Re: [PATCH v3 09/29] LoongArch: KVM: Implement vcpu get, vcpu set registers

From: maobibo
Date: Mon Mar 06 2023 - 21:24:03 EST

Next message: Po-Hsu Lin: "Re: [PATCH] selftests: net: devlink_port_split.py: skip test if no suitable device available"
Previous message: ChiYuan Huang: "Re: [PATCH v17 RESEND 1/3] leds: rgb: mt6370: Add MediaTek MT6370 current sink type LED Indicator support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

在 2023/2/28 15:00, Tianrui Zhao 写道:
> Implement loongarch vcpu get registers and set registers operations, it
> is called when user space use the ioctl interface to get or set regs.
>
> Signed-off-by: Tianrui Zhao <zhaotianrui@xxxxxxxxxxx>
> ---
> arch/loongarch/kvm/vcpu.c | 375 ++++++++++++++++++++++++++++++++++++++
> 1 file changed, 375 insertions(+)
>
> diff --git a/arch/loongarch/kvm/vcpu.c b/arch/loongarch/kvm/vcpu.c
> index e16b7ca852e3..dce89ace2750 100644
> --- a/arch/loongarch/kvm/vcpu.c
> +++ b/arch/loongarch/kvm/vcpu.c
> @@ -13,6 +13,381 @@
> #define CREATE_TRACE_POINTS
> #include "trace.h"
>
> +int _kvm_getcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 *v, int force)
> +{
> + struct loongarch_csrs *csr = vcpu->arch.csr;
> +
> + GET_HW_GCSR(id, LOONGARCH_CSR_CRMD, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PRMD, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_EUEN, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_MISC, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_ECFG, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_ESTAT, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_ERA, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_BADV, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_BADI, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_EENTRY, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBIDX, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBEHI, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBELO0, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBELO1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_ASID, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PGDL, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PGDH, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PWCTL0, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PWCTL1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_STLBPGSIZE, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_RVACFG, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_CPUID, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PRCFG1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PRCFG2, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_PRCFG3, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS0, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS2, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS3, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS4, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS5, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS6, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_KS7, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TMID, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TCFG, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TVAL, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_CNTC, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_LLBCTL, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRENTRY, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRBADV, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRERA, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRSAVE, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRELO0, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRELO1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBREHI, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_TLBRPRMD, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_DMWIN0, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_DMWIN1, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_DMWIN2, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_DMWIN3, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_MWPS, v);
> + GET_HW_GCSR(id, LOONGARCH_CSR_FWPS, v);

#define GET_HW_GCSR(id, csrid, v) \
do { \
if (csrid == id) { \
*v = (long)kvm_read_hw_gcsr(csrid); \
return 0; \
} \
} while (0)

With expanded macro, this piece of code shows like this:
if (LOONGARCH_CSR_CRMD == id) {
*v = (long)kvm_read_hw_gcsr(LOONGARCH_CSR_CRMD);
return 0;
}
if (LOONGARCH_CSR_PRMD == id) {
*v = (long)kvm_read_hw_gcsr(LOONGARCH_CSR_PRMD);
return 0;
}
if (LOONGARCH_CSR_EUEN == id) {
*v = (long)kvm_read_hw_gcsr(LOONGARCH_CSR_EUEN);
return 0;
}
...

It is strange since macro kvm_read_hw_gcsr must be constant variable.
Can we define generic csr reading function like this?
SYM_CODE_START(kvm_csrread)
ldpcrel t0, 0f
multi.d a0, a0, CSRFUNC_SIZE
add.d t0, t0, a0
jirl r0, t0, 0

csrnum=0
.rept NR_CSR_NUM
0 :
csrread a0, csrnum
jirl r0, ra, 0
/* Ensure that the above is CSRFUNC_ALIGN bytes max */
.fill 0b + CSRFUNC_SIZE - ., 1, 0xcc
csrnum = csrnum+1
.endr
SYM_CODE_END(kvm_csrread)

Wit this function, only one sentence can be used here.
v = kvm_csrread(id);

