On Wed, Jul 13, 2022 at 6:56 AM Atish Patra <atishp@xxxxxxxxxxxxxx> wrote:
>
> On Thu, Jul 7, 2022 at 7:53 AM Anup Patel <apatel@xxxxxxxxxxxxxxxx> wrote:
> >
> > The in-kernel AIA IMSIC support requires on-demand mapping / unmapping
> > of Guest IMSIC address to Host IMSIC guest files. To help achieve this,
> > we add kvm_riscv_stage2_ioremap() and kvm_riscv_stage2_iounmap()
> > functions. These new functions for updating G-stage page table mappings
> > will be called in atomic context so we have special "in_atomic" parameter
> > for this purpose.
> >
> > Signed-off-by: Anup Patel <apatel@xxxxxxxxxxxxxxxx>
> > ---
> > arch/riscv/include/asm/kvm_host.h | 5 +++++
> > arch/riscv/kvm/mmu.c | 18 ++++++++++++++----
> > 2 files changed, 19 insertions(+), 4 deletions(-)
> >
> > diff --git a/arch/riscv/include/asm/kvm_host.h b/arch/riscv/include/asm/kvm_host.h
> > index 59a0cf2ca7b9..60c517e4d576 100644
> > --- a/arch/riscv/include/asm/kvm_host.h
> > +++ b/arch/riscv/include/asm/kvm_host.h
> > @@ -284,6 +284,11 @@ void kvm_riscv_hfence_vvma_gva(struct kvm *kvm,
> > void kvm_riscv_hfence_vvma_all(struct kvm *kvm,
> > unsigned long hbase, unsigned long hmask);
> >
> > +int kvm_riscv_gstage_ioremap(struct kvm *kvm, gpa_t gpa,
> > + phys_addr_t hpa, unsigned long size,
> > + bool writable, bool in_atomic);
> > +void kvm_riscv_gstage_iounmap(struct kvm *kvm, gpa_t gpa,
> > + unsigned long size);
> > int kvm_riscv_gstage_map(struct kvm_vcpu *vcpu,
> > struct kvm_memory_slot *memslot,
> > gpa_t gpa, unsigned long hva, bool is_write);
> > diff --git a/arch/riscv/kvm/mmu.c b/arch/riscv/kvm/mmu.c
> > index b75d4e200064..f7862ca4c4c6 100644
> > --- a/arch/riscv/kvm/mmu.c
> > +++ b/arch/riscv/kvm/mmu.c
> > @@ -343,8 +343,9 @@ static void gstage_wp_memory_region(struct kvm *kvm, int slot)
> > kvm_flush_remote_tlbs(kvm);
> > }
> >
> > -static int gstage_ioremap(struct kvm *kvm, gpa_t gpa, phys_addr_t hpa,
> > - unsigned long size, bool writable)
> > +int kvm_riscv_gstage_ioremap(struct kvm *kvm, gpa_t gpa,
> > + phys_addr_t hpa, unsigned long size,
> > + bool writable, bool in_atomic)
> > {
> > pte_t pte;
> > int ret = 0;
> > @@ -353,6 +354,7 @@ static int gstage_ioremap(struct kvm *kvm, gpa_t gpa, phys_addr_t hpa,
> > struct kvm_mmu_memory_cache pcache;
> >
> > memset(&pcache, 0, sizeof(pcache));
> > + pcache.gfp_custom = (in_atomic) ? GFP_ATOMIC | __GFP_ACCOUNT : 0;
> > pcache.gfp_zero = __GFP_ZERO;
> >
> > end = (gpa + size + PAGE_SIZE - 1) & PAGE_MASK;
> > @@ -382,6 +384,13 @@ static int gstage_ioremap(struct kvm *kvm, gpa_t gpa, phys_addr_t hpa,
> > return ret;
> > }
> >
> > +void kvm_riscv_gstage_iounmap(struct kvm *kvm, gpa_t gpa, unsigned long size)
> > +{
> > + spin_lock(&kvm->mmu_lock);
> > + gstage_unmap_range(kvm, gpa, size, false);
> > + spin_unlock(&kvm->mmu_lock);
> > +}
> > +
> > void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
> > struct kvm_memory_slot *slot,
> > gfn_t gfn_offset,
> > @@ -517,8 +526,9 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
> > goto out;
> > }
> >
> > - ret = gstage_ioremap(kvm, gpa, pa,
> > - vm_end - vm_start, writable);
> > + ret = kvm_riscv_gstage_ioremap(kvm, gpa, pa,
> > + vm_end - vm_start,
> > + writable, false);
> > if (ret)
> > break;
> > }
> > --
> > 2.34.1
> >
>
> Reviewed-by: Atish Patra <atishp@xxxxxxxxxxxx>
Queued this patch for 5.20.
Thanks,
Anup
