Skip to contentSkip to conversation

KVM: s390: pci: fix GAIT table indexing due to double-scaling pointer arithmetic

PATCH Applied Junrui Luo 3 participants 5 messages Apr 15 – Apr 17

Subsystems (4)
lore
Share & tools
lore

Junrui Luo

author
#0 lore raw

PATCH 1/1 KVM: s390: pci: fix GAIT table indexing due to double-scaling pointer arithmetic

kvm_s390_pci_aif_enable(), kvm_s390_pci_aif_disable(), and aen_host_forward() index the GAIT by manually multiplying the index with sizeof(struct zpci_gaite).

Since aift->gait is already a struct zpci_gaite pointer, this double-scales the offset, accessing element aisb*16 instead of aisb.

This causes out-of-bounds accesses when aisb >= 32 (with ZPCI_NR_DEVICES=512)

Fix by removing the erroneous sizeof multiplication.

Fixes 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/disabling interrupt forwarding")
Fixes 73f91b004321 ("KVM: s390: pci: enable host forwarding of Adapter Event Notifications")
---
 arch/s390/kvm/interrupt.c | 3 +--
 arch/s390/kvm/pci.c       | 6 ++----
 2 files changed, 3 insertions(+), 6 deletions(-)
arch/s390/kvm/interrupt.c +1 -2
index 7cb8ce833b62..f48f25c7dc8f 100644
@@ -3307,8 +3307,7 @@ static void aen_host_forward(unsigned long si)
 	struct zpci_gaite *gaite;
 	struct kvm *kvm;
 
-	gaite = (struct zpci_gaite *)aift->gait +
-		(si * sizeof(struct zpci_gaite));
+	gaite = aift->gait + si;
 	if (gaite->count == 0)
 		return;
 	if (gaite->aisb != 0)
arch/s390/kvm/pci.c +2 -4
index 86d93e8dddae..eed45af1a92d 100644
@@ -290,8 +290,7 @@ static int kvm_s390_pci_aif_enable(struct zpci_dev *zdev, struct zpci_fib *fib,
 				    phys_to_virt(fib->fmt0.aibv));
 
 	spin_lock_irq(&aift->gait_lock);
-	gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb *
-						   sizeof(struct zpci_gaite));
+	gaite = aift->gait + zdev->aisb;
 
 	/* If assist not requested, host will get all alerts */
 	if (assist)
@@ -357,8 +356,7 @@ static int kvm_s390_pci_aif_disable(struct zpci_dev *zdev, bool force)
 	if (zdev->kzdev->fib.fmt0.aibv == 0)
 		goto out;
 	spin_lock_irq(&aift->gait_lock);
-	gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb *
-						   sizeof(struct zpci_gaite));
+	gaite = aift->gait + zdev->aisb;
 	isc = gaite->gisc;
 	gaite->count--;
 	if (gaite->count == 0) {

---
base-commit: 7aaa8047eafd0bd628065b15757d9b48c5f9c07d
change-id: 20260415-fixes-0a69953f011b

Best regards,

Signature · 1 line
Junrui Luo <[email protected]>
maintainer Reviewed-by on v1 1/1
#1 lore raw

Am 15.04.26 um 11:26 schrieb Junrui Luo:

Quoted 10 lines

kvm_s390_pci_aif_enable(), kvm_s390_pci_aif_disable(), and aen_host_forward() index the GAIT by manually multiplying the index with sizeof(struct zpci_gaite).

Since aift->gait is already a struct zpci_gaite pointer, this double-scales the offset, accessing element aisb*16 instead of aisb.

This causes out-of-bounds accesses when aisb >= 32 (with ZPCI_NR_DEVICES=512)

Quoted 7 lines

Fix by removing the erroneous sizeof multiplication.

Fixes 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/disabling interrupt forwarding")
Fixes 73f91b004321 ("KVM: s390: pci: enable host forwarding of Adapter Event Notifications")

looks good to me.

Out of interest, was this found by static code checking or AI by any chance?

@Matt, can you test/review this as well?

