Skip to contentSkip to conversation

KVM: arm64: Get rid of REG_HIDDEN_USER

PATCH Applied Marc Zyngier 2 participants 6 messages Sep 4, 2024 – Sep 11, 2024

Subsystems (3)
KVM arm64 KVM arm64
lore
Share & tools
lore

Marc Zyngier

author maintainer
#0 lore raw

cover KVM: arm64: Get rid of REG_HIDDEN_USER

REG_HIDDEN_USER was introduced as a way to deal with the ARMv8.3 flavour of FEAT_NV, where most EL12 sysreg accesses would trap and yet be mapped to a EL1 register (KVM doing in SW what FEAT_NV2 does in HW). This handling imposed that the EL12 register shouldn't be visible to userspace, hence the special REG_HIDDEN_USER visibility.

Since 4d4f52052ba8 ("KVM: arm64: nv: Drop EL12 register traps that are redirected to VNCR") and the admission that KVM would never be supporting the original FEAT_NV, REG_HIDDEN_USER only had a few users, all of which could either be replaced by a more ad-hoc mechanism, or removed altogether.

This series goes ahead and cleans it up for good, removing a tiny bit of unnecessary complexity.

Marc Zyngier (3):
KVM: arm64: Simplify handling of CNTKCTL_EL12
KVM: arm64: Simplify visibility handling of AArch32 SPSR_*
KVM: arm64: Get rid of REG_HIDDEN_USER visibility qualifier

 arch/arm64/kvm/sys_regs.c | 52 +++++++++++++++------------------------
 arch/arm64/kvm/sys_regs.h | 14 ++---------
 2 files changed, 22 insertions(+), 44 deletions(-)
Signature · 1 line
2.39.2
author maintainer
#1 lore raw

PATCH 1/3 KVM: arm64: Simplify handling of CNTKCTL_EL12

We go trough a great deal of effort to map CNTKCTL_EL12 to CNTKCTL_EL1 while hidding this mapping from userspace via a special visibility helper.

However, it would be far simpler to just provide an accessor doing the mapping job, removing the need for a visibility helper.

With that done, we can also remove the EL12_REG() macro which serves no purpose.

Signed-off-by Marc Zyngier <[email protected]>
---
 arch/arm64/kvm/sys_regs.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)
arch/arm64/kvm/sys_regs.c +13 -10
index 45c9e3b2acd4..5328fac8d547 100644
@@ -2105,15 +2105,6 @@ static unsigned int hidden_user_visibility(const struct kvm_vcpu *vcpu,
 	return REG_HIDDEN_USER;
 }
 
-#define EL12_REG(name, acc, rst, v) {		\
-	SYS_DESC(SYS_##name##_EL12),		\
-	.access = acc,				\
-	.reset = rst,				\
-	.reg = name##_EL1,			\
-	.val = v,				\
-	.visibility = hidden_user_visibility,	\
-}
-
 /*
  * Since reset() callback and field val are not used for idregs, they will be
  * used for specific purposes for idregs.

@@ -2221,6 +2212,18 @@ static bool access_spsr(struct kvm_vcpu *vcpu,
 	return true;
 }
 
+static bool access_cntkctl_el12(struct kvm_vcpu *vcpu,
+				struct sys_reg_params *p,
+				const struct sys_reg_desc *r)
+{
+	if (p->is_write)
+		__vcpu_sys_reg(vcpu, CNTKCTL_EL1) = p->regval;
+	else
+		p->regval = __vcpu_sys_reg(vcpu, CNTKCTL_EL1);
+
+	return true;
+}
+
 static u64 reset_hcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
 {
 	u64 val = r->val;

@@ -2825,7 +2828,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	EL2_REG_VNCR(CNTVOFF_EL2, reset_val, 0),
 	EL2_REG(CNTHCTL_EL2, access_rw, reset_val, 0),
 
-	EL12_REG(CNTKCTL, access_rw, reset_val, 0),
+	{ SYS_DESC(SYS_CNTKCTL_EL12), access_cntkctl_el12 },
 
 	EL2_REG(SP_EL2, NULL, reset_unknown, 0),
 };
Signature · 1 line
2.39.2
author maintainer
#2 lore raw

PATCH 2/3 KVM: arm64: Simplify visibility handling of AArch32 SPSR_*

In reply to Marc Zyngier

Since SPSR_* are not associated with any register in the sysreg array, nor do they have .get_user()/.set_user() helpers, they are invisible to userspace with that encoding.

