Message ID | 20230801222629.210929-1-dbarboza@ventanamicro.com (mailing list archive) |
---|---|
Headers | show |
Series | RISC-V: KVM: change get_reg/set_reg error codes | expand |
On Tue, Aug 01, 2023 at 07:26:20PM -0300, Daniel Henrique Barboza wrote: > Hi, > > In this new version 3 new patches (6, 7, 8) were added by Andrew's > request during the v1 review. > > We're now avoiding throwing an -EBUSY error if a reg write is done after > the vcpu started spinning if the value being written is the same as KVM > already uses. This follows the design choice made in patch 3, allowing > for userspace 'lazy write' of registers. > > I decided to add 3 patches instead of one because the no-op check made > in patches 6 and 8 aren't just a matter of doing reg_val = host_val. > They can be squashed in a single patch if required. > > Please check the version 1 cover-letter [1] for the motivation behind > this work. Patches were based on top of riscv_kvm_queue. > > Changes from v1: > - patches 6,7, 8 (new): > - make reg writes a no-op, regardless of vcpu->arch.ran_atleast_once > state, if the value being written is the same as the host > - v1 link: https://lore.kernel.org/kvm/20230731120420.91007-1-dbarboza@ventanamicro.com/ > > [1] https://lore.kernel.org/kvm/20230731120420.91007-1-dbarboza@ventanamicro.com/ > I found three missing conversions, which are in the diff below. Also, I saw that the vector registers were lacking good error returns, so I reworked that and attached a completely (not even compile) tested patch for them. Please work that patch into this series. Thanks, drew diff --git a/arch/riscv/kvm/vcpu_onereg.c b/arch/riscv/kvm/vcpu_onereg.c index c88b0c7f7f01..6ca90c04ba61 100644 --- a/arch/riscv/kvm/vcpu_onereg.c +++ b/arch/riscv/kvm/vcpu_onereg.c @@ -506,7 +506,7 @@ static int riscv_vcpu_get_isa_ext_multi(struct kvm_vcpu *vcpu, unsigned long i, ext_id, ext_val; if (reg_num > KVM_REG_RISCV_ISA_MULTI_REG_LAST) - return -EINVAL; + return -ENOENT; for (i = 0; i < BITS_PER_LONG; i++) { ext_id = i + reg_num * BITS_PER_LONG; @@ -529,7 +529,7 @@ static int riscv_vcpu_set_isa_ext_multi(struct kvm_vcpu *vcpu, unsigned long i, ext_id; if (reg_num > KVM_REG_RISCV_ISA_MULTI_REG_LAST) - return -EINVAL; + return -ENOENT; for_each_set_bit(i, ®_val, BITS_PER_LONG) { ext_id = i + reg_num * BITS_PER_LONG; @@ -644,7 +644,7 @@ int kvm_riscv_vcpu_set_reg(struct kvm_vcpu *vcpu, break; } - return -EINVAL; + return -ENOENT; } int kvm_riscv_vcpu_get_reg(struct kvm_vcpu *vcpu, From b1472501677ea6bd73689e0f947149f5f86da9cc Mon Sep 17 00:00:00 2001 From: Andrew Jones <ajones@ventanamicro.com> Date: Wed, 2 Aug 2023 11:52:04 +0300 Subject: [PATCH] riscv: KVM: Improve vector save/restore errors Content-type: text/plain --- arch/riscv/kvm/vcpu_vector.c | 60 ++++++++++++++++++++---------------- 1 file changed, 33 insertions(+), 27 deletions(-) diff --git a/arch/riscv/kvm/vcpu_vector.c b/arch/riscv/kvm/vcpu_vector.c index edd2eecbddc2..39c5bceb4d1b 100644 --- a/arch/riscv/kvm/vcpu_vector.c +++ b/arch/riscv/kvm/vcpu_vector.c @@ -91,44 +91,44 @@ void kvm_riscv_vcpu_free_vector_context(struct kvm_vcpu *vcpu) } #endif -static void *kvm_riscv_vcpu_vreg_addr(struct kvm_vcpu *vcpu, +static int kvm_riscv_vcpu_vreg_addr(struct kvm_vcpu *vcpu, unsigned long reg_num, - size_t reg_size) + size_t reg_size, + void **reg_val) { struct kvm_cpu_context *cntx = &vcpu->arch.guest_context; - void *reg_val; size_t vlenb = riscv_v_vsize / 32; if (reg_num < KVM_REG_RISCV_VECTOR_REG(0)) { if (reg_size != sizeof(unsigned long)) - return NULL; + return -EINVAL; switch (reg_num) { case KVM_REG_RISCV_VECTOR_CSR_REG(vstart): - reg_val = &cntx->vector.vstart; + *reg_val = &cntx->vector.vstart; break; case KVM_REG_RISCV_VECTOR_CSR_REG(vl): - reg_val = &cntx->vector.vl; + *reg_val = &cntx->vector.vl; break; case KVM_REG_RISCV_VECTOR_CSR_REG(vtype): - reg_val = &cntx->vector.vtype; + *reg_val = &cntx->vector.vtype; break; case KVM_REG_RISCV_VECTOR_CSR_REG(vcsr): - reg_val = &cntx->vector.vcsr; + *reg_val = &cntx->vector.vcsr; break; case KVM_REG_RISCV_VECTOR_CSR_REG(datap): default: - return NULL; + return -ENOENT; } } else if (reg_num <= KVM_REG_RISCV_VECTOR_REG(31)) { if (reg_size != vlenb) - return NULL; - reg_val = cntx->vector.datap + return -EINVAL; + *reg_val = cntx->vector.datap + (reg_num - KVM_REG_RISCV_VECTOR_REG(0)) * vlenb; } else { - return NULL; + return -ENOENT; } - return reg_val; + return 0; } int kvm_riscv_vcpu_get_reg_vector(struct kvm_vcpu *vcpu, @@ -141,17 +141,20 @@ int kvm_riscv_vcpu_get_reg_vector(struct kvm_vcpu *vcpu, unsigned long reg_num = reg->id & ~(KVM_REG_ARCH_MASK | KVM_REG_SIZE_MASK | rtype); - void *reg_val = NULL; size_t reg_size = KVM_REG_SIZE(reg->id); + void *reg_val; + int rc; - if (rtype == KVM_REG_RISCV_VECTOR && - riscv_isa_extension_available(isa, v)) { - reg_val = kvm_riscv_vcpu_vreg_addr(vcpu, reg_num, reg_size); - } - - if (!reg_val) + if (rtype != KVM_REG_RISCV_VECTOR) return -EINVAL; + if (!riscv_isa_extension_available(isa, v)) + return -ENOENT; + + rc = kvm_riscv_vcpu_vreg_addr(vcpu, reg_num, reg_size, ®_val); + if (rc) + return rc; + if (copy_to_user(uaddr, reg_val, reg_size)) return -EFAULT; @@ -168,17 +171,20 @@ int kvm_riscv_vcpu_set_reg_vector(struct kvm_vcpu *vcpu, unsigned long reg_num = reg->id & ~(KVM_REG_ARCH_MASK | KVM_REG_SIZE_MASK | rtype); - void *reg_val = NULL; size_t reg_size = KVM_REG_SIZE(reg->id); + void *reg_val; + int rc; - if (rtype == KVM_REG_RISCV_VECTOR && - riscv_isa_extension_available(isa, v)) { - reg_val = kvm_riscv_vcpu_vreg_addr(vcpu, reg_num, reg_size); - } - - if (!reg_val) + if (rtype != KVM_REG_RISCV_VECTOR) return -EINVAL; + if (!riscv_isa_extension_available(isa, v)) + return -ENOENT; + + rc = kvm_riscv_vcpu_vreg_addr(vcpu, reg_num, reg_size, ®_val); + if (rc) + return rc; + if (copy_from_user(reg_val, uaddr, reg_size)) return -EFAULT;
On Wed, Aug 02, 2023 at 12:04:25PM +0300, Andrew Jones wrote: > On Tue, Aug 01, 2023 at 07:26:20PM -0300, Daniel Henrique Barboza wrote: > > Hi, > > > > In this new version 3 new patches (6, 7, 8) were added by Andrew's > > request during the v1 review. > > > > We're now avoiding throwing an -EBUSY error if a reg write is done after > > the vcpu started spinning if the value being written is the same as KVM > > already uses. This follows the design choice made in patch 3, allowing > > for userspace 'lazy write' of registers. > > > > I decided to add 3 patches instead of one because the no-op check made > > in patches 6 and 8 aren't just a matter of doing reg_val = host_val. > > They can be squashed in a single patch if required. > > > > Please check the version 1 cover-letter [1] for the motivation behind > > this work. Patches were based on top of riscv_kvm_queue. > > > > Changes from v1: > > - patches 6,7, 8 (new): > > - make reg writes a no-op, regardless of vcpu->arch.ran_atleast_once > > state, if the value being written is the same as the host > > - v1 link: https://lore.kernel.org/kvm/20230731120420.91007-1-dbarboza@ventanamicro.com/ > > > > [1] https://lore.kernel.org/kvm/20230731120420.91007-1-dbarboza@ventanamicro.com/ > > > > I found three missing conversions, which are in the diff below. Also, I > saw that the vector registers were lacking good error returns, so I reworked > that and attached a completely (not even compile) tested patch for them. ^un