575 lines · plain
1What: /sys/bus/pci/drivers/.../bind2What: /sys/devices/pciX/.../bind3Date: December 20034Contact: linux-pci@vger.kernel.org5Description:6 Writing a device location to this file will cause7 the driver to attempt to bind to the device found at8 this location. This is useful for overriding default9 bindings. The format for the location is: DDDD:BB:DD.F.10 That is Domain:Bus:Device.Function and is the same as11 found in /sys/bus/pci/devices/. For example::12 13 # echo 0000:00:19.0 > /sys/bus/pci/drivers/foo/bind14 15 (Note: kernels before 2.6.28 may require echo -n).16 17What: /sys/bus/pci/drivers/.../unbind18What: /sys/devices/pciX/.../unbind19Date: December 200320Contact: linux-pci@vger.kernel.org21Description:22 Writing a device location to this file will cause the23 driver to attempt to unbind from the device found at24 this location. This may be useful when overriding default25 bindings. The format for the location is: DDDD:BB:DD.F.26 That is Domain:Bus:Device.Function and is the same as27 found in /sys/bus/pci/devices/. For example::28 29 # echo 0000:00:19.0 > /sys/bus/pci/drivers/foo/unbind30 31 (Note: kernels before 2.6.28 may require echo -n).32 33What: /sys/bus/pci/drivers/.../new_id34What: /sys/devices/pciX/.../new_id35Date: December 200336Contact: linux-pci@vger.kernel.org37Description:38 Writing a device ID to this file will attempt to39 dynamically add a new device ID to a PCI device driver.40 This may allow the driver to support more hardware than41 was included in the driver's static device ID support42 table at compile time. The format for the device ID is:43 VVVV DDDD SVVV SDDD CCCC MMMM PPPP. That is Vendor ID,44 Device ID, Subsystem Vendor ID, Subsystem Device ID,45 Class, Class Mask, and Private Driver Data. The Vendor ID46 and Device ID fields are required, the rest are optional.47 Upon successfully adding an ID, the driver will probe48 for the device and attempt to bind to it. For example::49 50 # echo "8086 10f5" > /sys/bus/pci/drivers/foo/new_id51 52What: /sys/bus/pci/drivers/.../remove_id53What: /sys/devices/pciX/.../remove_id54Date: February 200955Contact: Chris Wright <chrisw@sous-sol.org>56Description:57 Writing a device ID to this file will remove an ID58 that was dynamically added via the new_id sysfs entry.59 The format for the device ID is:60 VVVV DDDD SVVV SDDD CCCC MMMM. That is Vendor ID, Device61 ID, Subsystem Vendor ID, Subsystem Device ID, Class,62 and Class Mask. The Vendor ID and Device ID fields are63 required, the rest are optional. After successfully64 removing an ID, the driver will no longer support the65 device. This is useful to ensure auto probing won't66 match the driver to the device. For example::67 68 # echo "8086 10f5" > /sys/bus/pci/drivers/foo/remove_id69 70What: /sys/bus/pci/rescan71Date: January 200972Contact: Linux PCI developers <linux-pci@vger.kernel.org>73Description:74 Writing a non-zero value to this attribute will75 force a rescan of all PCI buses in the system, and76 re-discover previously removed devices.77 78What: /sys/bus/pci/devices/.../msi_bus79Date: September 201480Contact: Linux PCI developers <linux-pci@vger.kernel.org>81Description:82 Writing a zero value to this attribute disallows MSI and83 MSI-X for any future drivers of the device. If the device84 is a bridge, MSI and MSI-X will be disallowed for future85 drivers of all child devices under the bridge. Drivers86 must be reloaded for the new setting to take effect.87 88What: /sys/bus/pci/devices/.../msi_irqs/89Date: September, 201190Contact: Neil Horman <nhorman@tuxdriver.com>91Description:92 The /sys/devices/.../msi_irqs directory contains a variable set93 of files, with each file being named after a corresponding msi94 irq vector allocated to that device.95 96What: /sys/bus/pci/devices/.../msi_irqs/<N>97Date: September 201198Contact: Neil Horman <nhorman@tuxdriver.com>99Description:100 This attribute indicates the mode that the irq vector named by101 the file is in (msi vs. msix)102 103What: /sys/bus/pci/devices/.../irq104Date: August 