Skip to contentSkip to conversation

KVM: LoongArch: Fix wrong CPUCFG ID handling

PATCH Superseded WANG Xuerui 2 participants 8 messages Feb 14, 2024

Subsystems (3)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

WANG Xuerui

author reviewer
#0 lore raw

cover KVM: LoongArch: Fix wrong CPUCFG ID handling

From: WANG Xuerui <[email protected]>

Hi,

While trying to add loongarch to the Rust kvm-bindings crate, I accidentally discovered faulty logic in the handling of CPUCFG IDs ("leaves" for those more familiar with x86), that could result in incorrectly accepting every possible int for the ID; fortunately it is 6.8 material that hasn't seen a release yet, so a fix is possible.

The first two patches contain the fix, while the rest are general drive-by refactoring and comment cleanups.

(As it is currently the Chinese holiday season, it is probably best for this series to go through the kvm tree, instead of the loongarch one even though they seem to like prefer collecting every loongarch patch.)

WANG Xuerui (5):
KVM: LoongArch: Fix input value checking of _kvm_get_cpucfg
KVM: LoongArch: Fix kvm_check_cpucfg incorrectly accepting bad CPUCFG
IDs
KVM: LoongArch: Rename _kvm_get_cpucfg to _kvm_get_cpucfg_mask
KVM: LoongArch: Streamline control flow of kvm_check_cpucfg
KVM: LoongArch: Clean up comments of _kvm_get_cpucfg_mask and
kvm_check_cpucfg

 arch/loongarch/kvm/vcpu.c | 72 ++++++++++++++++-----------------------
 1 file changed, 29 insertions(+), 43 deletions(-)
Signature · 1 line
2.43.0
author reviewer
#1 lore raw

PATCH 1/5 KVM: LoongArch: Fix input value checking of _kvm_get_cpucfg

From: WANG Xuerui <[email protected]>

The range check for the CPUCFG ID is wrong (should have been a || instead of &&), and pointless, because the default case a few lines below already serves to deny all unhandled cases.

Furthermore, the juggling of the temp return value is unnecessary, because it is semantically equivalent and more readable to just return at every switch case's end. This is done too to avoid potential bugs in the future related to the unwanted complexity.

Fixes db1ecca22edf ("LoongArch: KVM: Add LSX (128bit SIMD) support")
---
 arch/loongarch/kvm/vcpu.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
arch/loongarch/kvm/vcpu.c +2 -9
index 27701991886d..c4a592623da6 100644
@@ -300,11 +300,6 @@ static int _kvm_setcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 val)
 
 static int _kvm_get_cpucfg(int id, u64 *v)
 {
-	int ret = 0;
-
-	if (id < 0 && id >= KVM_MAX_CPUCFG_REGS)
-		return -EINVAL;
-
 	switch (id) {
 	case 2:
 		/* Return CPUCFG2 features which have been supported by KVM */
@@ -324,12 +319,10 @@ static int _kvm_get_cpucfg(int id, u64 *v)
 		if (cpu_has_lasx)
 			*v |= CPUCFG2_LASX;
 
-		break;
+		return 0;
 	default:
-		ret = -EINVAL;
-		break;
+		return -EINVAL;
 	}
-	return ret;
 }
 
 static int kvm_check_cpucfg(int id, u64 val)
Signature · 1 line
2.43.0
author reviewer
#2 lore raw

PATCH 2/5 KVM: LoongArch: Fix kvm_check_cpucfg incorrectly accepting bad CPUCFG IDs

In reply to WANG Xuerui

From: WANG Xuerui <[email protected]>

The return value of _kvm_get_cpucfg is meant to be checked, but this was not done, so bad CPUCFG IDs wrongly fall back to the default case and 0 is returned.

Check the return value to fix the UAPI behavior.

While at it, also remove the redundant range check, because _kvm_get_cpucfg already rejects all unrecognized input CPUCFG IDs. It was also wrong and no-op, in the same way as the former identical check in _kvm_get_cpucfg.

