public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Michal Wajdeczko <michal.wajdeczko@intel.com>
To: "Michał Winiarski" <michal.winiarski@intel.com>,
	"Alex Williamson" <alex.williamson@redhat.com>,
	"Lucas De Marchi" <lucas.demarchi@intel.com>,
	"Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
	"Rodrigo Vivi" <rodrigo.vivi@intel.com>,
	"Jason Gunthorpe" <jgg@ziepe.ca>,
	"Yishai Hadas" <yishaih@nvidia.com>,
	"Kevin Tian" <kevin.tian@intel.com>,
	"Shameer Kolothum" <shameerali.kolothum.thodi@huawei.com>,
	intel-xe@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	kvm@vger.kernel.org
Cc: <dri-devel@lists.freedesktop.org>,
	Matthew Brost <matthew.brost@intel.com>,
	Jani Nikula <jani.nikula@linux.intel.com>,
	"Joonas Lahtinen" <joonas.lahtinen@linux.intel.com>,
	Tvrtko Ursulin <tursulin@ursulin.net>,
	David Airlie <airlied@gmail.com>, Simona Vetter <simona@ffwll.ch>,
	Lukasz Laguna <lukasz.laguna@intel.com>
Subject: Re: [PATCH 17/26] drm/xe/pf: Add helpers for VF GGTT migration data handling
Date: Mon, 13 Oct 2025 14:17:56 +0200	[thread overview]
Message-ID: <83f83d86-a89d-4315-aa01-9c48e782bd60@intel.com> (raw)
In-Reply-To: <20251011193847.1836454-18-michal.winiarski@intel.com>



On 10/11/2025 9:38 PM, Michał Winiarski wrote:
> In an upcoming change, the VF GGTT migration data will be handled as
> part of VF control state machine. Add the necessary helpers to allow the
> migration data transfer to/from the HW GGTT resource.
> 
> Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_ggtt.c               | 92 ++++++++++++++++++++++
>  drivers/gpu/drm/xe/xe_ggtt.h               |  2 +
>  drivers/gpu/drm/xe/xe_ggtt_types.h         |  2 +
>  drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c | 64 +++++++++++++++
>  drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h |  5 ++
>  5 files changed, 165 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c
> index aca7ae5489b91..89c0ad56c6a8a 100644
> --- a/drivers/gpu/drm/xe/xe_ggtt.c
> +++ b/drivers/gpu/drm/xe/xe_ggtt.c
> @@ -138,6 +138,14 @@ static void xe_ggtt_set_pte_and_flush(struct xe_ggtt *ggtt, u64 addr, u64 pte)
>  	ggtt_update_access_counter(ggtt);
>  }
>  
> +static u64 xe_ggtt_get_pte(struct xe_ggtt *ggtt, u64 addr)
> +{
> +	xe_tile_assert(ggtt->tile, !(addr & XE_PTE_MASK));
> +	xe_tile_assert(ggtt->tile, addr < ggtt->size);
> +
> +	return readq(&ggtt->gsm[addr >> XE_PTE_SHIFT]);
> +}
> +
>  static void xe_ggtt_clear(struct xe_ggtt *ggtt, u64 start, u64 size)
>  {
>  	u16 pat_index = tile_to_xe(ggtt->tile)->pat.idx[XE_CACHE_WB];
> @@ -220,16 +228,19 @@ void xe_ggtt_might_lock(struct xe_ggtt *ggtt)
>  static const struct xe_ggtt_pt_ops xelp_pt_ops = {
>  	.pte_encode_flags = xelp_ggtt_pte_flags,
>  	.ggtt_set_pte = xe_ggtt_set_pte,
> +	.ggtt_get_pte = xe_ggtt_get_pte,
>  };
>  
>  static const struct xe_ggtt_pt_ops xelpg_pt_ops = {
>  	.pte_encode_flags = xelpg_ggtt_pte_flags,
>  	.ggtt_set_pte = xe_ggtt_set_pte,
> +	.ggtt_get_pte = xe_ggtt_get_pte,
>  };
>  
>  static const struct xe_ggtt_pt_ops xelpg_pt_wa_ops = {
>  	.pte_encode_flags = xelpg_ggtt_pte_flags,
>  	.ggtt_set_pte = xe_ggtt_set_pte_and_flush,
> +	.ggtt_get_pte = xe_ggtt_get_pte,
>  };
>  
>  static void __xe_ggtt_init_early(struct xe_ggtt *ggtt, u32 reserved)
> @@ -914,6 +925,87 @@ void xe_ggtt_assign(const struct xe_ggtt_node *node, u16 vfid)
>  	xe_ggtt_assign_locked(node->ggtt, &node->base, vfid);
>  	mutex_unlock(&node->ggtt->lock);
>  }
> +
> +/**
> + * xe_ggtt_node_save - Save a &struct xe_ggtt_node to a buffer
> + * @node: the &struct xe_ggtt_node to be saved
> + * @dst: destination buffer

