^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) ====================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) Coherent Accelerator Interface (CXL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) ====================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) Introduction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) ============
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) The coherent accelerator interface is designed to allow the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) coherent connection of accelerators (FPGAs and other devices) to a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) POWER system. These devices need to adhere to the Coherent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) Accelerator Interface Architecture (CAIA).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) IBM refers to this as the Coherent Accelerator Processor Interface
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) or CAPI. In the kernel it's referred to by the name CXL to avoid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) confusion with the ISDN CAPI subsystem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) Coherent in this context means that the accelerator and CPUs can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) both access system memory directly and with the same effective
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) addresses.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) Hardware overview
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) =================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) POWER8/9 FPGA
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) +----------+ +---------+
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) | | | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) | CPU | | AFU |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) | | | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) | | | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) | | | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) +----------+ +---------+
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) | PHB | | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) | +------+ | PSL |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) | | CAPP |<------>| |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) +---+------+ PCIE +---------+
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) The POWER8/9 chip has a Coherently Attached Processor Proxy (CAPP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) unit which is part of the PCIe Host Bridge (PHB). This is managed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) by Linux by calls into OPAL. Linux doesn't directly program the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) CAPP.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) The FPGA (or coherently attached device) consists of two parts.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) The POWER Service Layer (PSL) and the Accelerator Function Unit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) (AFU). The AFU is used to implement specific functionality behind
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) the PSL. The PSL, among other things, provides memory address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) translation services to allow each AFU direct access to userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) memory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) The AFU is the core part of the accelerator (eg. the compression,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) crypto etc function). The kernel has no knowledge of the function
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) of the AFU. Only userspace interacts directly with the AFU.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) The PSL provides the translation and interrupt services that the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) AFU needs. This is what the kernel interacts with. For example, if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) the AFU needs to read a particular effective address, it sends
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) that address to the PSL, the PSL then translates it, fetches the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) data from memory and returns it to the AFU. If the PSL has a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) translation miss, it interrupts the kernel and the kernel services
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) the fault. The context to which this fault is serviced is based on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) who owns that acceleration function.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) - POWER8 and PSL Version 8 are compliant to the CAIA Version 1.0.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) - POWER9 and PSL Version 9 are compliant to the CAIA Version 2.0.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) This PSL Version 9 provides new features such as:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * Interaction with the nest MMU on the P9 chip.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) * Native DMA support.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * Supports sending ASB_Notify messages for host thread wakeup.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * Supports Atomic operations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) * etc.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) Cards with a PSL9 won't work on a POWER8 system and cards with a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) PSL8 won't work on a POWER9 system.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) AFU Modes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) =========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) There are two programming modes supported by the AFU. Dedicated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) and AFU directed. AFU may support one or both modes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) When using dedicated mode only one MMU context is supported. In
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) this mode, only one userspace process can use the accelerator at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) When using AFU directed mode, up to 16K simultaneous contexts can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) be supported. This means up to 16K simultaneous userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) applications may use the accelerator (although specific AFUs may
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) support fewer). In this mode, the AFU sends a 16 bit context ID
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) with each of its requests. This tells the PSL which context is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) associated with each operation. If the PSL can't translate an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) operation, the ID can also be accessed by the kernel so it can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) determine the userspace context associated with an operation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) MMIO space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) ==========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) A portion of the accelerator MMIO space can be directly mapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) from the AFU to userspace. Either the whole space can be mapped or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) just a per context portion. The hardware is self describing, hence
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) the kernel can determine the offset and size of the per context
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) portion.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) Interrupts
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) ==========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) AFUs may generate interrupts that are destined for userspace. These
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) are received by the kernel as hardware interrupts and passed onto
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) userspace by a read syscall documented below.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) Data storage faults and error interrupts are handled by the kernel
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) driver.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) Work Element Descriptor (WED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) =============================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) The WED is a 64-bit parameter passed to the AFU when a context is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) started. Its format is up to the AFU hence the kernel has no
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) knowledge of what it represents. Typically it will be the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) effective address of a work queue or status block where the AFU
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) and userspace can share control and status information.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) User API
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) ========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) 1. AFU character devices
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) ^^^^^^^^^^^^^^^^^^^^^^^^
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) For AFUs operating in AFU directed mode, two character device
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) files will be created. /dev/cxl/afu0.0m will correspond to a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) master context and /dev/cxl/afu0.0s will correspond to a slave
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) context. Master contexts have access to the full MMIO space an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) AFU provides. Slave contexts have access to only the per process