Fixes db1ecca22edf ("LoongArch: KVM: Add LSX (128bit SIMD) support")
---
 arch/loongarch/kvm/vcpu.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)
arch/loongarch/kvm/vcpu.c +3 -5
index c4a592623da6..cc2332b056ba 100644
@@ -328,12 +328,10 @@ static int _kvm_get_cpucfg(int id, u64 *v)
 static int kvm_check_cpucfg(int id, u64 val)
 {
 	u64 mask;
-	int ret = 0;
-
-	if (id < 0 && id >= KVM_MAX_CPUCFG_REGS)
-		return -EINVAL;
+	int ret;
 
-	if (_kvm_get_cpucfg(id, &mask))
+	ret = _kvm_get_cpucfg(id, &mask);
+	if (ret)
 		return ret;
 
 	switch (id) {
Signature · 1 line
2.43.0
author reviewer
#3 lore raw

PATCH 3/5 KVM: LoongArch: Rename _kvm_get_cpucfg to _kvm_get_cpucfg_mask

In reply to WANG Xuerui

From: WANG Xuerui <[email protected]>

The function is not actually a getter of guest CPUCFG, but rather validation of the input CPUCFG ID plus information about the supported bit flags of that CPUCFG leaf. So rename it to avoid confusion.

---
 arch/loongarch/kvm/vcpu.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
arch/loongarch/kvm/vcpu.c +3 -3
index cc2332b056ba..e973500611b4 100644
@@ -298,7 +298,7 @@ static int _kvm_setcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 val)
 	return ret;
 }
 
-static int _kvm_get_cpucfg(int id, u64 *v)
+static int _kvm_get_cpucfg_mask(int id, u64 *v)
 {
 	switch (id) {
 	case 2:
@@ -330,7 +330,7 @@ static int kvm_check_cpucfg(int id, u64 val)
 	u64 mask;
 	int ret;
 
-	ret = _kvm_get_cpucfg(id, &mask);
+	ret = _kvm_get_cpucfg_mask(id, &mask);
 	if (ret)
 		return ret;
 
@@ -557,7 +557,7 @@ static int kvm_loongarch_get_cpucfg_attr(struct kvm_vcpu *vcpu,
 	uint64_t val;
 	uint64_t __user *uaddr = (uint64_t __user *)attr->addr;
 
-	ret = _kvm_get_cpucfg(attr->attr, &val);
+	ret = _kvm_get_cpucfg_mask(attr->attr, &val);
 	if (ret)
 		return ret;
 
Signature · 1 line
2.43.0
author reviewer
#4 lore raw

PATCH 4/5 KVM: LoongArch: Streamline control flow of kvm_check_cpucfg

In reply to WANG Xuerui

From: WANG Xuerui <[email protected]>

All the checks currently done in kvm_check_cpucfg can be realized with early returns, so just do that to avoid extra cognitive burden related to the return value handling.

The default branch is unreachable because of the earlier validation by _kvm_get_cpucfg_mask, so mark it as such too to make things clearer.

---
 arch/loongarch/kvm/vcpu.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)
arch/loongarch/kvm/vcpu.c +11 -12
index e973500611b4..9e108ffaba30 100644
@@ -339,24 +339,23 @@ static int kvm_check_cpucfg(int id, u64 val)
 		/* CPUCFG2 features checking */
 		if (val & ~mask)
 			/* The unsupported features should not be set */
-			ret = -EINVAL;
-		else if (!(val & CPUCFG2_LLFTP))
+			return -EINVAL;
+		if (!(val & CPUCFG2_LLFTP))
 			/* The LLFTP must be set, as guest must has a constant timer */
-			ret = -EINVAL;
-		else if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
+			return -EINVAL;
+		if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
 			/* Single and double float point must both be set when enable FP */
-			ret = -EINVAL;
-		else if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
+			return -EINVAL;
+		if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
 			/* FP should be set when enable LSX */
-			ret = -EINVAL;
-		else if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))
+			return -EINVAL;
+		if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))
 			/* LSX, FP should be set when enable LASX, and FP has been checked before. */
-			ret = -EINVAL;
-		break;
+			return -EINVAL;
+		return 0;
 	default:
-		break;
+		unreachable();
 	}
-	return ret;
 }
 
 static int kvm_get_one_reg(struct kvm_vcpu *vcpu,
Signature · 1 line
2.43.0
maintainer
#5 lore raw

Re: [PATCH for-6.8 4/5] KVM: LoongArch: Streamline control flow of kvm_check_cpucfg

Hi, Xuerui,

Quoted 40 lines · WANG Xuerui

From: WANG Xuerui <[email protected]>

All the checks currently done in kvm_check_cpucfg can be realized with early returns, so just do that to avoid extra cognitive burden related to the return value handling.

