diff options
Diffstat (limited to 'drivers/gpu/drm/radeon/si.c')
| -rw-r--r-- | drivers/gpu/drm/radeon/si.c | 4114 |
1 files changed, 2335 insertions, 1779 deletions
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 234906709067..26197aceb001 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -21,65 +21,146 @@ * * Authors: Alex Deucher */ + #include <linux/firmware.h> -#include <linux/platform_device.h> -#include <linux/slab.h> #include <linux/module.h> -#include <drm/drmP.h> -#include "radeon.h" -#include "radeon_asic.h" +#include <linux/pci.h> +#include <linux/slab.h> + +#include <drm/drm_vblank.h> #include <drm/radeon_drm.h> -#include "sid.h" + #include "atom.h" -#include "si_blit_shaders.h" #include "clearstate_si.h" +#include "evergreen.h" +#include "r600.h" +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_audio.h" #include "radeon_ucode.h" +#include "si_blit_shaders.h" +#include "si.h" +#include "sid.h" MODULE_FIRMWARE("radeon/TAHITI_pfp.bin"); MODULE_FIRMWARE("radeon/TAHITI_me.bin"); MODULE_FIRMWARE("radeon/TAHITI_ce.bin"); MODULE_FIRMWARE("radeon/TAHITI_mc.bin"); +MODULE_FIRMWARE("radeon/TAHITI_mc2.bin"); MODULE_FIRMWARE("radeon/TAHITI_rlc.bin"); MODULE_FIRMWARE("radeon/TAHITI_smc.bin"); + +MODULE_FIRMWARE("radeon/tahiti_pfp.bin"); +MODULE_FIRMWARE("radeon/tahiti_me.bin"); +MODULE_FIRMWARE("radeon/tahiti_ce.bin"); +MODULE_FIRMWARE("radeon/tahiti_mc.bin"); +MODULE_FIRMWARE("radeon/tahiti_rlc.bin"); +MODULE_FIRMWARE("radeon/tahiti_smc.bin"); + MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin"); + +MODULE_FIRMWARE("radeon/pitcairn_pfp.bin"); +MODULE_FIRMWARE("radeon/pitcairn_me.bin"); +MODULE_FIRMWARE("radeon/pitcairn_ce.bin"); +MODULE_FIRMWARE("radeon/pitcairn_mc.bin"); +MODULE_FIRMWARE("radeon/pitcairn_rlc.bin"); +MODULE_FIRMWARE("radeon/pitcairn_smc.bin"); +MODULE_FIRMWARE("radeon/pitcairn_k_smc.bin"); + MODULE_FIRMWARE("radeon/VERDE_pfp.bin"); MODULE_FIRMWARE("radeon/VERDE_me.bin"); MODULE_FIRMWARE("radeon/VERDE_ce.bin"); MODULE_FIRMWARE("radeon/VERDE_mc.bin"); +MODULE_FIRMWARE("radeon/VERDE_mc2.bin"); MODULE_FIRMWARE("radeon/VERDE_rlc.bin"); MODULE_FIRMWARE("radeon/VERDE_smc.bin"); + +MODULE_FIRMWARE("radeon/verde_pfp.bin"); +MODULE_FIRMWARE("radeon/verde_me.bin"); +MODULE_FIRMWARE("radeon/verde_ce.bin"); +MODULE_FIRMWARE("radeon/verde_mc.bin"); +MODULE_FIRMWARE("radeon/verde_rlc.bin"); +MODULE_FIRMWARE("radeon/verde_smc.bin"); +MODULE_FIRMWARE("radeon/verde_k_smc.bin"); + MODULE_FIRMWARE("radeon/OLAND_pfp.bin"); MODULE_FIRMWARE("radeon/OLAND_me.bin"); MODULE_FIRMWARE("radeon/OLAND_ce.bin"); MODULE_FIRMWARE("radeon/OLAND_mc.bin"); +MODULE_FIRMWARE("radeon/OLAND_mc2.bin"); MODULE_FIRMWARE("radeon/OLAND_rlc.bin"); MODULE_FIRMWARE("radeon/OLAND_smc.bin"); + +MODULE_FIRMWARE("radeon/oland_pfp.bin"); +MODULE_FIRMWARE("radeon/oland_me.bin"); +MODULE_FIRMWARE("radeon/oland_ce.bin"); +MODULE_FIRMWARE("radeon/oland_mc.bin"); +MODULE_FIRMWARE("radeon/oland_rlc.bin"); +MODULE_FIRMWARE("radeon/oland_smc.bin"); +MODULE_FIRMWARE("radeon/oland_k_smc.bin"); + MODULE_FIRMWARE("radeon/HAINAN_pfp.bin"); MODULE_FIRMWARE("radeon/HAINAN_me.bin"); MODULE_FIRMWARE("radeon/HAINAN_ce.bin"); MODULE_FIRMWARE("radeon/HAINAN_mc.bin"); +MODULE_FIRMWARE("radeon/HAINAN_mc2.bin"); MODULE_FIRMWARE("radeon/HAINAN_rlc.bin"); MODULE_FIRMWARE("radeon/HAINAN_smc.bin"); +MODULE_FIRMWARE("radeon/hainan_pfp.bin"); +MODULE_FIRMWARE("radeon/hainan_me.bin"); +MODULE_FIRMWARE("radeon/hainan_ce.bin"); +MODULE_FIRMWARE("radeon/hainan_mc.bin"); +MODULE_FIRMWARE("radeon/hainan_rlc.bin"); +MODULE_FIRMWARE("radeon/hainan_smc.bin"); +MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/banks_k_2_smc.bin"); + +MODULE_FIRMWARE("radeon/si58_mc.bin"); + +static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); static void si_pcie_gen3_enable(struct radeon_device *rdev); static void si_program_aspm(struct radeon_device *rdev); -extern int r600_ih_ring_alloc(struct radeon_device *rdev); -extern void r600_ih_ring_fini(struct radeon_device *rdev); -extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev); -extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save); -extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save); -extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev); -extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev); -extern bool evergreen_is_display_hung(struct radeon_device *rdev); - -static const u32 verde_rlc_save_restore_register_list[] = -{ +extern void sumo_rlc_fini(struct radeon_device *rdev); +extern int sumo_rlc_init(struct radeon_device *rdev); +static void si_enable_gui_idle_interrupt(struct radeon_device *rdev, + bool enable); +static void si_init_pg(struct radeon_device *rdev); +static void si_init_cg(struct radeon_device *rdev); +static void si_fini_pg(struct radeon_device *rdev); +static void si_fini_cg(struct radeon_device *rdev); +static void si_rlc_stop(struct radeon_device *rdev); + +static const u32 crtc_offsets[] = { + EVERGREEN_CRTC0_REGISTER_OFFSET, + EVERGREEN_CRTC1_REGISTER_OFFSET, + EVERGREEN_CRTC2_REGISTER_OFFSET, + EVERGREEN_CRTC3_REGISTER_OFFSET, + EVERGREEN_CRTC4_REGISTER_OFFSET, + EVERGREEN_CRTC5_REGISTER_OFFSET +}; + +static const u32 si_disp_int_status[] = { + DISP_INTERRUPT_STATUS, + DISP_INTERRUPT_STATUS_CONTINUE, + DISP_INTERRUPT_STATUS_CONTINUE2, + DISP_INTERRUPT_STATUS_CONTINUE3, + DISP_INTERRUPT_STATUS_CONTINUE4, + DISP_INTERRUPT_STATUS_CONTINUE5 +}; + +#define DC_HPDx_CONTROL(x) (DC_HPD1_CONTROL + (x * 0xc)) +#define DC_HPDx_INT_CONTROL(x) (DC_HPD1_INT_CONTROL + (x * 0xc)) +#define DC_HPDx_INT_STATUS_REG(x) (DC_HPD1_INT_STATUS + (x * 0xc)) + +static const u32 verde_rlc_save_restore_register_list[] = { (0x8000 << 16) | (0x98f4 >> 2), 0x00000000, (0x8040 << 16) | (0x98f4 >> 2), @@ -300,8 +381,7 @@ static const u32 verde_rlc_save_restore_register_list[] = 0x00000000 }; -static const u32 tahiti_golden_rlc_registers[] = -{ +static const u32 tahiti_golden_rlc_registers[] = { 0xc424, 0xffffffff, 0x00601005, 0xc47c, 0xffffffff, 0x10104040, 0xc488, 0xffffffff, 0x0100000a, @@ -310,8 +390,7 @@ static const u32 tahiti_golden_rlc_registers[] = 0xf4a8, 0xffffffff, 0x00000000 }; -static const u32 tahiti_golden_registers[] = -{ +static const u32 tahiti_golden_registers[] = { 0x9a10, 0x00010000, 0x00018208, 0x9830, 0xffffffff, 0x00000000, 0x9834, 0xf00fffff, 0x00000400, @@ -345,13 +424,11 @@ static const u32 tahiti_golden_registers[] = 0x15c0, 0x000c0fc0, 0x000c0400 }; -static const u32 tahiti_golden_registers2[] = -{ +static const u32 tahiti_golden_registers2[] = { 0xc64, 0x00000001, 0x00000001 }; -static const u32 pitcairn_golden_rlc_registers[] = -{ +static const u32 pitcairn_golden_rlc_registers[] = { 0xc424, 0xffffffff, 0x00601004, 0xc47c, 0xffffffff, 0x10102020, 0xc488, 0xffffffff, 0x01000020, @@ -359,8 +436,7 @@ static const u32 pitcairn_golden_rlc_registers[] = 0xc30c, 0xffffffff, 0x800000a4 }; -static const u32 pitcairn_golden_registers[] = -{ +static const u32 pitcairn_golden_registers[] = { 0x9a10, 0x00010000, 0x00018208, 0x9830, 0xffffffff, 0x00000000, 0x9834, 0xf00fffff, 0x00000400, @@ -390,8 +466,7 @@ static const u32 pitcairn_golden_registers[] = 0x15c0, 0x000c0fc0, 0x000c0400 }; -static const u32 verde_golden_rlc_registers[] = -{ +static const u32 verde_golden_rlc_registers[] = { 0xc424, 0xffffffff, 0x033f1005, 0xc47c, 0xffffffff, 0x10808020, 0xc488, 0xffffffff, 0x00800008, @@ -399,8 +474,7 @@ static const u32 verde_golden_rlc_registers[] = 0xc30c, 0xffffffff, 0x80010014 }; -static const u32 verde_golden_registers[] = -{ +static const u32 verde_golden_registers[] = { 0x9a10, 0x00010000, 0x00018208, 0x9830, 0xffffffff, 0x00000000, 0x9834, 0xf00fffff, 0x00000400, @@ -455,8 +529,7 @@ static const u32 verde_golden_registers[] = 0x15c0, 0x000c0fc0, 0x000c0400 }; -static const u32 oland_golden_rlc_registers[] = -{ +static const u32 oland_golden_rlc_registers[] = { 0xc424, 0xffffffff, 0x00601005, 0xc47c, 0xffffffff, 0x10104040, 0xc488, 0xffffffff, 0x0100000a, @@ -464,8 +537,7 @@ static const u32 oland_golden_rlc_registers[] = 0xc30c, 0xffffffff, 0x800000f4 }; -static const u32 oland_golden_registers[] = -{ +static const u32 oland_golden_registers[] = { 0x9a10, 0x00010000, 0x00018208, 0x9830, 0xffffffff, 0x00000000, 0x9834, 0xf00fffff, 0x00000400, @@ -495,8 +567,7 @@ static const u32 oland_golden_registers[] = 0x15c0, 0x000c0fc0, 0x000c0400 }; -static const u32 hainan_golden_registers[] = -{ +static const u32 hainan_golden_registers[] = { 0x9a10, 0x00010000, 0x00018208, 0x9830, 0xffffffff, 0x00000000, 0x9834, 0xf00fffff, 0x00000400, @@ -524,13 +595,11 @@ static const u32 hainan_golden_registers[] = 0x15c0, 0x000c0fc0, 0x000c0400 }; -static const u32 hainan_golden_registers2[] = -{ +static const u32 hainan_golden_registers2[] = { 0x98f8, 0xffffffff, 0x02010001 }; -static const u32 tahiti_mgcg_cgcg_init[] = -{ +static const u32 tahiti_mgcg_cgcg_init[] = { 0xc400, 0xffffffff, 0xfffffffc, 0x802c, 0xffffffff, 0xe0000000, 0x9a60, 0xffffffff, 0x00000100, @@ -659,8 +728,7 @@ static const u32 tahiti_mgcg_cgcg_init[] = 0xd8c0, 0xfffffff0, 0x00000100 }; -static const u32 pitcairn_mgcg_cgcg_init[] = -{ +static const u32 pitcairn_mgcg_cgcg_init[] = { 0xc400, 0xffffffff, 0xfffffffc, 0x802c, 0xffffffff, 0xe0000000, 0x9a60, 0xffffffff, 0x00000100, @@ -757,8 +825,7 @@ static const u32 pitcairn_mgcg_cgcg_init[] = 0xd8c0, 0xfffffff0, 0x00000100 }; -static const u32 verde_mgcg_cgcg_init[] = -{ +static const u32 verde_mgcg_cgcg_init[] = { 0xc400, 0xffffffff, 0xfffffffc, 0x802c, 0xffffffff, 0xe0000000, 0x9a60, 0xffffffff, 0x00000100, @@ -857,8 +924,7 @@ static const u32 verde_mgcg_cgcg_init[] = 0xd8c0, 0xfffffff0, 0x00000100 }; -static const u32 oland_mgcg_cgcg_init[] = -{ +static const u32 oland_mgcg_cgcg_init[] = { 0xc400, 0xffffffff, 0xfffffffc, 0x802c, 0xffffffff, 0xe0000000, 0x9a60, 0xffffffff, 0x00000100, @@ -937,8 +1003,7 @@ static const u32 oland_mgcg_cgcg_init[] = 0xd8c0, 0xfffffff0, 0x00000100 }; -static const u32 hainan_mgcg_cgcg_init[] = -{ +static const u32 hainan_mgcg_cgcg_init[] = { 0xc400, 0xffffffff, 0xfffffffc, 0x802c, 0xffffffff, 0xe0000000, 0x9a60, 0xffffffff, 0x00000100, @@ -1014,8 +1079,7 @@ static const u32 hainan_mgcg_cgcg_init[] = 0xd8c0, 0xfffffff0, 0x00000100 }; -static u32 verde_pg_init[] = -{ +static u32 verde_pg_init[] = { 0x353c, 0xffffffff, 0x40000, 0x3538, 0xffffffff, 0x200010ff, 0x353c, 0xffffffff, 0x0, @@ -1210,6 +1274,36 @@ static void si_init_golden_registers(struct radeon_device *rdev) } } +/** + * si_get_allowed_info_register - fetch the register for the info ioctl + * + * @rdev: radeon_device pointer + * @reg: register offset in bytes + * @val: register value + * + * Returns 0 for success or -EINVAL for an invalid register + * + */ +int si_get_allowed_info_register(struct radeon_device *rdev, + u32 reg, u32 *val) +{ + switch (reg) { + case GRBM_STATUS: + case GRBM_STATUS2: + case GRBM_STATUS_SE0: + case GRBM_STATUS_SE1: + case SRBM_STATUS: + case SRBM_STATUS2: + case (DMA_STATUS_REG + DMA0_REGISTER_OFFSET): + case (DMA_STATUS_REG + DMA1_REGISTER_OFFSET): + case UVD_STATUS: + *val = RREG32(reg); + return 0; + default: + return -EINVAL; + } +} + #define PCIE_BUS_CLK 10000 #define TCLK (PCIE_BUS_CLK / 10) @@ -1223,7 +1317,7 @@ static void si_init_golden_registers(struct radeon_device *rdev) */ u32 si_get_xclk(struct radeon_device *rdev) { - u32 reference_clock = rdev->clock.spll.reference_freq; + u32 reference_clock = rdev->clock.spll.reference_freq; u32 tmp; tmp = RREG32(CG_CLKPIN_CNTL_2); @@ -1454,66 +1548,82 @@ static const u32 hainan_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = { }; /* ucode loading */ -static int si_mc_load_microcode(struct radeon_device *rdev) +int si_mc_load_microcode(struct radeon_device *rdev) { - const __be32 *fw_data; - u32 running, blackout = 0; - u32 *io_mc_regs; - int i, ucode_size, regs_size; + const __be32 *fw_data = NULL; + const __le32 *new_fw_data = NULL; + u32 running; + u32 *io_mc_regs = NULL; + const __le32 *new_io_mc_regs = NULL; + int i, regs_size, ucode_size; if (!rdev->mc_fw) return -EINVAL; - switch (rdev->family) { - case CHIP_TAHITI: - io_mc_regs = (u32 *)&tahiti_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; - regs_size = TAHITI_IO_MC_REGS_SIZE; - break; - case CHIP_PITCAIRN: - io_mc_regs = (u32 *)&pitcairn_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; - regs_size = TAHITI_IO_MC_REGS_SIZE; - break; - case CHIP_VERDE: - default: - io_mc_regs = (u32 *)&verde_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; - regs_size = TAHITI_IO_MC_REGS_SIZE; - break; - case CHIP_OLAND: - io_mc_regs = (u32 *)&oland_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; - regs_size = TAHITI_IO_MC_REGS_SIZE; - break; - case CHIP_HAINAN: - io_mc_regs = (u32 *)&hainan_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; - regs_size = TAHITI_IO_MC_REGS_SIZE; - break; + if (rdev->new_fw) { + const struct mc_firmware_header_v1_0 *hdr = + (const struct mc_firmware_header_v1_0 *)rdev->mc_fw->data; + + radeon_ucode_print_mc_hdr(&hdr->header); + regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2); + new_io_mc_regs = (const __le32 *) + (rdev->mc_fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes)); + ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; + new_fw_data = (const __le32 *) + (rdev->mc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + } else { + ucode_size = rdev->mc_fw->size / 4; + + switch (rdev->family) { + case CHIP_TAHITI: + io_mc_regs = (u32 *)&tahiti_io_mc_regs; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_PITCAIRN: + io_mc_regs = (u32 *)&pitcairn_io_mc_regs; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_VERDE: + default: + io_mc_regs = (u32 *)&verde_io_mc_regs; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_OLAND: + io_mc_regs = (u32 *)&oland_io_mc_regs; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_HAINAN: + io_mc_regs = (u32 *)&hainan_io_mc_regs; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + } + fw_data = (const __be32 *)rdev->mc_fw->data; } running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK; if (running == 0) { - if (running) { - blackout = RREG32(MC_SHARED_BLACKOUT_CNTL); - WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1); - } - /* reset the engine and set to writable */ WREG32(MC_SEQ_SUP_CNTL, 0x00000008); WREG32(MC_SEQ_SUP_CNTL, 0x00000010); /* load mc io regs */ for (i = 0; i < regs_size; i++) { - WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]); - WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]); + if (rdev->new_fw) { + WREG32(MC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(new_io_mc_regs++)); + WREG32(MC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++)); + } else { + WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]); + WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]); + } } /* load the MC ucode */ - fw_data = (const __be32 *)rdev->mc_fw->data; - for (i = 0; i < ucode_size; i++) - WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++)); + for (i = 0; i < ucode_size; i++) { + if (rdev->new_fw) + WREG32(MC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++)); + else + WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++)); + } /* put the engine back into the active state */ WREG32(MC_SEQ_SUP_CNTL, 0x00000008); @@ -1531,9 +1641,6 @@ static int si_mc_load_microcode(struct radeon_device *rdev) break; udelay(1); } - - if (running) - WREG32(MC_SHARED_BLACKOUT_CNTL, blackout); } return 0; @@ -1541,153 +1648,284 @@ static int si_mc_load_microcode(struct radeon_device *rdev) static int si_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *chip_name; - const char *rlc_chip_name; + const char *new_chip_name; size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size; - size_t smc_req_size; + size_t smc_req_size, mc2_req_size; char fw_name[30]; int err; + int new_fw = 0; + bool new_smc = false; + bool si58_fw = false; + bool banks2_fw = false; DRM_DEBUG("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_TAHITI: chip_name = "TAHITI"; - rlc_chip_name = "TAHITI"; + new_chip_name = "tahiti"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = TAHITI_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4); break; case CHIP_PITCAIRN: chip_name = "PITCAIRN"; - rlc_chip_name = "PITCAIRN"; + if ((rdev->pdev->revision == 0x81) && + ((rdev->pdev->device == 0x6810) || + (rdev->pdev->device == 0x6811))) + new_smc = true; + new_chip_name = "pitcairn"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = PITCAIRN_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4); break; case CHIP_VERDE: chip_name = "VERDE"; - rlc_chip_name = "VERDE"; + if (((rdev->pdev->device == 0x6820) && + ((rdev->pdev->revision == 0x81) || + (rdev->pdev->revision == 0x83))) || + ((rdev->pdev->device == 0x6821) && + ((rdev->pdev->revision == 0x83) || + (rdev->pdev->revision == 0x87))) || + ((rdev->pdev->revision == 0x87) && + ((rdev->pdev->device == 0x6823) || + (rdev->pdev->device == 0x682b)))) + new_smc = true; + new_chip_name = "verde"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = VERDE_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4); break; case CHIP_OLAND: chip_name = "OLAND"; - rlc_chip_name = "OLAND"; + if (((rdev->pdev->revision == 0x81) && + ((rdev->pdev->device == 0x6600) || + (rdev->pdev->device == 0x6604) || + (rdev->pdev->device == 0x6605) || + (rdev->pdev->device == 0x6610))) || + ((rdev->pdev->revision == 0x83) && + (rdev->pdev->device == 0x6610))) + new_smc = true; + new_chip_name = "oland"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4); break; case CHIP_HAINAN: chip_name = "HAINAN"; - rlc_chip_name = "HAINAN"; + if (((rdev->pdev->revision == 0x81) && + (rdev->pdev->device == 0x6660)) || + ((rdev->pdev->revision == 0x83) && + ((rdev->pdev->device == 0x6660) || + (rdev->pdev->device == 0x6663) || + (rdev->pdev->device == 0x6665) || + (rdev->pdev->device == 0x6667)))) + new_smc = true; + else if ((rdev->pdev->revision == 0xc3) && + (rdev->pdev->device == 0x6665)) + banks2_fw = true; + new_chip_name = "hainan"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4); break; - default: BUG(); + default: + BUG(); } - DRM_INFO("Loading %s Microcode\n", chip_name); + /* this memory configuration requires special firmware */ + if (((RREG32(MC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58) + si58_fw = true; - snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); - err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->pfp_fw->size != pfp_req_size) { - printk(KERN_ERR - "si_cp: Bogus length %zu in firmware \"%s\"\n", - rdev->pfp_fw->size, fw_name); - err = -EINVAL; - goto out; + DRM_INFO("Loading %s Microcode\n", new_chip_name); + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); + if (err) + goto out; + if (rdev->pfp_fw->size != pfp_req_size) { + pr_err("si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->pfp_fw->size, fw_name); + err = -EINVAL; + goto out; + } + } else { + err = radeon_ucode_validate(rdev->pfp_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->me_fw->size != me_req_size) { - printk(KERN_ERR - "si_cp: Bogus length %zu in firmware \"%s\"\n", - rdev->me_fw->size, fw_name); - err = -EINVAL; + snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", new_chip_name); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); + if (err) + goto out; + if (rdev->me_fw->size != me_req_size) { + pr_err("si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->me_fw->size, fw_name); + err = -EINVAL; + } + } else { + err = radeon_ucode_validate(rdev->me_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name); - err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->ce_fw->size != ce_req_size) { - printk(KERN_ERR - "si_cp: Bogus length %zu in firmware \"%s\"\n", - rdev->ce_fw->size, fw_name); - err = -EINVAL; + snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", new_chip_name); + err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name); + err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev); + if (err) + goto out; + if (rdev->ce_fw->size != ce_req_size) { + pr_err("si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->ce_fw->size, fw_name); + err = -EINVAL; + } + } else { + err = radeon_ucode_validate(rdev->ce_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name); - err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->rlc_fw->size != rlc_req_size) { - printk(KERN_ERR - "si_rlc: Bogus length %zu in firmware \"%s\"\n", - rdev->rlc_fw->size, fw_name); - err = -EINVAL; + snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", new_chip_name); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); + if (err) + goto out; + if (rdev->rlc_fw->size != rlc_req_size) { + pr_err("si_rlc: Bogus length %zu in firmware \"%s\"\n", + rdev->rlc_fw->size, fw_name); + err = -EINVAL; + } + } else { + err = radeon_ucode_validate(rdev->rlc_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); - err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { - printk(KERN_ERR - "si_mc: Bogus length %zu in firmware \"%s\"\n", - rdev->mc_fw->size, fw_name); - err = -EINVAL; + if (si58_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/si58_mc.bin"); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) + goto out; + } + if ((rdev->mc_fw->size != mc_req_size) && + (rdev->mc_fw->size != mc2_req_size)) { + pr_err("si_mc: Bogus length %zu in firmware \"%s\"\n", + rdev->mc_fw->size, fw_name); + err = -EINVAL; + } + DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size); + } else { + err = radeon_ucode_validate(rdev->mc_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); - err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->smc_fw->size != smc_req_size) { - printk(KERN_ERR - "si_smc: Bogus length %zu in firmware \"%s\"\n", - rdev->smc_fw->size, fw_name); - err = -EINVAL; + if (banks2_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/banks_k_2_smc.bin"); + else if (new_smc) + snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + pr_err("smc: error loading firmware \"%s\"\n", fw_name); + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; + err = 0; + } else if (rdev->smc_fw->size != smc_req_size) { + pr_err("si_smc: Bogus length %zu in firmware \"%s\"\n", + rdev->smc_fw->size, fw_name); + err = -EINVAL; + } + } else { + err = radeon_ucode_validate(rdev->smc_fw); + if (err) { + pr_err("si_cp: validation failed for firmware \"%s\"\n", + fw_name); + goto out; + } else { + new_fw++; + } } + if (new_fw == 0) { + rdev->new_fw = false; + } else if (new_fw < 6) { + pr_err("si_fw: mixing new and old firmware!\n"); + err = -EINVAL; + } else { + rdev->new_fw = true; + } out: - platform_device_unregister(pdev); - if (err) { if (err != -EINVAL) - printk(KERN_ERR - "si_cp: Failed to load firmware \"%s\"\n", + pr_err("si_cp: Failed to load firmware \"%s\"\n", fw_name); release_firmware(rdev->pfp_fw); rdev->pfp_fw = NULL; @@ -1711,7 +1949,8 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev, struct drm_display_mode *mode, struct drm_display_mode *other_mode) { - u32 tmp; + u32 tmp, buffer_alloc, i; + u32 pipe_offset = radeon_crtc->crtc_id * 0x20; /* * Line Buffer Setup * There are 3 line buffers, each one shared by 2 display controllers. @@ -1726,16 +1965,30 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev, * non-linked crtcs for maximum line buffer allocation. */ if (radeon_crtc->base.enabled && mode) { - if (other_mode) + if (other_mode) { tmp = 0; /* 1/2 */ - else + buffer_alloc = 1; + } else { tmp = 2; /* whole */ - } else + buffer_alloc = 2; + } + } else { tmp = 0; + buffer_alloc = 0; + } WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset, DC_LB_MEMORY_CONFIG(tmp)); + WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset, + DMIF_BUFFERS_ALLOCATED(buffer_alloc)); + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) & + DMIF_BUFFERS_ALLOCATED_COMPLETED) + break; + udelay(1); + } + if (radeon_crtc->base.enabled && mode) { switch (tmp) { case 0: @@ -1955,23 +2208,10 @@ static u32 dce6_latency_watermark(struct dce6_wm_params *wm) a.full = dfixed_const(available_bandwidth); b.full = dfixed_const(wm->num_heads); a.full = dfixed_div(a, b); + tmp = div_u64((u64) dmif_size * (u64) wm->disp_clk, mc_latency + 512); + tmp = min(dfixed_trunc(a), tmp); - b.full = dfixed_const(mc_latency + 512); - c.full = dfixed_const(wm->disp_clk); - b.full = dfixed_div(b, c); - - c.full = dfixed_const(dmif_size); - b.full = dfixed_div(c, b); - - tmp = min(dfixed_trunc(a), dfixed_trunc(b)); - - b.full = dfixed_const(1000); - c.full = dfixed_const(wm->disp_clk); - b.full = dfixed_div(c, b); - c.full = dfixed_const(wm->bytes_per_pixel); - b.full = dfixed_mul(b, c); - - lb_fill_bw = min(tmp, dfixed_trunc(b)); + lb_fill_bw = min(tmp, wm->disp_clk * wm->bytes_per_pixel / 1000); a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel); b.full = dfixed_const(1000); @@ -2038,7 +2278,7 @@ static void dce6_program_watermarks(struct radeon_device *rdev, struct drm_display_mode *mode = &radeon_crtc->base.mode; struct dce6_wm_params wm_low, wm_high; u32 dram_channels; - u32 pixel_period; + u32 active_time; u32 line_time = 0; u32 latency_watermark_a = 0, latency_watermark_b = 0; u32 priority_a_mark = 0, priority_b_mark = 0; @@ -2048,8 +2288,11 @@ static void dce6_program_watermarks(struct radeon_device *rdev, fixed20_12 a, b, c; if (radeon_crtc->base.enabled && num_heads && mode) { - pixel_period = 1000000 / (u32)mode->clock; - line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535); + active_time = (u32) div_u64((u64)mode->crtc_hdisplay * 1000000, + (u32)mode->clock); + line_time = (u32) div_u64((u64)mode->crtc_htotal * 1000000, + (u32)mode->clock); + line_time = min(line_time, (u32)65535); priority_a_cnt = 0; priority_b_cnt = 0; @@ -2071,7 +2314,7 @@ static void dce6_program_watermarks(struct radeon_device *rdev, wm_high.disp_clk = mode->clock; wm_high.src_width = mode->crtc_hdisplay; - wm_high.active_time = mode->crtc_hdisplay * pixel_period; + wm_high.active_time = active_time; wm_high.blank_time = line_time - wm_high.active_time; wm_high.interlaced = false; if (mode->flags & DRM_MODE_FLAG_INTERLACE) @@ -2098,7 +2341,7 @@ static void dce6_program_watermarks(struct radeon_device *rdev, wm_low.disp_clk = mode->clock; wm_low.src_width = mode->crtc_hdisplay; - wm_low.active_time = mode->crtc_hdisplay * pixel_period; + wm_low.active_time = active_time; wm_low.blank_time = line_time - wm_low.active_time; wm_low.interlaced = false; if (mode->flags & DRM_MODE_FLAG_INTERLACE) @@ -2159,6 +2402,9 @@ static void dce6_program_watermarks(struct radeon_device *rdev, c.full = dfixed_div(c, a); priority_b_mark = dfixed_trunc(c); priority_b_cnt |= priority_b_mark & PRIORITY_MARK_MASK; + + /* Save number of lines the linebuffer leads before the scanout */ + radeon_crtc->lb_vblank_lead_lines = DIV_ROUND_UP(lb_size, mode->crtc_hdisplay); } /* select wm A */ @@ -2198,6 +2444,9 @@ void dce6_bandwidth_update(struct radeon_device *rdev) u32 num_heads = 0, lb_size; int i; + if (!rdev->mode_info.mode_config_initialized) + return; + radeon_update_display_priority(rdev); for (i = 0; i < rdev->num_crtc; i++) { @@ -2219,8 +2468,10 @@ void dce6_bandwidth_update(struct radeon_device *rdev) */ static void si_tiling_mode_table_init(struct radeon_device *rdev) { - const u32 num_tile_mode_states = 32; - u32 reg_offset, gb_tile_moden, split_equal_to_row_size; + u32 *tile = rdev->config.si.tile_mode_array; + const u32 num_tile_mode_states = + ARRAY_SIZE(rdev->config.si.tile_mode_array); + u32 reg_offset, split_equal_to_row_size; switch (rdev->config.si.mem_row_size_in_kb) { case 1: @@ -2235,491 +2486,442 @@ static void si_tiling_mode_table_init(struct radeon_device *rdev) break; } - if ((rdev->family == CHIP_TAHITI) || - (rdev->family == CHIP_PITCAIRN)) { - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: /* non-AA compressed depth or any compressed stencil */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 1: /* 2xAA/4xAA compressed depth only */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 2: /* 8xAA compressed depth only */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 3: /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 4: /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 5: /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 6: /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 7: /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 8: /* 1D and 1D Array Surfaces */ - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 9: /* Displayable maps. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 10: /* Display 8bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 11: /* Display 16bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 12: /* Display 32bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 13: /* Thin. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 14: /* Thin 8 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 15: /* Thin 16 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 16: /* Thin 32 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 17: /* Thin 64 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - case 21: /* 8 bpp PRT. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 22: /* 16 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 23: /* 32 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 24: /* 64 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 25: /* 128 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | - NUM_BANKS(ADDR_SURF_8_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - default: - gb_tile_moden = 0; - break; - } - rdev->config.si.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden); - } - } else if ((rdev->family == CHIP_VERDE) || - (rdev->family == CHIP_OLAND) || - (rdev->family == CHIP_HAINAN)) { - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: /* non-AA compressed depth or any compressed stencil */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 1: /* 2xAA/4xAA compressed depth only */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 2: /* 8xAA compressed depth only */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 3: /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 4: /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 5: /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 6: /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 7: /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 8: /* 1D and 1D Array Surfaces */ - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 9: /* Displayable maps. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 10: /* Display 8bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 11: /* Display 16bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 12: /* Display 32bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 13: /* Thin. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 14: /* Thin 8 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 15: /* Thin 16 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 16: /* Thin 32 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 17: /* Thin 64 bpp. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P4_8x16) | - TILE_SPLIT(split_equal_to_row_size) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 21: /* 8 bpp PRT. */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 22: /* 16 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); - break; - case 23: /* 32 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 24: /* 64 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - NUM_BANKS(ADDR_SURF_16_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); - break; - case 25: /* 128 bpp PRT */ - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | - PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | - NUM_BANKS(ADDR_SURF_8_BANK) | - BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); - break; - default: - gb_tile_moden = 0; - break; - } - rdev->config.si.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden); - } - } else + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + tile[reg_offset] = 0; + + switch(rdev->family) { + case CHIP_TAHITI: + case CHIP_PITCAIRN: + /* non-AA compressed depth or any compressed stencil */ + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 2xAA/4xAA compressed depth only */ + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 8xAA compressed depth only */ + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ + tile[4] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ + tile[5] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ + tile[6] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ + tile[7] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 1D and 1D Array Surfaces */ + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Displayable maps. */ + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Display 8bpp. */ + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Display 16bpp. */ + tile[11] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Display 32bpp. */ + tile[12] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* Thin. */ + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin 8 bpp. */ + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* Thin 16 bpp. */ + tile[15] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* Thin 32 bpp. */ + tile[16] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* Thin 64 bpp. */ + tile[17] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + /* 8 bpp PRT. */ + tile[21] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 16 bpp PRT */ + tile[22] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 32 bpp PRT */ + tile[23] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 64 bpp PRT */ + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 128 bpp PRT */ + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | + NUM_BANKS(ADDR_SURF_8_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(GB_TILE_MODE0 + (reg_offset * 4), tile[reg_offset]); + break; + + case CHIP_VERDE: + case CHIP_OLAND: + case CHIP_HAINAN: + /* non-AA compressed depth or any compressed stencil */ + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 2xAA/4xAA compressed depth only */ + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 8xAA compressed depth only */ + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ + tile[4] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ + tile[5] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ + tile[6] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ + tile[7] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 1D and 1D Array Surfaces */ + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Displayable maps. */ + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Display 8bpp. */ + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* Display 16bpp. */ + tile[11] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Display 32bpp. */ + tile[12] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin. */ + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin 8 bpp. */ + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin 16 bpp. */ + tile[15] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin 32 bpp. */ + tile[16] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* Thin 64 bpp. */ + tile[17] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 8 bpp PRT. */ + tile[21] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 16 bpp PRT */ + tile[22] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + /* 32 bpp PRT */ + tile[23] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 64 bpp PRT */ + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + /* 128 bpp PRT */ + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | + NUM_BANKS(ADDR_SURF_8_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(GB_TILE_MODE0 + (reg_offset * 4), tile[reg_offset]); + break; + + default: DRM_ERROR("unknown asic: 0x%x\n", rdev->family); + } } static void si_select_se_sh(struct radeon_device *rdev, @@ -2795,7 +2997,7 @@ static void si_setup_spi(struct radeon_device *rdev, } static u32 si_get_rb_disabled(struct radeon_device *rdev, - u32 max_rb_num, u32 se_num, + u32 max_rb_num_per_se, u32 sh_per_se) { u32 data, mask; @@ -2809,14 +3011,14 @@ static u32 si_get_rb_disabled(struct radeon_device *rdev, data >>= BACKEND_DISABLE_SHIFT; - mask = si_create_bitmask(max_rb_num / se_num / sh_per_se); + mask = si_create_bitmask(max_rb_num_per_se / sh_per_se); return data & mask; } static void si_setup_rb(struct radeon_device *rdev, u32 se_num, u32 sh_per_se, - u32 max_rb_num) + u32 max_rb_num_per_se) { int i, j; u32 data, mask; @@ -2826,19 +3028,21 @@ static void si_setup_rb(struct radeon_device *rdev, for (i = 0; i < se_num; i++) { for (j = 0; j < sh_per_se; j++) { si_select_se_sh(rdev, i, j); - data = si_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se); + data = si_get_rb_disabled(rdev, max_rb_num_per_se, sh_per_se); disabled_rbs |= data << ((i * sh_per_se + j) * TAHITI_RB_BITMAP_WIDTH_PER_SH); } } si_select_se_sh(rdev, 0xffffffff, 0xffffffff); mask = 1; - for (i = 0; i < max_rb_num; i++) { + for (i = 0; i < max_rb_num_per_se * se_num; i++) { if (!(disabled_rbs & mask)) enabled_rbs |= mask; mask <<= 1; } + rdev->config.si.backend_enable_mask = enabled_rbs; + for (i = 0; i < se_num; i++) { si_select_se_sh(rdev, i, 0xffffffff); data = 0; @@ -2865,7 +3069,7 @@ static void si_setup_rb(struct radeon_device *rdev, static void si_gpu_init(struct radeon_device *rdev) { u32 gb_addr_config = 0; - u32 mc_shared_chmap, mc_arb_ramcfg; + u32 mc_arb_ramcfg; u32 sx_debug_1; u32 hdp_host_path_cntl; u32 tmp; @@ -2970,12 +3174,14 @@ static void si_gpu_init(struct radeon_device *rdev) } WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff)); + WREG32(SRBM_INT_CNTL, 1); + WREG32(SRBM_INT_ACK, 1); evergreen_fix_pci_max_read_req_size(rdev); WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN); - mc_shared_chmap = RREG32(MC_SHARED_CHMAP); + RREG32(MC_SHARED_CHMAP); mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG); rdev->config.si.num_tile_pipes = rdev->config.si.max_tile_pipes; @@ -3027,7 +3233,7 @@ static void si_gpu_init(struct radeon_device *rdev) /* XXX what about 12? */ rdev->config.si.tile_config |= (3 << 0); break; - } + } switch ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) { case 0: /* four banks */ rdev->config.si.tile_config |= 0 << 4; @@ -3067,6 +3273,13 @@ static void si_gpu_init(struct radeon_device *rdev) rdev->config.si.max_sh_per_se, rdev->config.si.max_cu_per_sh); + rdev->config.si.active_cus = 0; + for (i = 0; i < rdev->config.si.max_shader_engines; i++) { + for (j = 0; j < rdev->config.si.max_sh_per_se; j++) { + rdev->config.si.active_cus += + hweight32(si_get_cu_active_bitmap(rdev, i, j)); + } + } /* set HW defaults for 3D engine */ WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) | @@ -3155,7 +3368,7 @@ void si_fence_ring_emit(struct radeon_device *rdev, /* EVENT_WRITE_EOP - flush caches, send int */ radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); @@ -3167,6 +3380,7 @@ void si_fence_ring_emit(struct radeon_device *rdev, void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { struct radeon_ring *ring = &rdev->ring[ib->ring]; + unsigned vm_id = ib->vm ? ib->vm->ids[ib->ring].id : 0; u32 header; if (ib->is_const_ib) { @@ -3188,7 +3402,7 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); radeon_ring_write(ring, (1 << 8)); radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); radeon_ring_write(ring, next_rptr); } @@ -3202,14 +3416,13 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) #endif (ib->gpu_addr & 0xFFFFFFFC)); radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF); - radeon_ring_write(ring, ib->length_dw | - (ib->vm ? (ib->vm->id << 24) : 0)); + radeon_ring_write(ring, ib->length_dw | (vm_id << 24)); if (!ib->is_const_ib) { /* flush read cache over gart for this vmid */ radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, ib->vm ? ib->vm->id : 0); + radeon_ring_write(ring, vm_id); radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA | PACKET3_TC_ACTION_ENA | @@ -3229,7 +3442,8 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable) if (enable) WREG32(CP_ME_CNTL, 0); else { - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT)); WREG32(SCRATCH_UMSK, 0); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; @@ -3241,34 +3455,77 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable) static int si_cp_load_microcode(struct radeon_device *rdev) { - const __be32 *fw_data; int i; - if (!rdev->me_fw || !rdev->pfp_fw) + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw) return -EINVAL; si_cp_enable(rdev, false); - /* PFP */ - fw_data = (const __be32 *)rdev->pfp_fw->data; - WREG32(CP_PFP_UCODE_ADDR, 0); - for (i = 0; i < SI_PFP_UCODE_SIZE; i++) - WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++)); - WREG32(CP_PFP_UCODE_ADDR, 0); - - /* CE */ - fw_data = (const __be32 *)rdev->ce_fw->data; - WREG32(CP_CE_UCODE_ADDR, 0); - for (i = 0; i < SI_CE_UCODE_SIZE; i++) - WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++)); - WREG32(CP_CE_UCODE_ADDR, 0); - - /* ME */ - fw_data = (const __be32 *)rdev->me_fw->data; - WREG32(CP_ME_RAM_WADDR, 0); - for (i = 0; i < SI_PM4_UCODE_SIZE; i++) - WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++)); - WREG32(CP_ME_RAM_WADDR, 0); + if (rdev->new_fw) { + const struct gfx_firmware_header_v1_0 *pfp_hdr = + (const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data; + const struct gfx_firmware_header_v1_0 *ce_hdr = + (const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data; + const struct gfx_firmware_header_v1_0 *me_hdr = + (const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data; + const __le32 *fw_data; + u32 fw_size; + + radeon_ucode_print_gfx_hdr(&pfp_hdr->header); + radeon_ucode_print_gfx_hdr(&ce_hdr->header); + radeon_ucode_print_gfx_hdr(&me_hdr->header); + + /* PFP */ + fw_data = (const __le32 *) + (rdev->pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes)); + fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4; + WREG32(CP_PFP_UCODE_ADDR, 0); + for (i = 0; i < fw_size; i++) + WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++)); + WREG32(CP_PFP_UCODE_ADDR, 0); + + /* CE */ + fw_data = (const __le32 *) + (rdev->ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes)); + fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4; + WREG32(CP_CE_UCODE_ADDR, 0); + for (i = 0; i < fw_size; i++) + WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++)); + WREG32(CP_CE_UCODE_ADDR, 0); + + /* ME */ + fw_data = (const __be32 *) + (rdev->me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes)); + fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4; + WREG32(CP_ME_RAM_WADDR, 0); + for (i = 0; i < fw_size; i++) + WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++)); + WREG32(CP_ME_RAM_WADDR, 0); + } else { + const __be32 *fw_data; + + /* PFP */ + fw_data = (const __be32 *)rdev->pfp_fw->data; + WREG32(CP_PFP_UCODE_ADDR, 0); + for (i = 0; i < SI_PFP_UCODE_SIZE; i++) + WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_PFP_UCODE_ADDR, 0); + + /* CE */ + fw_data = (const __be32 *)rdev->ce_fw->data; + WREG32(CP_CE_UCODE_ADDR, 0); + for (i = 0; i < SI_CE_UCODE_SIZE; i++) + WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_CE_UCODE_ADDR, 0); + + /* ME */ + fw_data = (const __be32 *)rdev->me_fw->data; + WREG32(CP_ME_RAM_WADDR, 0); + for (i = 0; i < SI_PM4_UCODE_SIZE; i++) + WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_ME_RAM_WADDR, 0); + } WREG32(CP_PFP_UCODE_ADDR, 0); WREG32(CP_CE_UCODE_ADDR, 0); @@ -3301,7 +3558,7 @@ static int si_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE)); radeon_ring_write(ring, 0xc000); radeon_ring_write(ring, 0xe000); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); si_cp_enable(rdev, true); @@ -3330,17 +3587,21 @@ static int si_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) { ring = &rdev->ring[i]; r = radeon_ring_lock(rdev, ring, 2); + if (r) { + DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); + return r; + } /* clear the compute context state */ radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0)); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } return 0; @@ -3371,16 +3632,7 @@ static int si_cp_resume(struct radeon_device *rdev) u32 rb_bufsz; int r; - /* Reset cp; if cp is reset, then PA, SH, VGT also need to be reset */ - WREG32(GRBM_SOFT_RESET, (SOFT_RESET_CP | - SOFT_RESET_PA | - SOFT_RESET_VGT | - SOFT_RESET_SPI | - SOFT_RESET_SX)); - RREG32(GRBM_SOFT_RESET); - mdelay(15); - WREG32(GRBM_SOFT_RESET, 0); - RREG32(GRBM_SOFT_RESET); + si_enable_gui_idle_interrupt(rdev, false); WREG32(CP_SEM_WAIT_TIMER, 0x0); WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); @@ -3394,8 +3646,8 @@ static int si_cp_resume(struct radeon_device *rdev) /* ring 0 - compute and gfx */ /* Set ring buffer size */ ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3422,13 +3674,11 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB0_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB0_RPTR); - /* ring1 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3448,13 +3698,11 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB1_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB1_RPTR); - /* ring2 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3474,8 +3722,6 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB2_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB2_RPTR); - /* start the rings */ si_cp_start(rdev); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; @@ -3497,10 +3743,15 @@ static int si_cp_resume(struct radeon_device *rdev) rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; } + si_enable_gui_idle_interrupt(rdev, true); + + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + return 0; } -static u32 si_gpu_check_soft_reset(struct radeon_device *rdev) +u32 si_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -3598,6 +3849,13 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); + /* disable PG/CG */ + si_fini_pg(rdev); + si_fini_cg(rdev); + + /* stop the rlc */ + si_rlc_stop(rdev); + /* Disable CP parsing/prefetching */ WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); @@ -3706,19 +3964,131 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) evergreen_print_gpu_status_regs(rdev); } -int si_asic_reset(struct radeon_device *rdev) +static void si_set_clk_bypass_mode(struct radeon_device *rdev) +{ + u32 tmp, i; + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_BYPASS_EN; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL_2); + tmp |= SPLL_CTLREQ_CHG; + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(SPLL_STATUS) & SPLL_CHG_STATUS) + break; + udelay(1); + } + + tmp = RREG32(CG_SPLL_FUNC_CNTL_2); + tmp &= ~(SPLL_CTLREQ_CHG | SCLK_MUX_UPDATE); + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + tmp = RREG32(MPLL_CNTL_MODE); + tmp &= ~MPLL_MCLK_SEL; + WREG32(MPLL_CNTL_MODE, tmp); +} + +static void si_spll_powerdown(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32(SPLL_CNTL_MODE); + tmp |= SPLL_SW_DIR_CONTROL; + WREG32(SPLL_CNTL_MODE, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_RESET; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_SLEEP; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(SPLL_CNTL_MODE); + tmp &= ~SPLL_SW_DIR_CONTROL; + WREG32(SPLL_CNTL_MODE, tmp); +} + +static void si_gpu_pci_config_reset(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 tmp, i; + + dev_info(rdev->dev, "GPU pci config reset\n"); + + /* disable dpm? */ + + /* disable cg/pg */ + si_fini_pg(rdev); + si_fini_cg(rdev); + + /* Disable CP parsing/prefetching */ + WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); + /* dma0 */ + tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp); + /* dma1 */ + tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp); + /* XXX other engines? */ + + /* halt the rlc, disable cp internal ints */ + si_rlc_stop(rdev); + + udelay(50); + + /* disable mem access */ + evergreen_mc_stop(rdev, &save); + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timed out !