2021105Contact: Linux PCI developers <linux-pci@vger.kernel.org>106Description:107 If a driver has enabled MSI (not MSI-X), "irq" contains the108 IRQ of the first MSI vector. Otherwise "irq" contains the109 IRQ of the legacy INTx interrupt.110 111 "irq" being set to 0 indicates that the device isn't112 capable of generating legacy INTx interrupts.113 114What: /sys/bus/pci/devices/.../remove115Date: January 2009116Contact: Linux PCI developers <linux-pci@vger.kernel.org>117Description:118 Writing a non-zero value to this attribute will119 hot-remove the PCI device and any of its children.120 121What: /sys/bus/pci/devices/.../pci_bus/.../rescan122Date: May 2011123Contact: Linux PCI developers <linux-pci@vger.kernel.org>124Description:125 Writing a non-zero value to this attribute will126 force a rescan of the bus and all child buses,127 and re-discover devices removed earlier from this128 part of the device tree.129 130What: /sys/bus/pci/devices/.../rescan131Date: January 2009132Contact: Linux PCI developers <linux-pci@vger.kernel.org>133Description:134 Writing a non-zero value to this attribute will135 force a rescan of the device's parent bus and all136 child buses, and re-discover devices removed earlier137 from this part of the device tree.138 139What: /sys/bus/pci/devices/.../reset_method140Date: August 2021141Contact: Amey Narkhede <ameynarkhede03@gmail.com>142Description:143 Some devices allow an individual function to be reset144 without affecting other functions in the same slot.145 146 For devices that have this support, a file named147 reset_method is present in sysfs. Reading this file148 gives names of the supported and enabled reset methods and149 their ordering. Writing a space-separated list of names of150 reset methods sets the reset methods and ordering to be151 used when resetting the device. Writing an empty string152 disables the ability to reset the device. Writing153 "default" enables all supported reset methods in the154 default ordering.155 156What: /sys/bus/pci/devices/.../reset157Date: July 2009158Contact: Michael S. Tsirkin <mst@redhat.com>159Description:160 Some devices allow an individual function to be reset161 without affecting other functions in the same device.162 For devices that have this support, a file named reset163 will be present in sysfs. Writing 1 to this file164 will perform reset.165 166What: /sys/bus/pci/devices/.../vpd167Date: February 2008168Contact: Ben Hutchings <bwh@kernel.org>169Description:170 A file named vpd in a device directory will be a171 binary file containing the Vital Product Data for the172 device. It should follow the VPD format defined in173 PCI Specification 2.1 or 2.2, but users should consider174 that some devices may have incorrectly formatted data. 175 If the underlying VPD has a writable section then the176 corresponding section of this file will be writable.177 178What: /sys/bus/pci/devices/.../virtfn<N>179Date: March 2009180Contact: Yu Zhao <yu.zhao@intel.com>181Description:182 This symbolic link appears when hardware supports the SR-IOV183 capability and the Physical Function driver has enabled it.184 The symbolic link points to the PCI device sysfs entry of the185 Virtual Function whose index is N (0...MaxVFs-1).186 187What: /sys/bus/pci/devices/.../dep_link188Date: March 2009189Contact: Yu Zhao <yu.zhao@intel.com>190Description:191 This symbolic link appears when hardware supports the SR-IOV192 capability and the Physical Function driver has enabled it,193 and this device has vendor specific dependencies with others.194 The symbolic link points to the PCI device sysfs entry of195 Physical Function this device depends on.196 197What: /sys/bus/pci/devices/.../physfn198Date: March 2009199Contact: Yu Zhao <yu.zhao@intel.com>200Description:201 This symbolic link appears when a device is a Virtual Function.202 The symbolic link points to the PCI device sysfs entry of the203 Physical Function this device associates with.204 205What: /sys/bus/pci/devices/.../modalias206Date: May 2005207Contact: Greg Kroah-Hartman <gregkh@linuxfoundation.org>208Description:209 This attribute indicates the PCI ID of the device object.210 211 That is in the format:212 pci:vXXXXXXXXdXXXXXXXXsvXXXXXXXXsdXXXXXXXXbcXXscXXiXX,213 where:214 215 - vXXXXXXXX contains the vendor ID;216 - dXXXXXXXX contains the device ID;217 - svXXXXXXXX contains the sub-vendor ID;218 - sdXXXXXXXX contains the subsystem device ID;219 - bcXX contains the device class;220 - scXX contains the device subclass;221 - iXX contains the device class programming interface.222 223What: /sys/bus/pci/slots/.../module224Date: June 2009225Contact: linux-pci@vger.kernel.org226Description:227 This symbolic link points to the PCI hotplug controller driver228 module that manages the hotplug slot.229 230What: /sys/bus/pci/devices/.../label231Date: July 2010232Contact: Narendra K <narendra_k@dell.com>, linux-bugs@dell.com233Description:234 Reading this attribute will provide the firmware235 given name (SMBIOS type 41 string or ACPI _DSM string) of236 the PCI device. The attribute will be created only237 if the firmware has given a name to the PCI device.238 ACPI _DSM string name will be given priority if the239 system firmware provides SMBIOS type 41 string also.240Users:241 Userspace applications interested in knowing the242 firmware assigned name of the PCI device.243 244What: /sys/bus/pci/devices/.../index245Date: July 2010246Contact: Narendra K <narendra_k@dell.com>, linux-bugs@dell.com247Description:248 Reading this attribute will provide the firmware given instance249 number of the PCI device. Depending on the platform this can250 be for example the SMBIOS type 41 device type instance or the251 user-defined ID (UID) on s390. The attribute will be created252 only if the firmware has given an instance number to the PCI253 device and that number is guaranteed to uniquely identify the254 device in the system.255Users:256 Userspace applications interested in knowing the257 firmware assigned device type instance of the PCI258 device that can help in understanding the firmware259 intended order of the PCI device.260 261What: /sys/bus/pci/devices/.../acpi_index262Date: July 2010263Contact: Narendra K <narendra_k@dell.com>, linux-bugs@dell.com264Description:265 Reading this attribute will provide the firmware266 given instance (ACPI _DSM instance number) of the PCI device.267 The attribute will be created only if the firmware has given268 an instance number to the PCI device. ACPI _DSM instance number269 will be given priority if the system firmware provides SMBIOS270 type 41 device type instance also.271Users:272 Userspace applications interested in knowing the273 firmware assigned instance number of the PCI274 device that can help in understanding the firmware275 intended order of the PCI device.276 277What: /sys/bus/pci/devices/.../d3cold_allowed278Date: July 2012279Contact: Huang Ying <ying.huang@intel.com>280Description:281 d3cold_allowed is bit to control whether the corresponding PCI282 device can be put into D3Cold state. If it is cleared, the283 device will never be put into D3Cold state. If it is set, the284 device may be put into D3Cold state if other requirements are285 satisfied too. Reading this attribute will show the current286 value of d3cold_allowed bit. Writing this attribute will set287 the value of d3cold_allowed bit.288 289What: /sys/bus/pci/devices/.../sriov_totalvfs290Date: November 2012291Contact: Donald Dutile <ddutile@redhat.com>292Description:293 This file appears when a physical PCIe device supports SR-IOV.294 Userspace applications can read this file to determine the295 maximum number of Virtual Functions (VFs) a PCIe physical296 function (PF) can support. Typically, this is the value reported297 in the PF's SR-IOV extended capability structure's TotalVFs298 element. Drivers have the ability at probe time to reduce the299 value read from this file via the pci_sriov_set_totalvfs()300 function.301 302What: /sys/bus/pci/devices/.../sriov_numvfs303Date: November 2012304Contact: Donald Dutile <ddutile@redhat.com>305Description:306 