >> index 53b3c4a..21bc519 100644
> >> --- a/arch/arm/include/asm/kvm_asm.h
> >> +++ b/arch/arm/include/asm/kvm_asm.h
> >> @@ -78,6 +78,7 @@ extern char __kvm_hyp_code_end[];
> >>  
> >>  extern void __kvm_flush_vm_context(void);
> >>  extern void __kvm_tlb_flush_vmid_ipa(struct kvm *kvm, phys_addr_t ipa);
> >> +extern void __kvm_tlb_flush_vmid(struct kvm *kvm);
> >>  
> >>  extern int __kvm_vcpu_run(struct kvm_vcpu *vcpu);
> >>  #endif
> >> diff --git a/arch/arm/kvm/interrupts.S b/arch/arm/kvm/interrupts.S
> >> index 0d68d40..bddc66b 100644
> >> --- a/arch/arm/kvm/interrupts.S
> >> +++ b/arch/arm/kvm/interrupts.S
> >> @@ -66,6 +66,17 @@ ENTRY(__kvm_tlb_flush_vmid_ipa)
> >>  	bx	lr
> >>  ENDPROC(__kvm_tlb_flush_vmid_ipa)
> >>  
> >> +/**
> >> + * void __kvm_tlb_flush_vmid(struct kvm *kvm) - Flush per-VMID TLBs
> >> + *
> >> + * Reuses __kvm_tlb_flush_vmid_ipa() for ARMv7, without passing address
> >> + * parameter
> >> + */
> >> +
> >> +ENTRY(__kvm_tlb_flush_vmid)
> >> +	b	__kvm_tlb_flush_vmid_ipa
> >> +ENDPROC(__kvm_tlb_flush_vmid)
> >> +
> >>  /********************************************************************
> >>   * Flush TLBs and instruction caches of all CPUs inside the inner-shareable
> >>   * domain, for all VMIDs
> >> diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c
> >> index 2ac9588..ef29540 100644
> >> --- a/arch/arm/kvm/mmu.c
> >> +++ b/arch/arm/kvm/mmu.c
> >> @@ -56,6 +56,20 @@ static void kvm_tlb_flush_vmid_ipa(struct kvm *kvm, phys_addr_t ipa)
> >>  		kvm_call_hyp(__kvm_tlb_flush_vmid_ipa, kvm, ipa);
> >>  }
> >>  
> >> +/**
> >> + * kvm_flush_remote_tlbs() - flush all VM TLB entries
> >> + * @kvm:       pointer to kvm structure.
> >> + *
> >> + * Interface to HYP function to flush all VM TLB entries without address
> >> + * parameter. In HYP mode reuses __kvm_tlb_flush_vmid_ipa() function used by
> >> + * kvm_tlb_flush_vmid_ipa().
> >> + */
> >> +void kvm_flush_remote_tlbs(struct kvm *kvm)
> >> +{
> >> +	if (kvm)
> >> +		kvm_call_hyp(__kvm_tlb_flush_vmid, kvm);
> >> +}
> >> +
> >>  static int mmu_topup_memory_cache(struct kvm_mmu_memory_cache *cache,
> >>  				  int min, int max)
> >>  {
> >> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> >> index fa70c6e..ba25765 100644
> >> --- a/virt/kvm/kvm_main.c
> >> +++ b/virt/kvm/kvm_main.c
> >> @@ -184,7 +184,7 @@ static bool make_all_cpus_request(struct kvm *kvm, unsigned int req)
> >>  	return called;
> >>  }
> >>  
> >> -void kvm_flush_remote_tlbs(struct kvm *kvm)
> >> +void __weak kvm_flush_remote_tlbs(struct kvm *kvm)
> >>  {
> >>  	long dirty_count = kvm->tlbs_dirty;
> >>  
> >> -- 
> >> 1.7.9.5
> >>
> > 
> > This doesn't build or link on aarch64 :(
> > 
> > -Christoffer
> > 
> 
> I'll recompile and retest the dirty page logging portion on ARMv8 and
> resolve these issues, early next week.
> 
> In the meantime if it's ok with you, I'' move forward with the
> rest of the patches on ARMv7 to get through critical issues.
> 
> Would that work?

Your patches need to at least compile with other architectures, I
suggest fixing that up front.  You also eventually need to somehow test
(or ask maintainers to test) a branch with your patches on the
architectures that your code messes with.  This later point can be
addressed once we're close to a consensus on ARM.

-Christoffer