correct me: this is buffer for the PTEs

> + * @size: destination buffer size in bytes

and this is size of above buffer

> + *
> + * Return: 0 on success or a negative error code on failure.
> + */
> +int xe_ggtt_node_save(struct xe_ggtt_node *node, void *dst, size_t size)
> +{
> +	struct xe_ggtt *ggtt;
> +	u64 start, end;
> +	u64 *buf = dst;
> +
> +	if (!node || !node->ggtt)
> +		return -ENOENT;

hmm, non-NULL node must be initialized by xe_ggtt_node_init() which sets the .ggtt so this second check is redundant

> +
> +	mutex_lock(&node->ggtt->lock);

	guard(mutex)(&node->ggtt->lock);

> +
> +	ggtt = node->ggtt;
> +	start = node->base.start;
> +	end = start + node->base.size - 1;
> +
> +	if (node->base.size < size) {

so that's looks wrong, we are about to save 64bit PTEs of that node

we should compare size of all PTEs not the size of address space allocated by this node

> +		mutex_unlock(&node->ggtt->lock);
> +		return -EINVAL;
> +	}
> +
> +	while (start < end) {
> +		*buf++ = ggtt->pt_ops->ggtt_get_pte(ggtt, start) & ~GGTT_PTE_VFID;
> +		start += XE_PAGE_SIZE;
> +	}
> +
> +	mutex_unlock(&node->ggtt->lock);
> +
> +	return 0;
> +}
> +
> +/**
> + * xe_ggtt_node_load - Load a &struct xe_ggtt_node from a buffer
> + * @node: the &struct xe_ggtt_node to be loaded
> + * @src: source buffer
> + * @size: source buffer size in bytes
> + * @vfid: VF identifier
> + *
> + * Return: 0 on success or a negative error code on failure.
> + */
> +int xe_ggtt_node_load(struct xe_ggtt_node *node, const void *src, size_t size, u16 vfid)
> +{
> +	struct xe_ggtt *ggtt;
> +	u64 start, end;
> +	const u64 *buf = src;
> +	u64 vfid_pte = xe_encode_vfid_pte(vfid);

try to define vars in reverse xmas tree order

> +
> +	if (!node || !node->ggtt)
> +		return -ENOENT;
> +
> +	mutex_lock(&node->ggtt->lock);

use guard(mutex)

