| Message ID | 20240418142701.1493091-8-cleger@rivosinc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | riscv: Add support for Ssdbltrp extension | expand |
On Thu, Apr 18, 2024 at 04:26:46PM +0200, Clément Léger wrote: >When a double trap exception is generated from VS-mode, it should be >delivered to M-mode which might redirect it to S-mode. Currently, the >kvm double trap exception handling simply prints an error and returns >-EOPNOTSUPP to stop VM execution. In future, this might use KVM SBI SSE >extension implementation to actually send an SSE event to the guest VM. > >Signed-off-by: Clément Léger <cleger@rivosinc.com> >--- > arch/riscv/include/asm/kvm_host.h | 7 ++++--- > arch/riscv/include/uapi/asm/kvm.h | 1 + > arch/riscv/kvm/vcpu.c | 23 +++++++++------------ > arch/riscv/kvm/vcpu_exit.c | 33 +++++++++++++++++++++++++------ > arch/riscv/kvm/vcpu_insn.c | 15 +++++--------- > arch/riscv/kvm/vcpu_onereg.c | 2 ++ > arch/riscv/kvm/vcpu_sbi.c | 4 +--- > arch/riscv/kvm/vcpu_switch.S | 19 +++++++++++++++--- > 8 files changed, 65 insertions(+), 39 deletions(-) > >diff --git a/arch/riscv/include/asm/kvm_host.h b/arch/riscv/include/asm/kvm_host.h >index be60aaa07f57..1d699bf44c45 100644 >--- a/arch/riscv/include/asm/kvm_host.h >+++ b/arch/riscv/include/asm/kvm_host.h >@@ -358,12 +358,13 @@ unsigned long kvm_riscv_vcpu_unpriv_read(struct kvm_vcpu *vcpu, > bool read_insn, > unsigned long guest_addr, > struct kvm_cpu_trap *trap); >-void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >- struct kvm_cpu_trap *trap); >+int kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >+ struct kvm_cpu_trap *trap); > int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, > struct kvm_cpu_trap *trap); > >-void __kvm_riscv_switch_to(struct kvm_vcpu_arch *vcpu_arch); >+void __kvm_riscv_switch_to(struct kvm_vcpu_arch *vcpu_arch, >+ struct kvm_cpu_trap *trap); > > void kvm_riscv_vcpu_setup_isa(struct kvm_vcpu *vcpu); > unsigned long kvm_riscv_vcpu_num_regs(struct kvm_vcpu *vcpu); >diff --git a/arch/riscv/include/uapi/asm/kvm.h b/arch/riscv/include/uapi/asm/kvm.h >index fa3097da91c0..323f4e8380d2 100644 >--- a/arch/riscv/include/uapi/asm/kvm.h >+++ b/arch/riscv/include/uapi/asm/kvm.h >@@ -166,6 +166,7 @@ enum KVM_RISCV_ISA_EXT_ID { > KVM_RISCV_ISA_EXT_ZVFH, > KVM_RISCV_ISA_EXT_ZVFHMIN, > KVM_RISCV_ISA_EXT_ZFA, >+ KVM_RISCV_ISA_EXT_SSDBLTRP, > KVM_RISCV_ISA_EXT_MAX, > }; > >diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c >index 461ef60d4eda..89e663defe14 100644 >--- a/arch/riscv/kvm/vcpu.c >+++ b/arch/riscv/kvm/vcpu.c >@@ -121,6 +121,8 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu) > /* Setup reset state of shadow SSTATUS and HSTATUS CSRs */ > cntx = &vcpu->arch.guest_reset_context; > cntx->sstatus = SR_SPP | SR_SPIE; >+ if (riscv_isa_extension_available(vcpu->arch.isa, SSDBLTRP)) >+ cntx->sstatus |= SR_SDT; > cntx->hstatus = 0; > cntx->hstatus |= HSTATUS_VTW; > cntx->hstatus |= HSTATUS_SPVP; >@@ -579,6 +581,9 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) > csr->hvip = csr_read(CSR_HVIP); > csr->vsatp = csr_read(CSR_VSATP); > cfg->hedeleg = csr_read(CSR_HEDELEG); >+ cfg->henvcfg = csr_read(CSR_HENVCFG); >+ if (IS_ENABLED(CONFIG_32BIT)) >+ cfg->henvcfg = csr_read(CSR_HENVCFGH) << 32; > } > > static void kvm_riscv_check_vcpu_requests(struct kvm_vcpu *vcpu) >@@ -670,11 +675,12 @@ static __always_inline void kvm_riscv_vcpu_swap_in_host_state(struct kvm_vcpu *v > * This must be noinstr as instrumentation may make use of RCU, and this is not > * safe during the EQS. > */ >-static void noinstr kvm_riscv_vcpu_enter_exit(struct kvm_vcpu *vcpu) >+static void noinstr kvm_riscv_vcpu_enter_exit(struct kvm_vcpu *vcpu, >+ struct kvm_cpu_trap *trap) > { > kvm_riscv_vcpu_swap_in_guest_state(vcpu); > guest_state_enter_irqoff(); >- __kvm_riscv_switch_to(&vcpu->arch); >+ __kvm_riscv_switch_to(&vcpu->arch, trap); > vcpu->arch.last_exit_cpu = vcpu->cpu; > guest_state_exit_irqoff(); > kvm_riscv_vcpu_swap_in_host_state(vcpu); >@@ -789,22 +795,11 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) > > guest_timing_enter_irqoff(); > >- kvm_riscv_vcpu_enter_exit(vcpu); >+ kvm_riscv_vcpu_enter_exit(vcpu, &trap); > > vcpu->mode = OUTSIDE_GUEST_MODE; > vcpu->stat.exits++; > >- /* >- * Save SCAUSE, STVAL, HTVAL, and HTINST because we might >- * get an interrupt between __kvm_riscv_switch_to() and >- * local_irq_enable() which can potentially change CSRs. >- */ >- trap.sepc = vcpu->arch.guest_context.sepc; >- trap.scause = csr_read(CSR_SCAUSE); >- trap.stval = csr_read(CSR_STVAL); >- trap.htval = csr_read(CSR_HTVAL); >- trap.htinst = csr_read(CSR_HTINST); >- > /* Syncup interrupts state with HW */ > kvm_riscv_vcpu_sync_interrupts(vcpu); > >diff --git a/arch/riscv/kvm/vcpu_exit.c b/arch/riscv/kvm/vcpu_exit.c >index 2415722c01b8..892c6df97eaf 100644 >--- a/arch/riscv/kvm/vcpu_exit.c >+++ b/arch/riscv/kvm/vcpu_exit.c >@@ -126,17 +126,34 @@ unsigned long kvm_riscv_vcpu_unpriv_read(struct kvm_vcpu *vcpu, > return val; > } > >+static int kvm_riscv_double_trap(struct kvm_vcpu *vcpu, >+ struct kvm_cpu_trap *trap) >+{ >+ pr_err("Guest double trap"); >+ /* TODO: Implement SSE support */ >+ >+ return -EOPNOTSUPP; >+} >+ > /** > * kvm_riscv_vcpu_trap_redirect -- Redirect trap to Guest > * > * @vcpu: The VCPU pointer > * @trap: Trap details > */ >-void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >- struct kvm_cpu_trap *trap) >+int kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >+ struct kvm_cpu_trap *trap) > { > unsigned long vsstatus = csr_read(CSR_VSSTATUS); > >+ if (riscv_isa_extension_available(vcpu->arch.isa, SSDBLTRP)) { >+ if (vsstatus & SR_SDT) >+ return kvm_riscv_double_trap(vcpu, trap); >+ >+ /* Set Double Trap bit to enable double trap detection */ >+ vsstatus |= SR_SDT; I didn't get it. Why do this without checking if current config allows us to do so ? I am imagining we do this only when henvcfg for current vcpu says that DTE=1 for this this guest, right? >+ } >+ > /* Change Guest SSTATUS.SPP bit */ > vsstatus &= ~SR_SPP; > if (vcpu->arch.guest_context.sstatus & SR_SPP) >@@ -163,6 +180,8 @@ void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, > > /* Set Guest privilege mode to supervisor */ > vcpu->arch.guest_context.sstatus |= SR_SPP; >+ >+ return 1; > } > > /* >@@ -185,10 +204,8 @@ int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, > case EXC_INST_ILLEGAL: > case EXC_LOAD_MISALIGNED: > case EXC_STORE_MISALIGNED: >- if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) { >- kvm_riscv_vcpu_trap_redirect(vcpu, trap); >- ret = 1; >- } >+ if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) >+ ret = kvm_riscv_vcpu_trap_redirect(vcpu, trap); > break; > case EXC_VIRTUAL_INST_FAULT: > if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) >@@ -204,6 +221,10 @@ int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, > if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) > ret = kvm_riscv_vcpu_sbi_ecall(vcpu, run); > break; >+ case EXC_DOUBLE_TRAP: >+ if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) >+ ret = kvm_riscv_double_trap(vcpu, trap); >+ break; > default: > break; > } >diff --git a/arch/riscv/kvm/vcpu_insn.c b/arch/riscv/kvm/vcpu_insn.c >index 7a6abed41bc1..050e811204f2 100644 >--- a/arch/riscv/kvm/vcpu_insn.c >+++ b/arch/riscv/kvm/vcpu_insn.c >@@ -159,9 +159,8 @@ static int truly_illegal_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, > utrap.stval = insn; > utrap.htval = 0; > utrap.htinst = 0; >- kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > >- return 1; >+ return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > } > > static int truly_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, >@@ -175,9 +174,8 @@ static int truly_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, > utrap.stval = insn; > utrap.htval = 0; > utrap.htinst = 0; >- kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > >- return 1; >+ return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > } > > /** >@@ -422,8 +420,7 @@ int kvm_riscv_vcpu_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, > &utrap); > if (utrap.scause) { > utrap.sepc = ct->sepc; >- kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); >- return 1; >+ return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > } > } > if (INSN_IS_16BIT(insn)) >@@ -478,8 +475,7 @@ int kvm_riscv_vcpu_mmio_load(struct kvm_vcpu *vcpu, struct kvm_run *run, > if (utrap.scause) { > /* Redirect trap if we failed to read instruction */ > utrap.sepc = ct->sepc; >- kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); >- return 1; >+ return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > } > insn_len = INSN_LEN(insn); > } >@@ -604,8 +600,7 @@ int kvm_riscv_vcpu_mmio_store(struct kvm_vcpu *vcpu, struct kvm_run *run, > if (utrap.scause) { > /* Redirect trap if we failed to read instruction */ > utrap.sepc = ct->sepc; >- kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); >- return 1; >+ return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); > } > insn_len = INSN_LEN(insn); > } >diff --git a/arch/riscv/kvm/vcpu_onereg.c b/arch/riscv/kvm/vcpu_onereg.c >index 5f7355e96008..fece0043871c 100644 >--- a/arch/riscv/kvm/vcpu_onereg.c >+++ b/arch/riscv/kvm/vcpu_onereg.c >@@ -36,6 +36,7 @@ static const unsigned long kvm_isa_ext_arr[] = { > /* Multi letter extensions (alphabetically sorted) */ > KVM_ISA_EXT_ARR(SMSTATEEN), > KVM_ISA_EXT_ARR(SSAIA), >+ KVM_ISA_EXT_ARR(SSDBLTRP), > KVM_ISA_EXT_ARR(SSTC), > KVM_ISA_EXT_ARR(SVINVAL), > KVM_ISA_EXT_ARR(SVNAPOT), >@@ -153,6 +154,7 @@ static bool kvm_riscv_vcpu_isa_disable_allowed(unsigned long ext) > case KVM_RISCV_ISA_EXT_ZVKSED: > case KVM_RISCV_ISA_EXT_ZVKSH: > case KVM_RISCV_ISA_EXT_ZVKT: >+ case KVM_RISCV_ISA_EXT_SSDBLTRP: > return false; > /* Extensions which can be disabled using Smstateen */ > case KVM_RISCV_ISA_EXT_SSAIA: >diff --git a/arch/riscv/kvm/vcpu_sbi.c b/arch/riscv/kvm/vcpu_sbi.c >index 76901f0f34b7..b839d578dc26 100644 >--- a/arch/riscv/kvm/vcpu_sbi.c >+++ b/arch/riscv/kvm/vcpu_sbi.c >@@ -456,10 +456,8 @@ int kvm_riscv_vcpu_sbi_ecall(struct kvm_vcpu *vcpu, struct kvm_run *run) > > /* Handle special error cases i.e trap, exit or userspace forward */ > if (sbi_ret.utrap->scause) { >- /* No need to increment sepc or exit ioctl loop */ >- ret = 1; > sbi_ret.utrap->sepc = cp->sepc; >- kvm_riscv_vcpu_trap_redirect(vcpu, sbi_ret.utrap); >+ ret = kvm_riscv_vcpu_trap_redirect(vcpu, sbi_ret.utrap); > next_sepc = false; > goto ecall_done; > } >diff --git a/arch/riscv/kvm/vcpu_switch.S b/arch/riscv/kvm/vcpu_switch.S >index 0c26189aa01c..94d5eb9da788 100644 >--- a/arch/riscv/kvm/vcpu_switch.S >+++ b/arch/riscv/kvm/vcpu_switch.S >@@ -154,7 +154,6 @@ SYM_FUNC_START(__kvm_riscv_switch_to) > REG_L t2, (KVM_ARCH_HOST_SSCRATCH)(a0) > REG_L t3, (KVM_ARCH_HOST_SCOUNTEREN)(a0) > REG_L t4, (KVM_ARCH_HOST_HSTATUS)(a0) >- REG_L t5, (KVM_ARCH_HOST_SSTATUS)(a0) > > /* Save Guest SEPC */ > csrr t0, CSR_SEPC >@@ -171,8 +170,8 @@ SYM_FUNC_START(__kvm_riscv_switch_to) > /* Save Guest and Restore Host HSTATUS */ > csrrw t4, CSR_HSTATUS, t4 > >- /* Save Guest and Restore Host SSTATUS */ >- csrrw t5, CSR_SSTATUS, t5 >+ /* Save Guest SSTATUS */ >+ csrr t5, CSR_SSTATUS > > /* Store Guest CSR values */ > REG_S t0, (KVM_ARCH_GUEST_SEPC)(a0) >@@ -206,6 +205,20 @@ SYM_FUNC_START(__kvm_riscv_switch_to) > REG_L s10, (KVM_ARCH_HOST_S10)(a0) > REG_L s11, (KVM_ARCH_HOST_S11)(a0) > >+ csrr t1, CSR_SCAUSE >+ csrr t2, CSR_STVAL >+ csrr t3, CSR_HTVAL >+ csrr t4, CSR_HTINST >+ REG_S t0, (KVM_ARCH_TRAP_SEPC)(a1) >+ REG_S t1, (KVM_ARCH_TRAP_SCAUSE)(a1) >+ REG_S t2, (KVM_ARCH_TRAP_STVAL)(a1) >+ REG_S t3, (KVM_ARCH_TRAP_HTVAL)(a1) >+ REG_S t4, (KVM_ARCH_TRAP_HTINST)(a1) >+ >+ /* Restore Host SSTATUS */ >+ REG_L t5, (KVM_ARCH_HOST_SSTATUS)(a0) >+ csrw CSR_SSTATUS, t5 >+ > /* Return to C code */ > ret > SYM_FUNC_END(__kvm_riscv_switch_to) >-- >2.43.0 > >
