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/radeon/rs690.c | 879 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 879 insertions(+) create mode 100644 drivers/gpu/drm/radeon/rs690.c (limited to 'drivers/gpu/drm/radeon/rs690.c') diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c new file mode 100644 index 000000000..14fb0819b --- /dev/null +++ b/drivers/gpu/drm/radeon/rs690.c @@ -0,0 +1,879 @@ +/* + * Copyright 2008 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * 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. + * + * Authors: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ + +#include + +#include "atom.h" +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_audio.h" +#include "rs690d.h" + +int rs690_mc_wait_for_idle(struct radeon_device *rdev) +{ + unsigned i; + uint32_t tmp; + + for (i = 0; i < rdev->usec_timeout; i++) { + /* read MC_STATUS */ + tmp = RREG32_MC(R_000090_MC_SYSTEM_STATUS); + if (G_000090_MC_SYSTEM_IDLE(tmp)) + return 0; + udelay(1); + } + return -1; +} + +static void rs690_gpu_init(struct radeon_device *rdev) +{ + /* FIXME: is this correct ? */ + r420_pipes_init(rdev); + if (rs690_mc_wait_for_idle(rdev)) { + pr_warn("Failed to wait MC idle while programming pipes. Bad things might happen.\n"); + } +} + +union igp_info { + struct _ATOM_INTEGRATED_SYSTEM_INFO info; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_v2; +}; + +void rs690_pm_info(struct radeon_device *rdev) +{ + int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo); + union igp_info *info; + uint16_t data_offset; + uint8_t frev, crev; + fixed20_12 tmp; + + if (atom_parse_data_header(rdev->mode_info.atom_context, index, NULL, + &frev, &crev, &data_offset)) { + info = (union igp_info *)(rdev->mode_info.atom_context->bios + data_offset); + + /* Get various system informations from bios */ + switch (crev) { + case 1: + tmp.full = dfixed_const(100); + rdev->pm.igp_sideport_mclk.full = dfixed_const(le32_to_cpu(info->info.ulBootUpMemoryClock)); + rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); + if (le16_to_cpu(info->info.usK8MemoryClock)) + rdev->pm.igp_system_mclk.full = dfixed_const(le16_to_cpu(info->info.usK8MemoryClock)); + else if (rdev->clock.default_mclk) { + rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); + rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); + } else + rdev->pm.igp_system_mclk.full = dfixed_const(400); + rdev->pm.igp_ht_link_clk.full = dfixed_const(le16_to_cpu(info->info.usFSBClock)); + rdev->pm.igp_ht_link_width.full = dfixed_const(info->info.ucHTLinkWidth); + break; + case 2: + tmp.full = dfixed_const(100); + rdev->pm.igp_sideport_mclk.full = dfixed_const(le32_to_cpu(info->info_v2.ulBootUpSidePortClock)); + rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); + if (le32_to_cpu(info->info_v2.ulBootUpUMAClock)) + rdev->pm.igp_system_mclk.full = dfixed_const(le32_to_cpu(info->info_v2.ulBootUpUMAClock)); + else if (rdev->clock.default_mclk) + rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); + else + rdev->pm.igp_system_mclk.full = dfixed_const(66700); + rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); + rdev->pm.igp_ht_link_clk.full = dfixed_const(le32_to_cpu(info->info_v2.ulHTLinkFreq)); + rdev->pm.igp_ht_link_clk.full = dfixed_div(rdev->pm.igp_ht_link_clk, tmp); + rdev->pm.igp_ht_link_width.full = dfixed_const(le16_to_cpu(info->info_v2.usMinHTLinkWidth)); + break; + default: + /* We assume the slower possible clock ie worst case */ + rdev->pm.igp_sideport_mclk.full = dfixed_const(200); + rdev->pm.igp_system_mclk.full = dfixed_const(200); + rdev->pm.igp_ht_link_clk.full = dfixed_const(1000); + rdev->pm.igp_ht_link_width.full = dfixed_const(8); + DRM_ERROR("No integrated system info for your GPU, using safe default\n"); + break; + } + } else { + /* We assume the slower possible clock ie worst case */ + rdev->pm.igp_sideport_mclk.full = dfixed_const(200); + rdev->pm.igp_system_mclk.full = dfixed_const(200); + rdev->pm.igp_ht_link_clk.full = dfixed_const(1000); + rdev->pm.igp_ht_link_width.full = dfixed_const(8); + DRM_ERROR("No integrated system info for your GPU, using safe default\n"); + } + /* Compute various bandwidth */ + /* k8_bandwidth = (memory_clk / 2) * 2 * 8 * 0.5 = memory_clk * 4 */ + tmp.full = dfixed_const(4); + rdev->pm.k8_bandwidth.full = dfixed_mul(rdev->pm.igp_system_mclk, tmp); + /* ht_bandwidth = ht_clk * 2 * ht_width / 8 * 0.8 + * = ht_clk * ht_width / 5 + */ + tmp.full = dfixed_const(5); + rdev->pm.ht_bandwidth.full = dfixed_mul(rdev->pm.igp_ht_link_clk, + rdev->pm.igp_ht_link_width); + rdev->pm.ht_bandwidth.full = dfixed_div(rdev->pm.ht_bandwidth, tmp); + if (tmp.full < rdev->pm.max_bandwidth.full) { + /* HT link is a limiting factor */ + rdev->pm.max_bandwidth.full = tmp.full; + } + /* sideport_bandwidth = (sideport_clk / 2) * 2 * 2 * 0.7 + * = (sideport_clk * 14) / 10 + */ + tmp.full = dfixed_const(14); + rdev->pm.sideport_bandwidth.full = dfixed_mul(rdev->pm.igp_sideport_mclk, tmp); + tmp.full = dfixed_const(10); + rdev->pm.sideport_bandwidth.full = dfixed_div(rdev->pm.sideport_bandwidth, tmp); +} + +static void rs690_mc_init(struct radeon_device *rdev) +{ + u64 base; + uint32_t h_addr, l_addr; + unsigned long long k8_addr; + + rs400_gart_adjust_size(rdev); + rdev->mc.vram_is_ddr = true; + rdev->mc.vram_width = 128; + rdev->mc.real_vram_size = RREG32(RADEON_CONFIG_MEMSIZE); + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; + rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); + rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); + rdev->mc.visible_vram_size = rdev->mc.aper_size; + base = RREG32_MC(R_000100_MCCFG_FB_LOCATION); + base = G_000100_MC_FB_START(base) << 16; + rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + /* Some boards seem to be configured for 128MB of sideport memory, + * but really only have 64MB. Just skip the sideport and use + * UMA memory. + */ + if (rdev->mc.igp_sideport_enabled && + (rdev->mc.real_vram_size == (384 * 1024 * 1024))) { + base += 128 * 1024 * 1024; + rdev->mc.real_vram_size -= 128 * 1024 * 1024; + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; + } + + /* Use K8 direct mapping for fast fb access. */ + rdev->fastfb_working = false; + h_addr = G_00005F_K8_ADDR_EXT(RREG32_MC(R_00005F_MC_MISC_UMA_CNTL)); + l_addr = RREG32_MC(R_00001E_K8_FB_LOCATION); + k8_addr = ((unsigned long long)h_addr) << 32 | l_addr; +#if defined(CONFIG_X86_32) && !defined(CONFIG_X86_PAE) + if (k8_addr + rdev->mc.visible_vram_size < 0x100000000ULL) +#endif + { + /* FastFB shall be used with UMA memory. Here it is simply disabled when sideport + * memory is present. + */ + if (!rdev->mc.igp_sideport_enabled && radeon_fastfb == 1) { + DRM_INFO("Direct mapping: aper base at 0x%llx, replaced by direct mapping base 0x%llx.