Merge tag 'drm-xe-next-2025-01-07' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-next

UAPI Changes:
- OA new property: 'unblock after N reports' (Ashutosh)

i915 display Changes:
- UHBR rates for Thunderbolt (Kahola)

Driver Changes:
- IRQ related fixes and improvements (Ilia)
 - Revert some changes that break a mesa debug tool (John)
 - Fix migration issues (Nirmoy)
 - Enable GuC's WA_DUAL_QUEUE for newer platforms (Daniele)
 - Move shrink test out of xe_bo (Nirmoy)
 - SRIOV PF: Use correct function to check LMEM provisioning (Michal)
 - Fix a false-positive "Missing outer runtime PM protection" warning (Rodrigo)
 - Make GSCCS disabling message less alarming (Daniele)
 - Fix DG1 power gate sequence (Rodrigo)
 - Xe files fixes (Lucas)
 - Fix a potential TP_printk UAF (Thomas)
 - OA Fixes (Umesh)
 - Fix tlb invalidation when wedging (Lucas)
 - Documentation fix (Lucas)

Signed-off-by: Dave Airlie <[email protected]>

From: Rodrigo Vivi <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

Author: airlied

drivers/gpu/drm/i915/display/intel_cx0_phy.c

@@ -3070,7 +3070,10 @@ int intel_mtl_tbt_calc_port_clock(struct intel_encoder *encoder)
 
 	val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port));
 
-	clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val);
+	if (DISPLAY_VER(display) >= 30)
+		clock = REG_FIELD_GET(XE3_DDI_CLOCK_SELECT_MASK, val);
+	else
+		clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val);
 
 	drm_WARN_ON(display->drm, !(val & XELPDP_FORWARD_CLOCK_UNGATE));
 	drm_WARN_ON(display->drm, !(val & XELPDP_TBT_CLOCK_REQUEST));
@@ -3085,13 +3088,18 @@ int intel_mtl_tbt_calc_port_clock(struct intel_encoder *encoder)
 		return 540000;
 	case XELPDP_DDI_CLOCK_SELECT_TBT_810:
 		return 810000;
+	case XELPDP_DDI_CLOCK_SELECT_TBT_312_5:
+		return 1000000;
+	case XELPDP_DDI_CLOCK_SELECT_TBT_625:
+		return 2000000;
 	default:
 		MISSING_CASE(clock);
 		return 162000;
 	}
 }
 
-static int intel_mtl_tbt_clock_select(int clock)
+static int intel_mtl_tbt_clock_select(struct intel_display *display,
+				      int clock)
 {
 	switch (clock) {
 	case 162000:
@@ -3102,6 +3110,18 @@ static int intel_mtl_tbt_clock_select(int clock)
 		return XELPDP_DDI_CLOCK_SELECT_TBT_540;
 	case 810000:
 		return XELPDP_DDI_CLOCK_SELECT_TBT_810;
+	case 1000000:
+		if (DISPLAY_VER(display) < 30) {
+			drm_WARN_ON(display->drm, "UHBR10 not supported for the platform\n");
+			return XELPDP_DDI_CLOCK_SELECT_TBT_162;
+		}
+		return XELPDP_DDI_CLOCK_SELECT_TBT_312_5;
+	case 2000000:
+		if (DISPLAY_VER(display) < 30) {
+			drm_WARN_ON(display->drm, "UHBR20 not supported for the platform\n");
+			return XELPDP_DDI_CLOCK_SELECT_TBT_162;
+		}
+		return XELPDP_DDI_CLOCK_SELECT_TBT_625;
 	default:
 		MISSING_CASE(clock);
 		return XELPDP_DDI_CLOCK_SELECT_TBT_162;
@@ -3114,15 +3134,26 @@ static void intel_mtl_tbt_pll_enable(struct intel_encoder *encoder,
 	struct intel_display *display = to_intel_display(encoder);
 	enum phy phy = intel_encoder_to_phy(encoder);
 	u32 val = 0;
+	u32 mask;
 
 	/*
 	 * 1. Program PORT_CLOCK_CTL REGISTER to configure
 	 * clock muxes, gating and SSC
 	 */
-	val |= XELPDP_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(crtc_state->port_clock));
+
+	if (DISPLAY_VER(display) >= 30) {
+		mask = XE3_DDI_CLOCK_SELECT_MASK;
+		val |= XE3_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock));
+	} else {
+		mask = XELPDP_DDI_CLOCK_SELECT_MASK;
+		val |= XELPDP_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock));
+	}
+
+	mask |= XELPDP_FORWARD_CLOCK_UNGATE;
 	val |= XELPDP_FORWARD_CLOCK_UNGATE;
+
 	intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
-		     XELPDP_DDI_CLOCK_SELECT_MASK | XELPDP_FORWARD_CLOCK_UNGATE, val);
+		     mask, val);
 