Regards
Bibo, mao

> +
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_IMPCTL1, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_IMPCTL2, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRCTL, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRINFO1, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRINFO2, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRENTRY, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRERA, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRSAVE, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_CTAG, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_DEBUG, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_DERA, v);
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_DESAVE, v);
> +
> + GET_SW_GCSR(csr, id, LOONGARCH_CSR_TINTCLR, v);
> +
> + if (force && (id < CSR_ALL_SIZE)) {
> + *v = kvm_read_sw_gcsr(csr, id);
> + return 0;
> + }
> +
> + return -1;
> +}
> +
> +int _kvm_setcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 *v, int force)
> +{
> + struct loongarch_csrs *csr = vcpu->arch.csr;
> + int ret;
> +
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_CRMD, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_PRMD, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_EUEN, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_MISC, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_ECFG, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_ERA, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_BADV, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_BADI, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_EENTRY, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBIDX, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBEHI, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBELO0, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBELO1, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_ASID, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_PGDL, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_PGDH, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_PWCTL0, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_PWCTL1, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_STLBPGSIZE, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_RVACFG, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_CPUID, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS0, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS1, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS2, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS3, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS4, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS5, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS6, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_KS7, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TMID, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TCFG, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TVAL, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_CNTC, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_LLBCTL, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRENTRY, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRBADV, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRERA, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRSAVE, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRELO0, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRELO1, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBREHI, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_TLBRPRMD, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_DMWIN0, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_DMWIN1, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_DMWIN2, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_DMWIN3, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_MWPS, v);
> + SET_HW_GCSR(csr, id, LOONGARCH_CSR_FWPS, v);
> +
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_IMPCTL1, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_IMPCTL2, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRCTL, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRINFO1, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRINFO2, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRENTRY, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRERA, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_MERRSAVE, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_CTAG, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_DEBUG, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_DERA, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_DESAVE, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_PRCFG1, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_PRCFG2, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_PRCFG3, v);
> +
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_PGD, v);
> + SET_SW_GCSR(csr, id, LOONGARCH_CSR_TINTCLR, v);
> +
> + ret = -1;
> + switch (id) {
> + case LOONGARCH_CSR_ESTAT:
> + write_gcsr_estat(*v);
> + /* estat IP0~IP7 inject through guestexcept */
> + write_csr_gintc(((*v) >> 2) & 0xff);
> + ret = 0;
> + break;
> + default:
> + if (force && (id < CSR_ALL_SIZE)) {
> + kvm_set_sw_gcsr(csr, id, *v);
> + ret = 0;
> + }
> + break;
> + }
> +
> + return ret;
> +}
> +
> +static int _kvm_get_one_reg(struct kvm_vcpu *vcpu,
> + const struct kvm_one_reg *reg, s64 *v)
> +{
> + struct loongarch_csrs *csr = vcpu->arch.csr;
> + int reg_idx, ret;
> +
> + if ((reg->id & KVM_IOC_CSRID(0)) == KVM_IOC_CSRID(0)) {
> + reg_idx = KVM_GET_IOC_CSRIDX(reg->id);
> + ret = _kvm_getcsr(vcpu, reg_idx, v, 0);
> + if (ret == 0)
> + return ret;
> + }
> +
> + switch (reg->id) {
> + case KVM_REG_LOONGARCH_COUNTER:
> + *v = drdtime() + vcpu->kvm->arch.time_offset;
> + break;
> + default:
> + if ((reg->id & KVM_REG_LOONGARCH_MASK) != KVM_REG_LOONGARCH_CSR)
> + return -EINVAL;
> +
> + reg_idx = KVM_GET_IOC_CSRIDX(reg->id);
> + if (reg_idx < CSR_ALL_SIZE)
> + *v = kvm_read_sw_gcsr(csr, reg_idx);
> + else
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int _kvm_get_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg)
> +{
> + int ret;
> + s64 v;
> +
> + ret = _kvm_get_one_reg(vcpu, reg, &v);
> + if (ret)
> + return ret;
> +
> + ret = -EINVAL;
> + if ((reg->id & KVM_REG_SIZE_MASK) == KVM_REG_SIZE_U64) {
> + u64 __user *uaddr = (u64 __user *)(long)reg->addr;
> +
> + ret = put_user(v, uaddr);
> + } else if ((reg->id & KVM_REG_SIZE_MASK) == KVM_REG_SIZE_U32) {
> + u32 __user *uaddr = (u32 __user *)(long)reg->addr;
> + u32 v32 = (u32)v;
> +
> + ret = put_user(v32, uaddr);
> + }
> +
> + return ret;
> +}
> +
> +static int _kvm_set_one_reg(struct kvm_vcpu *vcpu,
> + const struct kvm_one_reg *reg,
> + s64 v)
> +{
> + struct loongarch_csrs *csr = vcpu->arch.csr;
> + int ret = 0;
> + unsigned long flags;
> + u64 val;
> + int reg_idx;
> +
> + val = v;
> + if ((reg->id & KVM_IOC_CSRID(0)) == KVM_IOC_CSRID(0)) {
> + reg_idx = KVM_GET_IOC_CSRIDX(reg->id);
> + ret = _kvm_setcsr(vcpu, reg_idx, &val, 0);
> + if (ret == 0)
> + return ret;
> + }
> +
> + switch (reg->id) {
> + case KVM_REG_LOONGARCH_COUNTER:
> + local_irq_save(flags);
> + /*
> + * gftoffset is relative with board, not vcpu
> + * only set for the first time for smp system
> + */
> + if (vcpu->vcpu_id == 0)
> + vcpu->kvm->arch.time_offset = (signed long)(v - drdtime());
> + write_csr_gcntc((ulong)vcpu->kvm->arch.time_offset);
> + local_irq_restore(flags);
> + break;
> + case KVM_REG_LOONGARCH_VCPU_RESET:
> + kvm_reset_timer(vcpu);
> + memset(&vcpu->arch.irq_pending, 0, sizeof(vcpu->arch.irq_pending));
> + memset(&vcpu->arch.irq_clear, 0, sizeof(vcpu->arch.irq_clear));
> + break;
> + default:
> + if ((reg->id & KVM_REG_LOONGARCH_MASK) != KVM_REG_LOONGARCH_CSR)
> + return -EINVAL;
> +
> + reg_idx = KVM_GET_IOC_CSRIDX(reg->id);
> + if (reg_idx < CSR_ALL_SIZE)
> + kvm_write_sw_gcsr(csr, reg_idx, v);
> + else
> + return -EINVAL;
> + }
> + return ret;
> +}
> +
> +static int _kvm_set_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg)
> +{
> + s64 v;
> + int ret;
> +
> + ret = -EINVAL;
> + if ((reg->id & KVM_REG_SIZE_MASK) == KVM_REG_SIZE_U64) {
> + u64 __user *uaddr;
> +
> + uaddr = (u64 __user *)(long)reg->addr;
> + ret = get_user(v, uaddr);
> + } else if ((reg->id & KVM_REG_SIZE_MASK) == KVM_REG_SIZE_U32) {
> + u32 __user *uaddr;
> + s32 v32;
> +
> + uaddr = (u32 __user *)(long)reg->addr;
> + ret = get_user(v32, uaddr);
> + v = (s64)v32;
> + }
> +
> + if (ret)
> + return -EFAULT;
> +
> + return _kvm_set_one_reg(vcpu, reg, v);
> +}
> +
> +int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
> + struct kvm_sregs *sregs)
> +{
> + return -ENOIOCTLCMD;
> +}
> +
> +int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
> + struct kvm_sregs *sregs)
> +{
> + return -ENOIOCTLCMD;
> +}
> +
> +int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
> +{
> + int i;
> +
> + vcpu_load(vcpu);
> +
> + for (i = 0; i < ARRAY_SIZE(vcpu->arch.gprs); i++)
> + regs->gpr[i] = vcpu->arch.gprs[i];
> +
> + regs->pc = vcpu->arch.pc;
> +
> + vcpu_put(vcpu);
> + return 0;
> +}
> +
> +int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
> +{
> + int i;
> +
> + vcpu_load(vcpu);
> +
> + for (i = 1; i < ARRAY_SIZE(vcpu->arch.gprs); i++)
> + vcpu->arch.gprs[i] = regs->gpr[i];
> + vcpu->arch.gprs[0] = 0; /* zero is special, and cannot be set. */
> + vcpu->arch.pc = regs->pc;
> +
> + vcpu_put(vcpu);
> + return 0;
> +}
> +
> +long kvm_arch_vcpu_ioctl(struct file *filp,
> + unsigned int ioctl, unsigned long arg)
> +{
> + struct kvm_vcpu *vcpu = filp->private_data;
> + void __user *argp = (void __user *)arg;
> + long r;
> +
> + vcpu_load(vcpu);
> +
> + switch (ioctl) {
> + case KVM_SET_ONE_REG:
> + case KVM_GET_ONE_REG: {
> + struct kvm_one_reg reg;
> +
> + r = -EFAULT;
> + if (copy_from_user(&reg, argp, sizeof(reg)))
> + break;
> + if (ioctl == KVM_SET_ONE_REG)
> + r = _kvm_set_reg(vcpu, &reg);
> + else
> + r = _kvm_get_reg(vcpu, &reg);
> + break;
> + }
> + default:
> + r = -ENOIOCTLCMD;
> + break;
> + }
> +
> + vcpu_put(vcpu);
> + return r;
> +}
> +
> int kvm_arch_vcpu_precreate(struct kvm *kvm, unsigned int id)
> {
> return 0;

Next message: Po-Hsu Lin: "Re: [PATCH] selftests: net: devlink_port_split.py: skip test if no suitable device available"
Previous message: ChiYuan Huang: "Re: [PATCH v17 RESEND 1/3] leds: rgb: mt6370: Add MediaTek MT6370 current sink type LED Indicator support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]