/* * Copyright 2005 Stephane Marchesin * All Rights Reserved. * * 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 (including the next * paragraph) 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 * PRECISION INSIGHT AND/OR ITS SUPPLIERS 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 "drmP.h" #include "drm.h" #include "drm_sarea.h" #include "nouveau_drv.h" #include "nouveau_drm.h" static int nouveau_init_card_mappings(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; int ret; /* resource 0 is mmio regs */ /* resource 1 is linear FB */ /* resource 2 is RAMIN (mmio regs + 0x1000000) */ /* resource 6 is bios */ /* map the mmio regs */ ret = drm_addmap(dev, drm_get_resource_start(dev, 0), drm_get_resource_len(dev, 0), _DRM_REGISTERS, _DRM_READ_ONLY, &dev_priv->mmio); if (ret) { DRM_ERROR("Unable to initialize the mmio mapping (%d). " "Please report your setup to " DRIVER_EMAIL "\n", ret); return -EINVAL; } DRM_DEBUG("regs mapped ok at 0x%lx\n", dev_priv->mmio->offset); /* map larger RAMIN aperture on NV40 cards */ dev_priv->ramin = NULL; if (dev_priv->card_type >= NV_40) { int ramin_resource = 2; if (drm_get_resource_len(dev, ramin_resource) == 0) ramin_resource = 3; ret = drm_addmap(dev, drm_get_resource_start(dev, ramin_resource), drm_get_resource_len(dev, ramin_resource), _DRM_REGISTERS, _DRM_READ_ONLY, &dev_priv->ramin); if (ret) { DRM_ERROR("Failed to init RAMIN mapping, " "limited instance memory available\n"); dev_priv->ramin = NULL; } } /* On older cards (or if the above failed), create a map covering * the BAR0 PRAMIN aperture */ if (!dev_priv->ramin) { ret = drm_addmap(dev, drm_get_resource_start(dev, 0) + NV_RAMIN, (1*1024*1024), _DRM_REGISTERS, _DRM_READ_ONLY, &dev_priv->ramin); if (ret) { DRM_ERROR("Failed to map BAR0 PRAMIN: %d\n", ret); return ret; } } return 0; } static int nouveau_stub_init(struct drm_device *dev) { return 0; } static void nouveau_stub_takedown(struct drm_device *dev) {} static int nouveau_init_engine_ptrs(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_engine *engine = &dev_priv->Engine; switch (dev_priv->chipset & 0xf0) { case 0x00: engine->instmem.init = nv04_instmem_init; engine->instmem.takedown= nv04_instmem_takedown; engine->instmem.populate = nv04_instmem_populate; engine->instmem.clear = nv04_instmem_clear; engine->instmem.bind = nv04_instmem_bind; engine->instmem.unbind = nv04_instmem_unbind; engine->mc.init = nv04_mc_init; engine->mc.takedown = nv04_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nv04_fb_init; engine->fb.takedown = nv04_fb_takedown; engine->graph.init = nv04_graph_init; engine->graph.takedown = nv04_graph_takedown; engine->graph.create_context = nv04_graph_create_context; engine->graph.destroy_context = nv04_graph_destroy_context; engine->graph.load_context = nv04_graph_load_context; engine->graph.save_context = nv04_graph_save_context; engine->fifo.channels = 16; engine->fifo.init = nouveau_fifo_init; engine->fifo.takedown = nouveau_stub_takedown; engine->fifo.channel_id = nv04_fifo_channel_id; engine->fifo.create_context = nv04_fifo_create_context; engine->fifo.destroy_context = nv04_fifo_destroy_context; engine->fifo.load_context = nv04_fifo_load_context; engine->fifo.save_context = nv04_fifo_save_context; break; case 0x10: engine->instmem.init = nv04_instmem_init; engine->instmem.takedown= nv04_instmem_takedown; engine->instmem.populate = nv04_instmem_populate; engine->instmem.clear = nv04_instmem_clear; engine->instmem.bind = nv04_instmem_bind; engine->instmem.unbind = nv04_instmem_unbind; engine->mc.init = nv04_mc_init; engine->mc.takedown = nv04_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nv10_fb_init; engine->fb.takedown = nv10_fb_takedown; engine->graph.init = nv10_graph_init; engine->graph.takedown = nv10_graph_takedown; engine->graph.create_context = nv10_graph_create_context; engine->graph.destroy_context = nv10_graph_destroy_context; engine->graph.load_context = nv10_graph_load_context; engine->graph.save_context = nv10_graph_save_context; engine->fifo.channels = 32; engine->fifo.init = nouveau_fifo_init; engine->fifo.takedown = nouveau_stub_takedown; engine->fifo.channel_id = nv10_fifo_channel_id; engine->fifo.create_context = nv10_fifo_create_context; engine->fifo.destroy_context = nv10_fifo_destroy_context; engine->fifo.load_context = nv10_fifo_load_context; engine->fifo.save_context = nv10_fifo_save_context; break; case 0x20: engine->instmem.init = nv04_instmem_init; engine->instmem.takedown= nv04_instmem_takedown; engine->instmem.populate = nv04_instmem_populate; engine->instmem.clear = nv04_instmem_clear; engine->instmem.bind = nv04_instmem_bind; engine->instmem.unbind = nv04_instmem_unbind; engine->mc.init = nv04_mc_init; engine->mc.takedown = nv04_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nv10_fb_init; engine->fb.takedown = nv10_fb_takedown; engine->graph.init = nv20_graph_init; engine->graph.takedown = nv20_graph_takedown; engine->graph.create_context = nv20_graph_create_context; engine->graph.destroy_context = nv20_graph_destroy_context; engine->graph.load_context = nv20_graph_load_context; engine->graph.save_context = nv20_graph_save_context; engine->fifo.channels = 32; engine->fifo.init = nouveau_fifo_init; engine->fifo.takedown = nouveau_stub_takedown; engine->fifo.channel_id = nv10_fifo_channel_id; engine->fifo.create_context = nv10_fifo_create_context; engine->fifo.destroy_context = nv10_fifo_destroy_context; engine->fifo.load_context = nv10_fifo_load_context; engine->fifo.save_context = nv10_fifo_save_context; break; case 0x30: engine->instmem.init = nv04_instmem_init; engine->instmem.takedown= nv04_instmem_takedown; engine->instmem.populate = nv04_instmem_populate; engine->instmem.clear = nv04_instmem_clear; engine->instmem.bind = nv04_instmem_bind; engine->instmem.unbind = nv04_instmem_unbind; engine->mc.init = nv04_mc_init; engine->mc.takedown = nv04_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nv10_fb_init; engine->fb.takedown = nv10_fb_takedown; engine->graph.init = nv30_graph_init; engine->graph.takedown = nv20_graph_takedown; engine->graph.create_context = nv20_graph_create_context; engine->graph.destroy_context = nv20_graph_destroy_context; engine->graph.load_context = nv20_graph_load_context; engine->graph.save_context = nv20_graph_save_context; engine->fifo.channels = 32; engine->fifo.init = nouveau_fifo_init; engine->fifo.takedown = nouveau_stub_takedown; engine->fifo.channel_id = nv10_fifo_channel_id; engine->fifo.create_context = nv10_fifo_create_context; engine->fifo.destroy_context = nv10_fifo_destroy_context; engine->fifo.load_context = nv10_fifo_load_context; engine->fifo.save_context = nv10_fifo_save_context; break; case 0x40: case 0x60: engine->instmem.init = nv04_instmem_init; engine->instmem.takedown= nv04_instmem_takedown; engine->instmem.populate = nv04_instmem_populate; engine->instmem.clear = nv04_instmem_clear; engine->instmem.bind = nv04_instmem_bind; engine->instmem.unbind = nv04_instmem_unbind; engine->mc.init = nv40_mc_init; engine->mc.takedown = nv40_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nv40_fb_init; engine->fb.takedown = nv40_fb_takedown; engine->graph.init = nv40_graph_init; engine->graph.takedown = nv40_graph_takedown; engine->graph.create_context = nv40_graph_create_context; engine->graph.destroy_context = nv40_graph_destroy_context; engine->graph.load_context = nv40_graph_load_context; engine->graph.save_context = nv40_graph_save_context; engine->fifo.channels = 32; engine->fifo.init = nv40_fifo_init; engine->fifo.takedown = nouveau_stub_takedown; engine->fifo.channel_id = nv10_fifo_channel_id; engine->fifo.create_context = nv40_fifo_create_context; engine->fifo.destroy_context = nv40_fifo_destroy_context; engine->fifo.load_context = nv40_fifo_load_context; engine->fifo.save_context = nv40_fifo_save_context; break; case 0x50: case 0x80: /* gotta love NVIDIA's consistency.. */ case 0x90: engine->instmem.init = nv50_instmem_init; engine->instmem.takedown= nv50_instmem_takedown; engine->instmem.populate = nv50_instmem_populate; engine->instmem.clear = nv50_instmem_clear; engine->instmem.bind = nv50_instmem_bind; engine->instmem.unbind = nv50_instmem_unbind; engine->mc.init = nv50_mc_init; engine->mc.takedown = nv50_mc_takedown; engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; engine->fb.init = nouveau_stub_init; engine->fb.takedown = nouveau_stub_takedown; engine->graph.init = nv50_graph_init; engine->graph.takedown = nv50_graph_takedown; engine->graph.create_context = nv50_graph_create_context; engine->graph.destroy_context = nv50_graph_destroy_context; engine->graph.load_context = nv50_graph_load_context; engine->graph.save_context = nv50_graph_save_context; engine->fifo.channels = 128; engine->fifo.init = nv50_fifo_init; engine->fifo.takedown = nv50_fifo_takedown; engine->fifo.channel_id = nv50_fifo_channel_id; engine->fifo.create_context = nv50_fifo_create_context; engine->fifo.destroy_context = nv50_fifo_destroy_context; engine->fifo.load_context = nv50_fifo_load_context; engine->fifo.save_context = nv50_fifo_save_context; break; default: DRM_ERROR("NV%02x unsupported\n", dev_priv->chipset); return 1; } return 0; } int nouveau_card_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_engine *engine; int ret; DRM_DEBUG("prev state = %d\n", dev_priv->init_state); if (dev_priv->init_state == NOUVEAU_CARD_INIT_DONE) return 0; dev_priv->ttm = 0; /* Determine exact chipset we're running on */ if (dev_priv->card_type < NV_10) dev_priv->chipset = dev_priv->card_type; else dev_priv->chipset = (NV_READ(NV03_PMC_BOOT_0) & 0x0ff00000) >> 20; /* Initialise internal driver API hooks */ ret = nouveau_init_engine_ptrs(dev); if (ret) return ret; engine = &dev_priv->Engine; dev_priv->init_state = NOUVEAU_CARD_INIT_FAILED; ret = nouveau_gpuobj_early_init(dev); if (ret) return ret; /* Initialise instance memory, must happen before mem_init so we * know exactly how much VRAM we're able to use for "normal" * purposes. */ ret = engine->instmem.init(dev); if (ret) return ret; /* Setup the memory manager */ if (dev_priv->ttm) { ret = nouveau_mem_init_ttm(dev); if (ret) return ret; } else { ret = nouveau_mem_init(dev); if (ret) return ret; } ret = nouveau_gpuobj_init(dev); if (ret) return ret; /* Parse BIOS tables / Run init tables? */ /* PMC */ ret = engine->mc.init(dev); if (ret) return ret; /* PTIMER */ ret = engine->timer.init(dev); if (ret) return ret; /* PFB */ ret = engine->fb.init(dev); if (ret) return ret; /* PGRAPH */ ret = engine->graph.init(dev); if (ret) return ret; /* PFIFO */ ret = engine->fifo.init(dev); if (ret) return ret; /* this call irq_preinstall, register irq handler and * call irq_postinstall */ ret = drm_irq_install(dev); if (ret) return ret; /* what about PVIDEO/PCRTC/PRAMDAC etc? */ ret = nouveau_dma_channel_init(dev); if (ret) return ret; dev_priv->init_state = NOUVEAU_CARD_INIT_DONE; return 0; } static void nouveau_card_takedown(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_engine *engine = &dev_priv->Engine; DRM_DEBUG("prev state = %d\n", dev_priv->init_state); if (dev_priv->init_state != NOUVEAU_CARD_INIT_DOWN) { nouveau_dma_channel_takedown(dev); engine->fifo.takedown(dev); engine->graph.takedown(dev); engine->fb.takedown(dev); engine->timer.takedown(dev); engine->mc.takedown(dev); nouveau_sgdma_nottm_hack_takedown(dev); nouveau_sgdma_takedown(dev); nouveau_gpuobj_takedown(dev); nouveau_gpuobj_del(dev, &dev_priv->vm_vram_pt); nouveau_mem_close(dev); engine->instmem.takedown(dev); drm_irq_uninstall(dev); nouveau_gpuobj_late_takedown(dev); dev_priv->init_state = NOUVEAU_CARD_INIT_DOWN; } } /* here a client dies, release the stuff that was allocated for its * file_priv */ void nouveau_preclose(struct drm_device *dev, struct drm_file *file_priv) { struct drm_nouveau_private *dev_priv = dev->dev_private; nouveau_fifo_cleanup(dev, file_priv); nouveau_mem_release(file_priv,dev_priv->fb_heap); nouveau_mem_release(file_priv,dev_priv->agp_heap); nouveau_mem_release(file_priv,dev_priv->pci_heap); } /* first module load, setup the mmio/fb mapping */ int nouveau_firstopen(struct drm_device *dev) { #if defined(__powerpc__) struct drm_nouveau_private *dev_priv = dev->dev_private; struct device_node *dn; #endif int ret; /* Map any PCI resources we need on the card */ ret = nouveau_init_card_mappings(dev); if (ret) return ret; #if defined(__powerpc__) /* Put the card in BE mode if it's not */ if (NV_READ(NV03_PMC_BOOT_1)) NV_WRITE(NV03_PMC_BOOT_1,0x00000001); DRM_MEMORYBARRIER(); #endif #if defined(__linux__) && defined(__powerpc__) /* if we have an OF card, copy vbios to RAMIN */ dn = pci_device_to_OF_node(dev->pdev); if (dn) { int size; #if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,22)) const uint32_t *bios = of_get_property(dn, "NVDA,BMP", &size); #else const uint32_t *bios = get_property(dn, "NVDA,BMP", &size); #endif if (bios) { int i; for(i=0;iflags = flags & NOUVEAU_FLAGS; dev_priv->init_state = NOUVEAU_CARD_INIT_DOWN; DRM_DEBUG("vendor: 0x%X device: 0x%X class: 0x%X\n", dev->pci_vendor, dev->pci_device, dev->pdev->class); /* Time to determine the card architecture */ regs = ioremap_nocache(pci_resource_start(dev->pdev, 0), 0x8); if (!regs) { DRM_ERROR("Could not ioremap to determine register\n"); return -ENOMEM; } reg0 = readl(regs+NV03_PMC_BOOT_0); reg1 = readl(regs+NV03_PMC_BOOT_1); #if defined(__powerpc__) if (reg1) reg0=___swab32(reg0); #endif /* We're dealing with >=NV10 */ if ((reg0 & 0x0f000000) > 0 ) { /* Bit 27-20 contain the architecture in hex */ architecture = (reg0 & 0xff00000) >> 20; /* NV04 or NV05 */ } else if ((reg0 & 0xff00fff0) == 0x20004000) { architecture = 0x04; } iounmap(regs); if (architecture >= 0x80) { dev_priv->card_type = NV_50; } else if (architecture >= 0x60) { /* FIXME we need to figure out who's who for NV6x */ dev_priv->card_type = NV_44; } else if (architecture >= 0x50) { dev_priv->card_type = NV_50; } else if (architecture >= 0x40) { uint8_t subarch = architecture & 0xf; /* Selection criteria borrowed from NV40EXA */ if (NV40_CHIPSET_MASK & (1 << subarch)) { dev_priv->card_type = NV_40; } else if (NV44_CHIPSET_MASK & (1 << subarch)) { dev_priv->card_type = NV_44; } else { dev_priv->card_type = NV_UNKNOWN; } } else if (architecture >= 0x30) { dev_priv->card_type = NV_30; } else if (architecture >= 0x20) { dev_priv->card_type = NV_20; } else if (architecture >= 0x17) { dev_priv->card_type = NV_17; } else if (architecture >= 0x11) { dev_priv->card_type = NV_11; } else if (architecture >= 0x10) { dev_priv->card_type = NV_10; } else if (architecture >= 0x04) { dev_priv->card_type = NV_04; } else { dev_priv->card_type = NV_UNKNOWN; } DRM_INFO("Detected an NV%d generation card (0x%08x)\n", dev_priv->card_type,reg0); if (dev_priv->card_type == NV_UNKNOWN) { return -EINVAL; } /* Special flags */ if (dev->pci_device == 0x01a0) { dev_priv->flags |= NV_NFORCE; } else if (dev->pci_device == 0x01f0) { dev_priv->flags |= NV_NFORCE2; } dev->dev_private = (void *)dev_priv; return 0; } void nouveau_lastclose(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; /* In the case of an error dev_priv