> +
> +	ggtt = node->ggtt;
> +	start = node->base.start;
> +	end = start + size - 1;
> +
> +	if (node->base.size != size) {
> +		mutex_unlock(&node->ggtt->lock);
> +		return -EINVAL;
> +	}
> +
> +	while (start < end) {
> +		ggtt->pt_ops->ggtt_set_pte(ggtt, start, (*buf & ~GGTT_PTE_VFID) | vfid_pte);
> +		start += XE_PAGE_SIZE;
> +		buf++;
> +	}
> +	xe_ggtt_invalidate(ggtt);
> +
> +	mutex_unlock(&node->ggtt->lock);
> +
> +	return 0;
> +}
> +
>  #endif
>  
>  /**
> diff --git a/drivers/gpu/drm/xe/xe_ggtt.h b/drivers/gpu/drm/xe/xe_ggtt.h
> index 75fc7a1efea76..469b3a6ca14b4 100644
> --- a/drivers/gpu/drm/xe/xe_ggtt.h
> +++ b/drivers/gpu/drm/xe/xe_ggtt.h
> @@ -43,6 +43,8 @@ u64 xe_ggtt_print_holes(struct xe_ggtt *ggtt, u64 alignment, struct drm_printer
>  
>  #ifdef CONFIG_PCI_IOV
>  void xe_ggtt_assign(const struct xe_ggtt_node *node, u16 vfid);
> +int xe_ggtt_node_save(struct xe_ggtt_node *node, void *dst, size_t size);
> +int xe_ggtt_node_load(struct xe_ggtt_node *node, const void *src, size_t size, u16 vfid);
>  #endif
>  
>  #ifndef CONFIG_LOCKDEP
> diff --git a/drivers/gpu/drm/xe/xe_ggtt_types.h b/drivers/gpu/drm/xe/xe_ggtt_types.h
> index c5e999d58ff2a..dacd796f81844 100644
> --- a/drivers/gpu/drm/xe/xe_ggtt_types.h
> +++ b/drivers/gpu/drm/xe/xe_ggtt_types.h
> @@ -78,6 +78,8 @@ struct xe_ggtt_pt_ops {
>  	u64 (*pte_encode_flags)(struct xe_bo *bo, u16 pat_index);
>  	/** @ggtt_set_pte: Directly write into GGTT's PTE */
>  	void (*ggtt_set_pte)(struct xe_ggtt *ggtt, u64 addr, u64 pte);
> +	/** @ggtt_get_pte: Directly read from GGTT's PTE */
> +	u64 (*ggtt_get_pte)(struct xe_ggtt *ggtt, u64 addr);
>  };
>  
>  #endif
> diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c
> index b2e5c52978e6a..51027921b2988 100644
> --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c
> +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c
> @@ -726,6 +726,70 @@ int xe_gt_sriov_pf_config_set_fair_ggtt(struct xe_gt *gt, unsigned int vfid,
>  	return xe_gt_sriov_pf_config_bulk_set_ggtt(gt, vfid, num_vfs, fair);
>  }
>  
> +/**
> + * xe_gt_sriov_pf_config_ggtt_save - Save a VF provisioned GGTT data into a buffer.
> + * @gt: the &struct xe_gt
> + * @vfid: VF identifier
> + * @buf: the GGTT data destination buffer
> + * @size: the size of the buffer
> + *
> + * This function can only be called on PF.
> + *
> + * Return: 0 on success or a negative error code on failure.
> + */
> +int xe_gt_sriov_pf_config_ggtt_save(struct xe_gt *gt, unsigned int vfid,
> +				    void *buf, size_t size)
> +{
> +	struct xe_gt_sriov_config *config;
> +	ssize_t ret;

int

> +
> +	xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt)));
> +	xe_gt_assert(gt, vfid);
> +	xe_gt_assert(gt, !(!buf ^ !size));

there seems to be no "query" option for this call, so both buf & size must be valid

> +
> +	mutex_lock(xe_gt_sriov_pf_master_mutex(gt));
> +	config = pf_pick_vf_config(gt, vfid);
> +	size = size / sizeof(u64) * XE_PAGE_SIZE;

?? something is wrong here - why do we have to change the size of the buf?

> +
> +	ret = xe_ggtt_node_save(config->ggtt_region, buf, size);
> +
> +	mutex_unlock(xe_gt_sriov_pf_master_mutex(gt));
> +
> +	return ret;
> +}
> +
> +/**
> + * xe_gt_sriov_pf_config_ggtt_restore - Restore a VF provisioned GGTT data from a buffer.
> + * @gt: the &struct xe_gt
> + * @vfid: VF identifier
> + * @buf: the GGTT data source buffer
> + * @size: the size of the buffer
> + *
> + * This function can only be called on PF.
> + *
> + * Return: 0 on success or a negative error code on failure.
> + */
> +int xe_gt_sriov_pf_config_ggtt_restore(struct xe_gt *gt, unsigned int vfid,
> +				       const void *buf, size_t size)
> +{
> +	struct xe_gt_sriov_config *config;
> +	ssize_t ret;
> +
> +	xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt)));
> +	xe_gt_assert(gt, vfid);
> +	xe_gt_assert(gt, !(!buf ^ !size));
> +
> +	mutex_lock(xe_gt_sriov_pf_master_mutex(gt));
> +	config = pf_pick_vf_config(gt, vfid);
> +	size = size / sizeof(u64) * XE_PAGE_SIZE;
> +
> +	ret = xe_ggtt_node_load(config->ggtt_region, buf, size, vfid);
> +
> +	mutex_unlock(xe_gt_sriov_pf_master_mutex(gt));
> +
> +	return ret;
> +}

