[PATCH 3/3] Revert "drm/xe/vf: Create contexts for CCS read write"
Satyanarayana K V P
satyanarayana.k.v.p at intel.com
Mon Jul 28 10:02:48 UTC 2025
The VF CCS save/restore series (patchwork #149108) has a dependency
on the migration framework. A recent migration update in commit
d65ff1ec8535 ("drm/xe: Split xe_migrate allocation from initialization")
caused a VM crash during XE driver release for iGPU devices.
This reverts commit f3009272ff2ec0498426001e4dd2a12c50234e08.
Signed-off-by: Satyanarayana K V P <satyanarayana.k.v.p at intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko at intel.com>
Cc: Matthew Brost <matthew.brost at intel.com>
Cc: Michał Winiarski <michal.winiarski at intel.com>
Cc: Piotr Piórkowski <piotr.piorkowski at intel.com>
---
drivers/gpu/drm/xe/Makefile | 1 -
drivers/gpu/drm/xe/xe_device.c | 4 -
drivers/gpu/drm/xe/xe_device_types.h | 4 -
drivers/gpu/drm/xe/xe_gt_debugfs.c | 39 ----
drivers/gpu/drm/xe/xe_sriov.c | 19 --
drivers/gpu/drm/xe/xe_sriov.h | 1 -
drivers/gpu/drm/xe/xe_sriov_vf_ccs.c | 208 ---------------------
drivers/gpu/drm/xe/xe_sriov_vf_ccs.h | 13 --
drivers/gpu/drm/xe/xe_sriov_vf_ccs_types.h | 45 -----
drivers/gpu/drm/xe/xe_sriov_vf_types.h | 6 -
10 files changed, 340 deletions(-)
delete mode 100644 drivers/gpu/drm/xe/xe_sriov_vf_ccs.c
delete mode 100644 drivers/gpu/drm/xe/xe_sriov_vf_ccs.h
delete mode 100644 drivers/gpu/drm/xe/xe_sriov_vf_ccs_types.h
diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
index 8e0c3412a757..47feda225086 100644
--- a/drivers/gpu/drm/xe/Makefile
+++ b/drivers/gpu/drm/xe/Makefile
@@ -152,7 +152,6 @@ xe-y += \
xe_memirq.o \
xe_sriov.o \
xe_sriov_vf.o \
- xe_sriov_vf_ccs.o \
xe_tile_sriov_vf.o
xe-$(CONFIG_PCI_IOV) += \
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index d04a0ae018e6..08f45777f797 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -940,10 +940,6 @@ int xe_device_probe(struct xe_device *xe)
xe_vsec_init(xe);
- err = xe_sriov_late_init(xe);
- if (err)
- goto err_unregister_display;
-
return devm_add_action_or_reset(xe->drm.dev, xe_device_sanitize, xe);
err_unregister_display:
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 38c8329b4d2c..eb6105523f23 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -23,7 +23,6 @@
#include "xe_sriov_pf_types.h"
#include "xe_sriov_types.h"
#include "xe_sriov_vf_types.h"
-#include "xe_sriov_vf_ccs_types.h"
#include "xe_step_types.h"
#include "xe_survivability_mode_types.h"
@@ -183,9 +182,6 @@ struct xe_tile {
struct {
/** @sriov.vf.ggtt_balloon: GGTT regions excluded from use. */
struct xe_ggtt_node *ggtt_balloon[2];
-
- /** @sriov.vf.ccs: CCS read and write contexts for VF. */
- struct xe_tile_vf_ccs ccs[XE_SRIOV_VF_CCS_CTX_COUNT];
