qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Jones <ajones@ventanamicro.com>
To: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Cc: qemu-devel@nongnu.org, qemu-riscv@nongnu.org,
	alistair.francis@wdc.com,  bmeng@tinylab.org,
	liweiwei@iscas.ac.cn, zhiwei_liu@linux.alibaba.com,
	 palmer@rivosinc.com
Subject: Re: [PATCH 13/16] target/riscv/kvm.c: add multi-letter extension KVM properties
Date: Wed, 7 Jun 2023 13:48:37 +0200	[thread overview]
Message-ID: <20230607-9a499522ac7de9ebcf0c8699@orel> (raw)
In-Reply-To: <20230530194623.272652-14-dbarboza@ventanamicro.com>

On Tue, May 30, 2023 at 04:46:20PM -0300, Daniel Henrique Barboza wrote:
> Let's add KVM user properties for the multi-letter extensions that KVM
> currently supports: zicbom, zicboz, zihintpause, zbb, ssaia, sstc,
> svinval and svpbmt.
> 
> As with the recently added MISA properties we're also going to add a
> 'user_set' flag in each of them. The flag will be set only if the user
> chose an option that's different from the host and will require extra
> handling from the KVM driver.
> 
> However, multi-letter CPUs have more cases to cover than MISA
> extensions, so we're adding an extra 'supported' flag as well. This flag
> will reflect if a given extension is supported by KVM, i.e. KVM knows
> how to handle it. This is determined during KVM extension discovery in
> kvm_riscv_init_multiext_cfg(), where we test for EINVAL errors. Any
> other error different from EINVAL will cause an abort.

I wish that was ENOENT, but I suppose that ship sailed.

> 
> The 'supported' flag will then be used later on to give an exception for
> users that are disabling multi-letter extensions that are unknown to
> KVM.
> 
> Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> ---
>  target/riscv/kvm.c | 136 +++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 136 insertions(+)
> 
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index bb1dafe263..b4193a10d8 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -202,6 +202,99 @@ static void kvm_riscv_update_cpu_misa_ext(RISCVCPU *cpu, CPUState *cs)
>      }
>  }
>  
> +typedef struct RISCVCPUMultiExtConfig {
> +    const char *name;

No description? I'd prefer we use the same cfg struct for single-letter
and multi-letter extensions. We can use a union to overlap cpu_cfg_offset
and misa_bit.

> +    int kvm_reg_id;
> +    int cpu_cfg_offset;
> +    bool supported;
> +    bool user_set;
> +} RISCVCPUMultiExtConfig;
> +
> +#define CPUCFG(_prop) offsetof(struct RISCVCPUConfig, _prop)
> +
> +/*
> + * KVM ISA Multi-letter extensions. We care about the order
> + * since it'll be used to create the ISA string later on.
> + * We follow the same ordering rules of isa_edata_arr[]
> + * from target/riscv/cpu.c.
> + */
> +static RISCVCPUMultiExtConfig kvm_multi_ext_cfgs[] = {
> +    {.name = "zicbom", .kvm_reg_id = KVM_RISCV_ISA_EXT_ZICBOM,
> +     .cpu_cfg_offset = CPUCFG(ext_icbom)},
> +    {.name = "zicboz", .kvm_reg_id = KVM_RISCV_ISA_EXT_ZICBOZ,
> +     .cpu_cfg_offset = CPUCFG(ext_icboz)},
> +    {.name = "zihintpause", .kvm_reg_id = KVM_RISCV_ISA_EXT_ZIHINTPAUSE,
> +     .cpu_cfg_offset = CPUCFG(ext_zihintpause)},
> +    {.name = "zbb", .kvm_reg_id = KVM_RISCV_ISA_EXT_ZBB,
> +     .cpu_cfg_offset = CPUCFG(ext_zbb)},
> +    {.name = "ssaia", .kvm_reg_id = KVM_RISCV_ISA_EXT_SSAIA,
> +     .cpu_cfg_offset = CPUCFG(ext_ssaia)},
> +    {.name = "sstc", .kvm_reg_id = KVM_RISCV_ISA_EXT_SSTC,
> +     .cpu_cfg_offset = CPUCFG(ext_sstc)},
> +    {.name = "svinval", .kvm_reg_id = KVM_RISCV_ISA_EXT_SVINVAL,
> +     .cpu_cfg_offset = CPUCFG(ext_svinval)},
> +    {.name = "svpbmt", .kvm_reg_id = KVM_RISCV_ISA_EXT_SVPBMT,
> +     .cpu_cfg_offset = CPUCFG(ext_svpbmt)},

As pointed out in the last patch, it'd be nice to share names (and
descriptions) with TCG.