Therefore hidden_user_visibility() serves no purpose here, and can be safely removed.

Signed-off-by Marc Zyngier <[email protected]>
---
 arch/arm64/kvm/sys_regs.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)
arch/arm64/kvm/sys_regs.c +4 -8
index 5328fac8d547..2586ad29eaa0 100644
@@ -2795,14 +2795,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
 	{ SYS_DESC(SYS_SP_EL1), access_sp_el1},
 
 	/* AArch32 SPSR_* are RES0 if trapped from a NV guest */
-	{ SYS_DESC(SYS_SPSR_irq), .access = trap_raz_wi,
-	  .visibility = hidden_user_visibility },
-	{ SYS_DESC(SYS_SPSR_abt), .access = trap_raz_wi,
-	  .visibility = hidden_user_visibility },
-	{ SYS_DESC(SYS_SPSR_und), .access = trap_raz_wi,
-	  .visibility = hidden_user_visibility },
-	{ SYS_DESC(SYS_SPSR_fiq), .access = trap_raz_wi,
-	  .visibility = hidden_user_visibility },
+	{ SYS_DESC(SYS_SPSR_irq), .access = trap_raz_wi },
+	{ SYS_DESC(SYS_SPSR_abt), .access = trap_raz_wi },
+	{ SYS_DESC(SYS_SPSR_und), .access = trap_raz_wi },
+	{ SYS_DESC(SYS_SPSR_fiq), .access = trap_raz_wi },
 
 	{ SYS_DESC(SYS_IFSR32_EL2), undef_access, reset_unknown, IFSR32_EL2 },
 	EL2_REG(AFSR0_EL2, access_rw, reset_val, 0),
Signature · 1 line
2.39.2
author maintainer
#3 lore raw

PATCH 3/3 KVM: arm64: Get rid of REG_HIDDEN_USER visibility qualifier

In reply to Marc Zyngier

Now that REG_HIDDEN_USER has no direct user anymore, remove it entirely and update all users of sysreg_hidden_user() to call sysreg_hidden() instead.

Signed-off-by Marc Zyngier <[email protected]>
---
 arch/arm64/kvm/sys_regs.c | 17 +++--------------
 arch/arm64/kvm/sys_regs.h | 14 ++------------
 2 files changed, 5 insertions(+), 26 deletions(-)
arch/arm64/kvm/sys_regs.c +3 -14
index 2586ad29eaa0..09a6a45efb49 100644
@@ -2094,17 +2094,6 @@ static bool bad_redir_trap(struct kvm_vcpu *vcpu,
 #define EL2_REG_VNCR(name, rst, v)	EL2_REG(name, bad_vncr_trap, rst, v)
 #define EL2_REG_REDIR(name, rst, v)	EL2_REG(name, bad_redir_trap, rst, v)
 
-/*
- * EL{0,1}2 registers are the EL2 view on an EL0 or EL1 register when
- * HCR_EL2.E2H==1, and only in the sysreg table for convenience of
- * handling traps. Given that, they are always hidden from userspace.
- */
-static unsigned int hidden_user_visibility(const struct kvm_vcpu *vcpu,
-					   const struct sys_reg_desc *rd)
-{
-	return REG_HIDDEN_USER;
-}
-
 /*
  * Since reset() callback and field val are not used for idregs, they will be
  * used for specific purposes for idregs.

@@ -4364,7 +4353,7 @@ int kvm_sys_reg_get_user(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg,
 	int ret;
 
 	r = id_to_sys_reg_desc(vcpu, reg->id, table, num);
-	if (!r || sysreg_hidden_user(vcpu, r))
+	if (!r || sysreg_hidden(vcpu, r))
 		return -ENOENT;
 
 	if (r->get_user) {

@@ -4408,7 +4397,7 @@ int kvm_sys_reg_set_user(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg,
 		return -EFAULT;
 
 	r = id_to_sys_reg_desc(vcpu, reg->id, table, num);
-	if (!r || sysreg_hidden_user(vcpu, r))
+	if (!r || sysreg_hidden(vcpu, r))
 		return -ENOENT;
 
 	if (sysreg_user_write_ignore(vcpu, r))

