On 06.12.2016 21:45, Bjorn Helgaas wrote:
On Tue, Dec 06, 2016 at 02:07:45PM +0100, Tomasz Nowicki wrote:
Hi Bjorn,
On 06.12.2016 00:26, Bjorn Helgaas wrote:
diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index d7e7c0a..1239a8e 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -240,7 +240,8 @@ config PCIE_QCOM
config PCI_HOST_THUNDER_PEM
bool "Cavium Thunder PCIe controller to off-chip devices"
- depends on OF && ARM64
+ depends on ARM64
+ depends on OF || (ACPI && PCI_QUIRKS)
select PCI_HOST_COMMON
help
Say Y here if you want PCIe support for CN88XX Cavium Thunder SoCs.
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 64845f0..97e6bfc 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -28,7 +28,7 @@ obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
obj-$(CONFIG_ARM64) += pcie-hisi.o
obj-$(CONFIG_PCIE_QCOM) += pcie-qcom.o
obj-$(CONFIG_PCI_HOST_THUNDER_ECAM) += pci-thunder-ecam.o
-obj-$(CONFIG_PCI_HOST_THUNDER_PEM) += pci-thunder-pem.o
+obj-$(CONFIG_ARM64) += pci-thunder-pem.o
Since we build pci-thunder-pem.o for ARM64 and ...
obj-$(CONFIG_PCIE_ARMADA_8K) += pcie-armada8k.o
obj-$(CONFIG_PCIE_ARTPEC6) += pcie-artpec6.o
obj-$(CONFIG_PCIE_ROCKCHIP) += pcie-rockchip.o
diff --git a/drivers/pci/host/pci-thunder-pem.c b/drivers/pci/host/pci-thunder-pem.c
index c3276ee..af722eb 100644
--- a/drivers/pci/host/pci-thunder-pem.c
+++ b/drivers/pci/host/pci-thunder-pem.c
@@ -18,8 +18,12 @@
#include <linux/init.h>
#include <linux/of_address.h>
#include <linux/of_pci.h>
+#include <linux/pci-acpi.h>
#include <linux/pci-ecam.h>
#include <linux/platform_device.h>
+#include "../pci.h"
+
+#if defined(CONFIG_PCI_HOST_THUNDER_PEM) || (defined(CONFIG_ACPI) && defined(CONFIG_PCI_QUIRKS))
have this ^^^ ifdef IMO we can drop Kconfig changes in
drivers/pci/host/Kconfig. The same for PCI_HOST_THUNDER_ECAM in
patch[14/15].
I think you're right that we *could* drop the Kconfig changes. I
didn't actually drop them because I think it might be slightly
confusing that with the previous Kconfig:
config PCI_HOST_THUNDER_PEM
bool "Cavium Thunder PCIe controller to off-chip devices"
depends on OF && ARM64
we still build pci-thunder-pem.o, even when CONFIG_OF isn't defined.
There's no question that this whole thing is confusing no matter what
we do. That reminds me that I meant to add some comments in the
Makefile. What do you think about the following:
commit 5d06f9125ec00215fa59b789491fdd51186a4ee1
Author: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
Date: Tue Dec 6 14:27:59 2016 -0600
PCI: Explain ARM64 ACPI/MCFG quirk Kconfig and build strategy
Add Makefile comments to explain the Kconfig and build strategy for ARM64
drivers that work around not-quite-ECAM issues. No functional change
intended.
Signed-off-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 6cc84b4..bfe3179 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -15,7 +15,6 @@ obj-$(CONFIG_PCIE_SPEAR13XX) += pcie-spear13xx.o
obj-$(CONFIG_PCI_KEYSTONE) += pci-keystone-dw.o pci-keystone.o
obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
obj-$(CONFIG_PCIE_XILINX_NWL) += pcie-xilinx-nwl.o
-obj-$(CONFIG_ARM64) += pci-xgene.o
obj-$(CONFIG_PCI_XGENE_MSI) += pci-xgene-msi.o
obj-$(CONFIG_PCI_LAYERSCAPE) += pci-layerscape.o
obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
@@ -25,11 +24,23 @@ obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
-obj-$(CONFIG_ARM64) += pcie-hisi.o
obj-$(CONFIG_PCIE_QCOM) += pcie-qcom.o
-obj-$(CONFIG_ARM64) += pci-thunder-ecam.o
-obj-$(CONFIG_ARM64) += pci-thunder-pem.o
obj-$(CONFIG_PCIE_ARMADA_8K) += pcie-armada8k.o
obj-$(CONFIG_PCIE_ARTPEC6) += pcie-artpec6.o
obj-$(CONFIG_PCIE_ROCKCHIP) += pcie-rockchip.o
obj-$(CONFIG_VMD) += vmd.o
+
+# The following drivers are for devices that use the generic ACPI
+# pci_root.c driver but don't support standard ECAM config access.
+# They contain MCFG quirks to replace the generic ECAM accessors with
+# device-specific ones that are shared with the DT driver.
+
+# The ACPI driver is generic and should not require driver-specific
+# config options to be enabled, so we always build these drivers on
+# ARM64 and use internal ifdefs to only build the pieces we need
+# depending on whether ACPI, the DT driver, or both are enabled.
+
+obj-$(CONFIG_ARM64) += pcie-hisi.o
+obj-$(CONFIG_ARM64) += pci-thunder-ecam.o
+obj-$(CONFIG_ARM64) += pci-thunder-pem.o
+obj-$(CONFIG_ARM64) += pci-xgene.o
Looks good to me.
Tomasz
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