} vf;
} sriov;
diff --git a/drivers/gpu/drm/xe/xe_gt_debugfs.c b/drivers/gpu/drm/xe/xe_gt_debugfs.c
index 8d4b66ca38b7..061a81f82e76 100644
--- a/drivers/gpu/drm/xe/xe_gt_debugfs.c
+++ b/drivers/gpu/drm/xe/xe_gt_debugfs.c
@@ -134,33 +134,6 @@ static int sa_info(struct xe_gt *gt, struct drm_printer *p)
return 0;
}
-static int sa_info_vf_ccs(struct xe_gt *gt, struct drm_printer *p)
-{
- struct xe_tile *tile = gt_to_tile(gt);
- struct xe_sa_manager *bb_pool;
- enum xe_sriov_vf_ccs_rw_ctxs ctx_id;
-
- if (!IS_VF_CCS_READY(gt_to_xe(gt)))
- return 0;
-
- xe_pm_runtime_get(gt_to_xe(gt));
-
- for_each_ccs_rw_ctx(ctx_id) {
- bb_pool = tile->sriov.vf.ccs[ctx_id].mem.ccs_bb_pool;
- if (!bb_pool)
- break;
-
- drm_printf(p, "ccs %s bb suballoc info\n", ctx_id ? "write" : "read");
- drm_printf(p, "-------------------------\n");
- drm_suballoc_dump_debug_info(&bb_pool->base, p, bb_pool->gpu_addr);
- drm_puts(p, "\n");
- }
-
- xe_pm_runtime_put(gt_to_xe(gt));
-
- return 0;
-}
-
static int topology(struct xe_gt *gt, struct drm_printer *p)
{
xe_pm_runtime_get(gt_to_xe(gt));
@@ -330,13 +303,6 @@ static const struct drm_info_list vf_safe_debugfs_list[] = {
{"hwconfig", .show = xe_gt_debugfs_simple_show, .data = hwconfig},
};
-/*
- * only for GT debugfs files which are valid on VF. Not valid on PF.
- */
-static const struct drm_info_list vf_only_debugfs_list[] = {
- {"sa_info_vf_ccs", .show = xe_gt_debugfs_simple_show, .data = sa_info_vf_ccs},
-};
-
/* everything else should be added here */
static const struct drm_info_list pf_only_debugfs_list[] = {
{"hw_engines", .show = xe_gt_debugfs_simple_show, .data = hw_engines},
@@ -458,11 +424,6 @@ void xe_gt_debugfs_register(struct xe_gt *gt)
drm_debugfs_create_files(pf_only_debugfs_list,
ARRAY_SIZE(pf_only_debugfs_list),
root, minor);
- else
- drm_debugfs_create_files(vf_only_debugfs_list,
- ARRAY_SIZE(vf_only_debugfs_list),
- root, minor);
-
xe_uc_debugfs_register(>->uc, root);
diff --git a/drivers/gpu/drm/xe/xe_sriov.c b/drivers/gpu/drm/xe/xe_sriov.c
index 87911fb4eea7..a0eab44c0e76 100644
--- a/drivers/gpu/drm/xe/xe_sriov.c
+++ b/drivers/gpu/drm/xe/xe_sriov.c
@@ -15,7 +15,6 @@
#include "xe_sriov.h"
#include "xe_sriov_pf.h"
#include "xe_sriov_vf.h"
-#include "xe_sriov_vf_ccs.h"
/**
* xe_sriov_mode_to_string - Convert enum value to string.
@@ -158,21 +157,3 @@ const char *xe_sriov_function_name(unsigned int n, char *buf, size_t size)
strscpy(buf, "PF", size);
return buf;
}
-
-/**
- * xe_sriov_late_init() - SR-IOV late initialization functions.
- * @xe: the &xe_device to initialize
- *
- * On VF this function will initialize code for CCS migration.
- *
- * Return: 0 on success or a negative error code on failure.