\n", + (unsigned long long)rdev->mc.aper_base, k8_addr); + rdev->mc.aper_base = (resource_size_t)k8_addr; + rdev->fastfb_working = true; + } + } + + rs690_pm_info(rdev); + radeon_vram_location(rdev, &rdev->mc, base); + rdev->mc.gtt_base_align = rdev->mc.gtt_size - 1; + radeon_gtt_location(rdev, &rdev->mc); + radeon_update_bandwidth_info(rdev); +} + +void rs690_line_buffer_adjust(struct radeon_device *rdev, + struct drm_display_mode *mode1, + struct drm_display_mode *mode2) +{ + u32 tmp; + + /* Guess line buffer size to be 8192 pixels */ + u32 lb_size = 8192; + + /* + * Line Buffer Setup + * There is a single line buffer shared by both display controllers. + * R_006520_DC_LB_MEMORY_SPLIT controls how that line buffer is shared between + * the display controllers. The paritioning can either be done + * manually or via one of four preset allocations specified in bits 1:0: + * 0 - line buffer is divided in half and shared between crtc + * 1 - D1 gets 3/4 of the line buffer, D2 gets 1/4 + * 2 - D1 gets the whole buffer + * 3 - D1 gets 1/4 of the line buffer, D2 gets 3/4 + * Setting bit 2 of R_006520_DC_LB_MEMORY_SPLIT controls switches to manual + * allocation mode. In manual allocation mode, D1 always starts at 0, + * D1 end/2 is specified in bits 14:4; D2 allocation follows D1. + */ + tmp = RREG32(R_006520_DC_LB_MEMORY_SPLIT) & C_006520_DC_LB_MEMORY_SPLIT; + tmp &= ~C_006520_DC_LB_MEMORY_SPLIT_MODE; + /* auto */ + if (mode1 && mode2) { + if (mode1->hdisplay > mode2->hdisplay) { + if (mode1->hdisplay > 2560) + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1_3Q_D2_1Q; + else + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; + } else if (mode2->hdisplay > mode1->hdisplay) { + if (mode2->hdisplay > 2560) + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q; + else + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; + } else + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; + } else if (mode1) { + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1_ONLY; + } else if (mode2) { + tmp |= V_006520_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q; + } + WREG32(R_006520_DC_LB_MEMORY_SPLIT, tmp); + + /* Save number of lines the linebuffer leads before the scanout */ + if (mode1) + rdev->mode_info.crtcs[0]->lb_vblank_lead_lines = DIV_ROUND_UP(lb_size, mode1->crtc_hdisplay); + + if (mode2) + rdev->mode_info.crtcs[1]->lb_vblank_lead_lines = DIV_ROUND_UP(lb_size, mode2->crtc_hdisplay); +} + +struct rs690_watermark { + u32 lb_request_fifo_depth; + fixed20_12 num_line_pair; + fixed20_12 estimated_width; + fixed20_12 worst_case_latency; + fixed20_12 consumption_rate; + fixed20_12 active_time; + fixed20_12 dbpp; + fixed20_12 priority_mark_max; + fixed20_12 priority_mark; + fixed20_12 sclk; +}; + +static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev, + struct radeon_crtc *crtc, + struct rs690_watermark *wm, + bool low) +{ + struct drm_display_mode *mode = &crtc->base.mode; + fixed20_12 a, b, c; + fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width; + fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency; + fixed20_12 sclk, core_bandwidth, max_bandwidth; + u32 selected_sclk; + + if (!crtc->base.enabled) { + /* FIXME: wouldn't it better to set priority mark to maximum */ + wm->lb_request_fifo_depth = 4; + return; + } + + if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) && + (rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) + selected_sclk = radeon_dpm_get_sclk(rdev, low); + else + selected_sclk = rdev->pm.current_sclk; + + /* sclk in Mhz */ + a.full = dfixed_const(100); + sclk.full = dfixed_const(selected_sclk); + sclk.full = dfixed_div(sclk, a); + + /* core_bandwidth = sclk(Mhz) * 16 */ + a.full = dfixed_const(16); + core_bandwidth.full = dfixed_div(rdev->pm.sclk, a); + + if (crtc->vsc.full > dfixed_const(2)) + wm->num_line_pair.full = dfixed_const(2); + else + wm->num_line_pair.full = dfixed_const(1); + + b.full = dfixed_const(mode->crtc_hdisplay); + c.full = dfixed_const(256); + a.full = dfixed_div(b, c); + request_fifo_depth.full = dfixed_mul(a, wm->num_line_pair); + request_fifo_depth.full = dfixed_ceil(request_fifo_depth); + if (a.full < dfixed_const(4)) { + wm->lb_request_fifo_depth = 4; + } else { + wm->lb_request_fifo_depth = dfixed_trunc(request_fifo_depth); + } + + /* Determine consumption rate + * pclk = pixel clock period(ns) = 1000 / (mode.clock / 1000) + * vtaps = number of vertical taps, + * vsc = vertical scaling ratio, defined as source/destination + * hsc = horizontal scaling ration, defined as source/destination + */ + a.full = dfixed_const(mode->clock); + b.full = dfixed_const(1000); + a.full = dfixed_div(a, b); + pclk.full = dfixed_div(b, a); + if (crtc->rmx_type != RMX_OFF) { + b.full = dfixed_const(2); + if (crtc->vsc.full > b.full) + b.full = crtc->vsc.full; + b.full = dfixed_mul(b, crtc->hsc); + c.full = dfixed_const(2); + b.full = dfixed_div(b, c); + consumption_time.full = dfixed_div(pclk, b); + } else { + consumption_time.full = pclk.full; + } + a.full = dfixed_const(1); + wm->consumption_rate.full = dfixed_div(a, consumption_time); + + + /* Determine line time + * LineTime = total time for one line of displayhtotal + * LineTime = total number of horizontal pixels + * pclk = pixel clock period(ns) + */ + a.full = dfixed_const(crtc->base.mode.crtc_htotal); + line_time.full = dfixed_mul(a, pclk); + + /* Determine active time + * ActiveTime = time of active region of display within one line, + * hactive = total number of horizontal active pixels + * htotal = total number of horizontal pixels + */ + a.full = dfixed_const(crtc->base.mode.crtc_htotal); + b.full = dfixed_const(crtc->base.mode.crtc_hdisplay); + wm->active_time.full = dfixed_mul(line_time, b); + wm->active_time.full = dfixed_div(wm->active_time, a); + + /* Maximun bandwidth is the minimun bandwidth of all component */ + max_bandwidth = core_bandwidth; + if (rdev->mc.igp_sideport_enabled) { + if (max_bandwidth.full > rdev->pm.sideport_bandwidth.full && + rdev->pm.sideport_bandwidth.full) + max_bandwidth = rdev->pm.sideport_bandwidth; + read_delay_latency.full = dfixed_const(370 * 800); + a.full = dfixed_const(1000); + b.full = dfixed_div(rdev->pm.igp_sideport_mclk, a); + read_delay_latency.full = dfixed_div(read_delay_latency, b); + read_delay_latency.full = dfixed_mul(read_delay_latency, a); + } else { + if (max_bandwidth.full > rdev->pm.k8_bandwidth.full && + rdev->pm.k8_bandwidth.full) + max_bandwidth = rdev->pm.k8_bandwidth; + if (max_bandwidth.full > rdev->pm.ht_bandwidth.full && + rdev->pm.ht_bandwidth.full) + max_bandwidth = rdev->pm.ht_bandwidth; + read_delay_latency.full = dfixed_const(5000); + } + + /* sclk = system clocks(ns) = 1000 / max_bandwidth / 16 */ + a.full = dfixed_const(16); + sclk.full = dfixed_mul(max_bandwidth, a); + a.full = dfixed_const(1000); + sclk.full = dfixed_div(a, sclk); + /* Determine chunk time + * ChunkTime = the time it takes the DCP to send one chunk of data + * to the LB which consists of pipeline delay and inter chunk gap + * sclk = system clock(ns) + */ + a.full = dfixed_const(256 * 13); + chunk_time.full = dfixed_mul(sclk, a); + a.full = dfixed_const(10); + chunk_time.full = dfixed_div(chunk_time, a); + + /* Determine the worst case latency + * NumLinePair = Number of line pairs to request(1=2 lines, 2=4 lines) + * WorstCaseLatency = worst case time from urgent to when the MC starts + * to return data + * READ_DELAY_IDLE_MAX = constant of 1us + * ChunkTime = time it takes the DCP to send one chunk of data to the LB + * which consists of pipeline delay and inter chunk gap + */ + if (dfixed_trunc(wm->num_line_pair) > 1) { + a.full = dfixed_const(3); + wm->worst_case_latency.full = dfixed_mul(a, chunk_time); + wm->worst_case_latency.full += read_delay_latency.full; + } else { + a.full = dfixed_const(2); + wm->worst_case_latency.full = dfixed_mul(a, chunk_time); + wm->worst_case_latency.full += read_delay_latency.full; + } + + /* Determine the tolerable latency + * TolerableLatency = Any given request has only 1 line time + * for the data to be returned + * LBRequestFifoDepth = Number of chunk requests the LB can + * put into the request FIFO for a display + * LineTime = total time for one line of display + * ChunkTime = the time it takes the DCP to send one chunk + * of data to the LB which consists of + * pipeline delay and inter chunk gap + */ + if ((2+wm->lb_request_fifo_depth) >= dfixed_trunc(request_fifo_depth)) { + tolerable_latency.full = line_time.full; + } else { + tolerable_latency.full = dfixed_const(wm->lb_request_fifo_depth - 2); + tolerable_latency.full = request_fifo_depth.full - tolerable_latency.full; + tolerable_latency.full = dfixed_mul(tolerable_latency, chunk_time); + tolerable_latency.full = line_time.full - tolerable_latency.full; + } + /* We assume worst case 32bits (4 bytes) */ + wm->dbpp.full = dfixed_const(4 * 8); + + /* Determine the maximum priority mark + * width = viewport width in pixels + */ + a.full = dfixed_const(16); + wm->priority_mark_max.full = dfixed_const(crtc->base.mode.crtc_hdisplay); + wm->priority_mark_max.full = dfixed_div(wm->priority_mark_max, a); + wm->priority_mark_max.full = dfixed_ceil(wm->priority_mark_max); + + /* Determine estimated width */ + estimated_width.full = tolerable_latency.full - wm->worst_case_latency.full; + estimated_width.full = dfixed_div(estimated_width, consumption_time); + if (dfixed_trunc(estimated_width) > crtc->base.mode.crtc_hdisplay) { + wm->priority_mark.full = dfixed_const(10); + } else { + a.full = dfixed_const(16); + wm->priority_mark.full = dfixed_div(estimated_width, a); + wm->priority_mark.full = dfixed_ceil(wm->priority_mark); + wm->priority_mark.full = wm->priority_mark_max.full - wm->priority_mark.full; + } +} + +static void rs690_compute_mode_priority(struct radeon_device *rdev, + struct rs690_watermark *wm0, + struct rs690_watermark *wm1, + struct drm_display_mode *mode0, + struct drm_display_mode *mode1, + u32 *d1mode_priority_a_cnt, + u32 *d2mode_priority_a_cnt) +{ + fixed20_12 priority_mark02, priority_mark12, fill_rate; + fixed20_12 a, b; + + *d1mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1); + *d2mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1); + + if (mode0 && mode1) { + if (dfixed_trunc(wm0->dbpp) > 64) + a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair); + else + a.full = wm0->num_line_pair.full; + if (dfixed_trunc(wm1->dbpp) > 64) + b.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair); + else + b.full = wm1->num_line_pair.full; + a.full += b.full; + fill_rate.full = dfixed_div(wm0->sclk, a); + if (wm0->consumption_rate.full > fill_rate.full) { + b.full = wm0->consumption_rate.full - fill_rate.full; + b.full = dfixed_mul(b, wm0->active_time); + a.full = dfixed_mul(wm0->worst_case_latency, + wm0->consumption_rate); + a.full = a.full + b.full; + b.full = dfixed_const(16 * 1000); + priority_mark02.full = dfixed_div(a, b); + } else { + a.full = dfixed_mul(wm0->worst_case_latency, + wm0->consumption_rate); + b.full = dfixed_const(16 * 1000); + priority_mark02.full = dfixed_div(a, b); + } + if (wm1->consumption_rate.full > fill_rate.full) { + b.full = wm1->consumption_rate.full - fill_rate.full; + b.full = dfixed_mul(b, wm1->active_time); + a.full = dfixed_mul(wm1->worst_case_latency, + wm1->consumption_rate); + a.full = a.full + b.full; + b.full = dfixed_const(16 * 1000); + priority_mark12.full = dfixed_div(a, b); + } else { + a.full = dfixed_mul(wm1->worst_case_latency, + wm1->consumption_rate); + b.full = dfixed_const(16 * 1000); + priority_mark12.full = dfixed_div(a, b); + } + if (wm0->priority_mark.full > priority_mark02.full) + priority_mark02.full = wm0->priority_mark.full; + if (wm0->priority_mark_max.full > priority_mark02.full) + priority_mark02.full = wm0->priority_mark_max.full; + if (wm1->priority_mark.full > priority_mark12.full) + priority_mark12.full = wm1->priority_mark.full; + if (wm1->priority_mark_max.full > priority_mark12.full) + priority_mark12.full = wm1->priority_mark_max.full; + *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); + *d2mode_priority_a_cnt = dfixed_trunc(priority_mark12); + if (rdev->disp_priority == 2) { + *d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1); + *d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1); + } + } else if (mode0) { + if (dfixed_trunc(wm0->dbpp) > 64) + a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair); + else + a.full = wm0->num_line_pair.full; + fill_rate.full = dfixed_div(wm0->sclk, a); + if (wm0->consumption_rate.full > fill_rate.full) { + b.full = wm0->consumption_rate.full - fill_rate.full; + b.full = dfixed_mul(b, wm0->active_time); + a.full = dfixed_mul(wm0->worst_case_latency, + wm0->consumption_rate); + a.full = a.full + b.full; + b.full = dfixed_const(16 * 1000); + priority_mark02.full = dfixed_div(a, b); + } else { + a.full = dfixed_mul(wm0->worst_case_latency, + wm0->consumption_rate); + b.full = dfixed_const(16 * 1000); + priority_mark02.full = dfixed_div(a, b); + } + if (wm0->priority_mark.full > priority_mark02.full) + priority_mark02.full = wm0->priority_mark.full; + if (wm0->priority_mark_max.full > priority_mark02.full) + priority_mark02.full = wm0->priority_mark_max.full; + *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); + if (rdev->disp_priority == 2) + *d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1); + } else if (mode1) { + if (dfixed_trunc(wm1->dbpp) > 64) + a.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair); + else + a.full = wm1->num_line_pair.full; + fill_rate.full = dfixed_div(wm1->sclk, a); + if (wm1->consumption_rate.full > fill_rate.full) { + b.full = wm1->consumption_rate.full - fill_rate.full; + b.full = dfixed_mul(b, wm1->active_time); + a.full = dfixed_mul(wm1->worst_case_latency, + wm1->consumption_rate); + a.full = a.full + b.full; + b.full = dfixed_const(16 * 1000); + priority_mark12.full = dfixed_div(a, b); + } else { + a.full = dfixed_mul(wm1->worst_case_latency, + wm1->consumption_rate); + b.full = dfixed_const(16 * 1000); + priority_mark12.full = dfixed_div(a, b); + } + if (wm1->priority_mark.full > priority_mark12.full) + priority_mark12.full = wm1->priority_mark.full; + if (wm1->priority_mark_max.full > priority_mark12.full) + priority_mark12.full = wm1->priority_mark_max.full; + *d2mode_priority_a_cnt = dfixed_trunc(priority_mark12); + if (rdev->disp_priority == 2) + *d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1); + } +} + +void rs690_bandwidth_update(struct radeon_device *rdev) +{ + struct drm_display_mode *mode0 = NULL; + struct drm_display_mode *mode1 = NULL; + struct rs690_watermark wm0_high, wm0_low; + struct rs690_watermark wm1_high, wm1_low; + u32 tmp; + u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt; + u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt; + + if (!rdev->mode_info.mode_config_initialized) + return; + + radeon_update_display_priority(rdev); + + if (rdev->mode_info.crtcs[0]->base.enabled) + mode0 = &rdev->mode_info.crtcs[0]->base.mode; + if (rdev->mode_info.crtcs[1]->base.enabled) + mode1 = &rdev->mode_info.crtcs[1]->base.mode; + /* + * Set