may not be be allocated yet */ if (dev_priv && dev_priv->card_type) { nouveau_card_takedown(dev); if(dev_priv->fb_mtrr>0) { drm_mtrr_del(dev_priv->fb_mtrr, drm_get_resource_start(dev, 1),nouveau_mem_fb_amount(dev), DRM_MTRR_WC); dev_priv->fb_mtrr=0; } } } int nouveau_unload(struct drm_device *dev) { drm_free(dev->dev_private, sizeof(*dev->dev_private), DRM_MEM_DRIVER); dev->dev_private = NULL; return 0; } int nouveau_ioctl_card_init(struct drm_device *dev, void *data, struct drm_file *file_priv) { return nouveau_card_init(dev); } int nouveau_ioctl_getparam(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct drm_nouveau_getparam *getparam = data; NOUVEAU_CHECK_INITIALISED_WITH_RETURN; switch (getparam->param) { case NOUVEAU_GETPARAM_CHIPSET_ID: getparam->value = dev_priv->chipset; break; case NOUVEAU_GETPARAM_PCI_VENDOR: getparam->value=dev->pci_vendor; break; case NOUVEAU_GETPARAM_PCI_DEVICE: getparam->value=dev->pci_device; break; case NOUVEAU_GETPARAM_BUS_TYPE: if (drm_device_is_agp(dev)) getparam->value=NV_AGP; else if (drm_device_is_pcie(dev)) getparam->value=NV_PCIE; else getparam->value=NV_PCI; break; case NOUVEAU_GETPARAM_FB_PHYSICAL: getparam->value=dev_priv->fb_phys; break; case NOUVEAU_GETPARAM_AGP_PHYSICAL: getparam->value=dev_priv->gart_info.aper_base; break; case NOUVEAU_GETPARAM_PCI_PHYSICAL: if ( dev -> sg ) getparam->value=(unsigned long)dev->sg->virtual; else { DRM_ERROR("Requested PCIGART address, while no PCIGART was created\n"); return -EINVAL; } break; case NOUVEAU_GETPARAM_FB_SIZE: getparam->value=dev_priv->fb_available_size; break; case NOUVEAU_GETPARAM_AGP_SIZE: getparam->value=dev_priv->gart_info.aper_size; break; default: DRM_ERROR("unknown parameter %lld\n", getparam->param); return -EINVAL; } return 0; } int nouveau_ioctl_setparam(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct drm_nouveau_setparam *setparam = data; NOUVEAU_CHECK_INITIALISED_WITH_RETURN; switch (setparam->param) { case NOUVEAU_SETPARAM_CMDBUF_LOCATION: switch (setparam->value) { case NOUVEAU_MEM_AGP: case NOUVEAU_MEM_FB: case NOUVEAU_MEM_PCI: case NOUVEAU_MEM_AGP | NOUVEAU_MEM_PCI_ACCEPTABLE: break; default: DRM_ERROR("invalid CMDBUF_LOCATION value=%lld\n", setparam->value); return -EINVAL; } dev_priv->config.cmdbuf.location = setparam->value; break; case NOUVEAU_SETPARAM_CMDBUF_SIZE: dev_priv->config.cmdbuf.size = setparam->value; break; default: DRM_ERROR("unknown parameter %lld\n", setparam->param); return -EINVAL; } return 0; } /* waits for idle */ void nouveau_wait_for_idle(struct drm_device *dev) { struct drm_nouveau_private *dev_priv=dev->dev_private; switch(dev_priv->card_type) { case NV_50: break; default: { /* This stuff is more or less a copy of what is seen * in nv28 kmmio dump. */ uint64_t started = dev_priv->Engine.timer.read(dev); uint64_t stopped = started; uint32_t status; do { uint32_t pmc_e = NV_READ(NV03_PMC_ENABLE); (void)pmc_e; status = NV_READ(NV04_PGRAPH_STATUS); if (!status) break; stopped = dev_priv->Engine.timer.read(dev); /* It'll never wrap anyway... */ } while (stopped - started < 1000000000ULL); if (status) DRM_ERROR("timed out with status 0x%08x\n", status); } } } 561'>561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 
/* r128_cce.c -- ATI Rage 128 driver -*- linux-c -*-
 * Created: Wed Apr  5 19:24:19 2000 by kevin@precisioninsight.com
 */
/*
 * Copyright 2000 Precision Insight, Inc., Cedar Park, Texas.
 * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
 * All Rights Reserved.
 *
 * 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 (including the next
 * paragraph) 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
 * PRECISION INSIGHT AND/OR ITS SUPPLIERS 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:
 *    Gareth Hughes <gareth@valinux.com>
 */

