[PATCH v2 5/7] drm/xe/vf: Rebase HWSP of all contexts after migration

Michal Wajdeczko michal.wajdeczko at intel.com
Fri May 16 17:01:27 UTC 2025



On 16.05.2025 00:18, Tomasz Lis wrote:
> All contexts require an update due to GGTT range shift, as that
> affects their HWSP.
> 
> The HW status page of a context contains GGTT references, which
> need to be shifted to a new range (or re-computed using the
> previously updated vma nodes). The references include ring start
> address and indirect state address.
> 
> v2: move some functions to better matched files
> 
> Signed-off-by: Tomasz Lis <tomasz.lis at intel.com>
> Cc: Michal Wajdeczko <michal.wajdeczko at intel.com>
> ---
>  drivers/gpu/drm/xe/xe_exec_queue.c |  8 ++++++++
>  drivers/gpu/drm/xe/xe_exec_queue.h |  2 ++
>  drivers/gpu/drm/xe/xe_guc_submit.c | 11 +++++++++++
>  drivers/gpu/drm/xe/xe_guc_submit.h |  2 ++
>  drivers/gpu/drm/xe/xe_lrc.c        | 16 ++++++++++++++++
>  drivers/gpu/drm/xe/xe_lrc.h        |  1 +
>  drivers/gpu/drm/xe/xe_sriov_vf.c   | 17 +++++++++++++++--
>  7 files changed, 55 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c
> index ce78cee5dec6..d03600519b27 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue.c
> +++ b/drivers/gpu/drm/xe/xe_exec_queue.c
> @@ -1041,3 +1041,11 @@ int xe_exec_queue_last_fence_test_dep(struct xe_exec_queue *q, struct xe_vm *vm)
>  
>  	return err;
>  }
> +

public functions require kernel-doc

> +void xe_exec_queue_contexts_hwsp_rebase(struct xe_exec_queue *eq)
> +{
> +	int i;
> +
> +	for (i = 0; i < eq->width; ++i)
> +		xe_lrc_update_hwctx_regs_with_address(eq->lrc[i]);
> +}
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue.h b/drivers/gpu/drm/xe/xe_exec_queue.h
> index 17bc50a7f05a..aeeecd38ba00 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue.h
> +++ b/drivers/gpu/drm/xe/xe_exec_queue.h
> @@ -90,4 +90,6 @@ int xe_exec_queue_last_fence_test_dep(struct xe_exec_queue *q,
>  				      struct xe_vm *vm);
>  void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q);
>  
> +void xe_exec_queue_contexts_hwsp_rebase(struct xe_exec_queue *eq);
> +
>  #endif
> diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> index 69ccfb2e1cff..4e1d8a86ed7e 100644
> --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> @@ -2424,3 +2424,14 @@ void xe_guc_submit_print(struct xe_guc *guc, struct drm_printer *p)
>  		guc_exec_queue_print(q, p);
>  	mutex_unlock(&guc->submission_state.lock);
>  }
> +

public functions require kernel-doc

> +void xe_guc_contexts_hwsp_rebase(struct xe_guc *guc)
> +{
> +	struct xe_exec_queue *eq;
> +	unsigned long index;
> +
> +	mutex_lock(&guc->submission_state.lock);
> +	xa_for_each(&guc->submission_state.exec_queue_lookup, index, eq)
> +		xe_exec_queue_contexts_hwsp_rebase(eq);
> +	mutex_unlock(&guc->submission_state.lock);
> +}
> diff --git a/drivers/gpu/drm/xe/xe_guc_submit.h b/drivers/gpu/drm/xe/xe_guc_submit.h
> index 2c2d2936440d..2cc44298465f 100644
> --- a/drivers/gpu/drm/xe/xe_guc_submit.h
> +++ b/drivers/gpu/drm/xe/xe_guc_submit.h
> @@ -44,4 +44,6 @@ void
>  xe_guc_exec_queue_snapshot_free(struct xe_guc_submit_exec_queue_snapshot *snapshot);
>  void xe_guc_submit_print(struct xe_guc *guc, struct drm_printer *p);
>  
> +void xe_guc_contexts_hwsp_rebase(struct xe_guc *guc);
> +
>  #endif
> diff --git a/drivers/gpu/drm/xe/xe_lrc.c b/drivers/gpu/drm/xe/xe_lrc.c
> index 61a2e87990a9..525565480aef 100644
> --- a/drivers/gpu/drm/xe/xe_lrc.c
> +++ b/drivers/gpu/drm/xe/xe_lrc.c
> @@ -1184,6 +1184,22 @@ void xe_lrc_destroy(struct kref *ref)
>  	kfree(lrc);
>  }
>  
> +/**
> + * xe_lrc_update_hwctx_regs_with_address - Re-compute GGTT references within given LRC.
> + * @lrc: the &xe_lrc struct instance
> + */
> +void xe_lrc_update_hwctx_regs_with_address(struct xe_lrc *lrc)
> +{
> +	if (xe_lrc_has_indirect_ring_state(lrc)) {
> +		xe_lrc_write_ctx_reg(lrc, CTX_INDIRECT_RING_STATE,
> +				     __xe_lrc_indirect_ring_ggtt_addr(lrc));
> +
> +		xe_lrc_write_indirect_ctx_reg(lrc, INDIRECT_CTX_RING_START,
> +					      __xe_lrc_ring_ggtt_addr(lrc));
> +	} else
> +		xe_lrc_write_ctx_reg(lrc, CTX_RING_START, __xe_lrc_ring_ggtt_addr(lrc));
> +}
> +
>  void xe_lrc_set_ring_tail(struct xe_lrc *lrc, u32 tail)
>  {
>  	if (xe_lrc_has_indirect_ring_state(lrc))
> diff --git a/drivers/gpu/drm/xe/xe_lrc.h b/drivers/gpu/drm/xe/xe_lrc.h
> index eb6e8de8c939..e7a99cfd0abe 100644
> --- a/drivers/gpu/drm/xe/xe_lrc.h
> +++ b/drivers/gpu/drm/xe/xe_lrc.h
> @@ -88,6 +88,7 @@ bool xe_lrc_ring_is_idle(struct xe_lrc *lrc);
>  u32 xe_lrc_indirect_ring_ggtt_addr(struct xe_lrc *lrc);
>  u32 xe_lrc_ggtt_addr(struct xe_lrc *lrc);
>  u32 *xe_lrc_regs(struct xe_lrc *lrc);
> +void xe_lrc_update_hwctx_regs_with_address(struct xe_lrc *lrc);
>  
>  u32 xe_lrc_read_ctx_reg(struct xe_lrc *lrc, int reg_nr);
>  void xe_lrc_write_ctx_reg(struct xe_lrc *lrc, int reg_nr, u32 val);
> diff --git a/drivers/gpu/drm/xe/xe_sriov_vf.c b/drivers/gpu/drm/xe/xe_sriov_vf.c
> index ad863be4783e..0e73b79b3a73 100644
> --- a/drivers/gpu/drm/xe/xe_sriov_vf.c
> +++ b/drivers/gpu/drm/xe/xe_sriov_vf.c
> @@ -7,12 +7,14 @@
>  
>  #include "xe_assert.h"
>  #include "xe_device.h"
> +#include "xe_exec_queue_types.h"
>  #include "xe_gt.h"
>  #include "xe_gt_sriov_printk.h"
>  #include "xe_gt_sriov_vf.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"
> @@ -223,6 +225,15 @@ static int vf_post_migration_requery_guc(struct xe_device *xe)
>  	return ret;
>  }
>  
> +static void vf_post_migration_fixup_contexts(struct xe_device *xe)
> +{
> +	struct xe_gt *gt;
> +	unsigned int id;
> +
> +	for_each_gt(gt, xe, id)
> +		xe_guc_contexts_hwsp_rebase(&gt->uc.guc);
> +}
> +
>  static void vf_post_migration_fixup_ctb(struct xe_device *xe)
>  {
>  	struct xe_gt *gt;
> @@ -303,9 +314,11 @@ static void vf_post_migration_recovery(struct xe_device *xe)
>  		goto fail;
>  
>  	need_fixups = vf_post_migration_fixup_ggtt_nodes(xe);
> -	/* FIXME: add the recovery steps */
> -	if (need_fixups)
> +	if (need_fixups) {
> +		vf_post_migration_fixup_contexts(xe);
> +		/* FIXME: add the recovery steps */
>  		vf_post_migration_fixup_ctb(xe);
> +	}
>  
>  	vf_post_migration_kickstart(xe);
>  	vf_post_migration_notify_resfix_done(xe);



More information about the Intel-xe mailing list