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) MMIO space an AFU provides.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) For AFUs operating in dedicated process mode, the driver will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) only create a single character device per AFU called
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) /dev/cxl/afu0.0d. This will have access to the entire MMIO space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) that the AFU provides (like master contexts in AFU directed).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) The types described below are defined in include/uapi/misc/cxl.h
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) The following file operations are supported on both slave and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) master devices.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) A userspace library libcxl is available here:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) https://github.com/ibm-capi/libcxl
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) This provides a C interface to this kernel API.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) open
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) ----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) Opens the device and allocates a file descriptor to be used with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) the rest of the API.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) A dedicated mode AFU only has one context and only allows the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) device to be opened once.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) An AFU directed mode AFU can have many contexts, the device can be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) opened once for each context that is available.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) When all available contexts are allocated the open call will fail
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) and return -ENOSPC.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) Note:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) IRQs need to be allocated for each context, which may limit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) the number of contexts that can be created, and therefore
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) how many times the device can be opened. The POWER8 CAPP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) supports 2040 IRQs and 3 are used by the kernel, so 2037 are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) left. If 1 IRQ is needed per context, then only 2037
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) contexts can be allocated. If 4 IRQs are needed per context,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) then only 2037/4 = 509 contexts can be allocated.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) ioctl
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) -----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) CXL_IOCTL_START_WORK:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) Starts the AFU context and associates it with the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) process. Once this ioctl is successfully executed, all memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) mapped into this process is accessible to this AFU context
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) using the same effective addresses. No additional calls are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) required to map/unmap memory. The AFU memory context will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) updated as userspace allocates and frees memory. This ioctl
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) returns once the AFU context is started.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) Takes a pointer to a struct cxl_ioctl_start_work
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) struct cxl_ioctl_start_work {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) __u64 flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) __u64 work_element_descriptor;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) __u64 amr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) __s16 num_interrupts;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) __s16 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) __s32 reserved2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) __u64 reserved3;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) __u64 reserved4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) __u64 reserved5;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) __u64 reserved6;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) Indicates which optional fields in the structure are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) valid.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) work_element_descriptor:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) The Work Element Descriptor (WED) is a 64-bit argument
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) defined by the AFU. Typically this is an effective
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) address pointing to an AFU specific structure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) describing what work to perform.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) amr:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) Authority Mask Register (AMR), same as the powerpc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) AMR. This field is only used by the kernel when the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) corresponding CXL_START_WORK_AMR value is specified in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) flags. If not specified the kernel will use a default
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) value of 0.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) num_interrupts:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) Number of userspace interrupts to request. This field
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) is only used by the kernel when the corresponding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) CXL_START_WORK_NUM_IRQS value is specified in flags.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) If not specified the minimum number required by the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) AFU will be allocated. The min and max number can be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) obtained from sysfs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) reserved fields:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) For ABI padding and future extensions
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) CXL_IOCTL_GET_PROCESS_ELEMENT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) Get the current context id, also known as the process element.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) The value is returned from the kernel as a __u32.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) mmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) ----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) An AFU may have an MMIO space to facilitate communication with the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) AFU. If it does, the MMIO space can be accessed via mmap. The size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) and contents of this area are specific to the particular AFU. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) size can be discovered via sysfs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) In AFU directed mode, master contexts are allowed to map all of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) the MMIO space and slave contexts are allowed to only map the per
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) process MMIO space associated with the context. In dedicated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) process mode the entire MMIO space can always be mapped.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) This mmap call must be done after the START_WORK ioctl.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) Care should be taken when accessing MMIO space. Only 32 and 64-bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) accesses are supported by POWER8. Also, the AFU will be designed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) with a specific endianness, so all MMIO accesses should consider
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) endianness (recommend endian(3) variants like: le64toh(),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) be64toh() etc). These endian issues equally apply to shared memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) queues the WED may describe.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) read
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) ----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) Reads events from the AFU. Blocks if no events are pending
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) (unless O_NONBLOCK is supplied). Returns -EIO in the case of an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) unrecoverable error or if the card is removed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) read() will always return an integral number of events.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) The buffer passed to read() must be at least 4K bytes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) The result of the read will be a buffer of one or more events,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) each event is of type struct cxl_event, of varying size::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) struct cxl_event {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) struct cxl_event_header header;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) union {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) struct cxl_event_afu_interrupt irq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) struct cxl_event_data_storage fault;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) struct cxl_event_afu_error afu_error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) The struct cxl_event_header is defined as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) struct cxl_event_header {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) __u16 type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) __u16 size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) __u16 process_element;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) __u16 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) type:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) This defines the type of event. The type determines how