#include "drmP.h"
#include "drm.h"
#include "r128_drm.h"
#include "r128_drv.h"

#define R128_FIFO_DEBUG		0

/* CCE microcode (from ATI) */
static u32 r128_cce_microcode[] = {
	0, 276838400, 0, 268449792, 2, 142, 2, 145, 0, 1076765731, 0,
	1617039951, 0, 774592877, 0, 1987540286, 0, 2307490946U, 0,
	599558925, 0, 589505315, 0, 596487092, 0, 589505315, 1,
	11544576, 1, 206848, 1, 311296, 1, 198656, 2, 912273422, 11,
	262144, 0, 0, 1, 33559837, 1, 7438, 1, 14809, 1, 6615, 12, 28,
	1, 6614, 12, 28, 2, 23, 11, 18874368, 0, 16790922, 1, 409600, 9,
	30, 1, 147854772, 16, 420483072, 3, 8192, 0, 10240, 1, 198656,
	1, 15630, 1, 51200, 10, 34858, 9, 42, 1, 33559823, 2, 10276, 1,
	15717, 1, 15718, 2, 43, 1, 15936948, 1, 570480831, 1, 14715071,
	12, 322123831, 1, 33953125, 12, 55, 1, 33559908, 1, 15718, 2,
	46, 4, 2099258, 1, 526336, 1, 442623, 4, 4194365, 1, 509952, 1,
	459007, 3, 0, 12, 92, 2, 46, 12, 176, 1, 15734, 1, 206848, 1,
	18432, 1, 133120, 1, 100670734, 1, 149504, 1, 165888, 1,
	15975928, 1, 1048576, 6, 3145806, 1, 15715, 16, 2150645232U, 2,
	268449859, 2, 10307, 12, 176, 1, 15734, 1, 15735, 1, 15630, 1,
	15631, 1, 5253120, 6, 3145810, 16, 2150645232U, 1, 15864, 2, 82,
	1, 343310, 1, 1064207, 2, 3145813, 1, 15728, 1, 7817, 1, 15729,
	3, 15730, 12, 92, 2, 98, 1, 16168, 1, 16167, 1, 16002, 1, 16008,
	1, 15974, 1, 15975, 1, 15990, 1, 15976, 1, 15977, 1, 15980, 0,
	15981, 1, 10240, 1, 5253120, 1, 15720, 1, 198656, 6, 110, 1,
	180224, 1, 103824738, 2, 112, 2, 3145839, 0, 536885440, 1,
	114880, 14, 125, 12, 206975, 1, 33559995, 12, 198784, 0,
	33570236, 1, 15803, 0, 15804, 3, 294912, 1, 294912, 3, 442370,
	1, 11544576, 0, 811612160, 1, 12593152, 1, 11536384, 1,
	14024704, 7, 310382726, 0, 10240, 1, 14796, 1, 14797, 1, 14793,
	1, 14794, 0, 14795, 1, 268679168, 1, 9437184, 1, 268449792, 1,
	198656, 1, 9452827, 1, 1075854602, 1, 1075854603, 1, 557056, 1,
	114880, 14, 159, 12, 198784, 1, 1109409213, 12, 198783, 1,
	1107312059, 12, 198784, 1, 1109409212, 2, 162, 1, 1075854781, 1,
	1073757627, 1, 1075854780, 1, 540672, 1, 10485760, 6, 3145894,
	16, 274741248, 9, 168, 3, 4194304, 3, 4209949, 0, 0, 0, 256, 14,
	174, 1, 114857, 1, 33560007, 12, 176, 0, 10240, 1, 114858, 1,
	33560018, 1, 114857, 3, 33560007, 1, 16008, 1, 114874, 1,
	33560360, 1, 114875, 1, 33560154, 0, 15963, 0, 256, 0, 4096, 1,
	409611, 9, 188, 0, 10240, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
};

static int R128_READ_PLL(struct drm_device * dev, int addr)
{
	drm_r128_private_t *dev_priv = dev->dev_private;

	R128_WRITE8(R128_CLOCK_CNTL_INDEX, addr & 0x1f);
	return R128_READ(R128_CLOCK_CNTL_DATA);
}

#if R128_FIFO_DEBUG
static void r128_status(drm_r128_private_t * dev_priv)
{
	printk("GUI_STAT           = 0x%08x\n",
	       (unsigned int)R128_READ(R128_GUI_STAT));
	printk("PM4_STAT           = 0x%08x\n",
	       (unsigned int)R128_READ(R128_PM4_STAT));
	printk("PM4_BUFFER_DL_WPTR = 0x%08x\n",
	       (unsigned int)R128_READ(R128_PM4_BUFFER_DL_WPTR));
	printk("PM4_BUFFER_DL_RPTR = 0x%08x\n",
	       (unsigned int)R128_READ(R128_PM4_BUFFER_DL_RPTR));
	printk("PM4_MICRO_CNTL     = 0x%08x\n",
	       (unsigned int)R128_READ(R128_PM4_MICRO_CNTL));
	printk("PM4_BUFFER_CNTL    = 0x%08x\n",
	       (unsigned int)R128_READ(R128_PM4_BUFFER_CNTL));
}
#endif

/* ================================================================
 * Engine, FIFO control
 */

static int r128_do_pixcache_flush(drm_r128_private_t * dev_priv)
{
	u32 tmp;
	int i;

	tmp = R128_READ(R128_PC_NGUI_CTLSTAT) | R128_PC_FLUSH_ALL;
	R128_WRITE(R128_PC_NGUI_CTLSTAT, tmp);

	for (i = 0; i < dev_priv->usec_timeout; i++) {
		if (!(R128_READ(R128_PC_NGUI_CTLSTAT) & R128_PC_BUSY)) {
			return 0;
		}
		DRM_UDELAY(1);
	}

#if R128_FIFO_DEBUG
	DRM_ERROR("failed!\n");
#endif
	return -EBUSY;
}

static int r128_do_wait_for_fifo(drm_r128_private_t * dev_priv, int entries)
{
	int i;

	for (i = 0; i < dev_priv->usec_timeout; i++) {
		int slots = R128_READ(R128_GUI_STAT) & R128_GUI_FIFOCNT_MASK;
		if (slots >= entries)
			return 0;
		DRM_UDELAY(1);
	}

