/*
* GV100 FB
*
* Copyright (c) 2017-2018, NVIDIA CORPORATION. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*/
#include <nvgpu/types.h>
#include <nvgpu/dma.h>
#include <nvgpu/log.h>
#include <nvgpu/enabled.h>
#include <nvgpu/gmmu.h>
#include <nvgpu/nvgpu_common.h>
#include <nvgpu/kmem.h>
#include <nvgpu/nvgpu_mem.h>
#include <nvgpu/acr/nvgpu_acr.h>
#include <nvgpu/firmware.h>
#include <nvgpu/pmu.h>
#include <nvgpu/falcon.h>
#include <nvgpu/io.h>
#include <nvgpu/utils.h>
#include <nvgpu/timers.h>
#include <nvgpu/gk20a.h>
#include "fb_gv100.h"
#include <nvgpu/hw/gv100/hw_fb_gv100.h>
#include <nvgpu/hw/gv100/hw_mc_gv100.h>
#define HW_SCRUB_TIMEOUT_DEFAULT 100 /* usec */
#define HW_SCRUB_TIMEOUT_MAX 2000000 /* usec */
#define MEM_UNLOCK_TIMEOUT 3500 /* msec */
void gv100_fb_reset(struct gk20a *g)
{
u32 val;
int retries = HW_SCRUB_TIMEOUT_MAX / HW_SCRUB_TIMEOUT_DEFAULT;
nvgpu_log_info(g, "reset gv100 fb");
/* wait for memory to be accessible */
do {
u32 w = gk20a_readl(g, fb_niso_scrub_status_r());
if (fb_niso_scrub_status_flag_v(w)) {
nvgpu_log_info(g, "done");
break;
}
nvgpu_udelay(HW_SCRUB_TIMEOUT_DEFAULT);
} while (--retries);
val = gk20a_readl(g, fb_mmu_priv_level_mask_r());
val &= ~fb_mmu_priv_level_mask_write_violation_m();
gk20a_writel(g, fb_mmu_priv_level_mask_r(), val);
}
void gv100_fb_enable_hub_intr(struct gk20a *g)
{
u32 mask = 0;
mask = fb_niso_intr_en_set_mmu_other_fault_notify_m() |
fb_niso_intr_en_set_mmu_nonreplayable_fault_notify_m() |
fb_niso_intr_en_set_mmu_nonreplayable_fault_overflow_m() |
fb_niso_intr_en_set_mmu_replayable_fault_notify_m() |
fb_niso_intr_en_set_mmu_replayable_fault_overflow_m();
gk20a_writel(g, fb_niso_intr_en_set_r(0),
mask);
}
void gv100_fb_disable_hub_intr(struct gk20a *g)
{
u32 mask = 0;
mask = fb_niso_intr_en_set_mmu_other_fault_notify_m() |
fb_niso_intr_en_set_mmu_nonreplayable_fault_notify_m() |
fb_niso_intr_en_set_mmu_nonreplayable_fault_overflow_m() |
fb_niso_intr_en_set_mmu_replayable_fault_notify_m() |
fb_niso_intr_en_set_mmu_replayable_fault_overflow_m();
gk20a_writel(g, fb_niso_intr_en_clr_r(0),
mask);
}
/*
* @brief Patch signatures into ucode image
*/
static int gv100_fb_acr_ucode_patch_sig(struct gk20a *g,
u32 *p_img,
u32 *p_prod_sig,
u32 *p_dbg_sig,
u32 *p_patch_loc,
u32 *p_patch_ind)
{
u32 *p_sig;
if (!g->ops.pmu.is_debug_mode_enabled(g)) {
p_sig = p_prod_sig;
} else {
p_sig = p_dbg_sig;
}
/* Patching logic. We have just one location to patch. */
p_img[(*p_patch_loc>>2)] = p_sig[(*p_patch_ind<<2)];
p_img[(*p_patch_loc>>2)+1U] = p_sig[(*p_patch_ind<<2)+1U];