display0/1 priority up in the memory controller for + * modes if the user specifies HIGH for displaypriority + * option. + */ + if ((rdev->disp_priority == 2) && + ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))) { + tmp = RREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER); + tmp &= C_000104_MC_DISP0R_INIT_LAT; + tmp &= C_000104_MC_DISP1R_INIT_LAT; + if (mode0) + tmp |= S_000104_MC_DISP0R_INIT_LAT(1); + if (mode1) + tmp |= S_000104_MC_DISP1R_INIT_LAT(1); + WREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER, tmp); + } + rs690_line_buffer_adjust(rdev, mode0, mode1); + + if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740)) + WREG32(R_006C9C_DCP_CONTROL, 0); + if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) + WREG32(R_006C9C_DCP_CONTROL, 2); + + rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_high, false); + rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_high, false); + + rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_low, true); + rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_low, true); + + tmp = (wm0_high.lb_request_fifo_depth - 1); + tmp |= (wm1_high.lb_request_fifo_depth - 1) << 16; + WREG32(R_006D58_LB_MAX_REQ_OUTSTANDING, tmp); + + rs690_compute_mode_priority(rdev, + &wm0_high, &wm1_high, + mode0, mode1, + &d1mode_priority_a_cnt, &d2mode_priority_a_cnt); + rs690_compute_mode_priority(rdev, + &wm0_low, &wm1_low, + mode0, mode1, + &d1mode_priority_b_cnt, &d2mode_priority_b_cnt); + + WREG32(R_006548_D1MODE_PRIORITY_A_CNT, d1mode_priority_a_cnt); + WREG32(R_00654C_D1MODE_PRIORITY_B_CNT, d1mode_priority_b_cnt); + WREG32(R_006D48_D2MODE_PRIORITY_A_CNT, d2mode_priority_a_cnt); + WREG32(R_006D4C_D2MODE_PRIORITY_B_CNT, d2mode_priority_b_cnt); +} + +uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg) +{ + unsigned long flags; + uint32_t r; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); + WREG32(R_000078_MC_INDEX, S_000078_MC_IND_ADDR(reg)); + r = RREG32(R_00007C_MC_DATA); + WREG32(R_000078_MC_INDEX, ~C_000078_MC_IND_ADDR); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); + return r; +} + +void rs690_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) +{ + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); + WREG32(R_000078_MC_INDEX, S_000078_MC_IND_ADDR(reg) | + S_000078_MC_IND_WR_EN(1)); + WREG32(R_00007C_MC_DATA, v); + WREG32(R_000078_MC_INDEX, 0x7F); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); +} + +static void rs690_mc_program(struct radeon_device *rdev) +{ + struct rv515_mc_save save; + + /* Stops all mc clients */ + rv515_mc_stop(rdev, &save); + + /* Wait for mc idle */ + if (rs690_mc_wait_for_idle(rdev)) + dev_warn(rdev->dev, "Wait MC idle timeout before updating MC.\n"); + /* Program MC, should be a 32bits limited address space */ + WREG32_MC(R_000100_MCCFG_FB_LOCATION, + S_000100_MC_FB_START(rdev->mc.vram_start >> 16) | + S_000100_MC_FB_TOP(rdev->mc.vram_end >> 16)); + WREG32(R_000134_HDP_FB_LOCATION, + S_000134_HDP_FB_START(rdev->mc.vram_start >> 16)); + + rv515_mc_resume(rdev, &save); +} + +static int rs690_startup(struct radeon_device *rdev) +{ + int r; + + rs690_mc_program(rdev); + /* Resume clock */ + rv515_clock_startup(rdev); + /* Initialize GPU configuration (# pipes, ...) */ + rs690_gpu_init(rdev); + /* Initialize GART (initialize after TTM so we can allocate + * memory through TTM but finalize after TTM) */ + r = rs400_gart_enable(rdev); + if (r) + return r; + + /* allocate wb buffer */ + r = radeon_wb_init(rdev); + if (r) + return r; + + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + /* Enable IRQ */ + if (!rdev->irq.installed) { + r = radeon_irq_kms_init(rdev); + if (r) + return r; + } + + rs600_irq_set(rdev); + rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); + /* 1M ring buffer */ + r = r100_cp_init(rdev, 1024 * 1024); + if (r) { + dev_err(rdev->dev, "failed initializing CP (%d).