gpu: nvgpu: gm20b: Use new delay APIs

Use platform agnostic delay functions instead of Linux kernel APIs.
This allows removing dependency to Linux header linux/delay.h. At the
same time remove #include lines for other unused Linux headers.

JIRA NVGPU-16

Change-Id: I05df9d72edaf4bb061febe0cb40fc8a7cf9f51c7
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1460114
Reviewed-by: svccoveritychecker <svccoveritychecker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Alex Waterman <alexw@nvidia.com>

14 files changed, 27 insertions, 59 deletions

diff --git a/drivers/gpu/nvgpu/gm20b/acr_gm20b.c b/drivers/gpu/nvgpu/gm20b/acr_gm20b.c
index 386b266d..c593dd7c 100644
--- a/drivers/gpu/nvgpu/gm20b/acr_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/acr_gm20b.c
@@ -11,12 +11,8 @@
  * more details.
  */
 
-#include <linux/delay.h>        /* for mdelay */
 #include <linux/firmware.h>
-#include <linux/module.h>
 #include <linux/debugfs.h>
-#include <linux/dma-mapping.h>
-#include <linux/io.h>
 
 #include <linux/platform/tegra/mc.h>
 
@@ -1501,7 +1497,7 @@ static int pmu_wait_for_halt(struct gk20a *g, unsigned int timeout_ms)
                         ret = 0;
                         break;
                 }
-                udelay(1);
+                nvgpu_udelay(1);
         } while (!nvgpu_timeout_expired(&timeout));
 
         if (ret) {
@@ -1543,7 +1539,7 @@ static int clear_halt_interrupt_status(struct gk20a *g, unsigned int timeout_ms)
                         /*halt irq is clear*/
                         return 0;
 
-                udelay(1);
+                nvgpu_udelay(1);
         } while (!nvgpu_timeout_expired(&timeout));
 
         return -ETIMEDOUT;
diff --git a/drivers/gpu/nvgpu/gm20b/bus_gm20b.c b/drivers/gpu/nvgpu/gm20b/bus_gm20b.c
index 08761ef7..7b414196 100644
--- a/drivers/gpu/nvgpu/gm20b/bus_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/bus_gm20b.c
@@ -13,8 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>
-
 #include <nvgpu/timers.h>
 
 #include "bus_gm20b.h"
@@ -46,7 +44,7 @@ static int gm20b_bus_bar1_bind(struct gk20a *g, struct nvgpu_mem *bar1_inst)
                 if (!pending && !outstanding)
                         break;
 
-                udelay(5);
+                nvgpu_udelay(5);
         } while (!nvgpu_timeout_expired(&timeout));
 
         if (nvgpu_timeout_peek_expired(&timeout))
diff --git a/drivers/gpu/nvgpu/gm20b/clk_gm20b.c b/drivers/gpu/nvgpu/gm20b/clk_gm20b.c
index 4d7470d4..358e7369 100644
--- a/drivers/gpu/nvgpu/gm20b/clk_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/clk_gm20b.c
@@ -18,8 +18,6 @@
 
 #include <linux/version.h>
 #include <linux/clk.h>
-#include <linux/delay.h>        /* for mdelay */
-#include <linux/module.h>
 #include <linux/debugfs.h>
 #include <linux/uaccess.h>
 #include <linux/clk/tegra.h>
@@ -403,7 +401,7 @@ static void clk_set_dfs_coeff(struct gk20a *g, u32 dfs_coeff)
         gk20a_writel(g, trim_sys_gpcpll_dvfs0_r(), data);
 
         data = gk20a_readl(g, trim_gpc_bcast_gpcpll_dvfs2_r());
-        udelay(1);
+        nvgpu_udelay(1);
         data &= ~DFS_EXT_STROBE;
         gk20a_writel(g, trim_gpc_bcast_gpcpll_dvfs2_r(), data);
 }
@@ -420,7 +418,7 @@ static void __maybe_unused clk_set_dfs_det_max(struct gk20a *g, u32 dfs_det_max)
         gk20a_writel(g, trim_sys_gpcpll_dvfs0_r(), data);
 
