| Message ID | 20191016200647.32050-18-robh@kernel.org (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Geert Uytterhoeven |
| Headers | show |
| Series | PCI host resource consolidation | expand |
On Wed, Oct 16, 2019 at 10:07 PM Rob Herring <robh@kernel.org> wrote: > PHYS_OFFSET is not universally defined on all arches and using it prevents > enabling COMPILE_TEST. PAGE_OFFSET and __pa() are always available, so use > them to get the physical start of memory address. > > This should have probably used 'dma-ranges' to get the address, but we > don't want to force a DT update to do that. At least in QEMU, the SMAP > registers have no effect (or perhaps the only value that is handled is 0). > > Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> > Cc: Andrew Murray <andrew.murray@arm.com> > Cc: Bjorn Helgaas <bhelgaas@google.com> > Signed-off-by: Rob Herring <robh@kernel.org> > --- > v2: > - New patch to fix build failure on some arches. Always wondered how to do this right, now I can fix other stuff! Acked-by: Linus Walleij <linus.walleij@linaro.org> Yours, Linus Walleij
diff --git a/drivers/pci/controller/pci-versatile.c b/drivers/pci/controller/pci-versatile.c index 18697f2ea345..eae1b859990b 100644 --- a/drivers/pci/controller/pci-versatile.c +++ b/drivers/pci/controller/pci-versatile.c @@ -99,7 +99,7 @@ static int versatile_pci_probe(struct platform_device *pdev) resource_list_for_each_entry(entry, &bridge->windows) { if (resource_type(entry->res) == IORESOURCE_MEM) { writel(entry->res->start >> 28, PCI_IMAP(mem)); - writel(PHYS_OFFSET >> 28, PCI_SMAP(mem)); + writel(__pa(PAGE_OFFSET) >> 28, PCI_SMAP(mem)); mem++; } } @@ -136,9 +136,9 @@ static int versatile_pci_probe(struct platform_device *pdev) /* * Configure the PCI inbound memory windows to be 1:1 mapped to SDRAM */ - writel(PHYS_OFFSET, local_pci_cfg_base + PCI_BASE_ADDRESS_0); - writel(PHYS_OFFSET, local_pci_cfg_base + PCI_BASE_ADDRESS_1); - writel(PHYS_OFFSET, local_pci_cfg_base + PCI_BASE_ADDRESS_2); + writel(__pa(PAGE_OFFSET), local_pci_cfg_base + PCI_BASE_ADDRESS_0); + writel(__pa(PAGE_OFFSET), local_pci_cfg_base + PCI_BASE_ADDRESS_1); + writel(__pa(PAGE_OFFSET), local_pci_cfg_base + PCI_BASE_ADDRESS_2); /* * For many years the kernel and QEMU were symbiotically buggy
PHYS_OFFSET is not universally defined on all arches and using it prevents enabling COMPILE_TEST. PAGE_OFFSET and __pa() are always available, so use them to get the physical start of memory address. This should have probably used 'dma-ranges' to get the address, but we don't want to force a DT update to do that. At least in QEMU, the SMAP registers have no effect (or perhaps the only value that is handled is 0). Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Cc: Andrew Murray <andrew.murray@arm.com> Cc: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Rob Herring <robh@kernel.org> --- v2: - New patch to fix build failure on some arches. drivers/pci/controller/pci-versatile.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) -- 2.20.1