From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: alistair.francis@wdc.com
Subject: [PATCH 07/22] target/riscv: add more RISCVCPUDef fields
Date: Fri, 28 Feb 2025 11:27:31 +0100 [thread overview]
Message-ID: <20250228102747.867770-8-pbonzini@redhat.com> (raw)
In-Reply-To: <20250228102747.867770-1-pbonzini@redhat.com>
Allow using RISCVCPUDef to replicate all the logic of custom .instance_init
functions. To simulate inheritance, merge the child's RISCVCPUDef with
the parent and then finally move it to the CPUState at the end of
TYPE_RISCV_CPU's own instance_init function.
STRUCT_FIELD is introduced here because I am not sure it is needed;
it is a bit ugly and I wanted not to have it in the patch that
introduces cpu_cfg_fields.h.inc. I don't really understand why satp_mode
is included in RISCVCPUConfig; therefore, the end of the series includes
a patch to move satp_mode directly in RISCVCPU, thus removing the need
for STRUCT_FIELD; it can be moved before this one in a non-RFC posting.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/riscv/cpu.h | 4 ++++
target/riscv/cpu.c | 30 +++++++++++++++++++++++++++++-
target/riscv/kvm/kvm-cpu.c | 2 ++
3 files changed, 35 insertions(+), 1 deletion(-)
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index a038122f80c..8c9e73c68cc 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -533,6 +533,10 @@ struct ArchCPU {
typedef struct RISCVCPUDef {
RISCVMXL misa_mxl_max; /* max mxl for this cpu */
+ uint32_t misa_ext;
+ int priv_spec;
+ int32_t vext_spec;
+ RISCVCPUConfig cfg;
} RISCVCPUDef;
/**
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 91dd63edc9f..c513d7ce32d 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -74,6 +74,13 @@ bool riscv_cpu_option_set(const char *optname)
return g_hash_table_contains(general_user_opts, optname);
}
+static void riscv_cpu_cfg_merge(RISCVCPUConfig *dest, const RISCVCPUConfig *src)
+{
+#define BOOL_FIELD(x) dest->x |= src->x;
+#define TYPED_FIELD(type, x) if (src->x) dest->x = src->x;
+#include "cpu_cfg_fields.h.inc"
+}
+
#define ISA_EXT_DATA_ENTRY(_name, _min_ver, _prop) \
{#_name, _min_ver, CPU_CFG_OFFSET(_prop)}
@@ -432,7 +439,7 @@ const char *satp_mode_str(uint8_t satp_mode, bool is_32_bit)
}
static void set_satp_mode_max_supported(RISCVCPU *cpu,
- uint8_t satp_mode)
+ int satp_mode)
{
bool rv32 = riscv_cpu_mxl(&cpu->env) == MXL_RV32;
const bool *valid_vm = rv32 ? valid_vm_1_10_32 : valid_vm_1_10_64;
@@ -1480,6 +1487,16 @@ static void riscv_cpu_init(Object *obj)
#ifndef CONFIG_USER_ONLY
cpu->cfg.max_satp_mode = -1;
#endif /* CONFIG_USER_ONLY */
+
+ env->misa_ext_mask = env->misa_ext = mcc->def->misa_ext;
+ riscv_cpu_cfg_merge(&cpu->cfg, &mcc->def->cfg);
+
+ if (mcc->def->priv_spec != RISCV_PROFILE_ATTR_UNUSED) {
+ cpu->env.priv_ver = mcc->def->priv_spec;
+ }
+ if (mcc->def->vext_spec != RISCV_PROFILE_ATTR_UNUSED) {
+ cpu->env.vext_ver = mcc->def->vext_spec;
+ }
}
static void riscv_bare_cpu_init(Object *obj)
@@ -2974,6 +2991,17 @@ static void riscv_cpu_class_base_init(ObjectClass *c, void *data)
assert(def->misa_mxl_max <= MXL_RV128);
mcc->def->misa_mxl_max = def->misa_mxl_max;
}
+ if (def->priv_spec != RISCV_PROFILE_ATTR_UNUSED) {
+ assert(def->priv_spec <= PRIV_VERSION_LATEST);
+ mcc->def->priv_spec = def->priv_spec;
+ }
+ if (def->vext_spec != RISCV_PROFILE_ATTR_UNUSED) {
+ assert(def->vext_spec != 0);
+ mcc->def->vext_spec = def->vext_spec;
+ }
+ mcc->def->misa_ext |= def->misa_ext;
+