         data = gk20a_readl(g, trim_gpc_bcast_gpcpll_dvfs2_r());
-        udelay(1);
+        nvgpu_udelay(1);
         data &= ~DFS_EXT_STROBE;
         gk20a_writel(g, trim_gpc_bcast_gpcpll_dvfs2_r(), data);
 }
@@ -435,7 +433,7 @@ static void clk_set_dfs_ext_cal(struct gk20a *g, u32 dfs_det_cal)
         gk20a_writel(g, trim_gpc_bcast_gpcpll_dvfs2_r(), data);
 
         data = gk20a_readl(g, trim_sys_gpcpll_dvfs1_r());
-        udelay(1);
+        nvgpu_udelay(1);
         if (~trim_sys_gpcpll_dvfs1_dfs_ctrl_v(data) & DFS_EXT_CAL_EN) {
                 data = set_field(data, trim_sys_gpcpll_dvfs1_dfs_ctrl_m(),
                                  trim_sys_gpcpll_dvfs1_dfs_ctrl_f(
@@ -460,7 +458,7 @@ static void clk_setup_dvfs_detection(struct gk20a *g, struct pll *gpll)
         gk20a_writel(g, trim_sys_gpcpll_dvfs0_r(), data);
 
         data = gk20a_readl(g, trim_gpc_bcast_gpcpll_dvfs2_r());
-        udelay(1);
+        nvgpu_udelay(1);
         data &= ~DFS_EXT_STROBE;
         gk20a_writel(g, trim_gpc_bcast_gpcpll_dvfs2_r(), data);
 
@@ -505,7 +503,7 @@ static int clk_enbale_pll_dvfs(struct gk20a *g)
                          trim_sys_gpcpll_cfg_iddq_power_on_v());
         gk20a_writel(g, trim_sys_gpcpll_cfg_r(), data);
         gk20a_readl(g, trim_sys_gpcpll_cfg_r());
-        udelay(delay);
+        nvgpu_udelay(delay);
 
         /*
          * Dynamic ramp setup based on update rate, which in DVFS mode on GM20b
@@ -530,7 +528,7 @@ static int clk_enbale_pll_dvfs(struct gk20a *g)
                 data = gk20a_readl(g, trim_sys_gpcpll_dvfs1_r());
                 if (trim_sys_gpcpll_dvfs1_dfs_cal_done_v(data))
                         break;
-                udelay(1);
+                nvgpu_udelay(1);
                 delay--;
         } while (delay > 0);
 
@@ -626,13 +624,13 @@ static int clk_slide_gpc_pll(struct gk20a *g, struct pll *gpll)
                 coeff = gk20a_readl(g, trim_sys_gpcpll_coeff_r());
                 coeff = set_field(coeff, trim_sys_gpcpll_coeff_ndiv_m(),
                         trim_sys_gpcpll_coeff_ndiv_f(gpll->dvfs.n_int));
-                udelay(1);
+                nvgpu_udelay(1);
                 gk20a_writel(g, trim_sys_gpcpll_coeff_r(), coeff);
         } else {
                 coeff = gk20a_readl(g, trim_sys_gpcpll_coeff_r());
                 coeff = set_field(coeff, trim_sys_gpcpll_coeff_ndiv_m(),
                                 trim_sys_gpcpll_coeff_ndiv_f(gpll->N));
-                udelay(1);
+                nvgpu_udelay(1);
                 gk20a_writel(g, trim_sys_gpcpll_coeff_r(), coeff);
         }
 
@@ -641,11 +639,11 @@ static int clk_slide_gpc_pll(struct gk20a *g, struct pll *gpll)
         data = set_field(data,
                         trim_sys_gpcpll_ndiv_slowdown_en_dynramp_m(),
                         trim_sys_gpcpll_ndiv_slowdown_en_dynramp_yes_f());
-        udelay(1);
+        nvgpu_udelay(1);
         gk20a_writel(g, trim_sys_gpcpll_ndiv_slowdown_r(), data);
 
         do {
-                udelay(1);
+                nvgpu_udelay(1);
                 ramp_timeout--;
                 data = gk20a_readl(
                         g, trim_gpc_bcast_gpcpll_ndiv_slowdown_debug_r());
@@ -692,14 +690,14 @@ static int clk_change_pldiv_under_bypass(struct gk20a *g, struct pll *gpll)
 
