On 21/03/17 11:05, Christoffer Dall wrote:
> There is no common code in the VGIC that uses the VMCR, so we have no
> use of the intermediate architecture-agnostic representation of the VMCR
> and might as well manipulate the bits specifically using the logic for
> the version of the GIC that the code supports.
>
> For GICv3, this means translating between the ICH_VMCR register format
> stored in memory and the ICC_X_EL1 registers exported to user space.
>
> Signed-off-by: Christoffer Dall <cdall@xxxxxxxxxx>
> ---
> arch/arm64/kvm/vgic-sys-reg-v3.c | 133 +++++++++++++++------------------------
> virt/kvm/arm/vgic/vgic-mmio.c | 4 +-
> virt/kvm/arm/vgic/vgic-v3.c | 38 -----------
> virt/kvm/arm/vgic/vgic.h | 2 -
> 4 files changed, 53 insertions(+), 124 deletions(-)
>
> diff --git a/arch/arm64/kvm/vgic-sys-reg-v3.c b/arch/arm64/kvm/vgic-sys-reg-v3.c
> index 33f111c..cd51433 100644
> --- a/arch/arm64/kvm/vgic-sys-reg-v3.c
> +++ b/arch/arm64/kvm/vgic-sys-reg-v3.c
> @@ -21,10 +21,9 @@
> static bool write_gic_ctlr(struct kvm_vcpu *vcpu, u32 val)
> {
> struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
> + u32 *vmcr = &vgic_cpu->vgic_v3.vgic_vmcr;
> u32 host_pri_bits, host_id_bits, host_seis, host_a3v, seis, a3v;
> - struct vgic_vmcr vmcr;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
> + u32 cbpr, eoimode;
>
> /*
> * Disallow restoring VM state if not supported by this
> @@ -57,24 +56,26 @@ static bool write_gic_ctlr(struct kvm_vcpu *vcpu, u32 val)
> if (host_a3v != a3v)
> return false;
>
> - /*
> - * Here set VMCR.CTLR in ICC_CTLR_EL1 layout.
> - * The vgic_set_vmcr() will convert to ICH_VMCR layout.
> - */
> - vmcr.ctlr = val & ICC_CTLR_EL1_CBPR_MASK;
> - vmcr.ctlr |= val & ICC_CTLR_EL1_EOImode_MASK;
> + /* Mask off any immutable bits from the ICC_CTLR_EL1 layout. */
> + eoimode = val & ICC_CTLR_EL1_EOImode_MASK;
> + cbpr = val & ICC_CTLR_EL1_CBPR_MASK;
>
> - vgic_set_vmcr(vcpu, &vmcr);
> + /* Convert the remaining bits to ICH_VMCR layout. */
> + *vmcr &= ~(ICH_VMCR_EOIM_MASK | ICH_VMCR_CBPR_MASK);
> + *vmcr |= (eoimode >> ICC_CTLR_EL1_EOImode_SHIFT) <<
> + ICH_VMCR_EOIM_SHIFT;
> + *vmcr |= (cbpr >> ICC_CTLR_EL1_CBPR_SHIFT) <<
> + ICH_VMCR_CBPR_SHIFT;
> +
> + *vmcr = val;
Something is wrong here. You can't possibly want to carefully clear and
insert bits, only to corrupt the whole thing later. Rework leftover?
> return true;
> }
>
> static u32 read_gic_ctlr(struct kvm_vcpu *vcpu)
> {
> struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
> + u32 vmcr = vgic_cpu->vgic_v3.vgic_vmcr;
> u32 val = 0;
> - struct vgic_vmcr vmcr;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
>
> val |= (vgic_cpu->num_pri_bits - 1) <<
> ICC_CTLR_EL1_PRI_BITS_SHIFT;
> @@ -85,12 +86,12 @@ static u32 read_gic_ctlr(struct kvm_vcpu *vcpu)
> val |= ((kvm_vgic_global_state.ich_vtr_el2 &
> ICH_VTR_A3V_MASK) >> ICH_VTR_A3V_SHIFT) <<
> ICC_CTLR_EL1_A3V_SHIFT;
> - /*
> - * The VMCR.CTLR value is in ICC_CTLR_EL1 layout.
