From 5b7c4cabbb65f5c469464da6c5f614cbd7f730f2 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 21 Feb 2023 18:24:12 -0800 Subject: Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next Pull networking updates from Jakub Kicinski: "Core: - Add dedicated kmem_cache for typical/small skb->head, avoid having to access struct page at kfree time, and improve memory use. - Introduce sysctl to set default RPS configuration for new netdevs. - Define Netlink protocol specification format which can be used to describe messages used by each family and auto-generate parsers. Add tools for generating kernel data structures and uAPI headers. - Expose all net/core sysctls inside netns. - Remove 4s sleep in netpoll if carrier is instantly detected on boot. - Add configurable limit of MDB entries per port, and port-vlan. - Continue populating drop reasons throughout the stack. - Retire a handful of legacy Qdiscs and classifiers. Protocols: - Support IPv4 big TCP (TSO frames larger than 64kB). - Add IP_LOCAL_PORT_RANGE socket option, to control local port range on socket by socket basis. - Track and report in procfs number of MPTCP sockets used. - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path manager. - IPv6: don't check net.ipv6.route.max_size and rely on garbage collection to free memory (similarly to IPv4). - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986). - ICMP: add per-rate limit counters. - Add support for user scanning requests in ieee802154. - Remove static WEP support. - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate reporting. - WiFi 7 EHT channel puncturing support (client & AP). BPF: - Add a rbtree data structure following the "next-gen data structure" precedent set by recently added linked list, that is, by using kfunc + kptr instead of adding a new BPF map type. - Expose XDP hints via kfuncs with initial support for RX hash and timestamp metadata. - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to better support decap on GRE tunnel devices not operating in collect metadata. - Improve x86 JIT's codegen for PROBE_MEM runtime error checks. - Remove the need for trace_printk_lock for bpf_trace_printk and bpf_trace_vprintk helpers. - Extend libbpf's bpf_tracing.h support for tracing arguments of kprobes/uprobes and syscall as a special case. - Significantly reduce the search time for module symbols by livepatch and BPF. - Enable cpumasks to be used as kptrs, which is useful for tracing programs tracking which tasks end up running on which CPUs in different time intervals. - Add support for BPF trampoline on s390x and riscv64. - Add capability to export the XDP features supported by the NIC. - Add __bpf_kfunc tag for marking kernel functions as kfuncs. - Add cgroup.memory=nobpf kernel parameter option to disable BPF memory accounting for container environments. Netfilter: - Remove the CLUSTERIP target. It has been marked as obsolete for years, and we still have WARN splats wrt races of the out-of-band /proc interface installed by this target. - Add 'destroy' commands to nf_tables. They are identical to the existing 'delete' commands, but do not return an error if the referenced object (set, chain, rule...) did not exist. Driver API: - Improve cpumask_local_spread() locality to help NICs set the right IRQ affinity on AMD platforms. - Separate C22 and C45 MDIO bus transactions more clearly. - Introduce new DCB table to control DSCP rewrite on egress. - Support configuration of Physical Layer Collision Avoidance (PLCA) Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of shared medium Ethernet. - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing preemption of low priority frames by high priority frames. - Add support for controlling MACSec offload using netlink SET. - Rework devlink instance refcounts to allow registration and de-registration under the instance lock. Split the code into multiple files, drop some of the unnecessarily granular locks and factor out common parts of netlink operation handling. - Add TX frame aggregation parameters (for USB drivers). - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning messages with notifications for debug. - Allow offloading of UDP NEW connections via act_ct. - Add support for per action HW stats in TC. - Support hardware miss to TC action (continue