+ riscv_cpu_cfg_merge(&mcc->def->cfg, &def->cfg);
}
if (!object_class_is_abstract(c)) {
diff --git a/target/riscv/kvm/kvm-cpu.c b/target/riscv/kvm/kvm-cpu.c
index f3ec4f33931..8259cb4deee 100644
--- a/target/riscv/kvm/kvm-cpu.c
+++ b/target/riscv/kvm/kvm-cpu.c
@@ -1992,10 +1992,12 @@ static const TypeInfo riscv_kvm_cpu_type_infos[] = {
#if defined(TARGET_RISCV32)
.class_data = &((const RISCVCPUDef) {
.misa_mxl_max = MXL_RV32,
+ .cfg.max_satp_mode = -1,
},
#elif defined(TARGET_RISCV64)
.class_data = &((const RISCVCPUDef) {
.misa_mxl_max = MXL_RV64,
+ .cfg.max_satp_mode = -1,
},
#endif
}
--
2.48.1
next prev parent reply other threads:[~2025-02-28 10:29 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-28 10:27 [PATCH v2 00/22] target/riscv: declarative CPU definitions Paolo Bonzini
2025-02-28 10:27 ` [PATCH 01/22] target/riscv: Declare RISCVCPUClass::misa_mxl_max as RISCVMXL Paolo Bonzini
2025-03-06 0:46 ` Alistair Francis
2025-02-28 10:27 ` [PATCH 02/22] target/riscv: Convert misa_mxl_max using GLib macros Paolo Bonzini
2025-03-06 0:47 ` Alistair Francis
2025-02-28 10:27 ` [PATCH 03/22] target/riscv: introduce RISCVCPUDef Paolo Bonzini
2025-02-28 17:19 ` Richard Henderson
2025-02-28 10:27 ` [PATCH 04/22] target/riscv: store RISCVCPUDef struct directly in the class Paolo Bonzini
2025-03-06 0:50 ` Alistair Francis
2025-02-28 10:27 ` [PATCH 05/22] target/riscv: merge riscv_cpu_class_init with the class_base function Paolo Bonzini
2025-03-06 0:52 ` Alistair Francis
2025-02-28 10:27 ` [PATCH 06/22] target/riscv: move RISCVCPUConfig fields to a header file Paolo Bonzini
2025-03-06 0:53 ` Alistair Francis
2025-02-28 10:27 ` Paolo Bonzini [this message]
2025-02-28 10:27 ` [PATCH 08/22] target/riscv: convert abstract CPU classes to RISCVCPUDef Paolo Bonzini
2025-02-28 10:27 ` [PATCH 09/22] target/riscv: do not make RISCVCPUConfig fields conditional Paolo Bonzini
2025-02-28 10:27 ` [PATCH 10/22] target/riscv: convert profile CPU models to RISCVCPUDef Paolo Bonzini
2025-02-28 10:27 ` [PATCH 11/22] target/riscv: convert bare " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 12/22] target/riscv: convert dynamic " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 13/22] target/riscv: convert SiFive E " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 14/22] target/riscv: convert ibex " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 15/22] target/riscv: convert SiFive U " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 16/22] target/riscv: th: make CSR insertion test a bit more intuitive Paolo Bonzini
2025-02-28 10:27 ` [PATCH 17/22] target/riscv: generalize custom CSR functionality Paolo Bonzini
2025-02-28 10:27 ` [PATCH 18/22] target/riscv: convert TT C906 to RISCVCPUDef Paolo Bonzini
2025-02-28 10:27 ` [PATCH 19/22] target/riscv: convert TT Ascalon " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 20/22] target/riscv: convert Ventana V1 " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 21/22] target/riscv: convert Xiangshan Nanhu " Paolo Bonzini
2025-02-28 10:27 ` [PATCH 22/22] target/riscv: remove .instance_post_init Paolo Bonzini
2025-04-17 13:40 ` [PATCH v2 00/22] target/riscv: declarative CPU definitions Paolo Bonzini
2025-04-24 10:37 ` Alistair Francis
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=20250228102747.867770-8-pbonzini@redhat.com \
--to=pbonzini@redhat.com \
--cc=alistair.francis@wdc.com \
--cc=qemu-devel@nongnu.org \
/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).