\n", r); + return r; + } + + r = radeon_ib_pool_init(rdev); + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + return r; + } + + r = radeon_audio_init(rdev); + if (r) { + dev_err(rdev->dev, "failed initializing audio\n"); + return r; + } + + return 0; +} + +int rs690_resume(struct radeon_device *rdev) +{ + int r; + + /* Make sur GART are not working */ + rs400_gart_disable(rdev); + /* Resume clock before doing reset */ + rv515_clock_startup(rdev); + /* Reset gpu before posting otherwise ATOM will enter infinite loop */ + if (radeon_asic_reset(rdev)) { + dev_warn(rdev->dev, "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", + RREG32(R_000E40_RBBM_STATUS), + RREG32(R_0007C0_CP_STAT)); + } + /* post */ + atom_asic_init(rdev->mode_info.atom_context); + /* Resume clock after posting */ + rv515_clock_startup(rdev); + /* Initialize surface registers */ + radeon_surface_init(rdev); + + rdev->accel_working = true; + r = rs690_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; +} + +int rs690_suspend(struct radeon_device *rdev) +{ + radeon_pm_suspend(rdev); + radeon_audio_fini(rdev); + r100_cp_disable(rdev); + radeon_wb_disable(rdev); + rs600_irq_disable(rdev); + rs400_gart_disable(rdev); + return 0; +} + +void rs690_fini(struct radeon_device *rdev) +{ + radeon_pm_fini(rdev); + radeon_audio_fini(rdev); + r100_cp_fini(rdev); + radeon_wb_fini(rdev); + radeon_ib_pool_fini(rdev); + radeon_gem_fini(rdev); + rs400_gart_fini(rdev); + radeon_irq_kms_fini(rdev); + radeon_fence_driver_fini(rdev); + radeon_bo_fini(rdev); + radeon_atombios_fini(rdev); + kfree(rdev->bios); + rdev->bios = NULL; +} + +int rs690_init(struct radeon_device *rdev) +{ + int r; + + /* Disable VGA */ + rv515_vga_render_disable(rdev); + /* Initialize scratch registers */ + radeon_scratch_init(rdev); + /* Initialize surface registers */ + radeon_surface_init(rdev); + /* restore some register to sane defaults */ + r100_restore_sanity(rdev); + /* TODO: disable VGA need to use VGA request */ + /* BIOS*/ + if (!radeon_get_bios(rdev)) { + if (ASIC_IS_AVIVO(rdev)) + return -EINVAL; + } + if (rdev->is_atom_bios) { + r = radeon_atombios_init(rdev); + if (r) + return r; + } else { + dev_err(rdev->dev, "Expecting atombios for RV515 GPU\n"); + return -EINVAL; + } + /* Reset gpu before posting otherwise ATOM will enter infinite loop */ + if (radeon_asic_reset(rdev)) { + dev_warn(rdev->dev, + "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", + RREG32(R_000E40_RBBM_STATUS), + RREG32(R_0007C0_CP_STAT)); + } + /* check if cards are posted or not */ + if (radeon_boot_test_post_card(rdev) == false) + return -EINVAL; + + /* Initialize clocks */ + radeon_get_clock_info(rdev->ddev); + /* initialize memory controller */ + rs690_mc_init(rdev); + rv515_debugfs(rdev); + /* Fence driver */ + radeon_fence_driver_init(rdev); + /* Memory manager */ + r = radeon_bo_init(rdev); + if (r) + return r; + r = rs400_gart_init(rdev); + if (r) + return r; + rs600_set_safe_registers(rdev); + + /* Initialize power management */ + radeon_pm_init(rdev); + + rdev->accel_working = true; + r = rs690_startup(rdev); + if (r) { + /* Somethings want wront with the accel init stop accel */ + dev_err(rdev->dev, "Disabling GPU acceleration\n"); + r100_cp_fini(rdev); + radeon_wb_fini(rdev); + radeon_ib_pool_fini(rdev); + rs400_gart_fini(rdev); + radeon_irq_kms_fini(rdev); + rdev->accel_working = false; + } + return 0; +} -- cgit v1.2.3