processing in SW from a specific point in the action chain). - Warn if old Wireless Extension user space interface is used with modern cfg80211/mac80211 drivers. Do not support Wireless Extensions for Wi-Fi 7 devices at all. Everyone should switch to using nl80211 interface instead. - Improve the CAN bit timing configuration. Use extack to return error messages directly to user space, update the SJW handling, including the definition of a new default value that will benefit CAN-FD controllers, by increasing their oscillator tolerance. New hardware / drivers: - Ethernet: - nVidia BlueField-3 support (control traffic driver) - Ethernet support for imx93 SoCs - Motorcomm yt8531 gigabit Ethernet PHY - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA) - Microchip LAN8841 PHY (incl. cable diagnostics and PTP) - Amlogic gxl MDIO mux - WiFi: - RealTek RTL8188EU (rtl8xxxu) - Qualcomm Wi-Fi 7 devices (ath12k) - CAN: - Renesas R-Car V4H Drivers: - Bluetooth: - Set Per Platform Antenna Gain (PPAG) for Intel controllers. - Ethernet NICs: - Intel (1G, igc): - support TSN / Qbv / packet scheduling features of i226 model - Intel (100G, ice): - use GNSS subsystem instead of TTY - multi-buffer XDP support - extend support for GPIO pins to E823 devices - nVidia/Mellanox: - update the shared buffer configuration on PFC commands - implement PTP adjphase function for HW offset control - TC support for Geneve and GRE with VF tunnel offload - more efficient crypto key management method - multi-port eswitch support - Netronome/Corigine: - add DCB IEEE support - support IPsec offloading for NFP3800 - Freescale/NXP (enetc): - support XDP_REDIRECT for XDP non-linear buffers - improve reconfig, avoid link flap and waiting for idle - support MAC Merge layer - Other NICs: - sfc/ef100: add basic devlink support for ef100 - ionic: rx_push mode operation (writing descriptors via MMIO) - bnxt: use the auxiliary bus abstraction for RDMA - r8169: disable ASPM and reset bus in case of tx timeout - cpsw: support QSGMII mode for J721e CPSW9G - cpts: support pulse-per-second output - ngbe: add an mdio bus driver - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing - r8152: handle devices with FW with NCM support - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation - virtio-net: support multi buffer XDP - virtio/vsock: replace virtio_vsock_pkt with sk_buff - tsnep: XDP support - Ethernet high-speed switches: - nVidia/Mellanox (mlxsw): - add support for latency TLV (in FW control messages) - Microchip (sparx5): - separate explicit and implicit traffic forwarding rules, make the implicit rules always active - add support for egress DSCP rewrite - IS0 VCAP support (Ingress Classification) - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS etc.) - ES2 VCAP support (Egress Access Control) - support for Per-Stream Filtering and Policing (802.1Q, 8.6.5.1) - Ethernet embedded switches: - Marvell (mv88e6xxx): - add MAB (port auth) offload support - enable PTP receive for mv88e6390 - NXP (ocelot): - support MAC Merge layer - support for the the vsc7512 internal copper phys - Microchip: - lan9303: convert to PHYLINK - lan966x: support TC flower filter statistics - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x - lan937x: support Credit Based Shaper configuration - ksz9477: support Energy Efficient Ethernet - other: - qca8k: convert to regmap read/write API, use bulk operations - rswitch: Improve TX timestamp accuracy - Intel WiFi (iwlwifi): - EHT (Wi-Fi 7) rate reporting - STEP equalizer support: transfer some STEP (connection to radio on platforms with integrated wifi) related parameters from the BIOS to the firmware. - Qualcomm 802.11ax WiFi (ath11k): - IPQ5018 support - Fine Timing Measurement (FTM) responder role support - channel 177 support - MediaTek WiFi (mt76): - per-PHY LED support - mt7996: EHT (Wi-Fi 7) support - Wireless Ethernet Dispatch (WED) reset support - switch to using page pool allocator - RealTek WiFi (rtw89): - support new version of Bluetooth co-existance - Mobile: - rmnet: support TX aggregation" * tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits) page_pool: add a comment explaining the fragment counter usage net: ethtool: fix __ethtool_dev_mm_supported() implementation ethtool: pse-pd: Fix double word in comments xsk: add linux/vmalloc.h