Skip to contentSkip to conversation

pci: remove explicit check to 64K ioport size

PATCH In review Hervé Poussineau 3 participants 4 messages Sep 13, 2013 – Sep 15, 2013

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 pci: remove explicit check to 64K ioport size

This check is useless, as bigger addresses will be ignored when added to 'io' MemoryRegion, which has a size of 64K.

However, some architectures don't use the 'io' MemoryRegion, like the alpha and versatile platforms. They create a PCI I/O region bigger than 64K, so let them handle PCI I/O BARs in the higher range.

---
 hw/pci/pci.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
hw/pci/pci.c +1 -2
index d00682e..a8e2b29 100644
@@ -1028,8 +1028,7 @@ static pcibus_t pci_bar_address(PCIDevice *d,
         }
         new_addr = pci_get_long(d->config + bar) & ~(size - 1);
         last_addr = new_addr + size - 1;
-        /* NOTE: we have only 64K ioports on PC */
-        if (last_addr <= new_addr || new_addr == 0 || last_addr > UINT16_MAX) {
+        if (last_addr <= new_addr || new_addr == 0) {
             return PCI_BAR_UNMAPPED;
         }
         return new_addr;
Signature · 1 line
1.7.10.4
Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · Hervé Poussineau

This check is useless, as bigger addresses will be ignored when added to 'io' MemoryRegion, which has a size of 64K.

However, some architectures don't use the 'io' MemoryRegion, like the alpha and versatile platforms. They create a PCI I/O region bigger than 64K, so let them handle PCI I/O BARs in the higher range.

---
 hw/pci/pci.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

r~

#2 lore raw
In reply to Hervé Poussineau
Quoted 28 lines · Hervé Poussineau

This check is useless, as bigger addresses will be ignored when added to 'io' MemoryRegion, which has a size of 64K.

However, some architectures don't use the 'io' MemoryRegion, like the alpha and versatile platforms. They create a PCI I/O region bigger than 64K, so let them handle PCI I/O BARs in the higher range.

---
 hw/pci/pci.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
hw/pci/pci.c +1 -2
index d00682e..a8e2b29 100644
@@ -1028,8 +1028,7 @@ static pcibus_t pci_bar_address(PCIDevice *d,
         }
         new_addr = pci_get_long(d->config + bar) & ~(size - 1);
         last_addr = new_addr + size - 1;
-        /* NOTE: we have only 64K ioports on PC */
-        if (last_addr <= new_addr || new_addr == 0 || last_addr > UINT16_MAX) {
+        if (last_addr <= new_addr || new_addr == 0) {
             return PCI_BAR_UNMAPPED;
         }
         return new_addr;
Signature · 1 line
1.7.10.4

We still didn't fix the overlap for PCI BARs properly
so it looks like the following is needed on top.
Could you please check whether this works for you on versatile/alpha?

hw/pci/pci.c +4 -1
index d452135..3118854 100644
@@ -1028,7 +1028,10 @@ static pcibus_t pci_bar_address(PCIDevice *d,
         }
         new_addr = pci_get_long(d->config + bar) & ~(size - 1);
         last_addr = new_addr + size - 1;
-        if (last_addr <= new_addr || new_addr == 0) {
+        /* Check if BAR is being sized explicitly.
+         * TODO: make priorities correct and remove this work around.
+         */
+        if (last_addr <= new_addr || new_addr == 0 || last_addr >= UINT32_MAX) {
             return PCI_BAR_UNMAPPED;
         }
         return new_addr;
#3 lore raw
Quoted 25 lines · Michael S. Tsirkin
Quoted 24 lines · Hervé Poussineau

This check is useless, as bigger addresses will be ignored when added to 'io' MemoryRegion, which has a size of 64K.

However, some architectures don't use the 'io' MemoryRegion, like the alpha and versatile platforms. They create a PCI I/O region bigger than 64K, so let them handle PCI I/O BARs in the higher range.

---
 hw/pci/pci.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
hw/pci/pci.c +1 -2
index d00682e..a8e2b29 100644
@@ -1028,8 +1028,7 @@ static pcibus_t pci_bar_address(PCIDevice *d,
         }
         new_addr = pci_get_long(d->config + bar) & ~(size - 1);
         last_addr = new_addr + size - 1;
-        /* NOTE: we have only 64K ioports on PC */
-        if (last_addr <= new_addr || new_addr == 0 || last_addr > UINT16_MAX) {
+        if (last_addr <= new_addr || new_addr == 0) {
             return PCI_BAR_UNMAPPED;
Quoted 8 lines
Quoted 4 lines
hw/pci/pci.c +0 -0
         }
         return new_addr;
Signature · 1 line
1.7.10.4

We still didn't fix the overlap for PCI BARs properly
so it looks like the following is needed on top.
Could you please check whether this works for you on versatile/alpha?

Yes, alpha and versatile still seem to work.

Hervé

Quoted 18 lines
hw/pci/pci.c +4 -1
index d452135..3118854 100644
@@ -1028,7 +1028,10 @@ static pcibus_t pci_bar_address(PCIDevice *d,
         }
         new_addr = pci_get_long(d->config + bar) & ~(size - 1);
         last_addr = new_addr + size - 1;
-        if (last_addr <= new_addr || new_addr == 0) {
+        /* Check if BAR is being sized explicitly.
+         * TODO: make priorities correct and remove this work around.
+         */
+        if (last_addr <= new_addr || new_addr == 0 || last_addr >= UINT32_MAX) {
             return PCI_BAR_UNMAPPED;
         }
         return new_addr;