Skip to contentSkip to conversation

pci: assert configuration access is within bounds

PATCH In review P J P 3 participants 4 messages Jun 4, 2020 – Jun 9, 2020

lore
Share & tools
lore

P J P

author
#0 lore raw

PATCH 1/1 pci: assert configuration access is within bounds

From: Prasad J Pandit <[email protected]>

While accessing PCI configuration bytes, assert that 'address + len' is within PCI configuration space.

Generally it is within bounds. This is more of a defensive assert, in case a buggy device was to send 'address' which may go out of bounds.

---
 hw/pci/pci.c | 4 ++++
 1 file changed, 4 insertions(+)

Update v1: defensive assert as separate patch
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00957.html
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00960.html

hw/pci/pci.c +4 -0
index 70c66965f5..7bf2ae6d92 100644
@@ -1381,6 +1381,8 @@ uint32_t pci_default_read_config(PCIDevice *d,
 {
     uint32_t val = 0;
 
+    assert(address + len <= pci_config_size(d));
+
     if (pci_is_express_downstream_port(d) &&
         ranges_overlap(address, len, d->exp.exp_cap + PCI_EXP_LNKSTA, 2)) {
         pcie_sync_bridge_lnk(d);
@@ -1394,6 +1396,8 @@ void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val_in, int
     int i, was_irq_disabled = pci_irq_disabled(d);
     uint32_t val = val_in;
 
+    assert(addr + l <= pci_config_size(d));
+
     for (i = 0; i < l; val >>= 8, ++i) {
         uint8_t wmask = d->wmask[addr + i];
         uint8_t w1cmask = d->w1cmask[addr + i];
Signature · 1 line
2.26.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · P J P

From: Prasad J Pandit <[email protected]>

While accessing PCI configuration bytes, assert that 'address + len' is within PCI configuration space.

Generally it is within bounds. This is more of a defensive assert, in case a buggy device was to send 'address' which may go out of bounds.

Quoted 32 lines
---
 hw/pci/pci.c | 4 ++++
 1 file changed, 4 insertions(+)

Update v1: defensive assert as separate patch
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00957.html
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00960.html

hw/pci/pci.c +4 -0
index 70c66965f5..7bf2ae6d92 100644
@@ -1381,6 +1381,8 @@ uint32_t pci_default_read_config(PCIDevice *d,
 {
     uint32_t val = 0;
 
+    assert(address + len <= pci_config_size(d));
+
     if (pci_is_express_downstream_port(d) &&
         ranges_overlap(address, len, d->exp.exp_cap + PCI_EXP_LNKSTA, 2)) {
         pcie_sync_bridge_lnk(d);
@@ -1394,6 +1396,8 @@ void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val_in, int
     int i, was_irq_disabled = pci_irq_disabled(d);
     uint32_t val = val_in;
 
+    assert(addr + l <= pci_config_size(d));
+
     for (i = 0; i < l; val >>= 8, ++i) {
         uint8_t wmask = d->wmask[addr + i];
         uint8_t w1cmask = d->w1cmask[addr + i];
Signature · 1 line
2.26.2
#2 lore raw
In reply to P J P
Quoted 8 lines · * P J P ([email protected]) wrote:

From: Prasad J Pandit <[email protected]>

While accessing PCI configuration bytes, assert that 'address + len' is within PCI configuration space.

Generally it is within bounds. This is more of a defensive assert, in case a buggy device was to send 'address' which may go out of bounds.

Can a guest trigger this assert?

Dave

Quoted 36 lines
---
 hw/pci/pci.c | 4 ++++
 1 file changed, 4 insertions(+)

Update v1: defensive assert as separate patch
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00957.html
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00960.html

hw/pci/pci.c +4 -0
index 70c66965f5..7bf2ae6d92 100644
@@ -1381,6 +1381,8 @@ uint32_t pci_default_read_config(PCIDevice *d,
 {
     uint32_t val = 0;
 
+    assert(address + len <= pci_config_size(d));
+
     if (pci_is_express_downstream_port(d) &&
         ranges_overlap(address, len, d->exp.exp_cap + PCI_EXP_LNKSTA, 2)) {
         pcie_sync_bridge_lnk(d);
@@ -1394,6 +1396,8 @@ void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val_in, int
     int i, was_irq_disabled = pci_irq_disabled(d);
     uint32_t val = val_in;
 
+    assert(addr + l <= pci_config_size(d));
+
     for (i = 0; i < l; val >>= 8, ++i) {
         uint8_t wmask = d->wmask[addr + i];
         uint8_t w1cmask = d->w1cmask[addr + i];
Signature · 1 line
2.26.2

--
Dr. David Alan Gilbert / [email protected] / Manchester, UK

#3 lore raw
Quoted 6 lines · Dr. David Alan Gilbert
Quoted 5 lines · * P J P ([email protected]) wrote:

From: Prasad J Pandit <[email protected]>

While accessing PCI configuration bytes, assert that 'address + len' is within PCI configuration space.

Quoted 7 lines
Quoted 3 lines

Generally it is within bounds. This is more of a defensive assert, in case a buggy device was to send 'address' which may go out of bounds.

Can a guest trigger this assert?

Dave

Shouldn't be able to, no.

Quoted 38 lines
Quoted 36 lines
---
 hw/pci/pci.c | 4 ++++
 1 file changed, 4 insertions(+)

Update v1: defensive assert as separate patch
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00957.html
-> https://lists.gnu.org/archive/html/qemu-devel/2020-06/msg00960.html

hw/pci/pci.c +4 -0
index 70c66965f5..7bf2ae6d92 100644
@@ -1381,6 +1381,8 @@ uint32_t pci_default_read_config(PCIDevice *d,
 {
     uint32_t val = 0;
 
+    assert(address + len <= pci_config_size(d));
+
     if (pci_is_express_downstream_port(d) &&
         ranges_overlap(address, len, d->exp.exp_cap + PCI_EXP_LNKSTA, 2)) {
         pcie_sync_bridge_lnk(d);
@@ -1394,6 +1396,8 @@ void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val_in, int
     int i, was_irq_disabled = pci_irq_disabled(d);
     uint32_t val = val_in;
 
+    assert(addr + l <= pci_config_size(d));
+
     for (i = 0; i < l; val >>= 8, ++i) {
         uint8_t wmask = d->wmask[addr + i];
         uint8_t w1cmask = d->w1cmask[addr + i];
Signature · 1 line
2.26.2

--
Dr. David Alan Gilbert / [email protected] / Manchester, UK