Linux KVM/arm64 development list
 help / color / mirror / Atom feed
From: Fuad Tabba <tabba@google.com>
To: kvmarm@lists.linux.dev, linux-arm-kernel@lists.infradead.org
Cc: maz@kernel.org, oliver.upton@linux.dev, will@kernel.org,
	 joey.gouly@arm.com, suzuki.poulose@arm.com,
	yuzenghui@huawei.com,  catalin.marinas@arm.com, tabba@google.com
Subject: [PATCH v1 2/4] KVM: arm64: Add compile-time type check for register in __vcpu_assign_sys_reg()
Date: Mon, 27 Oct 2025 11:39:41 +0000	[thread overview]
Message-ID: <20251027113943.1282568-3-tabba@google.com> (raw)
In-Reply-To: <20251027113943.1282568-1-tabba@google.com>

The 'r' (register) and 'val' (value) parameters in
__vcpu_assign_sys_reg() can be easily transposed. The register ID is an
enum, whereas the value is a u64. This has led to bugs in the past, and
the risk was increased by the historically inconsistent parameter
ordering of the vcpu_write_sys_reg() function.

To prevent this class of bugs, add a compile-time type compatibility check
to prevent the 'r' parameter from having a 'u64' type.

This directly catches the erroneous transposition (passing the 'u64'
value as 'r') while remaining flexible, as it does not force 'r' to
be a specific enum type.

This patch also updates several local variables in
arch/arm64/kvm/pmu-emul.c from 'u64' to 'enum vcpu_sysreg' to fix
build failures caused by the new check.

No functional change intended.

Signed-off-by: Fuad Tabba <tabba@google.com>
---
NOTE: In practice, 'enum vcpu_sysreg' cannot grow to become compatabile
with u64 since we are limited by the size of arrays that index sysregs.
If it does grow, other compile-time issues show up before this check
becomes an issue.

I verified that this would catch the bug that commit 798eb5978700
("KVM: arm64: Sync protected guest VBAR_EL1 on injecting an undef
exception") introduced, which was fixed later.
---
 arch/arm64/include/asm/kvm_host.h | 17 +++++++++--------
 arch/arm64/kvm/pmu-emul.c         |  7 ++++---
 2 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 1b8a420f9add..2c33ea5fdb1c 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -1136,14 +1136,15 @@ static inline u64 *___ctxt_sys_reg(const struct kvm_cpu_context *ctxt, int r)
 
 u64 kvm_vcpu_apply_reg_masks(const struct kvm_vcpu *, enum vcpu_sysreg, u64);
 
-#define __vcpu_assign_sys_reg(v, r, val)				\
-	do {								\
-		const struct kvm_cpu_context *ctxt = &(v)->arch.ctxt;	\
-		u64 __v = (val);					\
-		if (vcpu_has_nv((v)) && (r) >= __SANITISED_REG_START__)	\
-			__v = kvm_vcpu_apply_reg_masks((v), (r), __v);	\
-									\
-		ctxt_sys_reg(ctxt, (r)) = __v;				\
+#define __vcpu_assign_sys_reg(v, r, val)					\
+	do {									\
+		const struct kvm_cpu_context *ctxt = &(v)->arch.ctxt;		\
+		u64 __v = (val);						\
+		BUILD_BUG_ON_ZERO(__builtin_types_compatible_p(typeof(r), u64));\
+		if (vcpu_has_nv((v)) && (r) >= __SANITISED_REG_START__)		\
+			__v = kvm_vcpu_apply_reg_masks((v), (r), __v);		\
+										\
+		ctxt_sys_reg(ctxt, (r)) = __v;					\
 	} while (0)
 
 #define __vcpu_rmw_sys_reg(v, r, op, val)				\
diff --git a/arch/arm64/kvm/pmu-emul.c b/arch/arm64/kvm/pmu-emul.c
index b03dbda7f1ab..c7e2d9be77ae 100644
--- a/arch/arm64/kvm/pmu-emul.c
+++ b/arch/arm64/kvm/pmu-emul.c
@@ -160,7 +160,7 @@ u64 kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu, u64 select_idx)
 static void kvm_pmu_set_pmc_value(struct kvm_pmc *pmc, u64 val, bool force)
 {
 	struct kvm_vcpu *vcpu = kvm_pmc_to_vcpu(pmc);
-	u64 reg;
+	enum vcpu_sysreg reg;
 
 	kvm_pmu_release_perf_event(pmc);
 
@@ -230,7 +230,8 @@ static void kvm_pmu_release_perf_event(struct kvm_pmc *pmc)
 static void kvm_pmu_stop_counter(struct kvm_pmc *pmc)
 {
 	struct kvm_vcpu *vcpu = kvm_pmc_to_vcpu(pmc);
-	u64 reg, val;
+	enum vcpu_sysreg reg;
+	u64 val;
 
 	if (!pmc->perf_event)
 		return;
@@ -776,7 +777,7 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, u64 data,
 				    u64 select_idx)
 {
 	struct kvm_pmc *pmc = kvm_vcpu_idx_to_pmc(vcpu, select_idx);
-	u64 reg;
+	enum vcpu_sysreg reg;
 
 	reg = counter_index_to_evtreg(pmc->idx);
 	__vcpu_assign_sys_reg(vcpu, reg, (data & kvm_pmu_evtyper_mask(vcpu->kvm)));
-- 
2.51.1.838.g19442a804e-goog


  parent reply	other threads:[~2025-10-27 11:39 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-27 11:39 [PATCH v1 0/4] KVM: arm64: Prevent sysreg helper parameter transposition Fuad Tabba
2025-10-27 11:39 ` [PATCH v1 1/4] KVM: arm64: Switch reg and val parameter ordering in vcpu_write_sys_reg() Fuad Tabba
2025-10-28 10:38   ` Marc Zyngier
2025-10-28 10:40     ` Fuad Tabba
2025-10-27 11:39 ` Fuad Tabba [this message]
2025-10-28 10:51   ` [PATCH v1 2/4] KVM: arm64: Add compile-time type check for register in __vcpu_assign_sys_reg() Marc Zyngier
2025-10-28 10:58     ` Fuad Tabba
2025-10-28 17:05     ` Fuad Tabba
2025-10-30  9:53     ` Fuad Tabba
2025-10-27 11:39 ` [PATCH v1 3/4] KVM: arm64: Add compile-time type check to vcpu_write_sys_reg() Fuad Tabba
2025-10-27 11:39 ` [PATCH v1 4/4] KVM: arm64: Add compile-time type check for register in __vcpu_rmw_sys_reg() Fuad Tabba

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=20251027113943.1282568-3-tabba@google.com \
    --to=tabba@google.com \
    --cc=catalin.marinas@arm.com \
    --cc=joey.gouly@arm.com \
    --cc=kvmarm@lists.linux.dev \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=maz@kernel.org \
    --cc=oliver.upton@linux.dev \
    --cc=suzuki.poulose@arm.com \
    --cc=will@kernel.org \
    --cc=yuzenghui@huawei.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