#if R128_FIFO_DEBUG
	DRM_ERROR("failed!\n");
#endif
	return -EBUSY;
}

static int r128_do_wait_for_idle(drm_r128_private_t * dev_priv)
{
	int i, ret;

	ret = r128_do_wait_for_fifo(dev_priv, 64);
	if (ret)
		return ret;

	for (i = 0; i < dev_priv->usec_timeout; i++) {
		if (!(R128_READ(R128_GUI_STAT) & R128_GUI_ACTIVE)) {
			r128_do_pixcache_flush(dev_priv);
			return 0;
		}
		DRM_UDELAY(1);
	}

#if R128_FIFO_DEBUG
	DRM_ERROR("failed!\n");
#endif
	return -EBUSY;
}

/* ================================================================
 * CCE control, initialization
 */

/* Load the microcode for the CCE */
static void r128_cce_load_microcode(drm_r128_private_t * dev_priv)
{
	int i;

	DRM_DEBUG("\n");

	r128_do_wait_for_idle(dev_priv);

	R128_WRITE(R128_PM4_MICROCODE_ADDR, 0);
	for (i = 0; i < 256; i++) {
		R128_WRITE(R128_PM4_MICROCODE_DATAH, r128_cce_microcode[i * 2]);
		R128_WRITE(R128_PM4_MICROCODE_DATAL,
			   r128_cce_microcode[i * 2 + 1]);
	}
}

/* Flush any pending commands to the CCE.  This should only be used just
 * prior to a wait for idle, as it informs the engine that the command
 * stream is ending.
 */
static void r128_do_cce_flush(drm_r128_private_t * dev_priv)
{
	u32 tmp;

	tmp = R128_READ(R128_PM4_BUFFER_DL_WPTR) | R128_PM4_BUFFER_DL_DONE;
	R128_WRITE(R128_PM4_BUFFER_DL_WPTR, tmp);
}

/* Wait for the CCE to go idle.
 */
int r128_do_cce_idle(drm_r128_private_t * dev_priv)
{
	int i;

	for (i = 0; i < dev_priv->usec_timeout; i++) {
		if (GET_RING_HEAD(dev_priv) == dev_priv->ring.tail) {
			int pm4stat = R128_READ(R128_PM4_STAT);
			if (((pm4stat & R128_PM4_FIFOCNT_MASK) >=
			     dev_priv->cce_fifo_size) &&
			    !(pm4stat & (R128_PM4_BUSY |
					 R128_PM4_GUI_ACTIVE))) {
				return r128_do_pixcache_flush(dev_priv);
			}
		}
		DRM_UDELAY(1);
	}

#if R128_FIFO_DEBUG
	DRM_ERROR("failed!\n");
	r128_status(dev_priv);
#endif
	return -EBUSY;
}

/* Start the Concurrent Command Engine.
 */
static void r128_do_cce_start(drm_r128_private_t * dev_priv)
{
	r128_do_wait_for_idle(dev_priv);

	R128_WRITE(R128_PM4_BUFFER_CNTL,
		   dev_priv->cce_mode | dev_priv->ring.size_l2qw
		   | R128_PM4_BUFFER_CNTL_NOUPDATE);
	R128_READ(R128_PM4_BUFFER_ADDR);	/* as per the sample code */
	R128_WRITE(R128_PM4_MICRO_CNTL, R128_PM4_MICRO_FREERUN);

	dev_priv->cce_running = 1;
}

/* Reset the Concurrent Command Engine.  This will not flush any pending
 * commands, so you must wait for the CCE command stream to complete
 * before calling this routine.
 */
static void r128_do_cce_reset(drm_r128_private_t * dev_priv)
{
	R128_WRITE(R128_PM4_BUFFER_DL_WPTR, 0);
	R128_WRITE(R128_PM4_BUFFER_DL_RPTR, 0);
	dev_priv->ring.tail = 0;
}

/* Stop the Concurrent Command Engine.  This will not flush any pending
 * commands, so you must flush the command stream and wait for the CCE
 * to go idle before calling this routine.
 */
static void r128_do_cce_stop(drm_r128_private_t * dev_priv)
{
	R128_WRITE(R128_PM4_MICRO_CNTL, 0);
	R128_WRITE(R128_PM4_BUFFER_CNTL,
		   R128_PM4_NONPM4 | R128_PM4_BUFFER_CNTL_NOUPDATE);

	dev_priv->cce_running = 0;
}

/* Reset the engine.  This will stop the CCE if it is running.
 */
static int r128_do_engine_reset(struct drm_device * dev)
{
	drm_r128_private_t *dev_priv = dev->dev_private;
	u32 clock_cntl_index, mclk_cntl, gen_reset_cntl;

	r128_do_pixcache_flush(dev_priv);

	clock_cntl_index = R128_READ(R128_CLOCK_CNTL_INDEX);
	mclk_cntl = R128_READ_PLL(dev, R128_MCLK_CNTL);

