summaryrefslogtreecommitdiff
path: root/arch/x86/platform
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2022-12-14 20:54:10 +0300
committerLinus Torvalds <torvalds@linux-foundation.org>2022-12-14 20:54:10 +0300
commitc7020e1b346d5840e93b58cc4f2c67fc645d8df9 (patch)
treea73962eda2212fd73f1a50b122a96833693d0094 /arch/x86/platform
parenta0a6c76cf2a506c061072db6c187f8a641dc85de (diff)
parentf826afe5eae856b3834cbc65db6178cccd4a3142 (diff)
downloadlinux-c7020e1b346d5840e93b58cc4f2c67fc645d8df9.tar.xz
Merge tag 'pci-v6.2-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull PCI updates from Bjorn Helgaas: "Enumeration: - Squash portdrv_{core,pci}.c into portdrv.c to ease maintenance and make more things static. - Make portdrv bind to Switch Ports that have AER. Previously, if these Ports lacked MSI/MSI-X, portdrv failed to bind, which meant the Ports couldn't be suspended to low-power states. AER on these Ports doesn't use interrupts, and the AER driver doesn't need to claim them. - Assign PCI domain IDs using ida_alloc(), which makes host bridge add/remove work better. Resource management: - To work better with recent BIOSes that use EfiMemoryMappedIO for PCI host bridge apertures, remove those regions from the E820 map (E820 entries normally prevent us from allocating BARs). In v5.19, we added some quirks to disable E820 checking, but that's not very maintainable. EfiMemoryMappedIO means the OS needs to map the region for use by EFI runtime services; it shouldn't prevent OS from using it. PCIe native device hotplug: - Build pciehp by default if USB4 is enabled, since Thunderbolt/USB4 PCIe tunneling depends on native PCIe hotplug. - Enable Command Completed Interrupt only if supported to avoid user confusion from lspci output that says this is enabled but not supported. - Prevent pciehp from binding to Switch Upstream Ports; this happened because of interaction with acpiphp and caused devices below the Upstream Port to disappear. Power management: - Convert AGP drivers to generic power management. We hope to remove legacy power management from the PCI core eventually. Virtualization: - Fix pci_device_is_present(), which previously always returned "false" for VFs, causing virtio hangs when unbinding the driver. Miscellaneous: - Convert drivers to gpiod API to prepare for dropping some legacy code. - Fix DOE fencepost error for the maximum data object length. Baikal-T1 PCIe controller driver: - Add driver and DT bindings. Broadcom STB PCIe controller driver: - Enable Multi-MSI. - Delay 100ms after PERST# deassert to allow power and clocks to stabilize. - Configure Read Completion Boundary to 64 bytes. Freescale i.MX6 PCIe controller driver: - Initialize PHY before deasserting core reset to fix a regression in v6.0 on boards where the PHY provides the reference. - Fix imx6sx and imx8mq clock names in DT schema. Intel VMD host bridge driver: - Fix Secondary Bus Reset on VMD bridges, which allows reset of NVMe SSDs in VT-d pass-through scenarios. - Disable MSI remapping, which gets re-enabled by firmware during suspend/resume. MediaTek PCIe Gen3 controller driver: - Add MT7986 and MT8195 support. Qualcomm PCIe controller driver: - Add SC8280XP/SA8540P basic interconnect support. Rockchip DesignWare PCIe controller driver: - Base DT schema on common Synopsys schema. Synopsys DesignWare PCIe core: - Collect DT items shared between Root Port and Endpoint (PERST GPIO, PHY info, clocks, resets, link speed, number of lanes, number of iATU windows, interrupt info, etc) to snps,dw-pcie-common.yaml. - Add dma-ranges support for Root Ports and Endpoints. - Consolidate DT resource retrieval for "dbi", "dbi2", "atu", etc. to reduce code