> - * Extract it directly using ICC_CTLR_EL1 reg definitions.
> - */
> - val |= vmcr.ctlr & ICC_CTLR_EL1_CBPR_MASK;
> - val |= vmcr.ctlr & ICC_CTLR_EL1_EOImode_MASK;
> +
> + /* Convert the data in the ICH_VMCR_EL1 to the ICC_CTLR_EL1 layout */
> + val |= ((vmcr & ICH_VMCR_EOIM_MASK) >> ICH_VMCR_EOIM_SHIFT) <<
> + ICC_CTLR_EL1_EOImode_SHIFT;
> + val |= ((vmcr & ICH_VMCR_CBPR_MASK) >> ICH_VMCR_CBPR_SHIFT) <<
> + ICC_CTLR_EL1_CBPR_SHIFT;
>
> return val;
> }
> @@ -108,99 +109,67 @@ static bool access_gic_ctlr(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> return ret;
> }
>
> -static bool access_gic_pmr(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> - const struct sys_reg_desc *r)
> +static void access_vmcr_field(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> + u32 ich_mask, u32 ich_shift, u32 icc_shift)
> {
> - struct vgic_vmcr vmcr;
> + struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
> + u32 *vmcr = &vgic_cpu->vgic_v3.vgic_vmcr;
>
> - vgic_get_vmcr(vcpu, &vmcr);
> if (p->is_write) {
> - vmcr.pmr = (p->regval & ICC_PMR_EL1_MASK) >> ICC_PMR_EL1_SHIFT;
> - vgic_set_vmcr(vcpu, &vmcr);
> + *vmcr &= ~ich_mask;
> + *vmcr |= ((p->regval >> icc_shift) << ich_shift) & ich_mask;
> } else {
> - p->regval = (vmcr.pmr << ICC_PMR_EL1_SHIFT) & ICC_PMR_EL1_MASK;
> + p->regval = ((*vmcr & ich_mask) >> ich_shift) << icc_shift;
> }
> +}
>
> +static bool access_gic_pmr(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> + const struct sys_reg_desc *r)
> +{
> + access_vmcr_field(vcpu, p,
> + ICH_VMCR_PMR_MASK,
> + ICH_VMCR_PMR_SHIFT,
> + ICC_PMR_EL1_SHIFT);
> return true;
> }
>
> static bool access_gic_bpr0(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> const struct sys_reg_desc *r)
> {
> - struct vgic_vmcr vmcr;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
> - if (p->is_write) {
> - vmcr.bpr = (p->regval & ICC_BPR0_EL1_MASK) >>
> - ICC_BPR0_EL1_SHIFT;
> - vgic_set_vmcr(vcpu, &vmcr);
> - } else {
> - p->regval = (vmcr.bpr << ICC_BPR0_EL1_SHIFT) &
> - ICC_BPR0_EL1_MASK;
> - }
> -
> + access_vmcr_field(vcpu, p,
> + ICH_VMCR_BPR0_MASK,
> + ICH_VMCR_BPR0_SHIFT,
> + ICC_BPR0_EL1_SHIFT);
> return true;
> }
>
> static bool access_gic_bpr1(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> const struct sys_reg_desc *r)
> {
> - struct vgic_vmcr vmcr;
> -
> - if (!p->is_write)
> - p->regval = 0;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
> - if (!((vmcr.ctlr & ICH_VMCR_CBPR_MASK) >> ICH_VMCR_CBPR_SHIFT)) {
> - if (p->is_write) {
> - vmcr.abpr = (p->regval & ICC_BPR1_EL1_MASK) >>
> - ICC_BPR1_EL1_SHIFT;
> - vgic_set_vmcr(vcpu, &vmcr);
> - } else {
> - p->regval = (vmcr.abpr << ICC_BPR1_EL1_SHIFT) &
> - ICC_BPR1_EL1_MASK;
> - }
> - } else {
> - if (!p->is_write)
> - p->regval = min((vmcr.bpr + 1), 7U);
> - }
> -
> + access_vmcr_field(vcpu, p,
> + ICH_VMCR_BPR1_MASK,
> + ICH_VMCR_BPR1_SHIFT,
> + ICC_BPR1_EL1_SHIFT);
Aren't we loosing some of the existing semantics of reading BPR1 when
CBPR is set?
