1# SPDX-License-Identifier: GPL-2.0 2# 3# PCI configuration 4# 5 6# select this to offer the PCI prompt 7config HAVE_PCI 8 bool 9 10# select this to unconditionally force on PCI support 11config FORCE_PCI 12 bool 13 select HAVE_PCI 14 select PCI 15 16menuconfig PCI 17 bool "PCI support" 18 depends on HAVE_PCI 19 help 20 This option enables support for the PCI local bus, including 21 support for PCI-X and the foundations for PCI Express support. 22 Say 'Y' here unless you know what you are doing. 23 24if PCI 25 26config PCI_DOMAINS 27 bool 28 depends on PCI 29 30config PCI_DOMAINS_GENERIC 31 bool 32 select PCI_DOMAINS 33 34config PCI_SYSCALL 35 bool 36 37source "drivers/pci/pcie/Kconfig" 38 39config PCI_MSI 40 bool "Message Signaled Interrupts (MSI and MSI-X)" 41 select GENERIC_MSI_IRQ 42 help 43 This allows device drivers to enable MSI (Message Signaled 44 Interrupts). Message Signaled Interrupts enable a device to 45 generate an interrupt using an inbound Memory Write on its 46 PCI bus instead of asserting a device IRQ pin. 47 48 Use of PCI MSI interrupts can be disabled at kernel boot time 49 by using the 'pci=nomsi' option. This disables MSI for the 50 entire system. 51 52 If you don't know what to do here, say Y. 53 54config PCI_MSI_ARCH_FALLBACKS 55 bool 56 57config PCI_QUIRKS 58 default y 59 bool "Enable PCI quirk workarounds" if EXPERT 60 help 61 This enables workarounds for various PCI chipset bugs/quirks. 62 Disable this only if your target machine is unaffected by PCI 63 quirks. 64 65config PCI_DEBUG 66 bool "PCI Debugging" 67 depends on DEBUG_KERNEL 68 help 69 Say Y here if you want the PCI core to produce a bunch of debug 70 messages to the system log. Select this if you are having a 71 problem with PCI support and want to see more of what is going on. 72 73 When in doubt, say N. 74 75config PCI_REALLOC_ENABLE_AUTO 76 bool "Enable PCI resource re-allocation detection" 77 depends on PCI_IOV 78 help 79 Say Y here if you want the PCI core to detect if PCI resource 80 re-allocation needs to be enabled. You can always use pci=realloc=on 81 or pci=realloc=off to override it. It will automatically 82 re-allocate PCI resources if SR-IOV BARs have not been allocated by 83 the BIOS. 84 85 When in doubt, say N. 86 87config PCI_STUB 88 tristate "PCI Stub driver" 89 help 90 Say Y or M here if you want be able to reserve a PCI device 91 when it is going to be assigned to a guest operating system. 92 93 When in doubt, say N. 94 95config PCI_PF_STUB 96 tristate "PCI PF Stub driver" 97 depends on PCI_IOV 98 help 99 Say Y or M here if you want to enable support for devices that 100 require SR-IOV support, while at the same time the PF (Physical 101 Function) itself is not providing any actual services on the 102 host itself such as storage or networking. 103 104 When in doubt, say N. 105 106config XEN_PCIDEV_FRONTEND 107 tristate "Xen PCI Frontend" 108 depends on XEN_PV 109 select PCI_XEN 110 select XEN_XENBUS_FRONTEND 111 default y 112 help 113 The PCI device frontend driver allows the kernel to import arbitrary 114 PCI devices from a PCI backend to support PCI driver domains. 115 116config PCI_ATS 117 bool 118 119config PCI_DOE 120 bool 121 122config PCI_ECAM 123 bool 124 125config PCI_LOCKLESS_CONFIG 126 bool 127 128config PCI_BRIDGE_EMUL 129 bool 130 131config PCI_IOV 132 bool "PCI IOV support" 133 select PCI_ATS 134 help 135 I/O Virtualization is a PCI feature supported by some devices 136 which allows them to create virtual devices which share their 137 physical resources. 138 139 If unsure, say N. 140 141config PCI_PRI 142 bool "PCI PRI support" 143 select PCI_ATS 144 help 145 PRI is the PCI Page Request Interface. It allows PCI devices that are 146 behind an IOMMU to recover from page faults. 147 148 If unsure, say N. 149 150config PCI_PASID 151 bool "PCI PASID support" 152 select PCI_ATS 153 help 154 Process Address Space Identifiers (PASIDs) can be used by PCI devices 155 to access more than one IO address space at the same time. To make 156 use of this feature an IOMMU is required which also supports PASIDs. 157 Select this option if you have such an IOMMU and want to compile the 158 driver for it into your kernel. 159 160 If unsure, say N. 161 162config PCI_P2PDMA 163 bool "PCI peer-to-peer transfer support" 164 depends on ZONE_DEVICE 165 # 166 # The need for the scatterlist DMA bus address flag means PCI P2PDMA 167 # requires 64bit 168 # 169 depends on 64BIT 170 select GENERIC_ALLOCATOR 171 select NEED_SG_DMA_FLAGS 172 help 173 Enableѕ drivers to do PCI peer-to-peer transactions to and from 174 BARs that are exposed in other devices that are the part of 175 the hierarchy where peer-to-peer DMA is guaranteed by the PCI 176 specification to work (ie. anything below a single PCI bridge). 