path: root/drivers/gpu/drm/i915/vlv_suspend.c

// SPDX-License-Identifier: MIT
/*
 * Copyright © 2020 Intel Corporation
 */

#include <linux/string_helpers.h>
#include <linux/kernel.h>

#include <drm/drm_print.h>

#include "i915_drv.h"
#include "i915_reg.h"
#include "i915_trace.h"
#include "i915_utils.h"
#include "intel_clock_gating.h"
#include "vlv_suspend.h"

#include "gt/intel_gt_regs.h"

struct vlv_s0ix_state {
	/* GAM */
	u32 wr_watermark;
	u32 gfx_prio_ctrl;
	u32 arb_mode;
	u32 gfx_pend_tlb0;
	u32 gfx_pend_tlb1;
	u32 lra_limits[GEN7_LRA_LIMITS_REG_NUM];
	u32 media_max_req_count;
	u32 gfx_max_req_count;
	u32 render_hwsp;
	u32 ecochk;
	u32 bsd_hwsp;
	u32 blt_hwsp;
	u32 tlb_rd_addr;

	/* MBC */
	u32 g3dctl;
	u32 gsckgctl;
	u32 mbctl;

	/* GCP */
	u32 ucgctl1;
	u32 ucgctl3;
	u32 rcgctl1;
	u32 rcgctl2;
	u32 rstctl;
	u32 misccpctl;

	/* GPM */
	u32 gfxpause;
	u32 rpdeuhwtc;
	u32 rpdeuc;
	u32 ecobus;
	u32 pwrdwnupctl;
	u32 rp_down_timeout;
	u32 rp_deucsw;
	u32 rcubmabdtmr;
	u32 rcedata;
	u32 spare2gh;

	/* Display 1 CZ domain */
	u32 gt_imr;
	u32 gt_ier;
	u32 pm_imr;
	u32 pm_ier;
	u32 gt_scratch[GEN7_GT_SCRATCH_REG_NUM];

	/* GT SA CZ domain */
	u32 tilectl;
	u32 gt_fifoctl;
	u32 gtlc_wake_ctrl;
	u32 gtlc_survive;
	u32 pmwgicz;

	/* Display 2 CZ domain */
	u32 gu_ctl0;
	u32 gu_ctl1;
	u32 pcbr;
	u32 clock_gate_dis2;
};

/*
 * Save all Gunit registers that may be lost after a D3 and a subsequent
 * S0i[R123] transition. The list of registers needing a save/restore is
 * defined in the VLV2_S0IXRegs document. This documents marks all Gunit
 * registers in the following way:
 * - Driver: saved/restored by the driver
 * - Punit : saved/restored by the Punit firmware
 * - No, w/o marking: no need to save/restore, since the register is R/O or
 *                    used internally by the HW in a way that doesn't depend
 *                    keeping the content across a suspend/resume.
 * - Debug : used for debugging
 *
 * We save/restore all registers marked with 'Driver', with the following
 * exceptions:
 * - Registers out of use, including also registers marked with 'Debug'.
 *   These have no effect on the driver's operation, so we don't save/restore
 *   them to reduce the overhead.
 * - Registers that are fully setup by an initialization function called from
 *   the resume path. For example many clock gating and RPS/RC6 registers.
 * - Registers that provide the right functionality with their reset defaults.
 *
 * TODO: Except for registers that based on the above 3 criteria can be safely
 * ignored, we save/restore all others, practically treating the HW context as
 * a black-box for the driver. Further investigation is needed to reduce the
 * saved/restored registers even further, by following the same 3 criteria.
 */
static void vlv_save_gunit_s0ix_state(struct drm_i915_private *i915)
{
	struct vlv_s0ix_state *s = i915->vlv_s0ix_state;
	struct intel_uncore *uncore = &i915->uncore;
	int i;

	if (!s)
		return;

	/* GAM 0x4000-0x4770 */
	s->wr_watermark = intel_uncore_read(uncore, GEN7_WR_WATERMARK);
	s->gfx_prio_ctrl = intel_uncore_read(uncore, GEN7_GFX_PRIO_CTRL);
	s->arb_mode = intel_uncore_read(uncore, ARB_MODE);
	s->gfx_pend_tlb0 = intel_uncore_read(uncore, GEN7_GFX_PEND_TLB0);
	s->gfx_pend_tlb1 = intel_uncore_read(uncore, GEN7_GFX_PEND_TLB1);

	for (i = 0; i < ARRAY_SIZE(s->lra_limits); i++)
		s->lra_limits[i] = intel_uncore_read(uncore, GEN7_LRA_LIMITS(i));

	s->media_max_req_count = intel_uncore_read(uncore, GEN7_MEDIA_MAX_REQ_COUNT);
	s->gfx_max_req_count = intel_uncore_read(uncore, GEN7_GFX_MAX_REQ_COUNT);

	s->render_hwsp = intel_uncore_read(uncore, RENDER_HWS_PGA_GEN7);
	s->ecochk = intel_uncore_read(uncore, GAM_ECOCHK);
	s->bsd_hwsp = intel_uncore_read(uncore, BSD_HWS_PGA_GEN7);
	s->blt_hwsp = intel_uncore_read(uncore, BLT_HWS_PGA_GEN7);

	s->tlb_rd_addr = intel_uncore_read(uncore, GEN7_TLB_RD_ADDR);