to xsk.c sefltests: netdevsim: wait for devlink instance after netns removal selftest: fib_tests: Always cleanup before exit net/mlx5e: Align IPsec ASO result memory to be as required by hardware net/mlx5e: TC, Set CT miss to the specific ct action instance net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG net/mlx5: Refactor tc miss handling to a single function net/mlx5: Kconfig: Make tc offload depend on tc skb extension net/sched: flower: Support hardware miss to tc action net/sched: flower: Move filter handle initialization earlier net/sched: cls_api: Support hardware miss to tc action net/sched: Rename user cookie and act cookie sfc: fix builds without CONFIG_RTC_LIB sfc: clean up some inconsistent indentings net/mlx4_en: Introduce flexible array to silence overflow warning net: lan966x: Fix possible deadlock inside PTP net/ulp: Remove redundant ->clone() test in inet_clone_ulp(). ... --- drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c | 442 +++++++++++++++++++++++++++ 1 file changed, 442 insertions(+) create mode 100644 drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c (limited to 'drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c') diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c new file mode 100644 index 000000000..8aebe408c --- /dev/null +++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c @@ -0,0 +1,442 @@ +// SPDX-License-Identifier: GPL-2.0 OR MIT +/* + * Copyright 2014-2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "kfd_priv.h" +#include +#include +#include + +/* + * The primary memory I/O features being added for revisions of gfxip + * beyond 7.0 (Kaveri) are: + * + * Access to ATC/IOMMU mapped memory w/ associated extension of VA to 48b + * + * “Flat” shader memory access – These are new shader vector memory + * operations that do not reference a T#/V# so a “pointer” is what is + * sourced from the vector gprs for direct access to memory. + * This pointer space has the Shared(LDS) and Private(Scratch) memory + * mapped into this pointer space as apertures. + * The hardware then determines how to direct the memory request + * based on what apertures the request falls in. + * + * Unaligned support and alignment check + * + * + * System Unified Address - SUA + * + * The standard usage for GPU virtual addresses are that they are mapped by + * a set of page tables we call GPUVM and these page tables are managed by + * a combination of vidMM/driver software components. The current virtual + * address (VA) range for GPUVM is 40b. + * + * As of gfxip7.1 and beyond we’re adding the ability for compute memory + * clients (CP/RLC, DMA, SHADER(ifetch, scalar, and vector ops)) to access + * the same page tables used by host x86 processors and that are managed by + * the operating system. This is via a technique and hardware called ATC/IOMMU. + * The GPU has the capability of accessing both the GPUVM and ATC address + * spaces for a given VMID (process) simultaneously and we call this feature + * system unified address (SUA). + * + * There are three fundamental address modes of operation for a given VMID + * (process) on the GPU: + * + * HSA64 – 64b pointers and the default address space is ATC + * HSA32 – 32b pointers and the default address space is ATC + * GPUVM – 64b pointers and the default address space is GPUVM (driver + * model mode) + * + * + * HSA64 - ATC/IOMMU 64b + * + * A 64b pointer in the AMD64/IA64 CPU architecture is not fully utilized + * by the CPU so an AMD CPU can only access the high area + * (VA[63:47] == 0x1FFFF) and low area (VA[63:47 == 0) of the address space + * so the actual VA carried to translation is 48b. There is a “hole” in + * the middle of the 64b VA space. + * + * The GPU not only has access to all of the CPU accessible address space via + * ATC/IOMMU, but it also has access to the GPUVM address space. The “system + * unified address” feature (SUA) is the mapping of GPUVM and ATC address + * spaces into a unified pointer space. The method we take for 64b mode is + * to map the full 40b GPUVM address space into the hole of the 64b address + * space. + + * The GPUVM_Base/GPUVM_Limit defines the aperture in the 64b space where we + * direct requests to be translated via GPUVM page tables instead of the + * IOMMU path. + * + * + * 64b to 49b Address conversion + * + * Note that there are still significant portions of unused regions (holes) + * in the 64b address space even for the GPU. There are several places in + * the pipeline (sw and hw), we wish to compress the 64b virtual address + * to a 49b address. This 49b address is constituted of an “ATC” bit + * plus a 48b virtual address. This 49b address is what is passed to the + * translation hardware. ATC==0 means the 48b address is a GPUVM address + * (max of 2^40 – 1) intended to be translated via GPUVM page tables. + * ATC==1 means the 48b address is intended to be translated via IOMMU + * page tables. + * + * A 64b pointer is compared to the apertures that are defined (Base/Limit), in + * this case the GPUVM aperture (red) is defined and if a pointer falls in this + * aperture, we subtract the GPUVM_Base address and set the ATC bit to zero + * as part of the 64b to 49b conversion. + * + * Where this 64b to 49b conversion is done is a function of the usage. + * Most GPU memory access is via memory objects where the driver builds + * a descriptor which consists of a base address and a memory access by + * the GPU usually consists of some kind of an offset or Cartesian coordinate + * that references this memory descriptor. This is the case for shader + * instructions that reference the T# or V# constants, or for specified + * locations of assets (ex. the shader program location). In these cases + * the driver is what handles the 64b to 49b conversion and the base + * address in the descriptor (ex. V# or T# or shader program location) + * is defined as a 48b address w/ an ATC bit. For this usage a given + * memory object cannot straddle multiple apertures in the 64b address + * space. For example a shader program cannot jump in/out between ATC + * and GPUVM space. + * + * In some cases we wish to pass a 64b pointer to the GPU hardware and + * the GPU hw does the 64b to 49b conversion before passing memory + * requests to the cache/memory system. This is the case for the + * S_LOAD and FLAT_* shader memory instructions where we have 64b pointers + * in scalar and vector GPRs respectively. + * + * In all cases (no matter where the 64b -> 49b conversion is done), the gfxip + * hardware sends a 48b address along w/ an ATC bit, to the memory controller + * on the memory request interfaces. + * + * _MC_rdreq_atc // read request ATC bit + * + * 0 : _MC_rdreq_addr is a GPUVM VA + * + * 1 : _MC_rdreq_addr is a ATC VA + * + * + * “Spare” aperture (APE1) + * + * We use the GPUVM aperture to differentiate ATC vs. GPUVM, but we also use + * apertures to set the Mtype field for S_LOAD/FLAT_* ops which is input to the + * config tables for setting cache policies. The “spare” (APE1) aperture is + * motivated by getting a different Mtype from the default. + * The default aperture isn’t an actual base/limit aperture; it is just the + * address space that doesn’t hit any defined base/limit apertures. + * The following diagram is a complete picture of the gfxip7.x SUA apertures. + * The APE1 can be placed either below or above + * the hole (cannot be in the hole). + * + * + * General Aperture definitions and rules + * + * An aperture register definition consists of a Base, Limit, Mtype, and + * usually an ATC bit indicating which translation tables that aperture uses. + * In all cases (for SUA and DUA apertures discussed later), aperture base + * and limit definitions are 64KB aligned. + * + * _Base[63:0] = { _Base_register[63:16], 0x0000 } + * + * _Limit[63:0] = { _Limit_register[63:16], 0xFFFF } + * + * The base and limit are considered inclusive to an aperture so being + * inside an aperture means (address >= Base) AND (address <= Limit). + * + * In no case is a payload that straddles multiple apertures expected to work. + * For example a load_dword_x4 that starts in one aperture and ends in another, + * does not work. For the vector FLAT_* ops we have detection capability in + * the shader for reporting a “memory violation” back to the + * SQ block for use in traps. + * A memory violation results when an op falls into the hole, + * or a payload straddles multiple apertures. The S_LOAD instruction + * does not have this detection. + * + * Apertures cannot overlap. + * + * + * + * HSA32 - ATC/IOMMU 32b + * + * For HSA32 mode, the pointers are interpreted as 32 bits