[PATCH 7/8] ppc/pnv: using PCI core to do resource assignment
Gavin Shan
shangw at linux.vnet.ibm.com
Tue Jun 26 01:43:20 EST 2012
Currently, the PCI probe flags "PCI_PROBE_ONLY | PCI_REASSIGN_ALL_RSRC"
used on powernv platform. That means the platform has to do the PCI
resource assignment by itself.
The patch changes the PCI probe flag to "PCI_REASSIGN_ALL_RSRC" so
that the PCI core will do the resource assignment. Also, the I/O
and MMIO minimal alignment for P2P bridges have been configured
while doing fixup for the PHBs.
Signed-off-by: Gavin Shan <shangw at linux.vnet.ibm.com>
Reviewed-by: Ram Pai <linuxram at us.ibm.com>
Reviewed-by: Richard Yang <weiyang at linux.vnet.ibm.com>
---
arch/powerpc/platforms/powernv/pci-ioda.c | 46 ++++++++++++-----------------
1 file changed, 19 insertions(+), 27 deletions(-)
diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c
index ed8cf09..1368804 100644
--- a/arch/powerpc/platforms/powernv/pci-ioda.c
+++ b/arch/powerpc/platforms/powernv/pci-ioda.c
@@ -1130,29 +1130,19 @@ static void pnv_pci_init_ioda_msis(struct pnv_phb *phb) { }
*/
static void __devinit pnv_pci_ioda_fixup_phb(struct pci_controller *hose)
{
- resource_size_t size, align;
- struct pci_bus *child;
-
- /* Associate PEs per functions */
- pnv_ioda_setup_PEs(hose->bus);
-
- /* Calculate all resources */
- pnv_ioda_calc_bus(hose->bus, IORESOURCE_IO, &size, &align);
- pnv_ioda_calc_bus(hose->bus, IORESOURCE_MEM, &size, &align);
-
- /* Apply then to HW */
- pnv_ioda_update_resources(hose->bus);
-
- /* Setup DMA */
- pnv_ioda_setup_dma(hose->private_data);
+ struct pnv_phb *phb = hose->private_data;
+ struct pci_host_bridge *host_bridge;
- /* Configure PCI Express settings */
- list_for_each_entry(child, &hose->bus->children, node) {
- struct pci_dev *self = child->self;
- if (!self)
- continue;
- pcie_bus_configure_settings(child, self->pcie_mpss);
- }
+ host_bridge = pci_bus_host_bridge(hose->bus);
+ host_bridge->io_align_shift = __ffs(phb->ioda.io_segsize);
+ host_bridge->mem_align_shift = __ffs(phb->ioda.m32_segsize);
+ /*
+ * Note: We do not use the M64 window yet (64-bit MMIO window),
+ * so we leave the alignment setting alone. In the long run,
+ * we might use it for prefetchable memory regions in which
+ * case pmem_align_shift will have to be adjusted as well.
+ */
+ host_bridge->pmem_align_shift = host_bridge->mem_align_shift;
}
/*
@@ -1437,15 +1427,17 @@ void __init pnv_pci_init_ioda1_phb(struct device_node *np)
/* Setup MSI support */
pnv_pci_init_ioda_msis(phb);
- /* We set both PCI_PROBE_ONLY and PCI_REASSIGN_ALL_RSRC. This is an
- * odd combination which essentially means that we skip all resource
- * fixups and assignments in the generic code, and do it all
- * ourselves here
+ /*
+ * We pass the PCI probe flag PCI_REASSIGN_ALL_RSRC here
+ * to let the PCI core do resource assignment. It's supposed
+ * that the PCI core will do correct I/O and MMIO alignment
+ * for the P2P bridge bars so that each PCI bus (excluding
+ * the child P2P bridges) can form individual PE.
*/
ppc_md.pcibios_fixup_phb = pnv_pci_ioda_fixup_phb;
ppc_md.pcibios_fixup = pnv_pci_ioda_fixup;
ppc_md.pcibios_enable_device_hook = pnv_pci_enable_device_hook;
- pci_add_flags(PCI_PROBE_ONLY | PCI_REASSIGN_ALL_RSRC);
+ pci_add_flags(PCI_REASSIGN_ALL_RSRC);
/* Reset IODA tables to a clean state */
rc = opal_pci_reset(phb_id, OPAL_PCI_IODA_TABLE_RESET, OPAL_ASSERT_RESET);
--
1.7.9.5
More information about the Linuxppc-dev
mailing list