	/* MBC 0x9024-0x91D0, 0x8500 */
	s->g3dctl = intel_uncore_read(uncore, VLV_G3DCTL);
	s->gsckgctl = intel_uncore_read(uncore, VLV_GSCKGCTL);
	s->mbctl = intel_uncore_read(uncore, GEN6_MBCTL);

	/* GCP 0x9400-0x9424, 0x8100-0x810C */
	s->ucgctl1 = intel_uncore_read(uncore, GEN6_UCGCTL1);
	s->ucgctl3 = intel_uncore_read(uncore, GEN6_UCGCTL3);
	s->rcgctl1 = intel_uncore_read(uncore, GEN6_RCGCTL1);
	s->rcgctl2 = intel_uncore_read(uncore, GEN6_RCGCTL2);
	s->rstctl = intel_uncore_read(uncore, GEN6_RSTCTL);
	s->misccpctl = intel_uncore_read(uncore, GEN7_MISCCPCTL);

	/* GPM 0xA000-0xAA84, 0x8000-0x80FC */
	s->gfxpause = intel_uncore_read(uncore, GEN6_GFXPAUSE);
	s->rpdeuhwtc = intel_uncore_read(uncore, GEN6_RPDEUHWTC);
	s->rpdeuc = intel_uncore_read(uncore, GEN6_RPDEUC);
	s->ecobus = intel_uncore_read(uncore, ECOBUS);
	s->pwrdwnupctl = intel_uncore_read(uncore, VLV_PWRDWNUPCTL);
	s->rp_down_timeout = intel_uncore_read(uncore, GEN6_RP_DOWN_TIMEOUT);
	s->rp_deucsw = intel_uncore_read(uncore, GEN6_RPDEUCSW);
	s->rcubmabdtmr = intel_uncore_read(uncore, GEN6_RCUBMABDTMR);
	s->rcedata = intel_uncore_read(uncore, VLV_RCEDATA);
	s->spare2gh = intel_uncore_read(uncore, VLV_SPAREG2H);

	/* Display CZ domain, 0x4400C-0x4402C, 0x4F000-0x4F11F */
	s->gt_imr = intel_uncore_read(uncore, GTIMR);
	s->gt_ier = intel_uncore_read(uncore, GTIER);
	s->pm_imr = intel_uncore_read(uncore, GEN6_PMIMR);
	s->pm_ier = intel_uncore_read(uncore, GEN6_PMIER);

	for (i = 0; i < ARRAY_SIZE(s->gt_scratch); i++)
		s->gt_scratch[i] = intel_uncore_read(uncore, GEN7_GT_SCRATCH(i));

	/* GT SA CZ domain, 0x100000-0x138124 */
	s->tilectl = intel_uncore_read(uncore, TILECTL);
	s->gt_fifoctl = intel_uncore_read(uncore, GTFIFOCTL);
	s->gtlc_wake_ctrl = intel_uncore_read(uncore, VLV_GTLC_WAKE_CTRL);
	s->gtlc_survive = intel_uncore_read(uncore, VLV_GTLC_SURVIVABILITY_REG);
	s->pmwgicz = intel_uncore_read(uncore, VLV_PMWGICZ);

	/* Gunit-Display CZ domain, 0x182028-0x1821CF */
	s->gu_ctl0 = intel_uncore_read(uncore, VLV_GU_CTL0);
	s->gu_ctl1 = intel_uncore_read(uncore, VLV_GU_CTL1);
	s->pcbr = intel_uncore_read(uncore, VLV_PCBR);
	s->clock_gate_dis2 = intel_uncore_read(uncore, VLV_GUNIT_CLOCK_GATE2);

	/*
	 * Not saving any of:
	 * DFT,		0x9800-0x9EC0
	 * SARB,	0xB000-0xB1FC
	 * GAC,		0x5208-0x524C, 0x14000-0x14C000
	 * PCI CFG
	 */
}

static void vlv_restore_gunit_s0ix_state(struct drm_i915_private *i915)
{
	struct vlv_s0ix_state *s = i915->vlv_s0ix_state;
	struct intel_uncore *uncore = &i915->uncore;
	int i;

	if (!s)
		return;