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) the rest of the event is structured. These types are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) described below and defined by enum cxl_event_type.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) size:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) This is the size of the event in bytes including the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) struct cxl_event_header. The start of the next event can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) be found at this offset from the start of the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) event.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) process_element:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) Context ID of the event.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) reserved field:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) For future extensions and padding.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) If the event type is CXL_EVENT_AFU_INTERRUPT then the event
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) structure is defined as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) struct cxl_event_afu_interrupt {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) __u16 flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) __u16 irq; /* Raised AFU interrupt number */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) __u32 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) These flags indicate which optional fields are present
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) in this struct. Currently all fields are mandatory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) irq:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) The IRQ number sent by the AFU.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) reserved field:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) For future extensions and padding.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) If the event type is CXL_EVENT_DATA_STORAGE then the event
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) structure is defined as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) struct cxl_event_data_storage {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) __u16 flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) __u16 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) __u32 reserved2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) __u64 addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) __u64 dsisr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) __u64 reserved3;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) These flags indicate which optional fields are present in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) this struct. Currently all fields are mandatory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) address:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) The address that the AFU unsuccessfully attempted to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) access. Valid accesses will be handled transparently by the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) kernel but invalid accesses will generate this event.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) dsisr:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) This field gives information on the type of fault. It is a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) copy of the DSISR from the PSL hardware when the address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) fault occurred. The form of the DSISR is as defined in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) CAIA.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) reserved fields:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) For future extensions
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) If the event type is CXL_EVENT_AFU_ERROR then the event structure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) is defined as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) ::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) struct cxl_event_afu_error {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) __u16 flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) __u16 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) __u32 reserved2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) __u64 error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) These flags indicate which optional fields are present in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) this struct. Currently all fields are Mandatory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) error:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) Error status from the AFU. Defined by the AFU.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) reserved fields:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) For future extensions and padding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) 2. Card character device (powerVM guest only)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) In a powerVM guest, an extra character device is created for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) card. The device is only used to write (flash) a new image on the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) FPGA accelerator. Once the image is written and verified, the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) device tree is updated and the card is reset to reload the updated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) image.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) open
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) ----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) Opens the device and allocates a file descriptor to be used with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) the rest of the API. The device can only be opened once.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) ioctl
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) -----
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) CXL_IOCTL_DOWNLOAD_IMAGE / CXL_IOCTL_VALIDATE_IMAGE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) Starts and controls flashing a new FPGA image. Partial
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) reconfiguration is not supported (yet), so the image must contain
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) a copy of the PSL and AFU(s). Since an image can be quite large,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) the caller may have to iterate, splitting the image in smaller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) chunks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) Takes a pointer to a struct cxl_adapter_image::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) struct cxl_adapter_image {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) __u64 flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) __u64 data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) __u64 len_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) __u64 len_image;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) __u64 reserved1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) __u64 reserved2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) __u64 reserved3;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) __u64 reserved4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) These flags indicate which optional fields are present in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) this struct. Currently all fields are mandatory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) data:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) Pointer to a buffer with part of the image to write to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) card.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) len_data:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) Size of the buffer pointed to by data.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) len_image:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) Full size of the image.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) Sysfs Class
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) ===========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) A cxl sysfs class is added under /sys/class/cxl to facilitate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) enumeration and tuning of the accelerators. Its layout is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) described in Documentation/ABI/testing/sysfs-class-cxl
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) Udev rules
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) ==========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) The following udev rules could be used to create a symlink to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) most logical chardev to use in any programming mode (afuX.Yd for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) dedicated, afuX.Ys for afu directed), since the API is virtually
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) identical for each::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) SUBSYSTEM=="cxl", ATTRS{mode}=="dedicated_process", SYMLINK="cxl/%b"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) SUBSYSTEM=="cxl", ATTRS{mode}=="afu_directed", \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) KERNEL=="afu[0-9]*.[0-9]*s", SYMLINK="cxl/%b"
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags