diff options
Diffstat (limited to 'drivers/gpu/drm/xe/xe_sriov_vf.c')
| -rw-r--r-- | drivers/gpu/drm/xe/xe_sriov_vf.c | 191 |
1 files changed, 179 insertions, 12 deletions
diff --git a/drivers/gpu/drm/xe/xe_sriov_vf.c b/drivers/gpu/drm/xe/xe_sriov_vf.c index 26e243c28994..cdd9f8e78b2a 100644 --- a/drivers/gpu/drm/xe/xe_sriov_vf.c +++ b/drivers/gpu/drm/xe/xe_sriov_vf.c @@ -3,6 +3,7 @@ * Copyright © 2023-2024 Intel Corporation */ +#include <drm/drm_debugfs.h> #include <drm/drm_managed.h> #include "xe_assert.h" @@ -10,11 +11,16 @@ #include "xe_gt.h" #include "xe_gt_sriov_printk.h" #include "xe_gt_sriov_vf.h" +#include "xe_guc.h" #include "xe_guc_ct.h" +#include "xe_guc_submit.h" +#include "xe_irq.h" +#include "xe_lrc.h" #include "xe_pm.h" #include "xe_sriov.h" #include "xe_sriov_printk.h" #include "xe_sriov_vf.h" +#include "xe_sriov_vf_ccs.h" #include "xe_tile_sriov_vf.h" /** @@ -124,16 +130,66 @@ * | | | */ -static bool vf_migration_supported(struct xe_device *xe) +/** + * xe_sriov_vf_migration_supported - Report whether SR-IOV VF migration is + * supported or not. + * @xe: the &xe_device to check + * + * Returns: true if VF migration is supported, false otherwise. + */ +bool xe_sriov_vf_migration_supported(struct xe_device *xe) +{ + xe_assert(xe, IS_SRIOV_VF(xe)); + return xe->sriov.vf.migration.enabled; +} + +static void vf_disable_migration(struct xe_device *xe, const char *fmt, ...) +{ + struct va_format vaf; + va_list va_args; + + xe_assert(xe, IS_SRIOV_VF(xe)); + + va_start(va_args, fmt); + vaf.fmt = fmt; + vaf.va = &va_args; + xe_sriov_notice(xe, "migration disabled: %pV\n", &vaf); + va_end(va_args); + + xe->sriov.vf.migration.enabled = false; +} + +static void migration_worker_func(struct work_struct *w); + +static void vf_migration_init_early(struct xe_device *xe) { /* * TODO: Add conditions to allow specific platforms, when they're * supported at production quality. */ - return IS_ENABLED(CONFIG_DRM_XE_DEBUG); -} + if (!IS_ENABLED(CONFIG_DRM_XE_DEBUG)) + return vf_disable_migration(xe, + "experimental feature not available on production builds"); + + if (GRAPHICS_VER(xe) < 20) + return vf_disable_migration(xe, "requires gfx version >= 20, but only %u found", + GRAPHICS_VER(xe)); + + if (!IS_DGFX(xe)) { + struct xe_uc_fw_version guc_version; + + xe_gt_sriov_vf_guc_versions(xe_device_get_gt(xe, 0), NULL, &guc_version); + if (MAKE_GUC_VER_STRUCT(guc_version) < MAKE_GUC_VER(1, 23, 0)) + return vf_disable_migration(xe, + "CCS migration requires GuC ABI >= 1.23 but only %u.%u found", + guc_version.major, guc_version.minor); + } -static void migration_worker_func(struct work_struct *w); + INIT_WORK(&xe->sriov.vf.migration.worker, migration_worker_func); + + xe->sriov.vf.migration.enabled = true; + xe_sriov_dbg(xe, "migration support enabled\n"); +} /** * xe_sriov_vf_init_early - Initialize SR-IOV VF specific data. @@ -141,10 +197,57 @@ static void migration_worker_func(struct work_struct *w); */ void xe_sriov_vf_init_early(struct xe_device *xe) { - INIT_WORK(&xe->sriov.vf.migration.worker, migration_worker_func); + vf_migration_init_early(xe); +} + +/** + * vf_post_migration_shutdown - Stop the driver activities after VF migration. + * @xe: the &xe_device struct instance + * + * After this VM is migrated and assigned to a new VF, it is running on a new + * hardware, and therefore many hardware-dependent states and related structures + * require fixups. Without fixups, the hardware cannot do any work, and therefore + * all GPU pipelines are stalled. + * Stop some of kernel activities to make the fixup process faster. + */ +static void vf_post_migration_shutdown(struct xe_device *xe) +{ + struct xe_gt *gt; + unsigned int id; + int ret = 0; + + for_each_gt(gt, xe, id) { + xe_guc_submit_pause(>->uc.guc); + ret |= xe_guc_submit_reset_block(>->uc.guc); + } + + if (ret) + drm_info(&xe->drm, "migration recovery encountered ongoing reset\n"); +} + +/** + * vf_post_migration_kickstart - Re-start the driver activities under new hardware. + * @xe: the &xe_device struct instance + * + * After we have finished with all post-migration fixups, restart the driver + * activities to continue feeding the GPU with workloads. + */ +static void vf_post_migration_kickstart(struct xe_device *xe) +{ + struct xe_gt *gt; + unsigned int id; - if (!vf_migration_supported(xe)) - xe_sriov_info(xe, "migration not supported by this module version\n"); + /* + * Make sure interrupts on the new HW are properly set. The GuC IRQ + * must be working at this point, since the recovery did started, + * but the rest was not enabled using the procedure from spec. + */ + xe_irq_resume(xe); + + for_each_gt(gt, xe, id) { + xe_guc_submit_reset_unblock(>->uc.guc); + xe_guc_submit_unpause(>->uc.guc); + } } static bool gt_vf_post_migration_needed(struct xe_gt *gt) @@ -192,6 +295,11 @@ static int vf_get_next_migrated_gt_id(struct xe_device *xe) return -1; } +static size_t post_migration_scratch_size(struct xe_device *xe) +{ + return max(xe_lrc_reg_size(xe), LRC_WA_BB_SIZE); +} + /** * Perform post-migration fixups on a single GT. * @@ -208,19 +316,31 @@ static int vf_get_next_migrated_gt_id(struct xe_device *xe) static int gt_vf_post_migration_fixups(struct xe_gt *gt) { s64 shift; + void *buf; int err; + buf = kmalloc(post_migration_scratch_size(gt_to_xe(gt)), GFP_KERNEL); + if (!buf) + return -ENOMEM; + err = xe_gt_sriov_vf_query_config(gt); if (err) - return err; + goto out; shift = xe_gt_sriov_vf_ggtt_shift(gt); if (shift) { xe_tile_sriov_vf_fixup_ggtt_nodes(gt_to_tile(gt), shift); - /* FIXME: add the recovery steps */ + xe_gt_sriov_vf_default_lrcs_hwsp_rebase(gt); + err = xe_guc_contexts_hwsp_rebase(>->uc.guc, buf); + if (err) + goto out; + xe_guc_jobs_ring_rebase(>->uc.guc); xe_guc_ct_fixup_messages_with_ggtt(>->uc.guc.ct, shift); } - return 0; + +out: + kfree(buf); + return err; } static void vf_post_migration_recovery(struct xe_device *xe) @@ -230,9 +350,10 @@ static void vf_post_migration_recovery(struct xe_device *xe) drm_dbg(&xe->drm, "migration recovery in progress\n"); xe_pm_runtime_get(xe); + vf_post_migration_shutdown(xe); - if (!vf_migration_supported(xe)) { - xe_sriov_err(xe, "migration not supported by this module version\n"); + if (!xe_sriov_vf_migration_supported(xe)) { + xe_sriov_err(xe, "migration is not supported\n"); err = -ENOTRECOVERABLE; goto fail; } @@ -247,6 +368,7 @@ static void vf_post_migration_recovery(struct xe_device *xe) set_bit(id, &fixed_gts); } + vf_post_migration_kickstart(xe); err = vf_post_migration_notify_resfix_done(xe, fixed_gts); if (err) goto fail; @@ -306,3 +428,48 @@ void xe_sriov_vf_start_migration_recovery(struct xe_device *xe) drm_info(&xe->drm, "VF migration recovery %s\n", started ? "scheduled" : "already in progress"); } + +/** + * xe_sriov_vf_init_late() - SR-IOV VF late initialization functions. + * @xe: the &xe_device to initialize + * + * This function initializes code for CCS migration. + * + * Return: 0 on success or a negative error code on failure. + */ +int xe_sriov_vf_init_late(struct xe_device *xe) +{ + int err = 0; + + if (xe_sriov_vf_migration_supported(xe)) + err = xe_sriov_vf_ccs_init(xe); + + return err; +} + +static int sa_info_vf_ccs(struct seq_file *m, void *data) +{ + struct drm_info_node *node = m->private; + struct xe_device *xe = to_xe_device(node->minor->dev); + struct drm_printer p = drm_seq_file_printer(m); + + xe_sriov_vf_ccs_print(xe, &p); + return 0; +} + +static const struct drm_info_list debugfs_list[] = { + { .name = "sa_info_vf_ccs", .show = sa_info_vf_ccs }, +}; + +/** + * xe_sriov_vf_debugfs_register - Register VF debugfs attributes. + * @xe: the &xe_device + * @root: the root &dentry + * + * Prepare debugfs attributes exposed by the VF. + */ +void xe_sriov_vf_debugfs_register(struct xe_device *xe, struct dentry *root) +{ + drm_debugfs_create_files(debugfs_list, ARRAY_SIZE(debugfs_list), + root, xe->drm.primary); +} |