This file appears when a physical PCIe device supports SR-IOV.307 Userspace applications can read and write to this file to308 determine and control the enablement or disablement of Virtual309 Functions (VFs) on the physical function (PF). A read of this310 file will return the number of VFs that are enabled on this PF.311 A number written to this file will enable the specified312 number of VFs. A userspace application would typically read the313 file and check that the value is zero, and then write the number314 of VFs that should be enabled on the PF; the value written315 should be less than or equal to the value in the sriov_totalvfs316 file. A userspace application wanting to disable the VFs would317 write a zero to this file. The core ensures that valid values318 are written to this file, and returns errors when values are not319 valid. For example, writing a 2 to this file when sriov_numvfs320 is not 0 and not 2 already will return an error. Writing a 10321 when the value of sriov_totalvfs is 8 will return an error.322 323What: /sys/bus/pci/devices/.../driver_override324Date: April 2014325Contact: Alex Williamson <alex.williamson@redhat.com>326Description:327 This file allows the driver for a device to be specified which328 will override standard static and dynamic ID matching. When329 specified, only a driver with a name matching the value written330 to driver_override will have an opportunity to bind to the331 device. The override is specified by writing a string to the332 driver_override file (echo pci-stub > driver_override) and333 may be cleared with an empty string (echo > driver_override).334 This returns the device to standard matching rules binding.335 Writing to driver_override does not automatically unbind the336 device from its current driver or make any attempt to337 automatically load the specified driver. If no driver with a338 matching name is currently loaded in the kernel, the device339 will not bind to any driver. This also allows devices to340 opt-out of driver binding using a driver_override name such as341 "none". Only a single driver may be specified in the override,342 there is no support for parsing delimiters.343 344What: /sys/bus/pci/devices/.../numa_node345Date: Oct 2014346Contact: Prarit Bhargava <prarit@redhat.com>347Description:348 This file contains the NUMA node to which the PCI device is349 attached, or -1 if the node is unknown. The initial value350 comes from an ACPI _PXM method or a similar firmware351 source. If that is missing or incorrect, this file can be352 written to override the node. In that case, please report353 a firmware bug to the system vendor. Writing to this file354 taints the kernel with TAINT_FIRMWARE_WORKAROUND, which355 reduces the supportability of your system.356 357What: /sys/bus/pci/devices/.../revision358Date: November 2016359Contact: Emil Velikov <emil.l.velikov@gmail.com>360Description:361 This file contains the revision field of the PCI device.362 The value comes from device config space. The file is read only.363 364What: /sys/bus/pci/devices/.../sriov_drivers_autoprobe365Date: April 2017366Contact: Bodong Wang<bodong@mellanox.com>367Description:368 This file is associated with the PF of a device that369 supports SR-IOV. It determines whether newly-enabled VFs370 are immediately bound to a driver. It initially contains371 1, which means the kernel automatically binds VFs to a372 compatible driver immediately after they are enabled. If373 an application writes 0 to the file before enabling VFs,374 the kernel will not bind VFs to a driver.375 376 A typical use case is to write 0 to this file, then enable377 VFs, then assign the newly-created VFs to virtual machines.378 Note that changing this file does not affect already-379 enabled VFs. In this scenario, the user must first disable380 the VFs, write 0 to sriov_drivers_autoprobe, then re-enable381 the VFs.382 383 This is similar to /sys/bus/pci/drivers_autoprobe, but384 affects only the VFs associated with a specific PF.385 386What: /sys/bus/pci/devices/.../p2pmem/size387Date: November 2017388Contact: Logan Gunthorpe <logang@deltatee.com>389Description:390 If the device has any Peer-to-Peer memory registered, this391 file contains the total amount of memory that the device392 provides (in decimal).393 394What: /sys/bus/pci/devices/.../p2pmem/available395Date: November 2017396Contact: Logan Gunthorpe <logang@deltatee.com>397Description:398 If the device has any Peer-to-Peer memory registered, this399 file contains the amount of memory that has not been400 allocated (in decimal).401 402What: /sys/bus/pci/devices/.../p2pmem/published403Date: November 2017404Contact: Logan Gunthorpe <logang@deltatee.com>405Description:406 If the device has any Peer-to-Peer memory registered, this407 file contains a '1' if the memory has been published for408 use outside the driver that owns the device.409 410What: /sys/bus/pci/devices/.../p2pmem/allocate411Date: August 2022412Contact: Logan Gunthorpe <logang@deltatee.com>413Description:414 This file allows mapping p2pmem into userspace. For each415 mmap() call on this file, the kernel will allocate a chunk416 of Peer-to-Peer memory for use in Peer-to-Peer transactions.417 This memory can be used in O_DIRECT calls to NVMe backed418 files for Peer-to-Peer copies.419 420What: /sys/bus/pci/devices/.../link/clkpm421 /sys/bus/pci/devices/.../link/l0s_aspm422 /sys/bus/pci/devices/.../link/l1_aspm423 /sys/bus/pci/devices/.../link/l1_1_aspm424 /sys/bus/pci/devices/.../link/l1_2_aspm425 /sys/bus/pci/devices/.../link/l1_1_pcipm426 /sys/bus/pci/devices/.../link/l1_2_pcipm427Date: October 2019428Contact: Heiner Kallweit <hkallweit1@gmail.com>429Description: If ASPM is supported for an endpoint, these files can be430 used to disable or enable the individual power management431 states. Write y/1/on to enable, n/0/off to disable.432 433What: /sys/bus/pci/devices/.../power_state434Date: November 2020435Contact: Linux PCI developers <linux-pci@vger.kernel.org>436Description:437 This file contains the current PCI power state of the device.438 The value comes from the PCI kernel device state and can be one439 of: "unknown", "error", "D0", D1", "D2", "D3hot", "D3cold".440 The file is read only.441 442What: /sys/bus/pci/devices/.../sriov_vf_total_msix443Date: January 2021444Contact: Leon Romanovsky <leonro@nvidia.com>445Description:446 This file is associated with a SR-IOV physical function (PF).447 It contains the total number of MSI-X vectors available for448 assignment to all virtual functions (VFs) associated with PF.449 The value will be zero if the device doesn't support this450 functionality. For supported devices, the value will be451 constant and won't be changed after MSI-X vectors assignment.452 453What: /sys/bus/pci/devices/.../sriov_vf_msix_count454Date: January 2021455Contact: Leon Romanovsky <leonro@nvidia.com>456Description:457 This file is associated with a SR-IOV virtual function (VF).458 It allows configuration of the number of MSI-X vectors for459 the VF. This allows devices that have a global pool of MSI-X460 vectors to optimally divide them between VFs based on VF usage.461 462 The values accepted are:463 * > 0 - this number will be reported as the Table Size in the464 VF's MSI-X capability465 * < 0 - not valid466 * = 0 - will reset to the device default value467 468 The file is writable if the PF is bound to a driver that469 implements ->sriov_set_msix_vec_count().470 471What: /sys/bus/pci/devices/.../resourceN_resize472Date: September 2022473Contact: Alex Williamson <alex.williamson@redhat.com>474Description:475 These files provide an interface to PCIe Resizable BAR support.476 A file is created for each BAR resource (N) supported by the477 PCIe Resizable BAR extended capability of the device. Reading478 each file exposes the bitmap of available resource sizes:479 480 # cat resource1_resize481 00000000000001c0482 483 The bitmap represents supported resource sizes for the BAR,484 where bit0 = 1MB, bit1 = 2MB, bit2 = 4MB, etc. In the above485 example the device supports 64MB, 128MB, and 256MB BAR sizes.486 487 When writing the file, the user provides the bit position of488 the desired resource size, for example:489 490 # echo 7 > resource1_resize491 492 This indicates to set the size value corresponding to bit 7,493 128MB. The resulting size is 2 ^ (bit# + 20). This definition494 matches the PCIe specification of this capability.495 496 In order to make use of resource resizing, all PCI drivers must497 be unbound from the device and peer devices under the same498 parent bridge may need to be soft removed. In the case of499 VGA devices, writing a resize value will remove low level500 console drivers from the device. Raw users of pci-sysfs501 resourceN attributes must be terminated prior to resizing.502 Success of the resizing operation is not guaranteed.503 504What: /sys/bus/pci/devices/.../leds/*:enclosure:*/brightness505What: /sys/class/leds/*:enclosure:*/brightness506Date: August 2024507KernelVersion: 6.12508Description:509 LED indications on PCIe storage enclosures which are controlled510 through the NPEM interface (Native PCIe Enclosure Management,511 PCIe r6.1 sec 6.28) are accessible as led class devices, both512 below /sys/class/leds and below NPEM-capable PCI devices.513 514 Although these led class devices could be manipulated manually,515 in practice they are typically manipulated automatically by an516 application such as ledmon(8).517 518 The name of a led class device is as follows:519 <bdf>:enclosure:<indication>520 where:521 522 - <bdf> is the domain, bus, device and function number523 (e.g. 10000:02:05.0)524 - <indication> is a short description of the LED indication525 526 Valid indications per PCIe r6.1 table 6-27 are:527 528 - ok (drive is functioning normally)529 - locate (drive is being identified by an admin)530 - fail (drive is not functioning properly)531 - rebuild (drive is part of an array that is rebuilding)532 - pfa (drive is predicted to fail soon)533 - hotspare (drive is marked to be used as a replacement)534 - ica (drive is part of an array that is degraded)535 - ifa (drive is part of an array that is failed)536 - idt (drive is not the right type for the connector)537 - disabled (drive is disabled, removal is safe)538 - specific0 to specific7 (enclosure-specific indications)539 540 Broadly, the indications fall into one of these categories:541 542 - to signify drive state (ok, locate, fail, idt, disabled)543 - to signify drive role or state in a software RAID array544 (rebuild, pfa, hotspare, ica, ifa)545 - to signify any other role or state (specific0 to specific7)546 547 Mandatory indications per PCIe r6.1 sec 7.9.19.2 comprise:548 ok, locate, fail, rebuild. All others are optional.549 A led class device is only visible if the corresponding550 indication is supported by the device.551 552 To manipulate the indications, write 0 (LED_OFF) or 1 (LED_ON)553 to the "brightness" file. Note that manipulating an indication554 may implicitly manipulate other indications at the vendor's555 discretion. E.g. when the user lights up the "ok" indication,556 the vendor may choose to automatically turn off the "fail"557 indication. The current state of an indication can be558 retrieved by reading its "brightness" file.559 560 The PCIe Base Specification allows vendors leeway to choose561 different colors or blinking patterns for the indications,562 but they typically follow the IBPI standard. E.g. the "locate"563 indication is usually presented as one or two LEDs blinking at564 4 Hz frequency:565 https://en.wikipedia.org/wiki/International_Blinking_Pattern_Interpretation566 567 PCI Firmware Specification r3.3 sec 4.7 defines a DSM interface568 to facilitate shared access by operating system and platform569 firmware to a device's NPEM registers. The kernel will use570 this DSM interface where available, instead of accessing NPEM571 registers directly. The DSM interface does not support the572 enclosure-specific indications "specific0" to "specific7",573 hence the corresponding led class devices are unavailable if574 the DSM interface is used.575