- */
-int xe_sriov_late_init(struct xe_device *xe)
-{
- int err = 0;
-
- if (IS_VF_CCS_INIT_NEEDED(xe))
- err = xe_sriov_vf_ccs_init(xe);
-
- return err;
-}
diff --git a/drivers/gpu/drm/xe/xe_sriov.h b/drivers/gpu/drm/xe/xe_sriov.h
index 0e0c1abf2d14..688fbabf08f1 100644
--- a/drivers/gpu/drm/xe/xe_sriov.h
+++ b/drivers/gpu/drm/xe/xe_sriov.h
@@ -18,7 +18,6 @@ const char *xe_sriov_function_name(unsigned int n, char *buf, size_t len);
void xe_sriov_probe_early(struct xe_device *xe);
void xe_sriov_print_info(struct xe_device *xe, struct drm_printer *p);
int xe_sriov_init(struct xe_device *xe);
-int xe_sriov_late_init(struct xe_device *xe);
static inline enum xe_sriov_mode xe_device_sriov_mode(const struct xe_device *xe)
{
diff --git a/drivers/gpu/drm/xe/xe_sriov_vf_ccs.c b/drivers/gpu/drm/xe/xe_sriov_vf_ccs.c
deleted file mode 100644
index 697eea46659a..000000000000
--- a/drivers/gpu/drm/xe/xe_sriov_vf_ccs.c
+++ /dev/null
@@ -1,208 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright © 2025 Intel Corporation
- */
-
-#include "instructions/xe_mi_commands.h"
-#include "instructions/xe_gpu_commands.h"
-#include "xe_bo.h"
-#include "xe_device.h"
-#include "xe_migrate.h"
-#include "xe_sa.h"
-#include "xe_sriov_printk.h"
-#include "xe_sriov_vf_ccs.h"
-#include "xe_sriov_vf_ccs_types.h"
-
-/**
- * DOC: VF save/restore of compression Meta Data
- *
- * VF KMD registers two special contexts/LRCAs.
- *
- * Save Context/LRCA: contain necessary cmds+page table to trigger Meta data /
- * compression control surface (Aka CCS) save in regular System memory in VM.
- *
- * Restore Context/LRCA: contain necessary cmds+page table to trigger Meta data /
- * compression control surface (Aka CCS) Restore from regular System memory in
- * VM to corresponding CCS pool.
- *
- * Below diagram explain steps needed for VF save/Restore of compression Meta Data::
- *
- * CCS Save CCS Restore VF KMD Guc BCS
- * LRCA LRCA
- * | | | | |
- * | | | | |
- * | Create Save LRCA | | |
- * [ ]<----------------------------- [ ] | |
- * | | | | |
- * | | | | |
- * | | | Register save LRCA | |
- * | | | with Guc | |
- * | | [ ]--------------------------->[ ] |
- * | | | | |
- * | | Create restore LRCA | | |
- * | [ ]<------------------[ ] | |
- * | | | | |
- * | | | Register restore LRCA | |
- * | | | with Guc | |
- * | | [ ]--------------------------->[ ] |
- * | | | | |
- * | | | | |
- * | | [ ]------------------------- | |
- * | | [ ] Allocate main memory. | | |
- * | | [ ] Allocate CCS memory. | | |
- * | | [ ] Update Main memory & | | |
- * [ ]<------------------------------[ ] CCS pages PPGTT + BB | | |
- * | [ ]<------------------[ ] cmds to save & restore.| | |
- * | | [ ]<------------------------ | |
- * | | | | |
- * | | | | |
- * | | | | |
- * : : : : :
- * ---------------------------- VF Paused -------------------------------------
- * | | | | |
- * | | | | |
- * | | | |Schedule |
- * | | | |CCS Save |