         /* change PLDIV */
         coeff = gk20a_readl(g, trim_sys_gpcpll_coeff_r());
-        udelay(1);
+        nvgpu_udelay(1);
         coeff = set_field(coeff, trim_sys_gpcpll_coeff_pldiv_m(),
                           trim_sys_gpcpll_coeff_pldiv_f(gpll->PL));
         gk20a_writel(g, trim_sys_gpcpll_coeff_r(), coeff);
 
         /* put PLL back on vco */
         data = gk20a_readl(g, trim_sys_sel_vco_r());
-        udelay(1);
+        nvgpu_udelay(1);
         data = set_field(data, trim_sys_sel_vco_gpc2clk_out_m(),
                 trim_sys_sel_vco_gpc2clk_out_vco_f());
         gk20a_writel(g, trim_sys_sel_vco_r(), data);
@@ -718,14 +716,14 @@ static int clk_lock_gpc_pll_under_bypass(struct gk20a *g, struct pll *gpll)
         gk20a_writel(g, trim_sys_sel_vco_r(), data);
 
         cfg = gk20a_readl(g, trim_sys_gpcpll_cfg_r());
-        udelay(1);
+        nvgpu_udelay(1);
         if (trim_sys_gpcpll_cfg_iddq_v(cfg)) {
                 /* get out from IDDQ (1st power up) */
                 cfg = set_field(cfg, trim_sys_gpcpll_cfg_iddq_m(),
                                 trim_sys_gpcpll_cfg_iddq_power_on_v());
                 gk20a_writel(g, trim_sys_gpcpll_cfg_r(), cfg);
                 gk20a_readl(g, trim_sys_gpcpll_cfg_r());
-                udelay(gpc_pll_params.iddq_exit_delay);
+                nvgpu_udelay(gpc_pll_params.iddq_exit_delay);
         } else {
                 /* clear SYNC_MODE before disabling PLL */
                 cfg = set_field(cfg, trim_sys_gpcpll_cfg_sync_mode_m(),
@@ -769,7 +767,7 @@ static int clk_lock_gpc_pll_under_bypass(struct gk20a *g, struct pll *gpll)
         /* just delay in DVFS mode (lock cannot be used) */
         if (gpll->mode == GPC_PLL_MODE_DVFS) {
                 gk20a_readl(g, trim_sys_gpcpll_cfg_r());
-                udelay(gpc_pll_params.na_lock_delay);
+                nvgpu_udelay(gpc_pll_params.na_lock_delay);
                 gk20a_dbg_clk("NA config_pll under bypass: %u (%u) kHz %d mV",
                               gpll->freq, gpll->freq / 2,
                               (trim_sys_gpcpll_cfg3_dfs_testout_v(
@@ -791,7 +789,7 @@ static int clk_lock_gpc_pll_under_bypass(struct gk20a *g, struct pll *gpll)
         /* wait pll lock */
         timeout = gpc_pll_params.lock_timeout + 1;
         do {
-                udelay(1);
+                nvgpu_udelay(1);
                 cfg = gk20a_readl(g, trim_sys_gpcpll_cfg_r());
                 if (cfg & trim_sys_gpcpll_cfg_pll_lock_true_f())
                         goto pll_locked;
@@ -886,7 +884,7 @@ static int clk_program_gpc_pll(struct gk20a *g, struct pll *gpll_new,
                 /* Intentional 2nd write to assure linear divider operation */
                 gk20a_writel(g, trim_sys_gpc2clk_out_r(), data);
                 gk20a_readl(g, trim_sys_gpc2clk_out_r());
-                udelay(2);
+                nvgpu_udelay(2);
         }
 