On 27/04/2024 03:33, Deepak Gupta wrote: >> /** >> * kvm_riscv_vcpu_trap_redirect -- Redirect trap to Guest >> * >> * @vcpu: The VCPU pointer >> * @trap: Trap details >> */ >> -void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >> - struct kvm_cpu_trap *trap) >> +int kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, >> + struct kvm_cpu_trap *trap) >> { >> unsigned long vsstatus = csr_read(CSR_VSSTATUS); >> >> + if (riscv_isa_extension_available(vcpu->arch.isa, SSDBLTRP)) { >> + if (vsstatus & SR_SDT) >> + return kvm_riscv_double_trap(vcpu, trap); >> + >> + /* Set Double Trap bit to enable double trap detection */ >> + vsstatus |= SR_SDT; > > I didn't get it. > Why do this without checking if current config allows us to do so ? > I am imagining we do this only when henvcfg for current vcpu says that > DTE=1 > for this this guest, right? We actually do not really care since if not enabled, this bit will be masked by the hardware when writing it to the CSR. I could indeed add a check for that but in the end, it will yield the same result. In that case, just ORing it with the value seems more efficient than adding an "if" to check if the extension and DTE is enabled. Clément
diff --git a/arch/riscv/include/asm/kvm_host.h b/arch/riscv/include/asm/kvm_host.h index be60aaa07f57..1d699bf44c45 100644 --- a/arch/riscv/include/asm/kvm_host.h +++ b/arch/riscv/include/asm/kvm_host.h @@ -358,12 +358,13 @@ unsigned long kvm_riscv_vcpu_unpriv_read(struct kvm_vcpu *vcpu, bool read_insn, unsigned long guest_addr, struct kvm_cpu_trap *trap); -void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, - struct kvm_cpu_trap *trap); +int kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, + struct kvm_cpu_trap *trap); int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, struct kvm_cpu_trap *trap); -void __kvm_riscv_switch_to(struct kvm_vcpu_arch *vcpu_arch); +void __kvm_riscv_switch_to(struct kvm_vcpu_arch *vcpu_arch, + struct kvm_cpu_trap *trap); void kvm_riscv_vcpu_setup_isa(struct kvm_vcpu *vcpu); unsigned long kvm_riscv_vcpu_num_regs(struct kvm_vcpu *vcpu); diff --git a/arch/riscv/include/uapi/asm/kvm.h b/arch/riscv/include/uapi/asm/kvm.h index fa3097da91c0..323f4e8380d2 100644 --- a/arch/riscv/include/uapi/asm/kvm.h +++ b/arch/riscv/include/uapi/asm/kvm.h @@ -166,6 +166,7 @@ enum KVM_RISCV_ISA_EXT_ID { KVM_RISCV_ISA_EXT_ZVFH, KVM_RISCV_ISA_EXT_ZVFHMIN, KVM_RISCV_ISA_EXT_ZFA, + KVM_RISCV_ISA_EXT_SSDBLTRP, KVM_RISCV_ISA_EXT_MAX, }; diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c index 461ef60d4eda..89e663defe14 100644 --- a/arch/riscv/kvm/vcpu.c +++ b/arch/riscv/kvm/vcpu.c @@ -121,6 +121,8 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu) /* Setup reset state of shadow SSTATUS and HSTATUS CSRs */ cntx = &vcpu->arch.guest_reset_context; cntx->sstatus = SR_SPP | SR_SPIE; + if (riscv_isa_extension_available(vcpu->arch.isa, SSDBLTRP)) + cntx->sstatus |= SR_SDT; cntx->hstatus = 0; cntx->hstatus |= HSTATUS_VTW; cntx->hstatus |= HSTATUS_SPVP; @@ -579,6 +581,9 @@ void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) csr->hvip = csr_read(CSR_HVIP); csr->vsatp = csr_read(CSR_VSATP); cfg->hedeleg = csr_read(CSR_HEDELEG); + cfg->henvcfg = csr_read(CSR_HENVCFG); + if (IS_ENABLED(CONFIG_32BIT)) + cfg->henvcfg = csr_read(CSR_HENVCFGH) << 32; } static void kvm_riscv_check_vcpu_requests(struct kvm_vcpu *vcpu) @@ -670,11 +675,12 @@ static __always_inline void kvm_riscv_vcpu_swap_in_host_state(struct kvm_vcpu *v * This must be noinstr as instrumentation may make use of RCU, and this is not * safe during the EQS. */ -static void noinstr kvm_riscv_vcpu_enter_exit(struct kvm_vcpu *vcpu) +static void noinstr kvm_riscv_vcpu_enter_exit(struct kvm_vcpu *vcpu, + struct kvm_cpu_trap *trap) { kvm_riscv_vcpu_swap_in_guest_state(vcpu); guest_state_enter_irqoff(); - __kvm_riscv_switch_to(&vcpu->arch); + __kvm_riscv_switch_to(&vcpu->arch, trap); vcpu->arch.last_exit_cpu = vcpu->cpu; guest_state_exit_irqoff(); kvm_riscv_vcpu_swap_in_host_state(vcpu); @@ -789,22 +795,11 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) guest_timing_enter_irqoff(); - kvm_riscv_vcpu_enter_exit(vcpu); + kvm_riscv_vcpu_enter_exit(vcpu, &trap); vcpu->mode = OUTSIDE_GUEST_MODE; vcpu->stat.exits++; - /* - * Save SCAUSE, STVAL, HTVAL, and HTINST because we might - * get an interrupt between __kvm_riscv_switch_to() and - * local_irq_enable() which can potentially change CSRs. - */ - trap.sepc = vcpu->arch.guest_context.sepc; - trap.scause = csr_read(CSR_SCAUSE); - trap.stval = csr_read(CSR_STVAL); - trap.htval = csr_read(CSR_HTVAL); - trap.htinst = csr_read(CSR_HTINST); - /* Syncup interrupts state with HW */ kvm_riscv_vcpu_sync_interrupts(vcpu); diff --git a/arch/riscv/kvm/vcpu_exit.c b/arch/riscv/kvm/vcpu_exit.c index 2415722c01b8..892c6df97eaf 100644 --- a/arch/riscv/kvm/vcpu_exit.c +++ b/arch/riscv/kvm/vcpu_exit.c @@ -126,17 +126,34 @@ unsigned long kvm_riscv_vcpu_unpriv_read(struct kvm_vcpu *vcpu, return val; } +static int kvm_riscv_double_trap(struct kvm_vcpu *vcpu, + struct kvm_cpu_trap *trap) +{ + pr_err("Guest double trap"); + /* TODO: Implement SSE support */ + + return -EOPNOTSUPP; +} + /** * kvm_riscv_vcpu_trap_redirect -- Redirect trap to Guest * * @vcpu: The VCPU pointer * @trap: Trap details */ -void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, - struct kvm_cpu_trap *trap) +int kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, + struct kvm_cpu_trap *trap) { unsigned long vsstatus = csr_read(CSR_VSSTATUS); + if (riscv_isa_extension_available(vcpu->arch.isa, SSDBLTRP)) { + if (vsstatus & SR_SDT) + return kvm_riscv_double_trap(vcpu, trap); + + /* Set Double Trap bit to enable double trap detection */ + vsstatus |= SR_SDT; + } + /* Change Guest SSTATUS.SPP bit */ vsstatus &= ~SR_SPP; if (vcpu->arch.guest_context.sstatus & SR_SPP) @@ -163,6 +180,8 @@ void kvm_riscv_vcpu_trap_redirect(struct kvm_vcpu *vcpu, /* Set Guest privilege mode to supervisor */ vcpu->arch.guest_context.sstatus |= SR_SPP; + + return 1; } /* @@ -185,10 +204,8 @@ int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, case EXC_INST_ILLEGAL: case EXC_LOAD_MISALIGNED: case EXC_STORE_MISALIGNED: - if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) { - kvm_riscv_vcpu_trap_redirect(vcpu, trap); - ret = 1; - } + if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) + ret = kvm_riscv_vcpu_trap_redirect(vcpu, trap); break; case EXC_VIRTUAL_INST_FAULT: if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) @@ -204,6 +221,10 @@ int kvm_riscv_vcpu_exit(struct kvm_vcpu *vcpu, struct kvm_run *run, if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) ret = kvm_riscv_vcpu_sbi_ecall(vcpu, run); break; + case EXC_DOUBLE_TRAP: + if (vcpu->arch.guest_context.hstatus & HSTATUS_SPV) + ret = kvm_riscv_double_trap(vcpu, trap); + break; default: break; } diff --git a/arch/riscv/kvm/vcpu_insn.c b/arch/riscv/kvm/vcpu_insn.c index 7a6abed41bc1..050e811204f2 100644 --- a/arch/riscv/kvm/vcpu_insn.c +++ b/arch/riscv/kvm/vcpu_insn.c @@ -159,9 +159,8 @@ static int truly_illegal_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, utrap.stval = insn; utrap.htval = 0; utrap.htinst = 0; - kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); - return 1; + return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); } static int truly_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, @@ -175,9 +174,8 @@ static int truly_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, utrap.stval = insn; utrap.htval = 0; utrap.htinst = 0; - kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); - return 1; + return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); } /** @@ -422,8 +420,7 @@ int kvm_riscv_vcpu_virtual_insn(struct kvm_vcpu *vcpu, struct kvm_run *run, &utrap); if (utrap.scause) { utrap.sepc = ct->sepc; - kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); - return 1; + return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); } } if (INSN_IS_16BIT(insn)) @@ -478,8 +475,7 @@ int kvm_riscv_vcpu_mmio_load(struct kvm_vcpu *vcpu, struct kvm_run *run, if (utrap.scause) { /* Redirect trap if we failed to read instruction */ utrap.sepc = ct->sepc; - kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); - return 1; + return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); } insn_len = INSN_LEN(insn); } @@ -604,8 +600,7 @@ int kvm_riscv_vcpu_mmio_store(struct kvm_vcpu *vcpu, struct kvm_run *run, if (utrap.scause) { /* Redirect trap if we failed to read instruction */ utrap.sepc = ct->sepc; - kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); - return 1; + return kvm_riscv_vcpu_trap_redirect(vcpu, &utrap); } insn_len = INSN_LEN(insn); } diff --git a/arch/riscv/kvm/vcpu_onereg.c b/arch/riscv/kvm/vcpu_onereg.c index 5f7355e96008..fece0043871c 100644 --- a/arch/riscv/kvm/vcpu_onereg.c +++ b/arch/riscv/kvm/vcpu_onereg.c @@ -36,6 +36,7 @@ static const unsigned long kvm_isa_ext_arr[] = { /* Multi letter extensions (alphabetically sorted) */ KVM_ISA_EXT_ARR(SMSTATEEN), KVM_ISA_EXT_ARR(SSAIA), + KVM_ISA_EXT_ARR(SSDBLTRP), KVM_ISA_EXT_ARR(SSTC), KVM_ISA_EXT_ARR(SVINVAL), KVM_ISA_EXT_ARR(SVNAPOT), @@ -153,6 +154,7 @@ static bool