Quoted 44 lines
---
  arch/s390/kvm/interrupt.c | 3 +--
  arch/s390/kvm/pci.c       | 6 ++----
  2 files changed, 3 insertions(+), 6 deletions(-)
arch/s390/kvm/interrupt.c +1 -2
index 7cb8ce833b62..f48f25c7dc8f 100644
@@ -3307,8 +3307,7 @@ static void aen_host_forward(unsigned long si)
  	struct zpci_gaite *gaite;
  	struct kvm *kvm;
 
-	gaite = (struct zpci_gaite *)aift->gait +
-		(si * sizeof(struct zpci_gaite));
+	gaite = aift->gait + si;
  	if (gaite->count == 0)
  		return;
  	if (gaite->aisb != 0)
arch/s390/kvm/pci.c +2 -4
index 86d93e8dddae..eed45af1a92d 100644
@@ -290,8 +290,7 @@ static int kvm_s390_pci_aif_enable(struct zpci_dev *zdev, struct zpci_fib *fib,
  				    phys_to_virt(fib->fmt0.aibv));
 
  	spin_lock_irq(&aift->gait_lock);
-	gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb *
-						   sizeof(struct zpci_gaite));
+	gaite = aift->gait + zdev->aisb;
 
  	/* If assist not requested, host will get all alerts */
  	if (assist)
@@ -357,8 +356,7 @@ static int kvm_s390_pci_aif_disable(struct zpci_dev *zdev, bool force)
  	if (zdev->kzdev->fib.fmt0.aibv == 0)
  		goto out;
  	spin_lock_irq(&aift->gait_lock);
-	gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb *
-						   sizeof(struct zpci_gaite));
+	gaite = aift->gait + zdev->aisb;
  	isc = gaite->gisc;
  	gaite->count--;
  	if (gaite->count == 0) {
Reviewed-by Tested-by
#2 lore raw
Quoted 22 lines · Christian Borntraeger

Am 15.04.26 um 11:26 schrieb Junrui Luo:

Quoted 12 lines

kvm_s390_pci_aif_enable(), kvm_s390_pci_aif_disable(), and aen_host_forward() index the GAIT by manually multiplying the index with sizeof(struct zpci_gaite).

Since aift->gait is already a struct zpci_gaite pointer, this double-scales the offset, accessing element aisb*16 instead of aisb.

This causes out-of-bounds accesses when aisb >= 32 (with ZPCI_NR_DEVICES=512)

Fix by removing the erroneous sizeof multiplication.

Quoted 7 lines
Fixes 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/

disabling interrupt forwarding")

Fixes 73f91b004321 ("KVM: s390: pci: enable host forwarding of

Adapter Event Notifications")

looks good to me.

Quoted 9 lines

Out of interest, was this found by static code checking or AI by any chance?

@Matt, can you test/review this as well?

Thanks for the report and the fix.

I did some tracing to confirm the issue -- indeed, gaite is being indexed by 256B instead of 16B.

Because the incorrect offset calculation was used consistently in all 3 spots, the expected gaite is always found; a problem only arises once we attempt to access beyond the allocated number of pages for the gaite array.

When this can possibly happen is a factor of the configurable CONFIG_PCI_NR_FUNCTIONS -- but yes, as the commit message alludes to, the default for this is 512 which means once aisb reaches 32 * 256B we would cross beyond the 2 contiguous 4K pages that would have been allocated for 512 * 16B entries.

I also did some general regression testing with a mixture of ISM, mlx and NVMe devices with this patch applied.

author
#3 lore raw

Hi Christian, Matthew,

Thanks for the review and testing.

Quoted 1 line · Christian Borntraeger

Out of interest, was this found by static code checking or AI by any chance?

Yes, it was found with the help of an LLM, and we verified the issue manually.

Thanks,
Junrui Luo

maintainer Applied
#4 lore raw
In reply to Junrui Luo

Am 15.04.26 um 11:26 schrieb Junrui Luo:

Quoted 10 lines

kvm_s390_pci_aif_enable(), kvm_s390_pci_aif_disable(), and aen_host_forward() index the GAIT by manually multiplying the index with sizeof(struct zpci_gaite).

Since aift->gait is already a struct zpci_gaite pointer, this double-scales the offset, accessing element aisb*16 instead of aisb.

This causes out-of-bounds accesses when aisb >= 32 (with ZPCI_NR_DEVICES=512)

Quoted 7 lines

Fix by removing the erroneous sizeof multiplication.

Fixes 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/disabling interrupt forwarding")
Fixes 73f91b004321 ("KVM: s390: pci: enable host forwarding of Adapter Event Notifications")

thanks, applied.