\n"); + } + + /* set mclk/sclk to bypass */ + si_set_clk_bypass_mode(rdev); + /* powerdown spll */ + si_spll_powerdown(rdev); + /* disable BM */ + pci_clear_master(rdev->pdev); + /* reset */ + radeon_pci_config_reset(rdev); + /* wait for asic to come out of reset */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CONFIG_MEMSIZE) != 0xffffffff) + break; + udelay(1); + } +} + +int si_asic_reset(struct radeon_device *rdev, bool hard) { u32 reset_mask; + if (hard) { + si_gpu_pci_config_reset(rdev); + return 0; + } + reset_mask = si_gpu_check_soft_reset(rdev); if (reset_mask) r600_set_bios_scratch_engine_hung(rdev, true); + /* try soft reset */ si_gpu_soft_reset(rdev, reset_mask); reset_mask = si_gpu_check_soft_reset(rdev); + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + si_gpu_pci_config_reset(rdev); + + reset_mask = si_gpu_check_soft_reset(rdev); + if (!reset_mask) r600_set_bios_scratch_engine_hung(rdev, false); @@ -3741,39 +4111,9 @@ bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); - return false; - } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * si_dma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up. - * Returns true if the engine appears to be locked up, false if not. - */ -bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = si_gpu_check_soft_reset(rdev); - u32 mask; - - if (ring->idx == R600_RING_TYPE_DMA_INDEX) - mask = RADEON_RESET_DMA; - else - mask = RADEON_RESET_DMA1; - - if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -3894,8 +4234,15 @@ static int si_mc_init(struct radeon_device *rdev) rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); /* size in MB on si */ - rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; - rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; + tmp = RREG32(CONFIG_MEMSIZE); + /* some boards may have garbage in the upper 16 bits */ + if (tmp & 0xffff0000) { + DRM_INFO("Probable bad vram size: 0x%08x\n", tmp); + if (tmp & 0xffff) + tmp &= 0xffff; + } + rdev->mc.mc_vram_size = tmp * 1024ULL * 1024ULL; + rdev->mc.real_vram_size = rdev->mc.mc_vram_size; rdev->mc.visible_vram_size = rdev->mc.aper_size; si_vram_gtt_location(rdev, &rdev->mc); radeon_update_bandwidth_info(rdev); @@ -3926,23 +4273,25 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) r = radeon_gart_table_vram_pin(rdev); if (r) return r; - radeon_gart_restore(rdev); /* Setup TLB control */ WREG32(MC_VM_MX_L1_TLB_CNTL, (0xA << 7) | ENABLE_L1_TLB | + ENABLE_L1_FRAGMENT_PROCESSING | SYSTEM_ACCESS_MODE_NOT_IN_SYS | ENABLE_ADVANCED_DRIVER_MODEL | SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | + ENABLE_L2_FRAGMENT_PROCESSING | ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE | EFFECTIVE_L2_QUEUE_SIZE(7) | CONTEXT1_IDENTITY_ACCESS_MODE(1)); WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE); WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | - L2_CACHE_BIGK_FRAGMENT_SIZE(0)); + BANK_SELECT(4) | + L2_CACHE_BIGK_FRAGMENT_SIZE(4)); /* setup context0 */ WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12); @@ -3960,7 +4309,7 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) /* empty context1-15 */ /* set vm size, must be a multiple of 4 */ WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0); - WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn); + WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn - 1); /* Assign the pt base to something valid for now; the pts used for * the VMs are determined by the application and setup and assigned * on the fly in the vm part of radeon_gart.c @@ -3968,10 +4317,10 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) for (i = 1; i < 16; i++) { if (i < 8) WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); else WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); } /* enable context1-15 */ @@ -3979,6 +4328,7 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) (u32)(rdev->dummy_page.addr >> 12)); WREG32(VM_CONTEXT1_CNTL2, 4); WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) | + PAGE_TABLE_BLOCK_SIZE(radeon_vm_block_size - 9) | RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT | RANGE_PROTECTION_FAULT_ENABLE_DEFAULT | DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT | @@ -4002,6 +4352,17 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) static void si_pcie_gart_disable(struct radeon_device *rdev) { + unsigned i; + + for (i = 1; i < 16; ++i) { + uint32_t reg; + if (i < 8) + reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2); + else + reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2); + rdev->vm_manager.saved_table_addr[i] = RREG32(reg); + } + /* Disable all tables */ WREG32(VM_CONTEXT0_CNTL, 0); WREG32(VM_CONTEXT1_CNTL, 0); @@ -4033,6 +4394,10 @@ static bool si_vm_reg_valid(u32 reg) if (reg >= 0x28000) return true; + /* shader regs are also fine */ + if (reg >= 0xB000 && reg < 0xC000) + return true; + /* check config regs */ switch (reg) { case GRBM_GFX_INDEX: @@ -4061,6 +4426,7 @@ static bool si_vm_reg_valid(u32 reg) case SPI_CONFIG_CNTL: case SPI_CONFIG_CNTL_1: case TA_CNTL_AUX: + case TA_CS_BC_BASE_ADDR: return true; default: DRM_ERROR("Invalid register 0x%x in CS\n", reg); @@ -4090,13 +4456,64 @@ static int si_vm_packet3_ce_check(struct radeon_device *rdev, return 0; } +static int si_vm_packet3_cp_dma_check(u32 *ib, u32 idx) +{ + u32 start_reg, reg, i; + u32 command = ib[idx + 4]; + u32 info = ib[idx + 1]; + u32 idx_value = ib[idx]; + if (command & PACKET3_CP_DMA_CMD_SAS) { + /* src address space is register */ + if (((info & 0x60000000) >> 29) == 0) { + start_reg = idx_value << 2; + if (command & PACKET3_CP_DMA_CMD_SAIC) { + reg = start_reg; + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad SRC register\n"); + return -EINVAL; + } + } else { + for (i = 0; i < (command & 0x1fffff); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad SRC register\n"); + return -EINVAL; + } + } + } + } + } + if (command & PACKET3_CP_DMA_CMD_DAS) { + /* dst address space is register */ + if (((info & 0x00300000) >> 20) == 0) { + start_reg = ib[idx + 2]; + if (command & PACKET3_CP_DMA_CMD_DAIC) { + reg = start_reg; + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad DST register\n"); + return -EINVAL; + } + } else { + for (i = 0; i < (command & 0x1fffff); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad DST register\n"); + return -EINVAL; + } + } + } + } + } + return 0; +} + static int si_vm_packet3_gfx_check(struct radeon_device *rdev, u32 *ib, struct radeon_cs_packet *pkt) { + int r; u32 idx = pkt->idx + 1; u32 idx_value = ib[idx]; u32 start_reg, end_reg, reg, i; - u32 command, info; switch (pkt->opcode) { case PACKET3_NOP: @@ -4197,50 +4614,9 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev, } break; case PACKET3_CP_DMA: - command = ib[idx + 4]; - info = ib[idx + 1]; - if (command & PACKET3_CP_DMA_CMD_SAS) { - /* src address space is register */ - if (((info & 0x60000000) >> 29) == 0) { - start_reg = idx_value << 2; - if (command & PACKET3_CP_DMA_CMD_SAIC) { - reg = start_reg; - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad SRC register\n"); - return -EINVAL; - } - } else { - for (i = 0; i < (command & 0x1fffff); i++) { - reg = start_reg + (4 * i); - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad SRC register\n"); - return -EINVAL; - } - } - } - } - } - if (command & PACKET3_CP_DMA_CMD_DAS) { - /* dst address space is register */ - if (((info & 0x00300000) >> 20) == 0) { - start_reg = ib[idx + 2]; - if (command & PACKET3_CP_DMA_CMD_DAIC) { - reg = start_reg; - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad DST register\n"); - return -EINVAL; - } - } else { - for (i = 0; i < (command & 0x1fffff); i++) { - reg = start_reg + (4 * i); - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad DST register\n"); - return -EINVAL; - } - } - } - } - } + r = si_vm_packet3_cp_dma_check(ib, idx); + if (r) + return r; break; default: DRM_ERROR("Invalid GFX packet3: 0x%x\n", pkt->opcode); @@ -4252,6 +4628,7 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev, static int si_vm_packet3_compute_check(struct radeon_device *rdev, u32 *ib, struct radeon_cs_packet *pkt) { + int r; u32 idx = pkt->idx + 1; u32 idx_value = ib[idx]; u32 start_reg, reg, i; @@ -4324,6 +4701,11 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev, return -EINVAL; } break; + case PACKET3_CP_DMA: + r = si_vm_packet3_cp_dma_check(ib, idx); + if (r) + return r; + break; default: DRM_ERROR("Invalid Compute packet3: 0x%x\n", pkt->opcode); return -EINVAL; @@ -4334,7 +4716,7 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev, int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib) { int ret = 0; - u32 idx = 0; + u32 idx = 0, i; struct radeon_cs_packet pkt; do { @@ -4376,8 +4758,15 @@ int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib) ret = -EINVAL; break; } - if (ret) + if (ret) { + for (i = 0; i < ib->length_dw; i++) { + if (i == idx) + printk("\t0x%08x <---\n", ib->ptr[i]); + else + printk("\t0x%08x\n", ib->ptr[i]); + } break; + } } while (idx < ib->length_dw); return ret; @@ -4401,136 +4790,290 @@ void si_vm_fini(struct radeon_device *rdev) } /** - * si_vm_set_page - update the page tables using the CP + * si_vm_decode_fault - print human readable fault info * * @rdev: radeon_device pointer - * @ib: indirect buffer to fill with commands - * @pe: addr of the page entry - * @addr: dst addr to write into pe - * @count: number of page entries to update - * @incr: increase next addr by incr bytes - * @flags: access flags + * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value + * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value * - * Update the page tables using the CP (SI). + * Print human readable fault information (SI). */ -void si_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags) +static void si_vm_decode_fault(struct radeon_device *rdev, + u32 status, u32 addr) { - uint32_t r600_flags = cayman_vm_page_flags(rdev, flags); - uint64_t value; - unsigned ndw; - - if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) { - while (count) { - ndw = 2 + count * 2; - if (ndw > 0x3FFE) - ndw = 0x3FFE; - - ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw); - ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(1)); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe); - for (; ndw > 2; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } + u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT; + u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT; + u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT; + char *block; + + if (rdev->family == CHIP_TAHITI) { + switch (mc_id) { + case 160: + case 144: + case 96: + case 80: + case 224: + case 208: + case 32: + case 16: + block = "CB"; + break; + case 161: + case 145: + case 97: + case 81: + case 225: + case 209: + case 33: + case 17: + block = "CB_FMASK"; + break; + case 162: + case 146: + case 98: + case 82: + case 226: + case 210: + case 34: + case 18: + block = "CB_CMASK"; + break; + case 163: + case 147: + case 99: + case 83: + case 227: + case 211: + case 35: + case 19: + block = "CB_IMMED"; + break; + case 164: + case 148: + case 100: + case 84: + case 228: + case 212: + case 36: + case 20: + block = "DB"; + break; + case 165: + case 149: + case 101: + case 85: + case 229: + case 213: + case 37: + case 21: + block = "DB_HTILE"; + break; + case 167: + case 151: + case 103: + case 87: + case 231: + case 215: + case 39: + case 23: + block = "DB_STEN"; + break; + case 72: + case 68: + case 64: + case 8: + case 4: + case 0: + case 136: + case 132: + case 128: + case 200: + case 196: + case 192: + block = "TC"; + break; + case 112: + case 48: + block = "CP"; + break; + case 49: + case 177: + case 50: + case 178: + block = "SH"; + break; + case 53: + case 190: + block = "VGT"; + break; + case 117: + block = "IH"; + break; + case 51: + case 115: + block = "RLC"; + break; + case 119: + case 183: + block = "DMA0"; + break; + case 61: + block = "DMA1"; + break; + case 248: + case 120: + block = "HDP"; + break; + default: + block = "unknown"; + break; } } else { - /* DMA */ - if (flags & RADEON_VM_PAGE_SYSTEM) { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - /* for non-physically contiguous pages (system) */ - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - } else { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - if (flags & RADEON_VM_PAGE_VALID) - value = addr; - else - value = 0; - /* for physically contiguous pages (vram) */ - ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw); - ib->ptr[ib->length_dw++] = pe; /* dst addr */ - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - ib->ptr[ib->length_dw++] = r600_flags; /* mask */ - ib->ptr[ib->length_dw++] = 0; - ib->ptr[ib->length_dw++] = value; /* value */ - ib->ptr[ib->length_dw++] = upper_32_bits(value); - ib->ptr[ib->length_dw++] = incr; /* increment size */ - ib->ptr[ib->length_dw++] = 0; - pe += ndw * 4; - addr += (ndw / 2) * incr; - count -= ndw / 2; - } + switch (mc_id) { + case 32: + case 16: + case 96: + case 80: + case 160: + case 144: + case 224: + case 208: + block = "CB"; + break; + case 33: + case 17: + case 97: + case 81: + case 161: + case 145: + case 225: + case 209: + block = "CB_FMASK"; + break; + case 34: + case 18: + case 98: + case 82: + case 162: + case 146: + case 226: + case 210: + block = "CB_CMASK"; + break; + case 35: + case 19: + case 99: + case 83: + case 163: + case 147: + case 227: + case 211: + block = "CB_IMMED"; + break; + case 36: + case 20: + case 100: + case 84: + case 164: + case 148: + case 228: + case 212: + block = "DB"; + break; + case 37: + case 21: + case 101: + case 85: + case 165: + case 149: + case 229: + case 213: + block = "DB_HTILE"; + break; + case 39: + case 23: + case 103: + case 87: + case 167: + case 151: + case 231: + case 215: + block = "DB_STEN"; + break; + case 72: + case 68: + case 8: + case 4: + case 136: + case 132: + case 200: + case 196: + block = "TC"; + break; + case 112: + case 48: + block = "CP"; + break; + case 49: + case 177: + case 50: + case 178: + block = "SH"; + break; + case 53: + block = "VGT"; + break; + case 117: + block = "IH"; + break; + case 51: + case 115: + block = "RLC"; + break; + case 119: + case 183: + block = "DMA0"; + break; + case 61: + block = "DMA1"; + break; + case 248: + case 120: + block = "HDP"; + break; + default: + block = "unknown"; + break; } - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0); } + + printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n", + protections, vmid, addr, + (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read", + block, mc_id); } -void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +void si_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring, + unsigned vm_id, uint64_t pd_addr) { - struct radeon_ring *ring = &rdev->ring[ridx]; - - if (vm == NULL) - return; - /* write new base address */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) | WRITE_DATA_DST_SEL(0))); - if (vm->id < 8) { + if (vm_id < 8) { radeon_ring_write(ring, - (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2); + (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2); } else { radeon_ring_write(ring, - (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2); + (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2); } radeon_ring_write(ring, 0); - radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + radeon_ring_write(ring, pd_addr >> 12); /* flush hdp cache */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2); radeon_ring_write(ring, 0); @@ -4538,43 +5081,27 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) /* bits 0-15 are the VM contexts0-15 */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); radeon_ring_write(ring, 0); - radeon_ring_write(ring, 1 << vm->id); + radeon_ring_write(ring, 1 << vm_id); + + /* wait for the invalidate to complete */ + radeon_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5)); + radeon_ring_write(ring, (WAIT_REG_MEM_FUNCTION(0) | /* always */ + WAIT_REG_MEM_ENGINE(0))); /* me */ + radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); /* ref */ + radeon_ring_write(ring, 0); /* mask */ + radeon_ring_write(ring, 0x20); /* poll interval */ /* sync PFP to ME, otherwise we might get invalid PFP reads */ radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); radeon_ring_write(ring, 0x0); } -void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) -{ - struct radeon_ring *ring = &rdev->ring[ridx]; - - if (vm == NULL) - return; - - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - if (vm->id < 8) { - radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2)); - } else { - radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2)); - } - radeon_ring_write(ring, vm->pd_gpu_addr >> 12); - - /* flush hdp cache */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); - radeon_ring_write(ring, 1); - - /* bits 0-7 are the VM contexts0-7 */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2)); - radeon_ring_write(ring, 1 << vm->id); -} - /* * Power and clock gating */ @@ -4642,7 +5169,7 @@ static void si_set_uvd_dcm(struct radeon_device *rdev, WREG32_UVD_CTX(UVD_CGC_CTRL2, tmp2); } -static void si_init_uvd_internal_cg(struct radeon_device *rdev) +void si_init_uvd_internal_cg(struct radeon_device *rdev) { bool hw_mode = true; @@ -4685,7 +5212,7 @@ static void si_enable_dma_pg(struct radeon_device *rdev, bool enable) u32 data, orig; orig = data = RREG32(DMA_PG); - if (enable) + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA)) data |= PG_CNTL_ENABLE; else data &= ~PG_CNTL_ENABLE; @@ -4709,7 +5236,7 @@ static void si_enable_gfx_cgpg(struct radeon_device *rdev, { u32 tmp; - if (enable) { + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG)) { tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10); WREG32(RLC_TTOP_D, tmp); @@ -4812,9 +5339,9 @@ static void si_enable_cgcg(struct radeon_device *rdev, orig = data = RREG32(RLC_CGCG_CGLS_CTRL); - si_enable_gui_idle_interrupt(rdev, enable); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGCG)) { + si_enable_gui_idle_interrupt(rdev, true); - if (enable) { WREG32(RLC_GCPM_GENERAL_3, 0x00000080); tmp = si_halt_rlc(rdev); @@ -4831,6 +5358,8 @@ static void si_enable_cgcg(struct radeon_device *rdev, data |= CGCG_EN | CGLS_EN; } else { + si_enable_gui_idle_interrupt(rdev, false); + RREG32(CB_CGTT_SCLK_CTRL); RREG32(CB_CGTT_SCLK_CTRL); RREG32(CB_CGTT_SCLK_CTRL); @@ -4848,16 +5377,18 @@ static void si_enable_mgcg(struct radeon_device *rdev, { u32 data, orig, tmp = 0; - if (enable) { + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGCG)) { orig = data = RREG32(CGTS_SM_CTRL_REG); data = 0x96940200; if (orig != data) WREG32(CGTS_SM_CTRL_REG, data); - orig = data = RREG32(CP_MEM_SLP_CNTL); - data |= CP_MEM_LS_EN; - if (orig != data) - WREG32(CP_MEM_SLP_CNTL, data); + if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CP_LS) { + orig = data = RREG32(CP_MEM_SLP_CNTL); + data |= CP_MEM_LS_EN; + if (orig != data) + WREG32(CP_MEM_SLP_CNTL, data); + } orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE); data &= 0xffffffc0; @@ -4902,7 +5433,7 @@ static void si_enable_uvd_mgcg(struct radeon_device *rdev, { u32 orig, data, tmp; - if (enable) { + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_UVD_MGCG)) { tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL); tmp |= 0x3fff; WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp); @@ -4950,7 +5481,7 @@ static void si_enable_mc_ls(struct radeon_device *rdev, for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { orig = data = RREG32(mc_cg_registers[i]); - if (enable) + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_LS)) data |= MC_LS_ENABLE; else data &= ~MC_LS_ENABLE; @@ -4959,230 +5490,301 @@ static void si_enable_mc_ls(struct radeon_device *rdev, } } - -static void si_init_cg(struct radeon_device *rdev) +static void si_enable_mc_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_uvd = true; - - si_enable_mgcg(rdev, true); - si_enable_cgcg(rdev, true); - /* disable MC LS on Tahiti */ - if (rdev->family == CHIP_TAHITI) - si_enable_mc_ls(rdev, false); - if (has_uvd) { - si_enable_uvd_mgcg(rdev, true); - si_init_uvd_internal_cg(rdev); + int i; + u32 orig, data; + + for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { + orig = data = RREG32(mc_cg_registers[i]); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_MGCG)) + data |= MC_CG_ENABLE; + else + data &= ~MC_CG_ENABLE; + if (data != orig) + WREG32(mc_cg_registers[i], data); } } -static void si_fini_cg(struct radeon_device *rdev) +static void si_enable_dma_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_uvd = true; + u32 orig, data, offset; + int i; - if (has_uvd) - si_enable_uvd_mgcg(rdev, false); - si_enable_cgcg(rdev, false); - si_enable_mgcg(rdev, false); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_SDMA_MGCG)) { + for (i = 0; i < 2; i++) { + if (i == 0) + offset = DMA0_REGISTER_OFFSET; + else + offset = DMA1_REGISTER_OFFSET; + orig = data = RREG32(DMA_POWER_CNTL + offset); + data &= ~MEM_POWER_OVERRIDE; + if (data != orig) + WREG32(DMA_POWER_CNTL + offset, data); + WREG32(DMA_CLK_CTRL + offset, 0x00000100); + } + } else { + for (i = 0; i < 2; i++) { + if (i == 0) + offset = DMA0_REGISTER_OFFSET; + else + offset = DMA1_REGISTER_OFFSET; + orig = data = RREG32(DMA_POWER_CNTL + offset); + data |= MEM_POWER_OVERRIDE; + if (data != orig) + WREG32(DMA_POWER_CNTL + offset, data); + + orig = data = RREG32(DMA_CLK_CTRL + offset); + data = 0xff000000; + if (data != orig) + WREG32(DMA_CLK_CTRL + offset, data); + } + } } -static void si_init_pg(struct radeon_device *rdev) +static void si_enable_bif_mgls(struct radeon_device *rdev, + bool enable) { - bool has_pg = false; + u32 orig, data; - /* only cape verde supports PG */ - if (rdev->family == CHIP_VERDE) - has_pg = true; + orig = data = RREG32_PCIE(PCIE_CNTL2); - if (has_pg) { - si_init_ao_cu_mask(rdev); - si_init_dma_pg(rdev); - si_enable_dma_pg(rdev, true); - si_init_gfx_cgpg(rdev); - si_enable_gfx_cgpg(rdev, true); - } else { - WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); - WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); - } + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_BIF_LS)) + data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN; + else + data &= ~(SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN); + + if (orig != data) + WREG32_PCIE(PCIE_CNTL2, data); } -static void si_fini_pg(struct radeon_device *rdev) +static void si_enable_hdp_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_pg = false; + u32 orig, data; - /* only cape verde supports PG */ - if (rdev->family == CHIP_VERDE) - has_pg = true; + orig = data = RREG32(HDP_HOST_PATH_CNTL); - if (has_pg) { - si_enable_dma_pg(rdev, false); - si_enable_gfx_cgpg(rdev, false); - } + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_MGCG)) + data &= ~CLOCK_GATING_DIS; + else + data |= CLOCK_GATING_DIS; + + if (orig != data) + WREG32(HDP_HOST_PATH_CNTL, data); } -/* - * RLC - */ -void si_rlc_fini(struct radeon_device *rdev) +static void si_enable_hdp_ls(struct radeon_device *rdev, + bool enable) { - int r; + u32 orig, data; - /* save restore block */ - if (rdev->rlc.save_restore_obj) { - r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); - if (unlikely(r != 0)) - dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r); - radeon_bo_unpin(rdev->rlc.save_restore_obj); - radeon_bo_unreserve(rdev->rlc.save_restore_obj); + orig = data = RREG32(HDP_MEM_POWER_LS); - radeon_bo_unref(&rdev->rlc.save_restore_obj); - rdev->rlc.save_restore_obj = NULL; - } - - /* clear state block */ - if (rdev->rlc.clear_state_obj) { - r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); - if (unlikely(r != 0)) - dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r); - radeon_bo_unpin(rdev->rlc.clear_state_obj); - radeon_bo_unreserve(rdev->rlc.clear_state_obj); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_LS)) + data |= HDP_LS_ENABLE; + else + data &= ~HDP_LS_ENABLE; - radeon_bo_unref(&rdev->rlc.clear_state_obj); - rdev->rlc.clear_state_obj = NULL; - } + if (orig != data) + WREG32(HDP_MEM_POWER_LS, data); } -#define RLC_CLEAR_STATE_END_MARKER 0x00000001 - -int si_rlc_init(struct radeon_device *rdev) +static void si_update_cg(struct radeon_device *rdev, + u32 block, bool enable) { - volatile u32 *dst_ptr; - u32 dws, data, i, j, k, reg_num; - u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index; - u64 reg_list_mc_addr; - const struct cs_section_def *cs_data = si_cs_data; - int r; - - /* save restore block */ - if (rdev->rlc.save_restore_obj == NULL) { - r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, - &rdev->rlc.save_restore_obj); - if (r) { - dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r); - return r; + if (block & RADEON_CG_BLOCK_GFX) { + si_enable_gui_idle_interrupt(rdev, false); + /* order matters! */ + if (enable) { + si_enable_mgcg(rdev, true); + si_enable_cgcg(rdev, true); + } else { + si_enable_cgcg(rdev, false); + si_enable_mgcg(rdev, false); } + si_enable_gui_idle_interrupt(rdev, true); } - r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); - if (unlikely(r != 0)) { - si_rlc_fini(rdev); - return r; + if (block & RADEON_CG_BLOCK_MC) { + si_enable_mc_mgcg(rdev, enable); + si_enable_mc_ls(rdev, enable); } - r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.save_restore_gpu_addr); - if (r) { - radeon_bo_unreserve(rdev->rlc.save_restore_obj); - dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r); - si_rlc_fini(rdev); - return r; + + if (block & RADEON_CG_BLOCK_SDMA) { + si_enable_dma_mgcg(rdev, enable); } - if (rdev->family == CHIP_VERDE) { - r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r); - si_rlc_fini(rdev); - return r; - } - /* write the sr buffer */ - dst_ptr = rdev->rlc.sr_ptr; - for (i = 0; i < ARRAY_SIZE(verde_rlc_save_restore_register_list); i++) { - dst_ptr[i] = verde_rlc_save_restore_register_list[i]; - } - radeon_bo_kunmap(rdev->rlc.save_restore_obj); + if (block & RADEON_CG_BLOCK_BIF) { + si_enable_bif_mgls(rdev, enable); } - radeon_bo_unreserve(rdev->rlc.save_restore_obj); - /* clear state block */ - reg_list_num = 0; - dws = 0; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_list_num++; - dws += cs_data[i].section[j].reg_count; + if (block & RADEON_CG_BLOCK_UVD) { + if (rdev->has_uvd) { + si_enable_uvd_mgcg(rdev, enable); } } - reg_list_blk_index = (3 * reg_list_num + 2); - dws += reg_list_blk_index; - if (rdev->rlc.clear_state_obj == NULL) { - r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj); - if (r) { - dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; - } + if (block & RADEON_CG_BLOCK_HDP) { + si_enable_hdp_mgcg(rdev, enable); + si_enable_hdp_ls(rdev, enable); } - r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); - if (unlikely(r != 0)) { - si_rlc_fini(rdev); - return r; +} + +static void si_init_cg(struct radeon_device *rdev) +{ + si_update_cg(rdev, (RADEON_CG_BLOCK_GFX | + RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_HDP), true); + if (rdev->has_uvd) { + si_update_cg(rdev, RADEON_CG_BLOCK_UVD, true); + si_init_uvd_internal_cg(rdev); } - r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.clear_state_gpu_addr); - if (r) { +} - radeon_bo_unreserve(rdev->rlc.clear_state_obj); - dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; +static void si_fini_cg(struct radeon_device *rdev) +{ + if (rdev->has_uvd) { + si_update_cg(rdev, RADEON_CG_BLOCK_UVD, false); } - r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; + si_update_cg(rdev, (RADEON_CG_BLOCK_GFX | + RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_HDP), false); +} + +u32 si_get_csb_size(struct radeon_device *rdev) +{ + u32 count = 0; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return 0; + + /* begin clear state */ + count += 2; + /* context control state */ + count += 3; + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) + count += 2 + ext->reg_count; + else + return 0; + } } - /* set up the cs buffer */ - dst_ptr = rdev->rlc.cs_ptr; - reg_list_hdr_blk_index = 0; - reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4); - data = upper_32_bits(reg_list_mc_addr); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_num = cs_data[i].section[j].reg_count; - data = reg_list_mc_addr & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - data = 0x08000000 | (reg_num * 4); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - for (k = 0; k < reg_num; k++) { - data = cs_data[i].section[j].extent[k]; - dst_ptr[reg_list_blk_index + k] = data; + /* pa_sc_raster_config */ + count += 3; + /* end clear state */ + count += 2; + /* clear state */ + count += 2; + + return count; +} + +void si_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer) +{ + u32 count = 0, i; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return; + if (buffer == NULL) + return; + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); + buffer[count++] = cpu_to_le32(0x80000000); + buffer[count++] = cpu_to_le32(0x80000000); + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) { + buffer[count++] = + cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); + buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000); + for (i = 0; i < ext->reg_count; i++) + buffer[count++] = cpu_to_le32(ext->extent[i]); + } else { + return; } - reg_list_mc_addr += reg_num * 4; - reg_list_blk_index += reg_num; } } - dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER; - radeon_bo_kunmap(rdev->rlc.clear_state_obj); - radeon_bo_unreserve(rdev->rlc.clear_state_obj); + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + buffer[count++] = cpu_to_le32(PA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); + switch (rdev->family) { + case CHIP_TAHITI: + case CHIP_PITCAIRN: + buffer[count++] = cpu_to_le32(0x2a00126a); + break; + case CHIP_VERDE: + buffer[count++] = cpu_to_le32(0x0000124a); + break; + case CHIP_OLAND: + buffer[count++] = cpu_to_le32(0x00000082); + break; + case CHIP_HAINAN: + buffer[count++] = cpu_to_le32(0x00000000); + break; + default: + buffer[count++] = cpu_to_le32(0x00000000); + break; + } - return 0; + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); + buffer[count++] = cpu_to_le32(0); +} + +static void si_init_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + if (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA) { + si_init_dma_pg(rdev); + } + si_init_ao_cu_mask(rdev); + if (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG) { + si_init_gfx_cgpg(rdev); + } else { + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + } + si_enable_dma_pg(rdev, true); + si_enable_gfx_cgpg(rdev, true); + } else { + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + } } -static void si_rlc_reset(struct radeon_device *rdev) +static void si_fini_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + si_enable_dma_pg(rdev, false); + si_enable_gfx_cgpg(rdev, false); + } +} + +/* + * RLC + */ +void si_rlc_reset(struct radeon_device *rdev) { u32 tmp = RREG32(GRBM_SOFT_RESET); @@ -5243,7 +5845,6 @@ static void si_enable_lbpw(struct radeon_device *rdev, bool enable) static int si_rlc_resume(struct radeon_device *rdev) { u32 i; - const __be32 *fw_data; if (!rdev->rlc_fw) return -EINVAL; @@ -5266,10 +5867,26 @@ static int si_rlc_resume(struct radeon_device *rdev) WREG32(RLC_MC_CNTL, 0); WREG32(RLC_UCODE_CNTL, 0); - fw_data = (const __be32 *)rdev->rlc_fw->data; - for (i = 0; i < SI_RLC_UCODE_SIZE; i++) { - WREG32(RLC_UCODE_ADDR, i); - WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); + if (rdev->new_fw) { + const struct rlc_firmware_header_v1_0 *hdr = + (const struct rlc_firmware_header_v1_0 *)rdev->rlc_fw->data; + u32 fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; + const __le32 *fw_data = (const __le32 *) + (rdev->rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + + radeon_ucode_print_rlc_hdr(&hdr->header); + + for (i = 0; i < fw_size; i++) { + WREG32(RLC_UCODE_ADDR, i); + WREG32(RLC_UCODE_DATA, le32_to_cpup(fw_data++)); + } + } else { + const __be32 *fw_data = + (const __be32 *)rdev->rlc_fw->data; + for (i = 0; i < SI_RLC_UCODE_SIZE; i++) { + WREG32(RLC_UCODE_ADDR, i); + WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); + } } WREG32(RLC_UCODE_ADDR, 0); @@ -5310,9 +5927,12 @@ static void si_disable_interrupts(struct radeon_device *rdev) static void si_disable_interrupt_state(struct radeon_device *rdev) { + int i; u32 tmp; - WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + tmp = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + WREG32(CP_INT_CNTL_RING0, tmp); WREG32(CP_INT_CNTL_RING1, 0); WREG32(CP_INT_CNTL_RING2, 0); tmp = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE; @@ -5320,47 +5940,18 @@ static void si_disable_interrupt_state(struct radeon_device *rdev) tmp = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET) & ~TRAP_ENABLE; WREG32(DMA_CNTL + DMA1_REGISTER_OFFSET, tmp); WREG32(GRBM_INT_CNTL, 0); - if (rdev->num_crtc >= 2) { - WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); - WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); - } - if (rdev->num_crtc >= 4) { - WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); - WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); - } - if (rdev->num_crtc >= 6) { - WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); - WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); - } - - if (rdev->num_crtc >= 2) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); - } - if (rdev->num_crtc >= 4) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); - } - if (rdev->num_crtc >= 6) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); - } + WREG32(SRBM_INT_CNTL, 0); + for (i = 0; i < rdev->num_crtc; i++) + WREG32(INT_MASK + crtc_offsets[i], 0); + for (i = 0; i < rdev->num_crtc; i++) + WREG32(GRPH_INT_CONTROL + crtc_offsets[i], 0); if (!ASIC_IS_NODCE(rdev)) { - WREG32(DACA_AUTODETECT_INT_CONTROL, 0); - - tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD1_INT_CONTROL, tmp); - tmp = RREG32(DC_HPD2_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD2_INT_CONTROL, tmp); - tmp = RREG32(DC_HPD3_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD3_INT_CONTROL, tmp); - tmp = RREG32(DC_HPD4_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD4_INT_CONTROL, tmp); - tmp = RREG32(DC_HPD5_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD5_INT_CONTROL, tmp); - tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY; - WREG32(DC_HPD6_INT_CONTROL, tmp); + WREG32(DAC_AUTODETECT_INT_CONTROL, 0); + + for (i = 0; i < 6; i++) + WREG32_AND(DC_HPDx_INT_CONTROL(i), + DC_HPDx_INT_POLARITY); } } @@ -5386,8 +5977,8 @@ static int si_irq_init(struct radeon_device *rdev) } /* setup interrupt control */ - /* set dummy read address to ring address */ - WREG32(INTERRUPT_CNTL2, rdev->ih.gpu_addr >> 8); + /* set dummy read address to dummy page address */ + WREG32(INTERRUPT_CNTL2, rdev->dummy_page.addr >> 8); interrupt_cntl = RREG32(INTERRUPT_CNTL); /* IH_DUMMY_RD_OVERRIDE=0 - dummy read disabled with msi, enabled without msi * IH_DUMMY_RD_OVERRIDE=1 - dummy read controlled by IH_DUMMY_RD_EN @@ -5398,7 +5989,7 @@ static int si_irq_init(struct radeon_device *rdev) WREG32(INTERRUPT_CNTL, interrupt_cntl); WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8); - rb_bufsz = drm_order(rdev->ih.ring_size / 4); + rb_bufsz = order_base_2(rdev->ih.ring_size / 4); ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE | IH_WPTR_OVERFLOW_CLEAR | @@ -5435,14 +6026,13 @@ static int si_irq_init(struct radeon_device *rdev) return ret; } +/* The order we write back each register here is important */ int si_irq_set(struct radeon_device *rdev) { - u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE; + int i; + u32 cp_int_cntl; u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0; - u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; - u32 hpd1 = 0, hpd2 = 0, hpd3 = 0, hpd4 = 0, hpd5 = 0, hpd6 = 0; u32 grbm_int_cntl = 0; - u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0; u32 dma_cntl, dma_cntl1; u32 thermal_int = 0; @@ -5458,14 +6048,8 @@ int si_irq_set(struct radeon_device *rdev) return 0; } - if (!ASIC_IS_NODCE(rdev)) { - hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN; - hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN; - hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN; - hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN; - hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN; - hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN; - } + cp_int_cntl = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); dma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE; dma_cntl1 = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET) & ~TRAP_ENABLE; @@ -5495,60 +6079,6 @@ int si_irq_set(struct radeon_device *rdev) DRM_DEBUG("si_irq_set: sw int dma1\n"); dma_cntl1 |= TRAP_ENABLE; } - if (rdev->irq.crtc_vblank_int[0] || - atomic_read(&rdev->irq.pflip[0])) { - DRM_DEBUG("si_irq_set: vblank 0\n"); - crtc1 |= VBLANK_INT_MASK; - } - if (rdev->irq.crtc_vblank_int[1] || - atomic_read(&rdev->irq.pflip[1])) { - DRM_DEBUG("si_irq_set: vblank 1\n"); - crtc2 |= VBLANK_INT_MASK; - } - if (rdev->irq.crtc_vblank_int[2] || - atomic_read(&rdev->irq.pflip[2])) { - DRM_DEBUG("si_irq_set: vblank 2\n"); - crtc3 |= VBLANK_INT_MASK; - } - if (rdev->irq.crtc_vblank_int[3] || - atomic_read(&rdev->irq.pflip[3])) { - DRM_DEBUG("si_irq_set: vblank 3\n"); - crtc4 |= VBLANK_INT_MASK; - } - if (rdev->irq.crtc_vblank_int[4] || - atomic_read(&rdev->irq.pflip[4])) { - DRM_DEBUG("si_irq_set: vblank 4\n"); - crtc5 |= VBLANK_INT_MASK; - } - if (rdev->irq.crtc_vblank_int[5] || - atomic_read(&rdev->irq.pflip[5])) { - DRM_DEBUG("si_irq_set: vblank 5\n"); - crtc6 |= VBLANK_INT_MASK; - } - if (rdev->irq.hpd[0]) { - DRM_DEBUG("si_irq_set: hpd 1\n"); - hpd1 |= DC_HPDx_INT_EN; - } - if (rdev->irq.hpd[1]) { - DRM_DEBUG("si_irq_set: hpd 2\n"); - hpd2 |= DC_HPDx_INT_EN; - } - if (rdev->irq.hpd[2]) { - DRM_DEBUG("si_irq_set: hpd 3\n"); - hpd3 |= DC_HPDx_INT_EN; - } - if (rdev->irq.hpd[3]) { - DRM_DEBUG("si_irq_set: hpd 4\n"); - hpd4 |= DC_HPDx_INT_EN; - } - if (rdev->irq.hpd[4]) { - DRM_DEBUG("si_irq_set: hpd 5\n"); - hpd5 |= DC_HPDx_INT_EN; - } - if (rdev->irq.hpd[5]) { - DRM_DEBUG("si_irq_set: hpd 6\n"); - hpd6 |= DC_HPDx_INT_EN; - } WREG32(CP_INT_CNTL_RING0, cp_int_cntl); WREG32(CP_INT_CNTL_RING1, cp_int_cntl1); @@ -5564,142 +6094,75 @@ int si_irq_set(struct radeon_device *rdev) thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW; } - if (rdev->num_crtc >= 2) { - WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1); - WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2); - } - if (rdev->num_crtc >= 4) { - WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3); - WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4); - } - if (rdev->num_crtc >= 6) { - WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5); - WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6); + for (i = 0; i < rdev->num_crtc; i++) { + radeon_irq_kms_set_irq_n_enabled( + rdev, INT_MASK + crtc_offsets[i], VBLANK_INT_MASK, + rdev->irq.crtc_vblank_int[i] || + atomic_read(&rdev->irq.pflip[i]), "vblank", i); } - if (rdev->num_crtc >= 2) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2); - } - if (rdev->num_crtc >= 4) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4); - } - if (rdev->num_crtc >= 6) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6); - } + for (i = 0; i < rdev->num_crtc; i++) + WREG32(GRPH_INT_CONTROL + crtc_offsets[i], GRPH_PFLIP_INT_MASK); if (!ASIC_IS_NODCE(rdev)) { - WREG32(DC_HPD1_INT_CONTROL, hpd1); - WREG32(DC_HPD2_INT_CONTROL, hpd2); - WREG32(DC_HPD3_INT_CONTROL, hpd3); - WREG32(DC_HPD4_INT_CONTROL, hpd4); - WREG32(DC_HPD5_INT_CONTROL, hpd5); - WREG32(DC_HPD6_INT_CONTROL, hpd6); + for (i = 0; i < 6; i++) { + radeon_irq_kms_set_irq_n_enabled( + rdev, DC_HPDx_INT_CONTROL(i), + DC_HPDx_INT_EN | DC_HPDx_RX_INT_EN, + rdev->irq.hpd[i], "HPD", i); + } } WREG32(CG_THERMAL_INT, thermal_int); + /* posting read */ + RREG32(SRBM_STATUS); + return 0; } +/* The order we write back each register here is important */ static inline void si_irq_ack(struct radeon_device *rdev) { - u32 tmp; + int i, j; + u32 *disp_int = rdev->irq.stat_regs.evergreen.disp_int; + u32 *grph_int = rdev->irq.stat_regs.evergreen.grph_int; if (ASIC_IS_NODCE(rdev)) return; - rdev->irq.stat_regs.evergreen.disp_int = RREG32(DISP_INTERRUPT_STATUS); - rdev->irq.stat_regs.evergreen.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE); - rdev->irq.stat_regs.evergreen.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2); - rdev->irq.stat_regs.evergreen.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3); - rdev->irq.stat_regs.evergreen.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4); - rdev->irq.stat_regs.evergreen.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5); - rdev->irq.stat_regs.evergreen.d1grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET); - rdev->irq.stat_regs.evergreen.d2grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET); - if (rdev->num_crtc >= 4) { - rdev->irq.stat_regs.evergreen.d3grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET); - rdev->irq.stat_regs.evergreen.d4grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET); - } - if (rdev->num_crtc >= 6) { - rdev->irq.stat_regs.evergreen.d5grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET); - rdev->irq.stat_regs.evergreen.d6grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET); - } - - if (rdev->irq.stat_regs.evergreen.d1grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.d2grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK); - - if (rdev->num_crtc >= 4) { - if (rdev->irq.stat_regs.evergreen.d3grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.d4grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK); - } - - if (rdev->num_crtc >= 6) { - if (rdev->irq.stat_regs.evergreen.d5grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.d6grph_int & GRPH_PFLIP_INT_OCCURRED) - WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) - WREG32(VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK); - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) - WREG32(VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK); - } - - if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) { - tmp = RREG32(DC_HPD1_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD1_INT_CONTROL, tmp); - } - if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) { - tmp = RREG32(DC_HPD2_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD2_INT_CONTROL, tmp); - } - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) { - tmp = RREG32(DC_HPD3_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD3_INT_CONTROL, tmp); - } - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) { - tmp = RREG32(DC_HPD4_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD4_INT_CONTROL, tmp); - } - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) { - tmp = RREG32(DC_HPD5_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD5_INT_CONTROL, tmp); - } - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) { - tmp = RREG32(DC_HPD5_INT_CONTROL); - tmp |= DC_HPDx_INT_ACK; - WREG32(DC_HPD6_INT_CONTROL, tmp); + for (i = 0; i < 6; i++) { + disp_int[i] = RREG32(si_disp_int_status[i]); + if (i < rdev->num_crtc) + grph_int[i] = RREG32(GRPH_INT_STATUS + crtc_offsets[i]); + } + + /* We write back each interrupt register in pairs of two */ + for (i = 0; i < rdev->num_crtc; i += 2) { + for (j = i; j < (i + 2); j++) { + if (grph_int[j] & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + crtc_offsets[j], + GRPH_PFLIP_INT_CLEAR); + } + + for (j = i; j < (i + 2); j++) { + if (disp_int[j] & LB_D1_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + crtc_offsets[j], + VBLANK_ACK); + if (disp_int[j] & LB_D1_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + crtc_offsets[j], + VLINE_ACK); + } + } + + for (i = 0; i < 6; i++) { + if (disp_int[i] & DC_HPD1_INTERRUPT) + WREG32_OR(DC_HPDx_INT_CONTROL(i), DC_HPDx_INT_ACK); + } + + for (i = 0; i < 6; i++) { + if (disp_int[i] & DC_HPD1_RX_INTERRUPT) + WREG32_OR(DC_HPDx_INT_CONTROL(i), DC_HPDx_RX_INT_ACK); } } @@ -5734,12 +6197,13 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev) wptr = RREG32(IH_RB_WPTR); if (wptr & RB_OVERFLOW) { - /* When a ring buffer overflow happen start parsing interrupt + wptr &= ~RB_OVERFLOW; + /* When a ring buffer overflow happens, start parsing interrupts * from the last not overwritten vector (wptr + 16). Hopefully * this should allow us to catchup. */ - dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n", - wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask); + dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", + wptr, rdev->ih.rptr, (wptr + 16) & rdev->ih.ptr_mask); rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask; tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; @@ -5760,12 +6224,18 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev) */ int si_irq_process(struct radeon_device *rdev) { + u32 *disp_int = rdev->irq.stat_regs.evergreen.disp_int; + u32 crtc_idx, hpd_idx; + u32 mask; u32 wptr; u32 rptr; u32 src_id, src_data, ring_id; u32 ring_index; bool queue_hotplug = false; + bool queue_dp = false; bool queue_thermal = false; + u32 status, addr; + const char *event_name; if (!rdev->ih.enabled || rdev->shutdown) return IRQ_NONE; @@ -5795,219 +6265,102 @@ restart_ih: switch (src_id) { case 1: /* D1 vblank/vline */ - switch (src_data) { - case 0: /* D1 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[0]) { - drm_handle_vblank(rdev->ddev, 0); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); - rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D1 vblank\n"); - } - break; - case 1: /* D1 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VLINE_INTERRUPT; - DRM_DEBUG("IH: D1 vline\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; - } - break; case 2: /* D2 vblank/vline */ - switch (src_data) { - case 0: /* D2 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[1]) { - drm_handle_vblank(rdev->ddev, 1); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); - rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D2 vblank\n"); - } - break; - case 1: /* D2 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT; - DRM_DEBUG("IH: D2 vline\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; - } - break; case 3: /* D3 vblank/vline */ - switch (src_data) { - case 0: /* D3 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[2]) { - drm_handle_vblank(rdev->ddev, 2); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[2])) - radeon_crtc_handle_flip(rdev, 2); - rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D3 vblank\n"); - } - break; - case 1: /* D3 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT; - DRM_DEBUG("IH: D3 vline\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; - } - break; case 4: /* D4 vblank/vline */ - switch (src_data) { - case 0: /* D4 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[3]) { - drm_handle_vblank(rdev->ddev, 3); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[3])) - radeon_crtc_handle_flip(rdev, 3); - rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D4 vblank\n"); - } - break; - case 1: /* D4 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT; - DRM_DEBUG("IH: D4 vline\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; - } - break; case 5: /* D5 vblank/vline */ - switch (src_data) { - case 0: /* D5 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[4]) { - drm_handle_vblank(rdev->ddev, 4); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[4])) - radeon_crtc_handle_flip(rdev, 4); - rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D5 vblank\n"); - } - break; - case 1: /* D5 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT; - DRM_DEBUG("IH: D5 vline\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; - } - break; case 6: /* D6 vblank/vline */ - switch (src_data) { - case 0: /* D6 vblank */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) { - if (rdev->irq.crtc_vblank_int[5]) { - drm_handle_vblank(rdev->ddev, 5); - rdev->pm.vblank_sync = true; - wake_up(&rdev->irq.vblank_queue); - } - if (atomic_read(&rdev->irq.pflip[5])) - radeon_crtc_handle_flip(rdev, 5); - rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT; - DRM_DEBUG("IH: D6 vblank\n"); + crtc_idx = src_id - 1; + + if (src_data == 0) { /* vblank */ + mask = LB_D1_VBLANK_INTERRUPT; + event_name = "vblank"; + + if (rdev->irq.crtc_vblank_int[crtc_idx]) { + drm_handle_vblank(rdev_to_drm(rdev), crtc_idx); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); } - break; - case 1: /* D6 vline */ - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT; - DRM_DEBUG("IH: D6 vline\n"); + if (atomic_read(&rdev->irq.pflip[crtc_idx])) { + radeon_crtc_handle_vblank(rdev, + crtc_idx); } + + } else if (src_data == 1) { /* vline */ + mask = LB_D1_VLINE_INTERRUPT; + event_name = "vline"; + } else { + DRM_DEBUG("Unhandled interrupt: %d %d\n", + src_id, src_data); break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); - break; } + + if (!(disp_int[crtc_idx] & mask)) { + DRM_DEBUG("IH: D%d %s - IH event w/o asserted irq bit?\n", + crtc_idx + 1, event_name); + } + + disp_int[crtc_idx] &= ~mask; + DRM_DEBUG("IH: D%d %s\n", crtc_idx + 1, event_name); + + break; + case 8: /* D1 page flip */ + case 10: /* D2 page flip */ + case 12: /* D3 page flip */ + case 14: /* D4 page flip */ + case 16: /* D5 page flip */ + case 18: /* D6 page flip */ + DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1); + if (radeon_use_pflipirq > 0) + radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1); break; case 42: /* HPD hotplug */ - switch (src_data) { - case 0: - if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int &= ~DC_HPD1_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD1\n"); - } - break; - case 1: - if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont &= ~DC_HPD2_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD2\n"); - } - break; - case 2: - if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~DC_HPD3_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD3\n"); - } - break; - case 3: - if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~DC_HPD4_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD4\n"); - } - break; - case 4: - if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~DC_HPD5_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD5\n"); - } - break; - case 5: - if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) { - rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~DC_HPD6_INTERRUPT; - queue_hotplug = true; - DRM_DEBUG("IH: HPD6\n"); - } - break; - default: - DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + if (src_data <= 5) { + hpd_idx = src_data; + mask = DC_HPD1_INTERRUPT; + queue_hotplug = true; + event_name = "HPD"; + + } else if (src_data <= 11) { + hpd_idx = src_data - 6; + mask = DC_HPD1_RX_INTERRUPT; + queue_dp = true; + event_name = "HPD_RX"; + + } else { + DRM_DEBUG("Unhandled interrupt: %d %d\n", + src_id, src_data); break; } + + if (!(disp_int[hpd_idx] & mask)) + DRM_DEBUG("IH: IH event w/o asserted irq bit?\n"); + + disp_int[hpd_idx] &= ~mask; + DRM_DEBUG("IH: %s%d\n", event_name, hpd_idx + 1); + break; + case 96: + DRM_ERROR("SRBM_READ_ERROR: 0x%x\n", RREG32(SRBM_READ_ERROR)); + WREG32(SRBM_INT_ACK, 0x1); + break; + case 124: /* UVD */ + DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data); + radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX); break; case 146: case 147: + addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR); + status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS); + /* reset addr and status */ + WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + if (addr == 0x0 && status == 0x0) + break; dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR)); + addr); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); - /* reset addr and status */ - WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + status); + si_vm_decode_fault(rdev, status, addr); break; case 176: /* RINGID0 CP_INT */ radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); @@ -6061,13 +6414,15 @@ restart_ih: /* wptr/rptr are in bytes! */ rptr += 16; rptr &= rdev->ih.ptr_mask; + WREG32(IH_RB_RPTR, rptr); } + if (queue_dp) + schedule_work(&rdev->dp_work); if (queue_hotplug) - schedule_work(&rdev->hotplug_work); + schedule_delayed_work(&rdev->hotplug_work, 0); if (queue_thermal && rdev->pm.dpm_enabled) schedule_work(&rdev->pm.dpm.thermal.work); rdev->ih.rptr = rptr; - WREG32(IH_RB_RPTR, rdev->ih.rptr); atomic_set(&rdev->ih.lock, 0); /* make sure wptr hasn't changed while processing */ @@ -6078,83 +6433,162 @@ restart_ih: return IRQ_HANDLED; } -/** - * si_copy_dma - copy pages using the DMA engine - * - * @rdev: radeon_device pointer - * @src_offset: src GPU address - * @dst_offset: dst GPU address - * @num_gpu_pages: number of GPU pages to xfer - * @fence: radeon fence object - * - * Copy GPU paging using the DMA engine (SI). - * Used by the radeon ttm implementation to move pages if - * registered as the asic copy callback. +/* + * startup/shutdown callbacks */ -int si_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) +static void si_uvd_init(struct radeon_device *rdev) { - struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; - struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_bytes, cur_size_in_bytes; - int i, num_loops; - int r = 0; - - r = radeon_semaphore_create(rdev, &sem); + int r; + + if (!rdev->has_uvd) + return; + + r = radeon_uvd_init(rdev); if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; + dev_err(rdev->dev, "failed UVD (%d) init.\n", r); + /* + * At this point rdev->uvd.vcpu_bo is NULL which trickles down + * to early fails uvd_v2_2_resume() and thus nothing happens + * there. So it is pointless to try to go through that code + * hence why we disable uvd here. + */ + rdev->has_uvd = false; + return; } + rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX], 4096); +} + +static void si_uvd_start(struct radeon_device *rdev) +{ + int r; + + if (!rdev->has_uvd) + return; - size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); - num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff); - r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11); + r = uvd_v2_2_resume(rdev); if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - radeon_semaphore_free(rdev, &sem, NULL); - return r; + dev_err(rdev->dev, "failed UVD resume (%d).\n", r); + goto error; + } + r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing UVD fences (%d).\n", r); + goto error; } + return; - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); +error: + rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0; +} + +static void si_uvd_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int r; + + if (!rdev->has_uvd || !rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size) + return; + + ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, PACKET0(UVD_NO_OP, 0)); + if (r) { + dev_err(rdev->dev, "failed initializing UVD ring (%d).\n", r); + return; } + r = uvd_v1_0_init(rdev); + if (r) { + dev_err(rdev->dev, "failed initializing UVD (%d).\n", r); + return; + } +} - for (i = 0; i < num_loops; i++) { - cur_size_in_bytes = size_in_bytes; - if (cur_size_in_bytes > 0xFFFFF) - cur_size_in_bytes = 0xFFFFF; - size_in_bytes -= cur_size_in_bytes; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 1, 0, 0, cur_size_in_bytes)); - radeon_ring_write(ring, dst_offset & 0xffffffff); - radeon_ring_write(ring, src_offset & 0xffffffff); - radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); - radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); - src_offset += cur_size_in_bytes; - dst_offset += cur_size_in_bytes; +static void si_vce_init(struct radeon_device *rdev) +{ + int r; + + if (!rdev->has_vce) + return; + + r = radeon_vce_init(rdev); + if (r) { + dev_err(rdev->dev, "failed VCE (%d) init.\n", r); + /* + * At this point rdev->vce.vcpu_bo is NULL which trickles down + * to early fails si_vce_start() and thus nothing happens + * there. So it is pointless to try to go through that code + * hence why we disable vce here. + */ + rdev->has_vce = false; + return; } + rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[TN_RING_TYPE_VCE1_INDEX], 4096); + rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[TN_RING_TYPE_VCE2_INDEX], 4096); +} + +static void si_vce_start(struct radeon_device *rdev) +{ + int r; + + if (!rdev->has_vce) + return; - r = radeon_fence_emit(rdev, fence, ring->idx); + r = radeon_vce_resume(rdev); if (r) { - radeon_ring_unlock_undo(rdev, ring); - return r; + dev_err(rdev->dev, "failed VCE resume (%d).\n", r); + goto error; + } + r = vce_v1_0_resume(rdev); + if (r) { + dev_err(rdev->dev, "failed VCE resume (%d).\n", r); + goto error; + } + r = radeon_fence_driver_start_ring(rdev, TN_RING_TYPE_VCE1_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing VCE1 fences (%d).\n", r); + goto error; } + r = radeon_fence_driver_start_ring(rdev, TN_RING_TYPE_VCE2_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing VCE2 fences (%d).\n", r); + goto error; + } + return; - radeon_ring_unlock_commit(rdev, ring); - radeon_semaphore_free(rdev, &sem, *fence); +error: + rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size = 0; + rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_size = 0; +} - return r; +static void si_vce_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int r; + + if (!rdev->has_vce || !rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size) + return; + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, VCE_CMD_NO_OP); + if (r) { + dev_err(rdev->dev, "failed initializing VCE1 ring (%d).\n", r); + return; + } + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, VCE_CMD_NO_OP); + if (r) { + dev_err(rdev->dev, "failed initializing VCE1 ring (%d).\n", r); + return; + } + r = vce_v1_0_init(rdev); + if (r) { + dev_err(rdev->dev, "failed initializing VCE (%d).\n", r); + return; + } } -/* - * startup/shutdown callbacks - */ static int si_startup(struct radeon_device *rdev) { struct radeon_ring *ring; @@ -6165,33 +6599,34 @@ static int si_startup(struct radeon_device *rdev) /* enable aspm */ si_program_aspm(rdev); - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || - !rdev->rlc_fw || !rdev->mc_fw) { - r = si_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - - r = si_mc_load_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load MC firmware!\n"); - return r; - } - + /* scratch needs to be initialized before MC */ r = r600_vram_scratch_init(rdev); if (r) return r; si_mc_program(rdev); + + if (!rdev->pm.dpm_enabled) { + r = si_mc_load_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load MC firmware!\n"); + return r; + } + } + r = si_pcie_gart_enable(rdev); if (r) return r; si_gpu_init(rdev); /* allocate rlc buffers */ - r = si_rlc_init(rdev); + if (rdev->family == CHIP_VERDE) { + rdev->rlc.reg_list = verde_rlc_save_restore_register_list; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(verde_rlc_save_restore_register_list); + } + rdev->rlc.cs_data = si_cs_data; + r = sumo_rlc_init(rdev); if (r) { DRM_ERROR("Failed to init rlc BOs!