and use a single GPR + * instead of two for the S_LOAD and FLAT_* ops. The entire GPUVM space of 40b + * will not fit so there is only partial visibility to the GPUVM + * space (defined by the aperture) for S_LOAD and FLAT_* ops. + * There is no spare (APE1) aperture for HSA32 mode. + * + * + * GPUVM 64b mode (driver model) + * + * This mode is related to HSA64 in that the difference really is that + * the default aperture is GPUVM (ATC==0) and not ATC space. + * We have gfxip7.x hardware that has FLAT_* and S_LOAD support for + * SUA GPUVM mode, but does not support HSA32/HSA64. + * + * + * Device Unified Address - DUA + * + * Device unified address (DUA) is the name of the feature that maps the + * Shared(LDS) memory and Private(Scratch) memory into the overall address + * space for use by the new FLAT_* vector memory ops. The Shared and + * Private memories are mapped as apertures into the address space, + * and the hardware detects when a FLAT_* memory request is to be redirected + * to the LDS or Scratch memory when it falls into one of these apertures. + * Like the SUA apertures, the Shared/Private apertures are 64KB aligned and + * the base/limit is “in” the aperture. For both HSA64 and GPUVM SUA modes, + * the Shared/Private apertures are always placed in a limited selection of + * options in the hole of the 64b address space. For HSA32 mode, the + * Shared/Private apertures can be placed anywhere in the 32b space + * except at 0. + * + * + * HSA64 Apertures for FLAT_* vector ops + * + * For HSA64 SUA mode, the Shared and Private apertures are always placed + * in the hole w/ a limited selection of possible locations. The requests + * that fall in the private aperture are expanded as a function of the + * work-item id (tid) and redirected to the location of the + * “hidden private memory”. The hidden private can be placed in either GPUVM + * or ATC space. The addresses that fall in the shared aperture are + * re-directed to the on-chip LDS memory hardware. + * + * + * HSA32 Apertures for FLAT_* vector ops + * + * In HSA32 mode, the Private and Shared apertures can be placed anywhere + * in the 32b space except at 0 (Private or Shared Base at zero disables + * the apertures). If the base address of the apertures are non-zero + * (ie apertures exists), the size is always 64KB. + * + * + * GPUVM Apertures for FLAT_* vector ops + * + * In GPUVM mode, the Shared/Private apertures are specified identically + * to HSA64 mode where they are always in the hole at a limited selection + * of locations. + * + * + * Aperture Definitions for SUA and DUA + * + * The interpretation of the aperture register definitions for a given + * VMID is a function of the “SUA Mode” which is one of HSA64, HSA32, or + * GPUVM64 discussed in previous sections. The mode is first decoded, and + * then the remaining register decode is a function of the mode. + * + * + * SUA Mode Decode + * + * For the S_LOAD and FLAT_* shader operations, the SUA mode is decoded from + * the COMPUTE_DISPATCH_INITIATOR:DATA_ATC bit and + * the SH_MEM_CONFIG:PTR32 bits. + * + * COMPUTE_DISPATCH_INITIATOR:DATA_ATC SH_MEM_CONFIG:PTR32 Mode + * + * 1 0 HSA64 + * + * 1 1 HSA32 + * + * 0 X GPUVM64 + * + * In general the hardware will ignore the PTR32 bit and treat + * as “0” whenever DATA_ATC = “0”, but sw should set PTR32=0 + * when DATA_ATC=0. + * + * The DATA_ATC bit is only set for compute dispatches. + * All “Draw” dispatches are hardcoded to GPUVM64 mode + * for FLAT_* / S_LOAD operations. + */ + +#define MAKE_GPUVM_APP_BASE_VI(gpu_num) \ + (((uint64_t)(gpu_num) << 61) + 0x1000000000000L) + +#define MAKE_GPUVM_APP_LIMIT(base, size) \ + (((uint64_t)(base) & 0xFFFFFF0000000000UL) + (size) - 1) + +#define MAKE_SCRATCH_APP_BASE_VI() \ + (((uint64_t)(0x1UL) << 61) + 0x100000000L) + +#define MAKE_SCRATCH_APP_LIMIT(base) \ + (((uint64_t)base & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF) + +#define MAKE_LDS_APP_BASE_VI() \ + (((uint64_t)(0x1UL) << 61) + 0x0) +#define MAKE_LDS_APP_LIMIT(base) \ + (((uint64_t)(base) & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF) + +/* On GFXv9 the LDS and scratch apertures are programmed independently + * using the high 16 bits of the 64-bit virtual address. They must be + * in the hole, which will be the case as long as the high 16 bits are + * not 