The default branch is unreachable because of the earlier validation by _kvm_get_cpucfg_mask, so mark it as such too to make things clearer.

---
 arch/loongarch/kvm/vcpu.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)
arch/loongarch/kvm/vcpu.c +6 -8
index e973500611b4..9e108ffaba30 100644
@@ -339,24 +339,23 @@ static int kvm_check_cpucfg(int id, u64 val)
                /* CPUCFG2 features checking */
                if (val & ~mask)
                        /* The unsupported features should not be set */
-                       ret = -EINVAL;
-               else if (!(val & CPUCFG2_LLFTP))
+                       return -EINVAL;
+               if (!(val & CPUCFG2_LLFTP))
                        /* The LLFTP must be set, as guest must has a constant timer */
-                       ret = -EINVAL;
-               else if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
+                       return -EINVAL;
+               if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
                        /* Single and double float point must both be set when enable FP */
-                       ret = -EINVAL;
-               else if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
+                       return -EINVAL;
+               if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
                        /* FP should be set when enable LSX */
-                       ret = -EINVAL;
-               else if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))

arch/loongarch/kvm/vcpu.c static int kvm_check_cpucfg(int id, u64 val) · line 357

Quoted 10 lines
arch/loongarch/kvm/vcpu.c +5 -3
+                       return -EINVAL;
+               if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))
                        /* LSX, FP should be set when enable LASX, and FP has been checked before. */
-                       ret = -EINVAL;
-               break;
+                       return -EINVAL;
+               return 0;
        default:
-               break;
+               unreachable();

Maybe BUG() is better than unreachable()?

Huacai

Quoted 8 lines
arch/loongarch/kvm/vcpu.c +0 -2
        }
-       return ret;
 }
 
 static int kvm_get_one_reg(struct kvm_vcpu *vcpu,
--

2.43.0

author reviewer
#6 lore raw

PATCH 5/5 KVM: LoongArch: Clean up comments of _kvm_get_cpucfg_mask and kvm_check_cpucfg

In reply to WANG Xuerui

From: WANG Xuerui <[email protected]>

Remove comments that are merely restatement of the code nearby, and paraphrase the rest so they read more natural for English speakers (that lack understanding of Chinese grammar). No functional changes.

---
 arch/loongarch/kvm/vcpu.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)
arch/loongarch/kvm/vcpu.c +11 -15
index 9e108ffaba30..ff51d6ba59aa 100644
@@ -302,20 +302,14 @@ static int _kvm_get_cpucfg_mask(int id, u64 *v)
 {
 	switch (id) {
 	case 2:
-		/* Return CPUCFG2 features which have been supported by KVM */
+		/* CPUCFG2 features unconditionally supported by KVM */
 		*v = CPUCFG2_FP     | CPUCFG2_FPSP  | CPUCFG2_FPDP     |
 		     CPUCFG2_FPVERS | CPUCFG2_LLFTP | CPUCFG2_LLFTPREV |
 		     CPUCFG2_LAM;
-		/*
-		 * If LSX is supported by CPU, it is also supported by KVM,
-		 * as we implement it.
-		 */
+		/* If LSX is supported by the host, then it is also supported by KVM */
 		if (cpu_has_lsx)
 			*v |= CPUCFG2_LSX;
-		/*
-		 * if LASX is supported by CPU, it is also supported by KVM,
-		 * as we implement it.
-		 */
+		/* Same with LASX */
 		if (cpu_has_lasx)
 			*v |= CPUCFG2_LASX;
 
@@ -336,21 +330,23 @@ static int kvm_check_cpucfg(int id, u64 val)
 
 	switch (id) {
 	case 2:
-		/* CPUCFG2 features checking */
 		if (val & ~mask)
-			/* The unsupported features should not be set */
+			/* Unsupported features should not be set */
 			return -EINVAL;
 		if (!(val & CPUCFG2_LLFTP))
-			/* The LLFTP must be set, as guest must has a constant timer */
+			/* Guests must have a constant timer */
 			return -EINVAL;
 		if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
-			/* Single and double float point must both be set when enable FP */
+			/* Single and double float point must both be set when FP is enabled */
 			return -EINVAL;
 		if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
-			/* FP should be set when enable LSX */
+			/* LSX is architecturally defined to imply FP */
 			return -EINVAL;
 		if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))
-			/* LSX, FP should be set when enable LASX, and FP has been checked before. */
+			/*
+			 * LASX is architecturally defined to imply LSX and FP
+			 * FP is checked just above
+			 */
 			return -EINVAL;
 		return 0;
 	default:
Signature · 1 line
2.43.0
maintainer
#7 lore raw