 #if PLDIV_GLITCHLESS
@@ -942,7 +940,7 @@ set_pldiv:
             trim_sys_gpc2clk_out_vcodiv_by1_f()) {
                 data = set_field(data, trim_sys_gpc2clk_out_vcodiv_m(),
                                  trim_sys_gpc2clk_out_vcodiv_by1_f());
-                udelay(2);
+                nvgpu_udelay(2);
                 gk20a_writel(g, trim_sys_gpc2clk_out_r(), data);
                 /* Intentional 2nd write to assure linear divider operation */
                 gk20a_writel(g, trim_sys_gpc2clk_out_r(), data);
@@ -1787,10 +1785,10 @@ static int monitor_get(void *data, u64 *val)
         /* It should take less than 25us to finish 800 cycle of 38.4MHz.
            But longer than 100us delay is required here. */
         gk20a_readl(g, trim_gpc_clk_cntr_ncgpcclk_cfg_r(0));
-        udelay(200);
+        nvgpu_udelay(200);
 
         count1 = gk20a_readl(g, trim_gpc_clk_cntr_ncgpcclk_cnt_r(0));
-        udelay(100);
+        nvgpu_udelay(100);
         count2 = gk20a_readl(g, trim_gpc_clk_cntr_ncgpcclk_cnt_r(0));
         freq *= trim_gpc_clk_cntr_ncgpcclk_cnt_value_v(count2);
         do_div(freq, ncycle);
diff --git a/drivers/gpu/nvgpu/gm20b/fifo_gm20b.c b/drivers/gpu/nvgpu/gm20b/fifo_gm20b.c
index af2a8cd2..ed3a4b98 100644
--- a/drivers/gpu/nvgpu/gm20b/fifo_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/fifo_gm20b.c
@@ -13,8 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>
-
 #include "gk20a/gk20a.h"
 #include "gk20a/fifo_gk20a.h"
 
@@ -103,7 +101,7 @@ static void gm20b_fifo_trigger_mmu_fault(struct gk20a *g,
                         break;
                 }
 
-                usleep_range(delay, delay * 2);
+                nvgpu_usleep_range(delay, delay * 2);
                 delay = min_t(u32, delay << 1, GR_IDLE_CHECK_MAX);
         } while (!nvgpu_timeout_expired(&timeout));
 
diff --git a/drivers/gpu/nvgpu/gm20b/gr_gm20b.c b/drivers/gpu/nvgpu/gm20b/gr_gm20b.c
index 74c35c02..7400a9e4 100644
--- a/drivers/gpu/nvgpu/gm20b/gr_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/gr_gm20b.c
@@ -13,9 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>
-#include <linux/io.h>
-#include <linux/version.h>
 #include <soc/tegra/fuse.h>
 
 #include <dt-bindings/soc/gm20b-fuse.h>
diff --git a/drivers/gpu/nvgpu/gm20b/pmu_gm20b.c b/drivers/gpu/nvgpu/gm20b/pmu_gm20b.c
index 730c5eaf..109b3ca5 100644
--- a/drivers/gpu/nvgpu/gm20b/pmu_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/pmu_gm20b.c
@@ -13,7 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>        /* for udelay */
 #include <soc/tegra/fuse.h>
 
 #include "gk20a/gk20a.h"
@@ -191,7 +190,7 @@ static int pmu_gm20b_ctx_wait_lsf_ready(struct gk20a *g, u32 timeout_ms,
                 reg = gk20a_readl(g, gr_fecs_ctxsw_mailbox_r(0));
                 if (reg == val)
                         return 0;
-                udelay(delay);
+                nvgpu_udelay(delay);
         } while (!nvgpu_timeout_expired(&timeout));
 
         return -ETIMEDOUT;
diff --git a/drivers/gpu/nvgpu/gm20b/regops_gm20b.c b/drivers/gpu/nvgpu/gm20b/regops_gm20b.c
index e06ee869..45e7ec3d 100644
--- a/drivers/gpu/nvgpu/gm20b/regops_gm20b.c
+++ b/drivers/gpu/nvgpu/gm20b/regops_gm20b.c
@@ -16,9 +16,6 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <linux/err.h>
-#include <uapi/linux/nvgpu.h>
-
 #include "gk20a/gk20a.h"
 #include "gk20a/dbg_gpu_gk20a.h"
 #include "gk20a/regops_gk20a.h"
diff --git a/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c b/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
index bc9fc843..019b2471 100644
--- a/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/fifo_gp10b.c
@@ -13,8 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>
-
 #include <nvgpu/dma.h>
 #include <nvgpu/bug.h>
 
diff --git a/drivers/gpu/nvgpu/gp10b/gr_gp10b.c b/drivers/gpu/nvgpu/gp10b/gr_gp10b.c
index 6e6223c8..24e7ffad 100644
--- a/drivers/gpu/nvgpu/gp10b/gr_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/gr_gp10b.c
@@ -13,8 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>
-#include <linux/version.h>
 #include <soc/tegra/fuse.h>
 