- * | | | | LRCA |
- * | | | [ ]------>[ ]
- * | | | | |
- * | | | | |
- * | | | |CCS save |
- * | | | |completed|
- * | | | [ ]<------[ ]
- * | | | | |
- * : : : : :
- * ---------------------------- VM Migrated -----------------------------------
- * | | | | |
- * | | | | |
- * : : : : :
- * ---------------------------- VF Resumed ------------------------------------
- * | | | | |
- * | | | | |
- * | | [ ]-------------- | |
- * | | [ ] Fix up GGTT | | |
- * | | [ ]<------------- | |
- * | | | | |
- * | | | | |
- * | | | Notify VF_RESFIX_DONE | |
- * | | [ ]--------------------------->[ ] |
- * | | | | |
- * | | | |Schedule |
- * | | | |CCS |
- * | | | |Restore |
- * | | | |LRCA |
- * | | | [ ]------>[ ]
- * | | | | |
- * | | | | |
- * | | | |CCS |
- * | | | |restore |
- * | | | |completed|
- * | | | [ ]<------[ ]
- * | | | | |
- * | | | | |
- * | | | VF_RESFIX_DONE complete | |
- * | | | notification | |
- * | | [ ]<---------------------------[ ] |
- * | | | | |
- * | | | | |
- * : : : : :
- * ------------------------- Continue VM restore ------------------------------
- */
-
-static u64 get_ccs_bb_pool_size(struct xe_device *xe)
-{
- u64 sys_mem_size, ccs_mem_size, ptes, bb_pool_size;
- struct sysinfo si;
-
- si_meminfo(&si);
- sys_mem_size = si.totalram * si.mem_unit;
- ccs_mem_size = div64_u64(sys_mem_size, NUM_BYTES_PER_CCS_BYTE(xe));
- ptes = DIV_ROUND_UP_ULL(sys_mem_size + ccs_mem_size, XE_PAGE_SIZE);
-
- /**
- * We need below BB size to hold PTE mappings and some DWs for copy
- * command. In reality, we need space for many copy commands. So, let
- * us allocate double the calculated size which is enough to holds GPU
- * instructions for the whole region.
- */
- bb_pool_size = ptes * sizeof(u32);
-
- return round_up(bb_pool_size * 2, SZ_1M);
-}
-
-static int alloc_bb_pool(struct xe_tile *tile, struct xe_tile_vf_ccs *ctx)
-{
- struct xe_device *xe = tile_to_xe(tile);
- struct xe_sa_manager *sa_manager;
- u64 bb_pool_size;
- int offset, err;
-
- bb_pool_size = get_ccs_bb_pool_size(xe);
- xe_sriov_info(xe, "Allocating %s CCS BB pool size = %lldMB\n",
- ctx->ctx_id ? "Restore" : "Save", bb_pool_size / SZ_1M);
-
- sa_manager = xe_sa_bo_manager_init(tile, bb_pool_size, SZ_16);
-
- if (IS_ERR(sa_manager)) {
- xe_sriov_err(xe, "Suballocator init failed with error: %pe\n",
- sa_manager);
- err = PTR_ERR(sa_manager);
- return err;
- }
-
- offset = 0;
- xe_map_memset(xe, &sa_manager->bo->vmap, offset, MI_NOOP,
- bb_pool_size);
-
- offset = bb_pool_size - sizeof(u32);
- xe_map_wr(xe, &sa_manager->bo->vmap, offset, u32, MI_BATCH_BUFFER_END);
-
- ctx->mem.ccs_bb_pool = sa_manager;
-
- return 0;
-}
-
-/**
- * xe_sriov_vf_ccs_init - Setup LRCA for save & restore.
- * @xe: the &xe_device to start recovery on
- *
- * This function shall be called only by VF. It initializes
- * LRCA and suballocator needed for CCS save & restore.
- *
- * Return: 0 on success. Negative error code on failure.