@@ -4494,7 +4483,7 @@ static int walk_one_sys_reg(const struct kvm_vcpu *vcpu,
 	if (!(rd->reg || rd->get_user))
 		return 0;
 
-	if (sysreg_hidden_user(vcpu, rd))
+	if (sysreg_hidden(vcpu, rd))
 		return 0;
 
 	if (!copy_reg_to_user(rd, uind))
arch/arm64/kvm/sys_regs.h +2 -12
index dfb2ec83b284..1d94ed6efad2 100644
@@ -95,9 +95,8 @@ struct sys_reg_desc {
 };
 
 #define REG_HIDDEN		(1 << 0) /* hidden from userspace and guest */
-#define REG_HIDDEN_USER		(1 << 1) /* hidden from userspace only */
-#define REG_RAZ			(1 << 2) /* RAZ from userspace and guest */
-#define REG_USER_WI		(1 << 3) /* WI from userspace only */
+#define REG_RAZ			(1 << 1) /* RAZ from userspace and guest */
+#define REG_USER_WI		(1 << 2) /* WI from userspace only */
 
 static __printf(2, 3)
 inline void print_sys_reg_msg(const struct sys_reg_params *p,

@@ -165,15 +164,6 @@ static inline bool sysreg_hidden(const struct kvm_vcpu *vcpu,
 	return sysreg_visibility(vcpu, r) & REG_HIDDEN;
 }
 
-static inline bool sysreg_hidden_user(const struct kvm_vcpu *vcpu,
-				      const struct sys_reg_desc *r)
-{
-	if (likely(!r->visibility))
-		return false;
-
-	return r->visibility(vcpu, r) & (REG_HIDDEN | REG_HIDDEN_USER);
-}
-
 static inline bool sysreg_visible_as_raz(const struct kvm_vcpu *vcpu,
 					 const struct sys_reg_desc *r)
 {
Signature · 1 line
2.39.2
Reviewed-by on the v1 series
#4 lore raw
In reply to Marc Zyngier
Quoted 6 lines · Marc Zyngier

REG_HIDDEN_USER was introduced as a way to deal with the ARMv8.3 flavour of FEAT_NV, where most EL12 sysreg accesses would trap and yet be mapped to a EL1 register (KVM doing in SW what FEAT_NV2 does in HW). This handling imposed that the EL12 register shouldn't be visible to userspace, hence the special REG_HIDDEN_USER visibility.

Quoted 7 lines

Since 4d4f52052ba8 ("KVM: arm64: nv: Drop EL12 register traps that are redirected to VNCR") and the admission that KVM would never be supporting the original FEAT_NV, REG_HIDDEN_USER only had a few users, all of which could either be replaced by a more ad-hoc mechanism, or removed altogether.

This series goes ahead and cleans it up for good, removing a tiny bit of unnecessary complexity.

Yeah, let's toss it.

Reviewed-by Oliver Upton <[email protected]>
Signature · 2 lines
Thanks,
Oliver
author maintainer Applied
#5 lore raw
In reply to Marc Zyngier
Quoted 6 lines · Marc Zyngier

REG_HIDDEN_USER was introduced as a way to deal with the ARMv8.3 flavour of FEAT_NV, where most EL12 sysreg accesses would trap and yet be mapped to a EL1 register (KVM doing in SW what FEAT_NV2 does in HW). This handling imposed that the EL12 register shouldn't be visible to userspace, hence the special REG_HIDDEN_USER visibility.

Quoted 6 lines

Since 4d4f52052ba8 ("KVM: arm64: nv: Drop EL12 register traps that are redirected to VNCR") and the admission that KVM would never be supporting the original FEAT_NV, REG_HIDDEN_USER only had a few users, all of which could either be replaced by a more ad-hoc mechanism, or removed altogether.

[...]

Applied to next, thanks!

[1/3] KVM: arm64: Simplify handling of CNTKCTL_EL12

      commit: 989fce63b2cb5061701c9fa04711d992dfaff5c6
[2/3] KVM: arm64: Simplify visibility handling of AArch32 SPSR_*
      commit: 84ed45456cee7e77effea8407f4f32b262f2e2ea
[3/3] KVM: arm64: Get rid of REG_HIDDEN_USER visibility qualifier
      commit: 0746096faca01823021f662282e1f067a69b965b

Cheers,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.