> +};
> +
> +static void kvm_cpu_cfg_set(RISCVCPU *cpu, RISCVCPUMultiExtConfig *multi_ext,
> +                            uint32_t val)
> +{
> +    int cpu_cfg_offset = multi_ext->cpu_cfg_offset;
> +    bool *ext_enabled = (void *)&cpu->cfg + cpu_cfg_offset;
> +
> +    *ext_enabled = val;
> +}
> +
> +static uint32_t kvm_cpu_cfg_get(RISCVCPU *cpu,
> +                                RISCVCPUMultiExtConfig *multi_ext)
> +{
> +    int cpu_cfg_offset = multi_ext->cpu_cfg_offset;
> +    bool *ext_enabled = (void *)&cpu->cfg + cpu_cfg_offset;
> +
> +    return *ext_enabled;
> +}
> +
> +static void kvm_cpu_set_multi_ext_cfg(Object *obj, Visitor *v,
> +                                      const char *name,
> +                                      void *opaque, Error **errp)
> +{
> +    RISCVCPUMultiExtConfig *multi_ext_cfg = opaque;
> +    RISCVCPU *cpu = RISCV_CPU(obj);
> +    bool value, host_val;
> +
> +    if (!visit_type_bool(v, name, &value, errp)) {
> +        return;
> +    }
> +
> +    host_val = kvm_cpu_cfg_get(cpu, multi_ext_cfg);
> +
> +    /*
> +     * Ignore if the user is setting the same value
> +     * as the host.
> +     */
> +    if (value == host_val) {
> +        return;
> +    }
> +
> +    if (!multi_ext_cfg->supported) {
> +        /*
> +         * Error out if the user is trying to enable an
> +         * extension that KVM doesn't support. Ignore
> +         * option otherwise.
> +         */
> +        if (value) {
> +            error_setg(errp, "KVM does not support disabling extension %s",
> +                       multi_ext_cfg->name);
> +        }
> +
> +        return;
> +    }
> +
> +    multi_ext_cfg->user_set = true;
> +    kvm_cpu_cfg_set(cpu, multi_ext_cfg, value);
> +}
> +
>  static void kvm_riscv_add_cpu_user_properties(Object *cpu_obj)
>  {
>      int i;
> @@ -216,6 +309,15 @@ static void kvm_riscv_add_cpu_user_properties(Object *cpu_obj)
>          object_property_set_description(cpu_obj, misa_cfg->name,
>                                          misa_cfg->description);
>      }
> +
> +    for (i = 0; i < ARRAY_SIZE(kvm_multi_ext_cfgs); i++) {
> +        RISCVCPUMultiExtConfig *multi_cfg = &kvm_multi_ext_cfgs[i];
> +
> +        object_property_add(cpu_obj, multi_cfg->name, "bool",
> +                            NULL,

You have a getter function, so we might as well set it here, no?

> +                            kvm_cpu_set_multi_ext_cfg,
> +                            NULL, multi_cfg);
> +    }
>  }
>  
>  static int kvm_riscv_get_regs_core(CPUState *cs)
> @@ -531,6 +633,39 @@ static void kvm_riscv_init_misa_ext_mask(RISCVCPU *cpu,
>      env->misa_ext = env->misa_ext_mask;
>  }
>  
> +static void kvm_riscv_init_multiext_cfg(RISCVCPU *cpu, KVMScratchCPU *kvmcpu)
> +{
> +    CPURISCVState *env = &cpu->env;
> +    uint64_t val;
> +    int i, ret;
> +
> +    for (i = 0; i < ARRAY_SIZE(kvm_multi_ext_cfgs); i++) {
> +        RISCVCPUMultiExtConfig *multi_ext_cfg = &kvm_multi_ext_cfgs[i];
> +        struct kvm_one_reg reg;
> +
> +        reg.id = kvm_riscv_reg_id(env, KVM_REG_RISCV_ISA_EXT,
> +                                  multi_ext_cfg->kvm_reg_id);
> +        reg.addr = (uint64_t)&val;
> +        ret = ioctl(kvmcpu->cpufd, KVM_GET_ONE_REG, &reg);
> +        if (ret != 0) {
> +            if (ret == -EINVAL) {
> +                /* Silently default to 'false' if KVM does not support it. */
> +                multi_ext_cfg->supported = false;
> +                val = false;
> +            } else {
> +                error_report("Unable to read ISA_EXT KVM register %s, "
> +                             "error %d", multi_ext_cfg->name, ret);
> +                kvm_riscv_destroy_scratch_vcpu(kvmcpu);

I don't think we usually bother cleaning up when exiting, we just let exit
do it, but OK.