- */
-int xe_sriov_vf_ccs_init(struct xe_device *xe)
-{
- struct xe_tile *tile = xe_device_get_root_tile(xe);
- enum xe_sriov_vf_ccs_rw_ctxs ctx_id;
- struct xe_migrate *migrate;
- struct xe_tile_vf_ccs *ctx;
- int err;
-
- xe_assert(xe, IS_SRIOV_VF(xe));
- xe_assert(xe, !IS_DGFX(xe));
- xe_assert(xe, xe_device_has_flat_ccs(xe));
-
- for_each_ccs_rw_ctx(ctx_id) {
- ctx = &tile->sriov.vf.ccs[ctx_id];
- ctx->ctx_id = ctx_id;
-
- migrate = xe_migrate_init(tile);
- if (IS_ERR(migrate)) {
- err = PTR_ERR(migrate);
- goto err_ret;
- }
- ctx->migrate = migrate;
-
- err = alloc_bb_pool(tile, ctx);
- if (err)
- goto err_ret;
- }
-
- xe->sriov.vf.ccs.initialized = 1;
-
- return 0;
-
-err_ret:
- return err;
-}
diff --git a/drivers/gpu/drm/xe/xe_sriov_vf_ccs.h b/drivers/gpu/drm/xe/xe_sriov_vf_ccs.h
deleted file mode 100644
index 5df9ba028d14..000000000000
--- a/drivers/gpu/drm/xe/xe_sriov_vf_ccs.h
+++ /dev/null
@@ -1,13 +0,0 @@
-/* SPDX-License-Identifier: MIT */
-/*
- * Copyright © 2025 Intel Corporation
- */
-
-#ifndef _XE_SRIOV_VF_CCS_H_
-#define _XE_SRIOV_VF_CCS_H_
-
-struct xe_device;
-
-int xe_sriov_vf_ccs_init(struct xe_device *xe);
-
-#endif
diff --git a/drivers/gpu/drm/xe/xe_sriov_vf_ccs_types.h b/drivers/gpu/drm/xe/xe_sriov_vf_ccs_types.h
deleted file mode 100644
index 6dc279d206ec..000000000000
--- a/drivers/gpu/drm/xe/xe_sriov_vf_ccs_types.h
+++ /dev/null
@@ -1,45 +0,0 @@
-/* SPDX-License-Identifier: MIT */
-/*
- * Copyright © 2025 Intel Corporation
- */
-
-#ifndef _XE_SRIOV_VF_CCS_TYPES_H_
-#define _XE_SRIOV_VF_CCS_TYPES_H_
-
-#define for_each_ccs_rw_ctx(id__) \
- for ((id__) = 0; (id__) < XE_SRIOV_VF_CCS_CTX_COUNT; (id__)++)
-
-#define IS_VF_CCS_READY(xe) ({ \
- struct xe_device *___xe = (xe); \
- xe_assert(___xe, IS_SRIOV_VF(___xe)); \
- ___xe->sriov.vf.ccs.initialized; \
- })
-
-#define IS_VF_CCS_INIT_NEEDED(xe) ({\
- struct xe_device *___xe = (xe); \
- IS_SRIOV_VF(___xe) && !IS_DGFX(___xe) && \
- xe_device_has_flat_ccs(___xe) && GRAPHICS_VER(___xe) >= 20; \
- })
-
-enum xe_sriov_vf_ccs_rw_ctxs {
- XE_SRIOV_VF_CCS_READ_CTX,
- XE_SRIOV_VF_CCS_WRITE_CTX,
- XE_SRIOV_VF_CCS_CTX_COUNT
-};
-
-struct xe_migrate;
-struct xe_sa_manager;
-
-struct xe_tile_vf_ccs {
- /** @id: Id to which context it belongs to */
- enum xe_sriov_vf_ccs_rw_ctxs ctx_id;
- /** @migrate: Migration helper for save/restore of CCS data */
- struct xe_migrate *migrate;
-
- struct {
- /** @ccs_rw_bb_pool: Pool from which batch buffers are allocated. */
- struct xe_sa_manager *ccs_bb_pool;
- } mem;
-};
-
-#endif
diff --git a/drivers/gpu/drm/xe/xe_sriov_vf_types.h b/drivers/gpu/drm/xe/xe_sriov_vf_types.h
index 24a873c50c49..8300416a6226 100644
--- a/drivers/gpu/drm/xe/xe_sriov_vf_types.h
+++ b/drivers/gpu/drm/xe/xe_sriov_vf_types.h
@@ -36,12 +36,6 @@ struct xe_device_vf {
/** @migration.gt_flags: Per-GT request flags for VF migration recovery */
unsigned long gt_flags;
} migration;
-
- /** @ccs: VF CCS state data */
- struct {
- /** @ccs.initialized: Initilalization of VF CCS is completed or not */
- bool initialized;
- } ccs;
};
#endif
--
2.43.0
More information about the Intel-xe
mailing list