Re: [PATCH for-6.8 5/5] KVM: LoongArch: Clean up comments of _kvm_get_cpucfg_mask and kvm_check_cpucfg

Hi, Xuerui,

Quoted 27 lines · WANG Xuerui

From: WANG Xuerui <[email protected]>

Remove comments that are merely restatement of the code nearby, and paraphrase the rest so they read more natural for English speakers (that lack understanding of Chinese grammar). No functional changes.

---
 arch/loongarch/kvm/vcpu.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)
arch/loongarch/kvm/vcpu.c +1 -3
index 9e108ffaba30..ff51d6ba59aa 100644
@@ -302,20 +302,14 @@ static int _kvm_get_cpucfg_mask(int id, u64 *v)
 {
        switch (id) {
        case 2:
-               /* Return CPUCFG2 features which have been supported by KVM */
+               /* CPUCFG2 features unconditionally supported by KVM */
                *v = CPUCFG2_FP     | CPUCFG2_FPSP  | CPUCFG2_FPDP     |
                     CPUCFG2_FPVERS | CPUCFG2_LLFTP | CPUCFG2_LLFTPREV |
                     CPUCFG2_LAM;
-               /*
-                * If LSX is supported by CPU, it is also supported by KVM,

arch/loongarch/kvm/vcpu.c static int _kvm_get_cpucfg_mask(int id, u64 *v) · line 312

Quoted 10 lines
arch/loongarch/kvm/vcpu.c +2 -6
-                * as we implement it.
-                */
+               /* If LSX is supported by the host, then it is also supported by KVM */
                if (cpu_has_lsx)
                        *v |= CPUCFG2_LSX;
-               /*
-                * if LASX is supported by CPU, it is also supported by KVM,
-                * as we implement it.
-                */
+               /* Same with LASX */

Consider a full description "If LASX is supported by the host, then it is also supported by KVM"?

Quoted 19 lines
arch/loongarch/kvm/vcpu.c +3 -4
                if (cpu_has_lasx)
                        *v |= CPUCFG2_LASX;
 
@@ -336,21 +330,23 @@ static int kvm_check_cpucfg(int id, u64 val)
 
        switch (id) {
        case 2:
-               /* CPUCFG2 features checking */
                if (val & ~mask)
-                       /* The unsupported features should not be set */
+                       /* Unsupported features should not be set */
                        return -EINVAL;
                if (!(val & CPUCFG2_LLFTP))
-                       /* The LLFTP must be set, as guest must has a constant timer */
+                       /* Guests must have a constant timer */
                        return -EINVAL;
                if ((val & CPUCFG2_FP) && (!(val & CPUCFG2_FPSP) || !(val & CPUCFG2_FPDP)))
-                       /* Single and double float point must both be set when enable FP */
+                       /* Single and double float point must both be set when FP is enabled */

arch/loongarch/kvm/vcpu.c static int kvm_check_cpucfg(int id, u64 val) · line 348

Quoted 10 lines
arch/loongarch/kvm/vcpu.c +4 -2
                        return -EINVAL;
                if ((val & CPUCFG2_LSX) && !(val & CPUCFG2_FP))
-                       /* FP should be set when enable LSX */
+                       /* LSX is architecturally defined to imply FP */
                        return -EINVAL;
                if ((val & CPUCFG2_LASX) && !(val & CPUCFG2_LSX))
-                       /* LSX, FP should be set when enable LASX, and FP has been checked before. */
+                       /*
+                        * LASX is architecturally defined to imply LSX and FP
+                        * FP is checked just above

I think "LASX is architecturally defined to imply LSX and FP" is enough here.

arch/loongarch/kvm/vcpu.c static int kvm_check_cpucfg(int id, u64 val) · line 352

Quoted 4 lines
arch/loongarch/kvm/vcpu.c +1 -0
+                        */
                        return -EINVAL;
                return 0;
        default:

And I prefer to squash the last two patches together.

Huacai

Quoted 4 lines

--
2.43.0