p_img[(*p_patch_loc>>2)+2U] = p_sig[(*p_patch_ind<<2)+2U];
p_img[(*p_patch_loc>>2)+3U] = p_sig[(*p_patch_ind<<2)+3U];
return 0;
}
int gv100_fb_memory_unlock(struct gk20a *g)
{
struct nvgpu_firmware *mem_unlock_fw = NULL;
struct bin_hdr *hsbin_hdr = NULL;
struct acr_fw_header *fw_hdr = NULL;
u32 *mem_unlock_ucode = NULL;
u32 *mem_unlock_ucode_header = NULL;
u32 sec_imem_dest = 0;
u32 val = 0;
int err = 0;
nvgpu_log_fn(g, " ");
nvgpu_log_info(g, "fb_mmu_vpr_info = 0x%08x",
gk20a_readl(g, fb_mmu_vpr_info_r()));
/*
* mem_unlock.bin should be written to install
* traps even if VPR isn’t actually supported
*/
mem_unlock_fw = nvgpu_request_firmware(g, "mem_unlock.bin", 0);
if (!mem_unlock_fw) {
nvgpu_err(g, "mem unlock ucode get fail");
err = -ENOENT;
goto exit;
}
/* Enable nvdec */
g->ops.mc.enable(g, mc_enable_nvdec_enabled_f());
/* nvdec falcon reset */
nvgpu_flcn_reset(&g->nvdec_flcn);
hsbin_hdr = (struct bin_hdr *)mem_unlock_fw->data;
fw_hdr = (struct acr_fw_header *)(mem_unlock_fw->data +
hsbin_hdr->header_offset);
mem_unlock_ucode_header = (u32 *)(mem_unlock_fw->data +
fw_hdr->hdr_offset);
mem_unlock_ucode = (u32 *)(mem_unlock_fw->data +
hsbin_hdr->data_offset);
/* Patch Ucode signatures */
if (gv100_fb_acr_ucode_patch_sig(g, mem_unlock_ucode,
(u32 *)(mem_unlock_fw->data + fw_hdr->sig_prod_offset),
(u32 *)(mem_unlock_fw->data + fw_hdr->sig_dbg_offset),
(u32 *)(mem_unlock_fw->data + fw_hdr->patch_loc),
(u32 *)(mem_unlock_fw->data + fw_hdr->patch_sig)) < 0) {
nvgpu_err(g, "mem unlock patch signatures fail");
err = -EPERM;
goto exit;
}
/* Clear interrupts */
nvgpu_flcn_set_irq(&g->nvdec_flcn, false, 0x0, 0x0);
/* Copy Non Secure IMEM code */
nvgpu_flcn_copy_to_imem(&g->nvdec_flcn, 0,
(u8 *)&mem_unlock_ucode[
mem_unlock_ucode_header[OS_CODE_OFFSET] >> 2],
mem_unlock_ucode_header[OS_CODE_SIZE], 0, false,
GET_IMEM_TAG(mem_unlock_ucode_header[OS_CODE_OFFSET]));
/* Put secure code after non-secure block */
sec_imem_dest = GET_NEXT_BLOCK(mem_unlock_ucode_header[OS_CODE_SIZE]);
nvgpu_flcn_copy_to_imem(&g->nvdec_flcn, sec_imem_dest,
(u8 *)&mem_unlock_ucode[
mem_unlock_ucode_header[APP_0_CODE_OFFSET] >> 2],
mem_unlock_ucode_header[APP_0_CODE_SIZE], 0, true,
GET_IMEM_TAG(mem_unlock_ucode_header[APP_0_CODE_OFFSET]));
/* load DMEM: ensure that signatures are patched */
nvgpu_flcn_copy_to_dmem(&g->nvdec_flcn, 0, (u8 *)&mem_unlock_ucode[
mem_unlock_ucode_header[OS_DATA_OFFSET] >> 2],
mem_unlock_ucode_header[OS_DATA_SIZE], 0);
/* Write non-zero value to mailbox register which is updated by
* mem_unlock bin to denote its return status.