> +                exit(EXIT_FAILURE);
> +            }
> +        } else {
> +            multi_ext_cfg->supported = true;
> +        }
> +
> +        kvm_cpu_cfg_set(cpu, multi_ext_cfg, val);
> +    }
> +}
> +
>  void kvm_riscv_init_user_properties(Object *cpu_obj)
>  {
>      RISCVCPU *cpu = RISCV_CPU(cpu_obj);
> @@ -543,6 +678,7 @@ void kvm_riscv_init_user_properties(Object *cpu_obj)
>      kvm_riscv_add_cpu_user_properties(cpu_obj);
>      kvm_riscv_init_machine_ids(cpu, &kvmcpu);
>      kvm_riscv_init_misa_ext_mask(cpu, &kvmcpu);
> +    kvm_riscv_init_multiext_cfg(cpu, &kvmcpu);
>  
>      kvm_riscv_destroy_scratch_vcpu(&kvmcpu);
>  }
> -- 
> 2.40.1
> 
>

Thanks,
drew


  reply	other threads:[~2023-06-07 11:49 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-30 19:46 [PATCH 00/16] target/riscv, KVM: fixes and enhancements Daniel Henrique Barboza
2023-05-30 19:46 ` [PATCH 01/16] target/riscv: skip features setup for KVM CPUs Daniel Henrique Barboza
2023-06-02  4:17   ` Alistair Francis
2023-06-02 14:52   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 02/16] hw/riscv/virt.c: skip 'mmu-type' FDT if satp mode not set Daniel Henrique Barboza
2023-06-06 13:13   ` Andrew Jones
2023-06-06 20:07     ` Daniel Henrique Barboza
2023-06-12  3:53   ` Alistair Francis
2023-05-30 19:46 ` [PATCH 03/16] target/riscv/cpu.c: restrict 'mvendorid' value Daniel Henrique Barboza
2023-06-06 13:19   ` Andrew Jones
2023-06-06 20:06     ` Daniel Henrique Barboza
2023-06-12  3:56   ` Alistair Francis
2023-06-12 18:52     ` Daniel Henrique Barboza
2023-06-13  6:46       ` Alistair Francis
2023-05-30 19:46 ` [PATCH 04/16] target/riscv/cpu.c: restrict 'mimpid' value Daniel Henrique Barboza
2023-06-06 15:31   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 05/16] target/riscv/cpu.c: restrict 'marchid' value Daniel Henrique Barboza
2023-06-06 15:33   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 06/16] target/riscv: use KVM scratch CPUs to init KVM properties Daniel Henrique Barboza
2023-06-06 15:46   ` Andrew Jones
2023-06-12  3:59   ` Alistair Francis
2023-05-30 19:46 ` [PATCH 07/16] target/riscv: read marchid/mimpid in kvm_riscv_init_machine_ids() Daniel Henrique Barboza
2023-06-06 15:47   ` Andrew Jones
2023-06-12  4:05   ` Alistair Francis
2023-05-30 19:46 ` [PATCH 08/16] target/riscv: handle mvendorid/marchid/mimpid for KVM CPUs Daniel Henrique Barboza
2023-06-06 15:51   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 09/16] linux-headers: Update to v6.4-rc1 Daniel Henrique Barboza
2023-05-30 19:46 ` [PATCH 10/16] target/riscv/kvm.c: init 'misa_ext_mask' with scratch CPU Daniel Henrique Barboza
2023-06-06 15:54   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 11/16] target/riscv: add KVM specific MISA properties Daniel Henrique Barboza
2023-06-07 11:33   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 12/16] target/riscv/kvm.c: update KVM MISA bits Daniel Henrique Barboza
2023-06-07 12:05   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 13/16] target/riscv/kvm.c: add multi-letter extension KVM properties Daniel Henrique Barboza
2023-06-07 11:48   ` Andrew Jones [this message]
2023-06-07 19:59     ` Daniel Henrique Barboza
2023-06-08  6:02       ` Andrew Jones
2023-06-12 19:24     ` Daniel Henrique Barboza
2023-05-30 19:46 ` [PATCH 14/16] target/riscv: adapt 'riscv_isa_string' for KVM Daniel Henrique Barboza
2023-06-07 12:21   ` Andrew Jones
2023-06-13 10:29     ` Daniel Henrique Barboza
2023-06-13 18:19       ` Daniel Henrique Barboza
2023-05-30 19:46 ` [PATCH 15/16] target/riscv: update multi-letter extension KVM properties Daniel Henrique Barboza
2023-06-07 12:30   ` Andrew Jones
2023-05-30 19:46 ` [PATCH 16/16] target/riscv/kvm.c: read/write (cbom|cboz)_blocksize in KVM Daniel Henrique Barboza
2023-06-07 13:01   ` Andrew Jones
2023-06-07 20:37     ` Daniel Henrique Barboza
2023-06-08  6:39       ` Andrew Jones

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230607-9a499522ac7de9ebcf0c8699@orel \
    --to=ajones@ventanamicro.com \
    --cc=alistair.francis@wdc.com \
    --cc=bmeng@tinylab.org \
    --cc=dbarboza@ventanamicro.com \
    --cc=liweiwei@iscas.ac.cn \
    --cc=palmer@rivosinc.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-riscv@nongnu.org \
    --cc=zhiwei_liu@linux.alibaba.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).