177 178 Many PCIe root complexes do not support P2P transactions and 179 it's hard to tell which support it at all, so at this time, 180 P2P DMA transactions must be between devices behind the same root 181 port. 182 183 Enabling this option will reduce the entropy of x86 KASLR memory 184 regions. For example - on a 46 bit system, the entropy goes down 185 from 16 bits to 15 bits. The actual reduction in entropy depends 186 on the physical address bits, on processor features, kernel config 187 (5 level page table) and physical memory present on the system. 188 189 If unsure, say N. 190 191config PCI_LABEL 192 def_bool y if (DMI || ACPI) 193 select NLS 194 195config PCI_HYPERV 196 tristate "Hyper-V PCI Frontend" 197 depends on ((X86 && X86_64) || ARM64) && HYPERV && PCI_MSI && SYSFS 198 select PCI_HYPERV_INTERFACE 199 help 200 The PCI device frontend driver allows the kernel to import arbitrary 201 PCI devices from a PCI backend to support PCI driver domains. 202 203config PCI_DYNAMIC_OF_NODES 204 bool "Create Device tree nodes for PCI devices" 205 depends on OF_IRQ 206 select OF_DYNAMIC 207 help 208 This option enables support for generating device tree nodes for some 209 PCI devices. Thus, the driver of this kind can load and overlay 210 flattened device tree for its downstream devices. 211 212 Once this option is selected, the device tree nodes will be generated 213 for all PCI bridges. 214 215choice 216 prompt "PCI Express hierarchy optimization setting" 217 default PCIE_BUS_DEFAULT 218 depends on PCI && EXPERT 219 help 220 MPS (Max Payload Size) and MRRS (Max Read Request Size) are PCIe 221 device parameters that affect performance and the ability to 222 support hotplug and peer-to-peer DMA. 223 224 The following choices set the MPS and MRRS optimization strategy 225 at compile-time. The choices are the same as those offered for 226 the kernel command-line parameter 'pci', i.e., 227 'pci=pcie_bus_tune_off', 'pci=pcie_bus_safe', 228 'pci=pcie_bus_perf', and 'pci=pcie_bus_peer2peer'. 229 230 This is a compile-time setting and can be overridden by the above 231 command-line parameters. If unsure, choose PCIE_BUS_DEFAULT. 232 233config PCIE_BUS_TUNE_OFF 234 bool "Tune Off" 235 depends on PCI 236 help 237 Use the BIOS defaults; don't touch MPS at all. This is the same 238 as booting with 'pci=pcie_bus_tune_off'. 239 240config PCIE_BUS_DEFAULT 241 bool "Default" 242 depends on PCI 243 help 244 Default choice; ensure that the MPS matches upstream bridge. 245 246config PCIE_BUS_SAFE 247 bool "Safe" 248 depends on PCI 249 help 250 Use largest MPS that boot-time devices support. If you have a 251 closed system with no possibility of adding new devices, this 252 will use the largest MPS that's supported by all devices. This 253 is the same as booting with 'pci=pcie_bus_safe'. 254 255config PCIE_BUS_PERFORMANCE 256 bool "Performance" 257 depends on PCI 258 help 259 Use MPS and MRRS for best performance. Ensure that a given 260 device's MPS is no larger than its parent MPS, which allows us to 261 keep all switches/bridges to the max MPS supported by their 262 parent. This is the same as booting with 'pci=pcie_bus_perf'. 263 264config PCIE_BUS_PEER2PEER 265 bool "Peer2peer" 266 depends on PCI 267 help 268 Set MPS = 128 for all devices. MPS configuration effected by the 269 other options could cause the MPS on one root port to be 270 different than that of the MPS on another, which may cause 271 hot-added devices or peer-to-peer DMA to fail. Set MPS to the 272 smallest possible value (128B) system-wide to avoid these issues. 273 This is the same as booting with 'pci=pcie_bus_peer2peer'. 274 275endchoice 276 277config VGA_ARB 278 bool "VGA Arbitration" if EXPERT 279 default y 280 depends on (PCI && !S390) 281 help 282 Some "legacy" VGA devices implemented on PCI typically have the same 283 hard-decoded addresses as they did on ISA. When multiple PCI devices 284 are accessed at same time they need some kind of coordination. Please 285 see Documentation/gpu/vgaarbiter.rst for more details. Select this to 286 enable VGA arbiter. 287 288config VGA_ARB_MAX_GPUS 289 int "Maximum number of GPUs" 290 default 16 291 depends on VGA_ARB 292 help 293 Reserves space in the kernel to maintain resource locking for 294 multiple GPUS. The overhead for each GPU is very small. 295 296source "drivers/pci/hotplug/Kconfig" 297source "drivers/pci/controller/Kconfig" 298source "drivers/pci/endpoint/Kconfig" 299source "drivers/pci/switch/Kconfig" 300 301endif 302