> return true;
> }
>
> static bool access_gic_grpen0(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> const struct sys_reg_desc *r)
> {
> - struct vgic_vmcr vmcr;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
> - if (p->is_write) {
> - vmcr.grpen0 = (p->regval & ICC_IGRPEN0_EL1_MASK) >>
> - ICC_IGRPEN0_EL1_SHIFT;
> - vgic_set_vmcr(vcpu, &vmcr);
> - } else {
> - p->regval = (vmcr.grpen0 << ICC_IGRPEN0_EL1_SHIFT) &
> - ICC_IGRPEN0_EL1_MASK;
> - }
> -
> + access_vmcr_field(vcpu, p,
> + ICH_VMCR_ENG0_MASK,
> + ICH_VMCR_ENG0_SHIFT,
> + ICC_IGRPEN0_EL1_SHIFT);
> return true;
> }
>
> static bool access_gic_grpen1(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
> const struct sys_reg_desc *r)
> {
> - struct vgic_vmcr vmcr;
> -
> - vgic_get_vmcr(vcpu, &vmcr);
> - if (p->is_write) {
> - vmcr.grpen1 = (p->regval & ICC_IGRPEN1_EL1_MASK) >>
> - ICC_IGRPEN1_EL1_SHIFT;
> - vgic_set_vmcr(vcpu, &vmcr);
> - } else {
> - p->regval = (vmcr.grpen1 << ICC_IGRPEN1_EL1_SHIFT) &
> - ICC_IGRPEN1_EL1_MASK;
> - }
> -
> + access_vmcr_field(vcpu, p,
> + ICH_VMCR_ENG1_MASK,
> + ICH_VMCR_ENG1_SHIFT,
> + ICC_IGRPEN1_EL1_SHIFT);
> return true;
> }
>
> diff --git a/virt/kvm/arm/vgic/vgic-mmio.c b/virt/kvm/arm/vgic/vgic-mmio.c
> index 3654b4c..b53c66e 100644
> --- a/virt/kvm/arm/vgic/vgic-mmio.c
> +++ b/virt/kvm/arm/vgic/vgic-mmio.c
> @@ -444,7 +444,7 @@ void vgic_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr)
> if (kvm_vgic_global_state.type == VGIC_V2)
> vgic_v2_set_vmcr(vcpu, vmcr);
> else
> - vgic_v3_set_vmcr(vcpu, vmcr);
> + BUG();
> }
>
> void vgic_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr)
> @@ -452,7 +452,7 @@ void vgic_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr)
> if (kvm_vgic_global_state.type == VGIC_V2)
> vgic_v2_get_vmcr(vcpu, vmcr);
> else
> - vgic_v3_get_vmcr(vcpu, vmcr);
> + BUG();
> }
>
> /*
> diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c
> index edc6ee2..4697f5d 100644
> --- a/virt/kvm/arm/vgic/vgic-v3.c
> +++ b/virt/kvm/arm/vgic/vgic-v3.c
> @@ -171,44 +171,6 @@ void vgic_v3_clear_lr(struct kvm_vcpu *vcpu, int lr)
> vcpu->arch.vgic_cpu.vgic_v3.vgic_lr[lr] = 0;
> }
>
> -void vgic_v3_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcrp)
> -{
> - u32 vmcr;
> -
> - /*
> - * Ignore the FIQen bit, because GIC emulation always implies
> - * SRE=1 which means the vFIQEn bit is also RES1.