	/* GAM 0x4000-0x4770 */
	intel_uncore_write(uncore, GEN7_WR_WATERMARK, s->wr_watermark);
	intel_uncore_write(uncore, GEN7_GFX_PRIO_CTRL, s->gfx_prio_ctrl);
	intel_uncore_write(uncore, ARB_MODE, s->arb_mode | (0xffff << 16));
	intel_uncore_write(uncore, GEN7_GFX_PEND_TLB0, s->gfx_pend_tlb0);
	intel_uncore_write(uncore, GEN7_GFX_PEND_TLB1, s->gfx_pend_tlb1);

	for (i = 0; i < ARRAY_SIZE(s->lra_limits); i++)
		intel_uncore_write(uncore, GEN7_LRA_LIMITS(i), s->lra_limits[i]);

	intel_uncore_write(uncore, GEN7_MEDIA_MAX_REQ_COUNT, s->media_max_req_count);
	intel_uncore_write(uncore, GEN7_GFX_MAX_REQ_COUNT, s->gfx_max_req_count);

	intel_uncore_write(uncore, RENDER_HWS_PGA_GEN7, s->render_hwsp);
	intel_uncore_write(uncore, GAM_ECOCHK, s->ecochk);
	intel_uncore_write(uncore, BSD_HWS_PGA_GEN7, s->bsd_hwsp);
	intel_uncore_write(uncore, BLT_HWS_PGA_GEN7, s->blt_hwsp);

	intel_uncore_write(uncore, GEN7_TLB_RD_ADDR, s->tlb_rd_addr);

	/* MBC 0x9024-0x91D0, 0x8500 */
	intel_uncore_write(uncore, VLV_G3DCTL, s->g3dctl);
	intel_uncore_write(uncore, VLV_GSCKGCTL, s->gsckgctl);
	intel_uncore_write(uncore, GEN6_MBCTL, s->mbctl);

	/* GCP 0x9400-0x9424, 0x8100-0x810C */
	intel_uncore_write(uncore, GEN6_UCGCTL1, s->ucgctl1);
	intel_uncore_write(uncore, GEN6_UCGCTL3, s->ucgctl3);
	intel_uncore_write(uncore, GEN6_RCGCTL1, s->rcgctl1);
	intel_uncore_write(uncore, GEN6_RCGCTL2, s->rcgctl2);
	intel_uncore_write(uncore, GEN6_RSTCTL, s->rstctl);
	intel_uncore_write(uncore, GEN7_MISCCPCTL, s->misccpctl);

	/* GPM 0xA000-0xAA84, 0x8000-0x80FC */
	intel_uncore_write(uncore, GEN6_GFXPAUSE, s->gfxpause);
	intel_uncore_write(uncore, GEN6_RPDEUHWTC, s->rpdeuhwtc);
	intel_uncore_write(uncore, GEN6_RPDEUC, s->rpdeuc);
	intel_uncore_write(uncore, ECOBUS, s->ecobus);
	intel_uncore_write(uncore, VLV_PWRDWNUPCTL, s->pwrdwnupctl);
	intel_uncore_write(uncore, GEN6_RP_DOWN_TIMEOUT, s->rp_down_timeout);
	intel_uncore_write(uncore, GEN6_RPDEUCSW, s->rp_deucsw);
	intel_uncore_write(uncore, GEN6_RCUBMABDTMR, s->rcubmabdtmr);
	intel_uncore_write(uncore, VLV_RCEDATA, s->rcedata);
	intel_uncore_write(uncore, VLV_SPAREG2H, s->spare2gh);

	/* Display CZ domain, 0x4400C-0x4402C, 0x4F000-0x4F11F */
	intel_uncore_write(uncore, GTIMR, s->gt_imr);
	intel_uncore_write(uncore, GTIER, s->gt_ier);
	intel_uncore_write(uncore, GEN6_PMIMR, s->pm_imr);
	intel_uncore_write(uncore, GEN6_PMIER, s->pm_ier);

	for (i = 0; i < ARRAY_SIZE(s->gt_scratch); i++)
		intel_uncore_write(uncore, GEN7_GT_SCRATCH(i), s->gt_scratch[i]);

	/* GT SA CZ domain, 0x100000-0x138124 */
	intel_uncore_write(uncore, TILECTL, s->tilectl);
	intel_uncore_write(uncore, GTFIFOCTL, s->gt_fifoctl);
	/*
	 * Preserve the GT allow wake and GFX force clock bit, they are not
	 * be restored, as they are used to control the s0ix suspend/resume
	 * sequence by the caller.
	 */
	intel_uncore_rmw(uncore, VLV_GTLC_WAKE_CTRL, ~VLV_GTLC_ALLOWWAKEREQ,
			 s->gtlc_wake_ctrl & ~VLV_GTLC_ALLOWWAKEREQ);

	intel_uncore_rmw(uncore, VLV_GTLC_SURVIVABILITY_REG, ~VLV_GFX_CLK_FORCE_ON_BIT,
			 s->gtlc_survive & ~VLV_GFX_CLK_FORCE_ON_BIT);

	intel_uncore_write(uncore, VLV_PMWGICZ, s->pmwgicz);

	/* Gunit-Display CZ domain, 0x182028-0x1821CF */
	intel_uncore_write(uncore, VLV_GU_CTL0, s->