	R128_WRITE_PLL(R128_MCLK_CNTL,
		       mclk_cntl | R128_FORCE_GCP | R128_FORCE_PIPE3D_CP);

	gen_reset_cntl = R128_READ(R128_GEN_RESET_CNTL);

	/* Taken from the sample code - do not change */
	R128_WRITE(R128_GEN_RESET_CNTL, gen_reset_cntl | R128_SOFT_RESET_GUI);
	R128_READ(R128_GEN_RESET_CNTL);
	R128_WRITE(R128_GEN_RESET_CNTL, gen_reset_cntl & ~R128_SOFT_RESET_GUI);
	R128_READ(R128_GEN_RESET_CNTL);

	R128_WRITE_PLL(R128_MCLK_CNTL, mclk_cntl);
	R128_WRITE(R128_CLOCK_CNTL_INDEX, clock_cntl_index);
	R128_WRITE(R128_GEN_RESET_CNTL, gen_reset_cntl);

	/* Reset the CCE ring */
	r128_do_cce_reset(dev_priv);

	/* The CCE is no longer running after an engine reset */
	dev_priv->cce_running = 0;

	/* Reset any pending vertex, indirect buffers */
	r128_freelist_reset(dev);

	return 0;
}

static void r128_cce_init_ring_buffer(struct drm_device * dev,
				      drm_r128_private_t * dev_priv)
{
	u32 ring_start;
	u32 tmp;

	DRM_DEBUG("\n");

	/* The manual (p. 2) says this address is in "VM space".  This
	 * means it's an offset from the start of AGP space.
	 */
#if __OS_HAS_AGP
	if (!dev_priv->is_pci)
		ring_start = dev_priv->cce_ring->offset - dev->agp->base;
	else
#endif
		ring_start = dev_priv->cce_ring->offset -
				(unsigned long)dev->sg->virtual;

	R128_WRITE(R128_PM4_BUFFER_OFFSET, ring_start | R128_AGP_OFFSET);

	R128_WRITE(R128_PM4_BUFFER_DL_WPTR, 0);
	R128_WRITE(R128_PM4_BUFFER_DL_RPTR, 0);

	/* Set watermark control */
	R128_WRITE(R128_PM4_BUFFER_WM_CNTL,
		   ((R128_WATERMARK_L / 4) << R128_WMA_SHIFT)
		   | ((R128_WATERMARK_M / 4) << R128_WMB_SHIFT)
		   | ((R128_WATERMARK_N / 4) << R128_WMC_SHIFT)
		   | ((R128_WATERMARK_K / 64) << R128_WB_WM_SHIFT));

	/* Force read.  Why?  Because it's in the examples... */
	R128_READ(R128_PM4_BUFFER_ADDR);

	/* Turn on bus mastering */
	tmp = R128_READ(R128_BUS_CNTL) & ~R128_BUS_MASTER_DIS;
	R128_WRITE(R128_BUS_CNTL, tmp);
}

static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
{
	drm_r128_private_t *dev_priv;

	DRM_DEBUG("\n");

	dev_priv = drm_alloc(sizeof(drm_r128_private_t), DRM_MEM_DRIVER);
	if (dev_priv == NULL)
		return -ENOMEM;

	memset(dev_priv, 0, sizeof(drm_r128_private_t));

	dev_priv->is_pci = init->is_pci;

	if (dev_priv->is_pci && !dev->sg) {
		DRM_ERROR("PCI GART memory not allocated!\n");
		dev->dev_private = (void *)dev_priv;
		r128_do_cleanup_cce(dev);
		return -EINVAL;
	}

	dev_priv->usec_timeout = init->usec_timeout;
	if (dev_priv->usec_timeout < 1 ||
	    dev_priv->usec_timeout > R128_MAX_USEC_TIMEOUT) {
		DRM_DEBUG("TIMEOUT problem!\n");
		dev->dev_private = (void *)dev_priv;
		r128_do_cleanup_cce(dev);
		return -EINVAL;
	}

	dev_priv->cce_mode = init->cce_mode;

	/* GH: Simple idle check.
	 */
	atomic_set(&dev_priv->idle_count, 0);

	/* We don't support anything other than bus-mastering ring mode,
	 * but the ring can be in either AGP or PCI space for the ring
	 * read pointer.
	 */
	if ((init->cce_mode != R128_PM4_192BM) &&
	    (init->cce_mode != R128_PM4_128BM_64INDBM) &&