> - */
> - vmcr = ((vmcrp->ctlr >> ICC_CTLR_EL1_EOImode_SHIFT) <<
> - ICH_VMCR_EOIM_SHIFT) & ICH_VMCR_EOIM_MASK;
> - vmcr |= (vmcrp->ctlr << ICH_VMCR_CBPR_SHIFT) & ICH_VMCR_CBPR_MASK;
> - vmcr |= (vmcrp->abpr << ICH_VMCR_BPR1_SHIFT) & ICH_VMCR_BPR1_MASK;
> - vmcr |= (vmcrp->bpr << ICH_VMCR_BPR0_SHIFT) & ICH_VMCR_BPR0_MASK;
> - vmcr |= (vmcrp->pmr << ICH_VMCR_PMR_SHIFT) & ICH_VMCR_PMR_MASK;
> - vmcr |= (vmcrp->grpen0 << ICH_VMCR_ENG0_SHIFT) & ICH_VMCR_ENG0_MASK;
> - vmcr |= (vmcrp->grpen1 << ICH_VMCR_ENG1_SHIFT) & ICH_VMCR_ENG1_MASK;
> -
> - vcpu->arch.vgic_cpu.vgic_v3.vgic_vmcr = vmcr;
> -}
> -
> -void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcrp)
> -{
> - u32 vmcr = vcpu->arch.vgic_cpu.vgic_v3.vgic_vmcr;
> -
> - /*
> - * Ignore the FIQen bit, because GIC emulation always implies
> - * SRE=1 which means the vFIQEn bit is also RES1.
> - */
> - vmcrp->ctlr = ((vmcr >> ICH_VMCR_EOIM_SHIFT) <<
> - ICC_CTLR_EL1_EOImode_SHIFT) & ICC_CTLR_EL1_EOImode_MASK;
> - vmcrp->ctlr |= (vmcr & ICH_VMCR_CBPR_MASK) >> ICH_VMCR_CBPR_SHIFT;
> - vmcrp->abpr = (vmcr & ICH_VMCR_BPR1_MASK) >> ICH_VMCR_BPR1_SHIFT;
> - vmcrp->bpr = (vmcr & ICH_VMCR_BPR0_MASK) >> ICH_VMCR_BPR0_SHIFT;
> - vmcrp->pmr = (vmcr & ICH_VMCR_PMR_MASK) >> ICH_VMCR_PMR_SHIFT;
> - vmcrp->grpen0 = (vmcr & ICH_VMCR_ENG0_MASK) >> ICH_VMCR_ENG0_SHIFT;
> - vmcrp->grpen1 = (vmcr & ICH_VMCR_ENG1_MASK) >> ICH_VMCR_ENG1_SHIFT;
> -}
> -
> #define INITIAL_PENDBASER_VALUE \
> (GIC_BASER_CACHEABILITY(GICR_PENDBASER, INNER, RaWb) | \
> GIC_BASER_CACHEABILITY(GICR_PENDBASER, OUTER, SameAsInner) | \
> diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h
> index db28f7c..5652983 100644
> --- a/virt/kvm/arm/vgic/vgic.h
> +++ b/virt/kvm/arm/vgic/vgic.h
> @@ -143,8 +143,6 @@ void vgic_v3_fold_lr_state(struct kvm_vcpu *vcpu);
> void vgic_v3_populate_lr(struct kvm_vcpu *vcpu, struct vgic_irq *irq, int lr);
> void vgic_v3_clear_lr(struct kvm_vcpu *vcpu, int lr);
> void vgic_v3_set_underflow(struct kvm_vcpu *vcpu);
> -void vgic_v3_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
> -void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
> void vgic_v3_enable(struct kvm_vcpu *vcpu);
> int vgic_v3_probe(const struct gic_kvm_info *info);
> int vgic_v3_map_resources(struct kvm *kvm);
>
Thanks,
M.
--
Jazz is not dead. It just smells funny...