0. + * + * The aperture sizes are still 4GB implicitly. + * + * A GPUVM aperture is not applicable on GFXv9. + */ +#define MAKE_LDS_APP_BASE_V9() ((uint64_t)(0x1UL) << 48) +#define MAKE_SCRATCH_APP_BASE_V9() ((uint64_t)(0x2UL) << 48) + +/* User mode manages most of the SVM aperture address space. The low + * 16MB are reserved for kernel use (CWSR trap handler and kernel IB + * for now). + */ +#define SVM_USER_BASE (u64)(KFD_CWSR_TBA_TMA_SIZE + 2*PAGE_SIZE) +#define SVM_CWSR_BASE (SVM_USER_BASE - KFD_CWSR_TBA_TMA_SIZE) +#define SVM_IB_BASE (SVM_CWSR_BASE - PAGE_SIZE) + +static void kfd_init_apertures_vi(struct kfd_process_device *pdd, uint8_t id) +{ + /* + * node id couldn't be 0 - the three MSB bits of + * aperture shouldn't be 0 + */ + pdd->lds_base = MAKE_LDS_APP_BASE_VI(); + pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base); + + if (!pdd->dev->use_iommu_v2) { + /* dGPUs: SVM aperture starting at 0 + * with small reserved space for kernel. + * Set them to CANONICAL addresses. + */ + pdd->gpuvm_base = SVM_USER_BASE; + pdd->gpuvm_limit = + pdd->dev->shared_resources.gpuvm_size - 1; + } else { + /* set them to non CANONICAL addresses, and no SVM is + * allocated. + */ + pdd->gpuvm_base = MAKE_GPUVM_APP_BASE_VI(id + 1); + pdd->gpuvm_limit = MAKE_GPUVM_APP_LIMIT(pdd->gpuvm_base, + pdd->dev->shared_resources.gpuvm_size); + } + + pdd->scratch_base = MAKE_SCRATCH_APP_BASE_VI(); + pdd->scratch_limit = MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base); +} + +static void kfd_init_apertures_v9(struct kfd_process_device *pdd, uint8_t id) +{ + pdd->lds_base = MAKE_LDS_APP_BASE_V9(); + pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base); + + /* Raven needs SVM to support graphic handle, etc. Leave the small + * reserved space before SVM on Raven as well, even though we don't + * have to. + * Set gpuvm_base and gpuvm_limit to CANONICAL addresses so that they + * are used in Thunk to reserve SVM. + */ + pdd->gpuvm_base = SVM_USER_BASE; + pdd->gpuvm_limit = + pdd->dev->shared_resources.gpuvm_size - 1; + + pdd->scratch_base = MAKE_SCRATCH_APP_BASE_V9(); + pdd->scratch_limit = MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base); +} + +int kfd_init_apertures(struct kfd_process *process) +{ + uint8_t id = 0; + struct kfd_dev *dev; + struct kfd_process_device *pdd; + + /*Iterating over all devices*/ + while (kfd_topology_enum_kfd_devices(id, &dev) == 0) { + if (!dev || kfd_devcgroup_check_permission(dev)) { + /* Skip non GPU devices and devices to which the + * current process have no access to. Access can be + * limited by placing the process in a specific + * cgroup hierarchy + */ + id++; + continue; + } + + pdd = kfd_create_process_device_data(dev, process); + if (!pdd) { + pr_err("Failed to create process device data\n"); + return -ENOMEM; + } + /* + * For 64 bit process apertures will be statically reserved in + * the x86_64 non canonical process address space + * amdkfd doesn't currently support apertures for 32 bit process + */ + if (process->is_32bit_user_mode) { + pdd->lds_base = pdd->lds_limit = 0; + pdd->gpuvm_base = pdd->gpuvm_limit = 0; + pdd->scratch_base = pdd->scratch_limit = 0; + } else { + switch (dev->adev->asic_type) { + case CHIP_KAVERI: + case CHIP_HAWAII: + case CHIP_CARRIZO: + case CHIP_TONGA: + case CHIP_FIJI: + case CHIP_POLARIS10: + case CHIP_POLARIS11: + case CHIP_POLARIS12: + case CHIP_VEGAM: + kfd_init_apertures_vi(pdd, id); + break; + default: + if (KFD_GC_VERSION(dev) >= IP_VERSION(9, 0, 1)) + kfd_init_apertures_v9(pdd, id); + else { + WARN(1, "Unexpected ASIC family %u", + dev->adev->asic_type); + return -EINVAL; + } + } + + if (!dev->use_iommu_v2) { + /* dGPUs: the reserved space for kernel + * before SVM + */ + pdd->qpd.cwsr_base = SVM_CWSR_BASE; + pdd->qpd.ib_base = SVM_IB_BASE; + } + } + + dev_dbg(kfd_device, "node id %u\n", id); + dev_dbg(kfd_device, "gpu id %u\n", pdd->dev->id); + dev_dbg(kfd_device, "lds_base %llX\n", pdd->lds_base); + dev_dbg(kfd_device, "lds_limit %llX\n", pdd->lds_limit); + dev_dbg(kfd_device, "gpuvm_base %llX\n", pdd->gpuvm_base); + dev_dbg(kfd_device, "gpuvm_limit %llX\n", pdd->gpuvm_limit); + dev_dbg(kfd_device, "scratch_base %llX\n", pdd->scratch_base); + dev_dbg(kfd_device, "scratch_limit %llX\n", pdd->scratch_limit); + + id++; + } + + return 0; +} -- cgit v1.2.3