\n"); return r; @@ -6232,17 +6667,8 @@ static int si_startup(struct radeon_device *rdev) return r; } - if (rdev->has_uvd) { - r = rv770_uvd_resume(rdev); - if (!r) { - r = radeon_fence_driver_start_ring(rdev, - R600_RING_TYPE_UVD_INDEX); - if (r) - dev_err(rdev->dev, "UVD fences init error (%d).\n", r); - } - if (r) - rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0; - } + si_uvd_start(rdev); + si_vce_start(rdev); /* Enable IRQ */ if (!rdev->irq.installed) { @@ -6261,38 +6687,31 @@ static int si_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - CP_RB0_RPTR, CP_RB0_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET, - CP_RB1_RPTR, CP_RB1_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET, - CP_RB2_RPTR, CP_RB2_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR + DMA0_REGISTER_OFFSET, - DMA_RB_WPTR + DMA0_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET, - DMA_RB_RPTR + DMA1_REGISTER_OFFSET, - DMA_RB_WPTR + DMA1_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); if (r) return r; @@ -6307,19 +6726,8 @@ static int si_startup(struct radeon_device *rdev) if (r) return r; - if (rdev->has_uvd) { - ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; - if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); - if (!r) - r = r600_uvd_init(rdev); - if (r) - DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); - } - } + si_uvd_resume(rdev); + si_vce_resume(rdev); r = radeon_ib_pool_init(rdev); if (r) { @@ -6333,6 +6741,10 @@ static int si_startup(struct radeon_device *rdev) return r; } + r = radeon_audio_init(rdev); + if (r) + return r; + return 0; } @@ -6350,6 +6762,9 @@ int si_resume(struct radeon_device *rdev) /* init golden registers */ si_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = si_startup(rdev); if (r) { @@ -6364,13 +6779,19 @@ int si_resume(struct radeon_device *rdev) int si_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); + radeon_audio_fini(rdev); radeon_vm_manager_fini(rdev); si_cp_enable(rdev, false); cayman_dma_stop(rdev); if (rdev->has_uvd) { - r600_uvd_rbc_stop(rdev); radeon_uvd_suspend(rdev); + uvd_v1_0_fini(rdev); } + if (rdev->has_vce) + radeon_vce_suspend(rdev); + si_fini_pg(rdev); + si_fini_cg(rdev); si_irq_suspend(rdev); radeon_wb_disable(rdev); si_pcie_gart_disable(rdev); @@ -6418,12 +6839,10 @@ int si_init(struct radeon_device *rdev) /* Initialize surface registers */ radeon_surface_init(rdev); /* Initialize clocks */ - radeon_get_clock_info(rdev->ddev); + radeon_get_clock_info(rdev_to_drm(rdev)); /* Fence driver */ - r = radeon_fence_driver_init(rdev); - if (r) - return r; + radeon_fence_driver_init(rdev); /* initialize memory controller */ r = si_mc_init(rdev); @@ -6434,6 +6853,18 @@ int si_init(struct radeon_device *rdev) if (r) return r; + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || + !rdev->rlc_fw || !rdev->mc_fw) { + r = si_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); @@ -6454,14 +6885,8 @@ int si_init(struct radeon_device *rdev) ring->ring_obj = NULL; r600_ring_init(rdev, ring, 64 * 1024); - if (rdev->has_uvd) { - r = radeon_uvd_init(rdev); - if (!r) { - ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; - ring->ring_obj = NULL; - r600_ring_init(rdev, ring, 4096); - } - } + si_uvd_init(rdev); + si_vce_init(rdev); rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -6477,7 +6902,7 @@ int si_init(struct radeon_device *rdev) si_cp_fini(rdev); cayman_dma_fini(rdev); si_irq_fini(rdev); - si_rlc_fini(rdev); + sumo_rlc_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); radeon_vm_manager_fini(rdev); @@ -6500,18 +6925,23 @@ int si_init(struct radeon_device *rdev) void si_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); si_cp_fini(rdev); cayman_dma_fini(rdev); - si_irq_fini(rdev); - si_rlc_fini(rdev); - si_fini_cg(rdev); si_fini_pg(rdev); + si_fini_cg(rdev); + si_irq_fini(rdev); + sumo_rlc_fini(rdev); radeon_wb_fini(rdev); radeon_vm_manager_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); - if (rdev->has_uvd) + if (rdev->has_uvd) { + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + } + if (rdev->has_vce) + radeon_vce_fini(rdev); si_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); @@ -6537,7 +6967,7 @@ uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev) mutex_lock(&rdev->gpu_clock_mutex); WREG32(RLC_CAPTURE_GPU_CLOCK_COUNT, 1); clock = (uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_LSB) | - ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL); + ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL); mutex_unlock(&rdev->gpu_clock_mutex); return clock; } @@ -6556,8 +6986,7 @@ int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~UPLL_BYPASS_EN_MASK); if (!vclk || !dclk) { - /* keep the Bypass mode, put PLL to sleep */ - WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK); + /* keep the Bypass mode */ return 0; } @@ -6573,8 +7002,7 @@ int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) /* set VCO_MODE to 1 */ WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_VCO_MODE_MASK, ~UPLL_VCO_MODE_MASK); - /* toggle UPLL_SLEEP to 1 then back to 0 */ - WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK); + /* disable sleep mode */ WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_SLEEP_MASK); /* deassert UPLL_RESET */ @@ -6636,11 +7064,14 @@ int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) static void si_pcie_gen3_enable(struct radeon_device *rdev) { struct pci_dev *root = rdev->pdev->bus->self; - int bridge_pos, gpu_pos; - u32 speed_cntl, mask, current_data_rate; - int ret, i; + enum pci_bus_speed speed_cap; + u32 speed_cntl, current_data_rate; + int i; u16 tmp16; + if (pci_is_root_bus(rdev->pdev->bus)) + return; + if (radeon_pcie_gen2 == 0) return; @@ -6650,23 +7081,24 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) if (!(rdev->flags & RADEON_IS_PCIE)) return; - ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask); - if (ret != 0) + speed_cap = pcie_get_speed_cap(root); + if (speed_cap == PCI_SPEED_UNKNOWN) return; - if (!(mask & (DRM_PCIE_SPEED_50 | DRM_PCIE_SPEED_80))) + if ((speed_cap != PCIE_SPEED_8_0GT) && + (speed_cap != PCIE_SPEED_5_0GT)) return; speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >> LC_CURRENT_DATA_RATE_SHIFT; - if (mask & DRM_PCIE_SPEED_80) { + if (speed_cap == PCIE_SPEED_8_0GT) { if (current_data_rate == 2) { DRM_INFO("PCIE gen 3 link speeds already enabled\n"); return; } DRM_INFO("enabling PCIE gen 3 link speeds, disable with radeon.pcie_gen2=0\n"); - } else if (mask & DRM_PCIE_SPEED_50) { + } else if (speed_cap == PCIE_SPEED_5_0GT) { if (current_data_rate == 1) { DRM_INFO("PCIE gen 2 link speeds already enabled\n"); return; @@ -6674,29 +7106,18 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n"); } - bridge_pos = pci_pcie_cap(root); - if (!bridge_pos) - return; - - gpu_pos = pci_pcie_cap(rdev->pdev); - if (!gpu_pos) + if (!pci_is_pcie(root) || !pci_is_pcie(rdev->pdev)) return; - if (mask & DRM_PCIE_SPEED_80) { + if (speed_cap == PCIE_SPEED_8_0GT) { /* re-try equalization if gen3 is not already enabled */ if (current_data_rate != 2) { u16 bridge_cfg, gpu_cfg; u16 bridge_cfg2, gpu_cfg2; u32 max_lw, current_lw, tmp; - pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg); - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg); - - tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD; - pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16); - - tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD; - pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16); + pcie_capability_set_word(root, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD); + pcie_capability_set_word(rdev->pdev, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD); tmp = RREG32_PCIE(PCIE_LC_STATUS1); max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT; @@ -6714,15 +7135,23 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) for (i = 0; i < 10; i++) { /* check status */ - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_DEVSTA, &tmp16); + pcie_capability_read_word(rdev->pdev, + PCI_EXP_DEVSTA, + &tmp16); if (tmp16 & PCI_EXP_DEVSTA_TRPND) break; - pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg); - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg); + pcie_capability_read_word(root, PCI_EXP_LNKCTL, + &bridge_cfg); + pcie_capability_read_word(rdev->pdev, + PCI_EXP_LNKCTL, + &gpu_cfg); - pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &bridge_cfg2); - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &gpu_cfg2); + pcie_capability_read_word(root, PCI_EXP_LNKCTL2, + &bridge_cfg2); + pcie_capability_read_word(rdev->pdev, + PCI_EXP_LNKCTL2, + &gpu_cfg2); tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); tmp |= LC_SET_QUIESCE; @@ -6732,29 +7161,31 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) tmp |= LC_REDO_EQ; WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); - mdelay(100); + msleep(100); /* linkctl */ - pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &tmp16); - tmp16 &= ~PCI_EXP_LNKCTL_HAWD; - tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD); - pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16); - - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &tmp16); - tmp16 &= ~PCI_EXP_LNKCTL_HAWD; - tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD); - pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16); + pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL, + PCI_EXP_LNKCTL_HAWD, + bridge_cfg & + PCI_EXP_LNKCTL_HAWD); + pcie_capability_clear_and_set_word(rdev->pdev, PCI_EXP_LNKCTL, + PCI_EXP_LNKCTL_HAWD, + gpu_cfg & + PCI_EXP_LNKCTL_HAWD); /* linkctl2 */ - pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &tmp16); - tmp16 &= ~((1 << 4) | (7 << 9)); - tmp16 |= (bridge_cfg2 & ((1 << 4) | (7 << 9))); - pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, tmp16); - - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16); - tmp16 &= ~((1 << 4) | (7 << 9)); - tmp16 |= (gpu_cfg2 & ((1 << 4) | (7 << 9))); - pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16); + pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL2, + PCI_EXP_LNKCTL2_ENTER_COMP | + PCI_EXP_LNKCTL2_TX_MARGIN, + bridge_cfg2 & + (PCI_EXP_LNKCTL2_ENTER_COMP | + PCI_EXP_LNKCTL2_TX_MARGIN)); + pcie_capability_clear_and_set_word(rdev->pdev, PCI_EXP_LNKCTL2, + PCI_EXP_LNKCTL2_ENTER_COMP | + PCI_EXP_LNKCTL2_TX_MARGIN, + gpu_cfg2 & + (PCI_EXP_LNKCTL2_ENTER_COMP | + PCI_EXP_LNKCTL2_TX_MARGIN)); tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); tmp &= ~LC_SET_QUIESCE; @@ -6768,15 +7199,15 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE; WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); - pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16); - tmp16 &= ~0xf; - if (mask & DRM_PCIE_SPEED_80) - tmp16 |= 3; /* gen3 */ - else if (mask & DRM_PCIE_SPEED_50) - tmp16 |= 2; /* gen2 */ + tmp16 = 0; + if (speed_cap == PCIE_SPEED_8_0GT) + tmp16 |= PCI_EXP_LNKCTL2_TLS_8_0GT; /* gen3 */ + else if (speed_cap == PCIE_SPEED_5_0GT) + tmp16 |= PCI_EXP_LNKCTL2_TLS_5_0GT; /* gen2 */ else - tmp16 |= 1; /* gen1 */ - pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16); + tmp16 |= PCI_EXP_LNKCTL2_TLS_2_5GT; /* gen1 */ + pcie_capability_clear_and_set_word(rdev->pdev, PCI_EXP_LNKCTL2, + PCI_EXP_LNKCTL2_TLS, tmp16); speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE; @@ -6796,6 +7227,9 @@ static void si_program_aspm(struct radeon_device *rdev) bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false; bool disable_clkreq = false; + if (radeon_aspm == 0) + return; + if (!(rdev->flags & RADEON_IS_PCIE)) return; @@ -6915,7 +7349,8 @@ static void si_program_aspm(struct radeon_device *rdev) if (orig != data) WREG32_PIF_PHY1(PB1_PIF_CNTL, data); - if (!disable_clkreq) { + if (!disable_clkreq && + !pci_is_root_bus(rdev->pdev->bus)) { struct pci_dev *root = rdev->pdev->bus->self; u32 lnkcap; @@ -6990,3 +7425,124 @@ static void si_program_aspm(struct radeon_device *rdev) } } } + +static int si_vce_send_vcepll_ctlreq(struct radeon_device *rdev) +{ + unsigned i; + + /* make sure VCEPLL_CTLREQ is deasserted */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~UPLL_CTLREQ_MASK); + + mdelay(10); + + /* assert UPLL_CTLREQ */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, UPLL_CTLREQ_MASK, ~UPLL_CTLREQ_MASK); + + /* wait for CTLACK and CTLACK2 to get asserted */ + for (i = 0; i < 100; ++i) { + uint32_t mask = UPLL_CTLACK_MASK | UPLL_CTLACK2_MASK; + if ((RREG32_SMC(CG_VCEPLL_FUNC_CNTL) & mask) == mask) + break; + mdelay(10); + } + + /* deassert UPLL_CTLREQ */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~UPLL_CTLREQ_MASK); + + if (i == 100) { + DRM_ERROR("Timeout setting UVD clocks!\n"); + return -ETIMEDOUT; + } + + return 0; +} + +int si_set_vce_clocks(struct radeon_device *rdev, u32 evclk, u32 ecclk) +{ + unsigned fb_div = 0, evclk_div = 0, ecclk_div = 0; + int r; + + /* bypass evclk and ecclk with bclk */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL_2, + EVCLK_SRC_SEL(1) | ECCLK_SRC_SEL(1), + ~(EVCLK_SRC_SEL_MASK | ECCLK_SRC_SEL_MASK)); + + /* put PLL in bypass mode */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_BYPASS_EN_MASK, + ~VCEPLL_BYPASS_EN_MASK); + + if (!evclk || !ecclk) { + /* keep the Bypass mode, put PLL to sleep */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_SLEEP_MASK, + ~VCEPLL_SLEEP_MASK); + return 0; + } + + r = radeon_uvd_calc_upll_dividers(rdev, evclk, ecclk, 125000, 250000, + 16384, 0x03FFFFFF, 0, 128, 5, + &fb_div, &evclk_div, &ecclk_div); + if (r) + return r; + + /* set RESET_ANTI_MUX to 0 */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL_5, 0, ~RESET_ANTI_MUX_MASK); + + /* set VCO_MODE to 1 */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_VCO_MODE_MASK, + ~VCEPLL_VCO_MODE_MASK); + + /* toggle VCEPLL_SLEEP to 1 then back to 0 */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_SLEEP_MASK, + ~VCEPLL_SLEEP_MASK); + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~VCEPLL_SLEEP_MASK); + + /* deassert VCEPLL_RESET */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~VCEPLL_RESET_MASK); + + mdelay(1); + + r = si_vce_send_vcepll_ctlreq(rdev); + if (r) + return r; + + /* assert VCEPLL_RESET again */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_RESET_MASK, ~VCEPLL_RESET_MASK); + + /* disable spread spectrum. */ + WREG32_SMC_P(CG_VCEPLL_SPREAD_SPECTRUM, 0, ~SSEN_MASK); + + /* set feedback divider */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL_3, VCEPLL_FB_DIV(fb_div), ~VCEPLL_FB_DIV_MASK); + + /* set ref divider to 0 */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~VCEPLL_REF_DIV_MASK); + + /* set PDIV_A and PDIV_B */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL_2, + VCEPLL_PDIV_A(evclk_div) | VCEPLL_PDIV_B(ecclk_div), + ~(VCEPLL_PDIV_A_MASK | VCEPLL_PDIV_B_MASK)); + + /* give the PLL some time to settle */ + mdelay(15); + + /* deassert PLL_RESET */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~VCEPLL_RESET_MASK); + + mdelay(15); + + /* switch from bypass mode to normal mode */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~VCEPLL_BYPASS_EN_MASK); + + r = si_vce_send_vcepll_ctlreq(rdev); + if (r) + return r; + + /* switch VCLK and DCLK selection */ + WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL_2, + EVCLK_SRC_SEL(16) | ECCLK_SRC_SEL(16), + ~(EVCLK_SRC_SEL_MASK | ECCLK_SRC_SEL_MASK)); + + mdelay(100); + + return 0; +} |