 	/* 2. Read back PORT_CLOCK_CTL REGISTER */
 	val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port));

drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h

@@ -187,14 +187,18 @@
 #define   XELPDP_TBT_CLOCK_REQUEST			REG_BIT(19)
 #define   XELPDP_TBT_CLOCK_ACK				REG_BIT(18)
 #define   XELPDP_DDI_CLOCK_SELECT_MASK			REG_GENMASK(15, 12)
+#define   XE3_DDI_CLOCK_SELECT_MASK			REG_GENMASK(16, 12)
 #define   XELPDP_DDI_CLOCK_SELECT(val)			REG_FIELD_PREP(XELPDP_DDI_CLOCK_SELECT_MASK, val)
+#define   XE3_DDI_CLOCK_SELECT(val)			REG_FIELD_PREP(XE3_DDI_CLOCK_SELECT_MASK, val)
 #define   XELPDP_DDI_CLOCK_SELECT_NONE			0x0
 #define   XELPDP_DDI_CLOCK_SELECT_MAXPCLK		0x8
 #define   XELPDP_DDI_CLOCK_SELECT_DIV18CLK		0x9
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_162		0xc
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_270		0xd
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_540		0xe
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_810		0xf
+#define   XELPDP_DDI_CLOCK_SELECT_TBT_312_5		0x18
+#define   XELPDP_DDI_CLOCK_SELECT_TBT_625		0x19
 #define   XELPDP_FORWARD_CLOCK_UNGATE			REG_BIT(10)
 #define   XELPDP_LANE1_PHY_CLOCK_SELECT			REG_BIT(8)
 #define   XELPDP_SSC_ENABLE_PLLA			REG_BIT(1)

drivers/gpu/drm/xe/regs/xe_engine_regs.h

@@ -83,6 +83,8 @@
 #define RING_IMR(base)				XE_REG((base) + 0xa8)
 #define RING_INT_STATUS_RPT_PTR(base)		XE_REG((base) + 0xac)
 
+#define CS_INT_VEC(base)			XE_REG((base) + 0x1b8)
+
 #define RING_EIR(base)				XE_REG((base) + 0xb0)
 #define RING_EMR(base)				XE_REG((base) + 0xb4)
 #define RING_ESR(base)				XE_REG((base) + 0xb8)
@@ -138,6 +140,7 @@
 
 #define RING_MODE(base)				XE_REG((base) + 0x29c)
 #define   GFX_DISABLE_LEGACY_MODE		REG_BIT(3)
+#define   GFX_MSIX_INTERRUPT_ENABLE		REG_BIT(13)
 
 #define RING_TIMESTAMP(base)			XE_REG((base) + 0x358)
 

drivers/gpu/drm/xe/regs/xe_lrc_layout.h

@@ -25,6 +25,9 @@
 #define CTX_INT_SRC_REPORT_REG		(CTX_LRI_INT_REPORT_PTR + 3)
 #define CTX_INT_SRC_REPORT_PTR		(CTX_LRI_INT_REPORT_PTR + 4)
 
+#define CTX_CS_INT_VEC_REG		0x5a
+#define CTX_CS_INT_VEC_DATA		(CTX_CS_INT_VEC_REG + 1)
+
 #define INDIRECT_CTX_RING_HEAD		(0x02 + 1)
 #define INDIRECT_CTX_RING_TAIL		(0x04 + 1)
 #define INDIRECT_CTX_RING_START		(0x06 + 1)

drivers/gpu/drm/xe/tests/xe_bo.c

@@ -606,8 +606,6 @@ static void xe_bo_shrink_kunit(struct kunit *test)
 static struct kunit_case xe_bo_tests[] = {
 	KUNIT_CASE_PARAM(xe_ccs_migrate_kunit, xe_pci_live_device_gen_param),
 	KUNIT_CASE_PARAM(xe_bo_evict_kunit, xe_pci_live_device_gen_param),
-	KUNIT_CASE_PARAM_ATTR(xe_bo_shrink_kunit, xe_pci_live_device_gen_param,
-			      {.speed = KUNIT_SPEED_SLOW}),
 	{}
 };
 
@@ -618,3 +616,17 @@ struct kunit_suite xe_bo_test_suite = {
 	.init = xe_kunit_helper_xe_device_live_test_init,
 };
 EXPORT_SYMBOL_IF_KUNIT(xe_bo_test_suite);
+
+static struct kunit_case xe_bo_shrink_test[] = {
+	KUNIT_CASE_PARAM_ATTR(xe_bo_shrink_kunit, xe_pci_live_device_gen_param,
+			      {.speed = KUNIT_SPEED_SLOW}),
+	{}
+};
+
+VISIBLE_IF_KUNIT
+struct kunit_suite xe_bo_shrink_test_suite = {
+	.name = "xe_bo_shrink",
+	.test_cases = xe_bo_shrink_test,
+	.init = xe_kunit_helper_xe_device_live_test_init,
+};
+EXPORT_SYMBOL_IF_KUNIT(xe_bo_shrink_test_suite);

drivers/gpu/drm/xe/tests/xe_live_test_mod.c

@@ -6,11 +6,13 @@
 #include <kunit/test.h>
 
 extern struct kunit_suite xe_bo_test_suite;
+extern struct kunit_suite xe_bo_shrink_test_suite;
 extern struct kunit_suite xe_dma_buf_test_suite;
 extern struct kunit_suite xe_migrate_test_suite;
 extern struct kunit_suite xe_mocs_test_suite;
 
 kunit_test_suite(xe_bo_test_suite);
+kunit_test_suite(xe_bo_shrink_test_suite);
 kunit_test_suite(xe_dma_buf_test_suite);
 kunit_test_suite(xe_migrate_test_suite);
 kunit_test_suite(xe_mocs_test_suite);

drivers/gpu/drm/xe/xe_bo.c

@@ -733,7 +733,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 	    new_mem->mem_type == XE_PL_SYSTEM) {
 		long timeout = dma_resv_wait_timeout(ttm_bo->base.resv,
 						     DMA_RESV_USAGE_BOOKKEEP,
-						     true,
+						     false,
 						     MAX_SCHEDULE_TIMEOUT);
 		if (timeout < 0) {
 			ret = timeout;
@@ -857,8 +857,16 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 
 out:
 	if ((!ttm_bo->resource || ttm_bo->resource->mem_type == XE_PL_SYSTEM) &&
-	    ttm_bo->ttm)
+	    ttm_bo->ttm) {
+		long timeout = dma_resv_wait_timeout(ttm_bo->base.resv,
+						     DMA_RESV_USAGE_KERNEL,
+						     false,
+						     MAX_SCHEDULE_TIMEOUT);
+		if (timeout < 0)
+			ret = timeout;
+
 		xe_tt_unmap_sg(ttm_bo->ttm);
+	}
 
 	return ret;
 }

drivers/gpu/drm/xe/xe_devcoredump.c

@@ -119,7 +119,11 @@ static ssize_t __xe_devcoredump_read(char *buffer, size_t count,
 	drm_puts(&p, "\n**** GuC CT ****\n");
 	xe_guc_ct_snapshot_print(ss->guc.ct, &p);
 
-	drm_puts(&p, "\n**** Contexts ****\n");
+	/*
+	 * Don't add a new section header here because the mesa debug decoder
+	 * tool expects the context information to be in the 'GuC CT' section.
+	 */
+	/* drm_puts(&p, "\n**** Contexts ****\n"); */
 	xe_guc_exec_queue_snapshot_print(ss->ge, &p);
 
 	drm_puts(&p, "\n**** Job ****\n");
@@ -416,6 +420,15 @@ void xe_print_blob_ascii85(struct drm_printer *p, const char *prefix,
 	char buff[ASCII85_BUFSZ], *line_buff;
 	size_t line_pos = 0;
 
+	/*
+	 * Splitting blobs across multiple lines is not compatible with the mesa
+	 * debug decoder tool. Note that even dropping the explicit '\n' below
+	 * doesn't help because the GuC log is so big some underlying implementation
+	 * still splits the lines at 512K characters. So just bail completely for
+	 * the moment.
+	 */
+	return;
+
 #define DMESG_MAX_LINE_LEN	800
 #define MIN_SPACE		(ASCII85_BUFSZ + 2)		/* 85 + "\n\0" */
 

drivers/gpu/drm/xe/xe_device.c

@@ -325,7 +325,9 @@ struct xe_device *xe_device_create(struct pci_dev *pdev,
 	xe->info.revid = pdev->revision;
 	xe->info.force_execlist = xe_modparam.force_execlist;
 
-	spin_lock_init(&xe->irq.lock);
+	err = xe_irq_init(xe);
+	if (err)
+		goto err;
 
 	init_waitqueue_head(&xe->ufence_wq);
 

drivers/gpu/drm/xe/xe_device.h

@@ -157,8 +157,7 @@ static inline bool xe_device_has_sriov(struct xe_device *xe)
 
 static inline bool xe_device_has_msix(struct xe_device *xe)
 {
-	/* TODO: change this when MSI-X support is fully integrated */
-	return false;
+	return xe->irq.msix.nvec > 0;
 }
 
 static inline bool xe_device_has_memirq(struct xe_device *xe)