kvm_riscv_vcpu_isa_disable_allowed(unsigned long ext) case KVM_RISCV_ISA_EXT_ZVKSED: case KVM_RISCV_ISA_EXT_ZVKSH: case KVM_RISCV_ISA_EXT_ZVKT: + case KVM_RISCV_ISA_EXT_SSDBLTRP: return false; /* Extensions which can be disabled using Smstateen */ case KVM_RISCV_ISA_EXT_SSAIA: diff --git a/arch/riscv/kvm/vcpu_sbi.c b/arch/riscv/kvm/vcpu_sbi.c index 76901f0f34b7..b839d578dc26 100644 --- a/arch/riscv/kvm/vcpu_sbi.c +++ b/arch/riscv/kvm/vcpu_sbi.c @@ -456,10 +456,8 @@ int kvm_riscv_vcpu_sbi_ecall(struct kvm_vcpu *vcpu, struct kvm_run *run) /* Handle special error cases i.e trap, exit or userspace forward */ if (sbi_ret.utrap->scause) { - /* No need to increment sepc or exit ioctl loop */ - ret = 1; sbi_ret.utrap->sepc = cp->sepc; - kvm_riscv_vcpu_trap_redirect(vcpu, sbi_ret.utrap); + ret = kvm_riscv_vcpu_trap_redirect(vcpu, sbi_ret.utrap); next_sepc = false; goto ecall_done; } diff --git a/arch/riscv/kvm/vcpu_switch.S b/arch/riscv/kvm/vcpu_switch.S index 0c26189aa01c..94d5eb9da788 100644 --- a/arch/riscv/kvm/vcpu_switch.S +++ b/arch/riscv/kvm/vcpu_switch.S @@ -154,7 +154,6 @@ SYM_FUNC_START(__kvm_riscv_switch_to) REG_L t2, (KVM_ARCH_HOST_SSCRATCH)(a0) REG_L t3, (KVM_ARCH_HOST_SCOUNTEREN)(a0) REG_L t4, (KVM_ARCH_HOST_HSTATUS)(a0) - REG_L t5, (KVM_ARCH_HOST_SSTATUS)(a0) /* Save Guest SEPC */ csrr t0, CSR_SEPC @@ -171,8 +170,8 @@ SYM_FUNC_START(__kvm_riscv_switch_to) /* Save Guest and Restore Host HSTATUS */ csrrw t4, CSR_HSTATUS, t4 - /* Save Guest and Restore Host SSTATUS */ - csrrw t5, CSR_SSTATUS, t5 + /* Save Guest SSTATUS */ + csrr t5, CSR_SSTATUS /* Store Guest CSR values */ REG_S t0, (KVM_ARCH_GUEST_SEPC)(a0) @@ -206,6 +205,20 @@ SYM_FUNC_START(__kvm_riscv_switch_to) REG_L s10, (KVM_ARCH_HOST_S10)(a0) REG_L s11, (KVM_ARCH_HOST_S11)(a0) + csrr t1, CSR_SCAUSE + csrr t2, CSR_STVAL + csrr t3, CSR_HTVAL + csrr t4, CSR_HTINST + REG_S t0, (KVM_ARCH_TRAP_SEPC)(a1) + REG_S t1, (KVM_ARCH_TRAP_SCAUSE)(a1) + REG_S t2, (KVM_ARCH_TRAP_STVAL)(a1) + REG_S t3, (KVM_ARCH_TRAP_HTVAL)(a1) + REG_S t4, (KVM_ARCH_TRAP_HTINST)(a1) + + /* Restore Host SSTATUS */ + REG_L t5, (KVM_ARCH_HOST_SSTATUS)(a0) + csrw CSR_SSTATUS, t5 + /* Return to C code */ ret SYM_FUNC_END(__kvm_riscv_switch_to)
When a double trap exception is generated from VS-mode, it should be delivered to M-mode which might redirect it to S-mode. Currently, the kvm double trap exception handling simply prints an error and returns -EOPNOTSUPP to stop VM execution. In future, this might use KVM SBI SSE extension implementation to actually send an SSE event to the guest VM. Signed-off-by: Clément Léger <cleger@rivosinc.com> --- arch/riscv/include/asm/kvm_host.h | 7 ++++--- arch/riscv/include/uapi/asm/kvm.h | 1 + arch/riscv/kvm/vcpu.c | 23 +++++++++------------ arch/riscv/kvm/vcpu_exit.c | 33 +++++++++++++++++++++++++------ arch/riscv/kvm/vcpu_insn.c | 15 +++++--------- arch/riscv/kvm/vcpu_onereg.c | 2 ++ arch/riscv/kvm/vcpu_sbi.c | 4 +--- arch/riscv/kvm/vcpu_switch.S | 19 +++++++++++++++--- 8 files changed, 65 insertions(+), 39 deletions(-)