duplication. - Add generic names for clocks and resets to encourage more consistent naming across drivers using DesignWare IP. - Stop advertising PTM Responder role for Endpoints, which aren't allowed to be responders. TI J721E PCIe driver: - Add j721s2 host mode ID to DT schema. - Add interrupt properties to DT schema. Toshiba Visconti PCIe controller driver: - Fix interrupts array max constraints in DT schema" * tag 'pci-v6.2-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (95 commits) x86/PCI: Use pr_info() when possible x86/PCI: Fix log message typo x86/PCI: Tidy E820 removal messages PCI: Skip allocate_resource() if too little space available efi/x86: Remove EfiMemoryMappedIO from E820 map PCI/portdrv: Allow AER service only for Root Ports & RCECs PCI: xilinx-nwl: Fix coding style violations PCI: mvebu: Switch to using gpiod API PCI: pciehp: Enable Command Completed Interrupt only if supported PCI: aardvark: Switch to using devm_gpiod_get_optional() dt-bindings: PCI: mediatek-gen3: add support for mt7986 dt-bindings: PCI: mediatek-gen3: add SoC based clock config dt-bindings: PCI: qcom: Allow 'dma-coherent' property PCI: mt7621: Add sentinel to quirks table PCI: vmd: Fix secondary bus reset for Intel bridges PCI: endpoint: pci-epf-vntb: Fix sparse ntb->reg build warning PCI: endpoint: pci-epf-vntb: Fix sparse build warning for epf_db PCI: endpoint: pci-epf-vntb: Replace hardcoded 4 with sizeof(u32) PCI: endpoint: pci-epf-vntb: Remove unused epf_db_phy struct member PCI: endpoint: pci-epf-vntb: Fix call pci_epc_mem_free_addr() in error path ...
Diffstat (limited to 'arch/x86/platform')
-rw-r--r--arch/x86/platform/efi/efi.c46
1 files changed, 46 insertions, 0 deletions
diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index 7e51c14a1ef0..55d9caf66401 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -305,6 +305,50 @@ static void __init efi_clean_memmap(void)
}
}
+/*
+ * Firmware can use EfiMemoryMappedIO to request that MMIO regions be
+ * mapped by the OS so they can be accessed by EFI runtime services, but
+ * should have no other significance to the OS (UEFI r2.10, sec 7.2).
+ * However, most bootloaders and EFI stubs convert EfiMemoryMappedIO
+ * regions to E820_TYPE_RESERVED entries, which prevent Linux from
+ * allocating space from them (see remove_e820_regions()).
+ *
+ * Some platforms use EfiMemoryMappedIO entries for PCI MMCONFIG space and
+ * PCI host bridge windows, which means Linux can't allocate BAR space for
+ * hot-added devices.
+ *
+ * Remove large EfiMemoryMappedIO regions from the E820 map to avoid this
+ * problem.
+ *
+ * Retain small EfiMemoryMappedIO regions because on some platforms, these
+ * describe non-window space that's included in host bridge _CRS. If we
+ * assign that space to PCI devices, they don't work.
+ */
+static void __init efi_remove_e820_mmio(void)
+{
+ efi_memory_desc_t *md;
+ u64 size, start, end;
+ int i = 0;
+
+ for_each_efi_memory_desc(md) {
+ if (md->type == EFI_MEMORY_MAPPED_IO) {
+ size = md->num_pages << EFI_PAGE_SHIFT;
+ start = md->phys_addr;
+ end = start + size - 1;
+ if (size >= 256*1024) {
+ pr_info("Remove mem%02u: MMIO range=[0x%08llx-0x%08llx] (%lluMB) from e820 map\n",
+ i, start, end, size >> 20);
+ e820__range_remove(start, size,
+ E820_TYPE_RESERVED, 1);
+ } else {
+ pr_info("Not removing mem%02u: MMIO range=[0x%08llx-0x%08llx] (%lluKB) from e820 map\n",
+ i, start, end, size >> 10);
+ }
+ }
+ i++;
+ }
+}
+
void __init efi_print_memmap(void)
{
efi_memory_desc_t *md;
@@ -476,6 +520,8 @@ void __init efi_init(void)
set_bit(EFI_RUNTIME_SERVICES, &efi.flags);
efi_clean_memmap();
+ efi_remove_e820_mmio();
+
if (efi_enabled(EFI_DBG))
efi_print_memmap();
}