 #include <dt-bindings/soc/gm20b-fuse.h>
@@ -1393,7 +1391,7 @@ static int gr_gp10b_wait_empty(struct gk20a *g, unsigned long duration_ms,
                         return 0;
                 }
 
-                usleep_range(delay, delay * 2);
+                nvgpu_usleep_range(delay, delay * 2);
                 delay = min_t(u32, delay << 1, GR_IDLE_CHECK_MAX);
         } while (!nvgpu_timeout_expired(&timeout));
 
@@ -2044,7 +2042,7 @@ static int gr_gp10b_suspend_contexts(struct gk20a *g,
                         if (!gr_ctx->t18x.cilp_preempt_pending)
                                 break;
 
-                        usleep_range(delay, delay * 2);
+                        nvgpu_usleep_range(delay, delay * 2);
                         delay = min_t(u32, delay << 1, GR_IDLE_CHECK_MAX);
                 } while (!nvgpu_timeout_expired(&timeout));
 
diff --git a/drivers/gpu/nvgpu/gp10b/hal_gp10b.c b/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
index ebba76b2..6990b4ef 100644
--- a/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/hal_gp10b.c
@@ -13,9 +13,6 @@
  * more details.
  */
 
-#include <linux/printk.h>
-#include <linux/version.h>
-
 #include "gk20a/gk20a.h"
 #include "gk20a/dbg_gpu_gk20a.h"
 #include "gk20a/css_gr_gk20a.h"
diff --git a/drivers/gpu/nvgpu/gp10b/mm_gp10b.c b/drivers/gpu/nvgpu/gp10b/mm_gp10b.c
index 2f894435..c2e0fddf 100644
--- a/drivers/gpu/nvgpu/gp10b/mm_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/mm_gp10b.c
@@ -13,9 +13,6 @@
  * more details.
  */
 
-#include <linux/pm_runtime.h>
-#include <linux/dma-mapping.h>
-
 #include <nvgpu/dma.h>
 
 #include "gk20a/gk20a.h"
diff --git a/drivers/gpu/nvgpu/gp10b/pmu_gp10b.c b/drivers/gpu/nvgpu/gp10b/pmu_gp10b.c
index 2d9882c9..1167006d 100644
--- a/drivers/gpu/nvgpu/gp10b/pmu_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/pmu_gp10b.c
@@ -13,7 +13,6 @@
  * more details.
  */
 
-#include <linux/delay.h>        /* for udelay */
 #include <soc/tegra/fuse.h>
 
 #include "gk20a/gk20a.h"
diff --git a/drivers/gpu/nvgpu/gp10b/regops_gp10b.c b/drivers/gpu/nvgpu/gp10b/regops_gp10b.c
index 33e331a4..885221df 100644
--- a/drivers/gpu/nvgpu/gp10b/regops_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/regops_gp10b.c
@@ -16,7 +16,6 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <linux/err.h>
 #include <uapi/linux/nvgpu.h>
 
 #include "gk20a/gk20a.h"
diff --git a/drivers/gpu/nvgpu/gp10b/rpfb_gp10b.c b/drivers/gpu/nvgpu/gp10b/rpfb_gp10b.c
index 77c6853c..0fb4248c 100644
--- a/drivers/gpu/nvgpu/gp10b/rpfb_gp10b.c
+++ b/drivers/gpu/nvgpu/gp10b/rpfb_gp10b.c
@@ -13,9 +13,6 @@
  * more details.
  */
 
-#include <linux/pm_runtime.h>
-#include <linux/dma-mapping.h>
-
 #include <nvgpu/dma.h>
 
 #include "gk20a/gk20a.h"