*/
nvgpu_flcn_mailbox_write(&g->nvdec_flcn, 0, 0xdeadbeef);
/* set BOOTVEC to start of non-secure code */
nvgpu_flcn_bootstrap(&g->nvdec_flcn, 0);
/* wait for complete & halt */
nvgpu_flcn_wait_for_halt(&g->nvdec_flcn, MEM_UNLOCK_TIMEOUT);
/* check mem unlock status */
val = nvgpu_flcn_mailbox_read(&g->nvdec_flcn, 0);
if (val) {
nvgpu_err(g, "memory unlock failed, err %x", val);
nvgpu_flcn_dump_stats(&g->nvdec_flcn);
err = -1;
goto exit;
}
exit:
if (mem_unlock_fw) {
nvgpu_release_firmware(g, mem_unlock_fw);
}
nvgpu_log_fn(g, "done, status - %d", err);
return err;
}
int gv100_fb_init_nvlink(struct gk20a *g)
{
u32 data;
u32 mask = g->nvlink.enabled_links;
/* Map enabled link to SYSMEM */
data = nvgpu_readl(g, fb_hshub_config0_r());
data = set_field(data, fb_hshub_config0_sysmem_nvlink_mask_m(),
fb_hshub_config0_sysmem_nvlink_mask_f(mask));
nvgpu_writel(g, fb_hshub_config0_r(), data);
return 0;
}
int gv100_fb_enable_nvlink(struct gk20a *g)
{
u32 data;
nvgpu_log(g, gpu_dbg_nvlink|gpu_dbg_info, "enabling nvlink");
/* Enable nvlink for NISO FBHUB */
data = nvgpu_readl(g, fb_niso_cfg1_r());
data = set_field(data, fb_niso_cfg1_sysmem_nvlink_m(),
fb_niso_cfg1_sysmem_nvlink_enabled_f());
nvgpu_writel(g, fb_niso_cfg1_r(), data);
/* Setup atomics */
data = nvgpu_readl(g, fb_mmu_ctrl_r());
data = set_field(data, fb_mmu_ctrl_atomic_capability_mode_m(),
fb_mmu_ctrl_atomic_capability_mode_rmw_f());
nvgpu_writel(g, fb_mmu_ctrl_r(), data);
data = nvgpu_readl(g, fb_hsmmu_pri_mmu_ctrl_r());
data = set_field(data, fb_hsmmu_pri_mmu_ctrl_atomic_capability_mode_m(),
fb_hsmmu_pri_mmu_ctrl_atomic_capability_mode_rmw_f());
nvgpu_writel(g, fb_hsmmu_pri_mmu_ctrl_r(), data);
data = nvgpu_readl(g, fb_fbhub_num_active_ltcs_r());
data = set_field(data, fb_fbhub_num_active_ltcs_hub_sys_atomic_mode_m(),
fb_fbhub_num_active_ltcs_hub_sys_atomic_mode_use_rmw_f());
nvgpu_writel(g, fb_fbhub_num_active_ltcs_r(), data);
data = nvgpu_readl(g, fb_hshub_num_active_ltcs_r());
data = set_field(data, fb_hshub_num_active_ltcs_hub_sys_atomic_mode_m(),
fb_hshub_num_active_ltcs_hub_sys_atomic_mode_use_rmw_f());
nvgpu_writel(g, fb_hshub_num_active_ltcs_r(), data);
return 0;
}
size_t gv100_fb_get_vidmem_size(struct gk20a *g)
{
u32 range = gk20a_readl(g, fb_mmu_local_memory_range_r());
u32 mag = fb_mmu_local_memory_range_lower_mag_v(range);
u32 scale = fb_mmu_local_memory_range_lower_scale_v(range);
u32 ecc = fb_mmu_local_memory_range_ecc_mode_v(range);
size_t bytes = ((size_t)mag << scale) * SZ_1M;
if (ecc) {
bytes = bytes / 16U * 15U;
}
return bytes;
}