ditto

> +
>  static u32 pf_get_min_spare_ctxs(struct xe_gt *gt)
>  {
>  	/* XXX: preliminary */
> diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h
> index 513e6512a575b..6916b8f58ebf2 100644
> --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h
> +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h
> @@ -61,6 +61,11 @@ ssize_t xe_gt_sriov_pf_config_save(struct xe_gt *gt, unsigned int vfid, void *bu
>  int xe_gt_sriov_pf_config_restore(struct xe_gt *gt, unsigned int vfid,
>  				  const void *buf, size_t size);
>  
> +int xe_gt_sriov_pf_config_ggtt_save(struct xe_gt *gt, unsigned int vfid,
> +				    void *buf, size_t size);
> +int xe_gt_sriov_pf_config_ggtt_restore(struct xe_gt *gt, unsigned int vfid,
> +				       const void *buf, size_t size);
> +
>  bool xe_gt_sriov_pf_config_is_empty(struct xe_gt *gt, unsigned int vfid);
>  
>  int xe_gt_sriov_pf_config_init(struct xe_gt *gt);


  reply	other threads:[~2025-10-13 12:18 UTC|newest]

Thread overview: 82+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-11 19:38 [PATCH 00/26] vfio/xe: Add driver variant for Xe VF migration Michał Winiarski
2025-10-11 19:38 ` [PATCH 01/26] drm/xe/pf: Remove GuC version check for migration support Michał Winiarski
2025-10-12 18:31   ` Michal Wajdeczko
2025-10-20 14:46     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 02/26] drm/xe: Move migration support to device-level struct Michał Winiarski
2025-10-12 18:58   ` Michal Wajdeczko
2025-10-20 14:48     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 03/26] drm/xe/pf: Add save/restore control state stubs and connect to debugfs Michał Winiarski
2025-10-12 20:09   ` Michal Wajdeczko
2025-10-11 19:38 ` [PATCH 04/26] drm/xe/pf: Extract migration mutex out of its struct Michał Winiarski
2025-10-12 19:08   ` Matthew Brost
2025-10-20 14:50     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 05/26] drm/xe/pf: Add data structures and handlers for migration rings Michał Winiarski
2025-10-12 21:06   ` Michal Wajdeczko
2025-10-20 14:56     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 06/26] drm/xe/pf: Add helpers for migration data allocation / free Michał Winiarski
2025-10-12 19:12   ` Matthew Brost
2025-10-21  0:26     ` Michał Winiarski
2025-10-13 10:15   ` Michal Wajdeczko
2025-10-21  0:01     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 07/26] drm/xe/pf: Add support for encap/decap of bitstream to/from packet Michał Winiarski
2025-10-11 22:28   ` kernel test robot
2025-10-13 10:46   ` Michal Wajdeczko
2025-10-21  0:25     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 08/26] drm/xe/pf: Add minimalistic migration descriptor Michał Winiarski
2025-10-11 22:52   ` kernel test robot
2025-10-13 10:56   ` Michal Wajdeczko
2025-10-21  0:31     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 09/26] drm/xe/pf: Expose VF migration data size over debugfs Michał Winiarski
2025-10-12 19:15   ` Matthew Brost
2025-10-21  0:37     ` Michał Winiarski
2025-10-13 11:04   ` Michal Wajdeczko
2025-10-21  0:42     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 10/26] drm/xe: Add sa/guc_buf_cache sync interface Michał Winiarski
2025-10-12 18:06   ` Matthew Brost
2025-10-21  0:45     ` Michał Winiarski
2025-10-13 11:20   ` Michal Wajdeczko
2025-10-21  0:44     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 11/26] drm/xe: Allow the caller to pass guc_buf_cache size Michał Winiarski
2025-10-11 23:35   ` kernel test robot
2025-10-13 11:08   ` Michal Wajdeczko
2025-10-21  0:47     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 12/26] drm/xe/pf: Increase PF GuC Buffer Cache size and use it for VF migration Michał Winiarski
2025-10-13 11:27   ` Michal Wajdeczko
2025-10-21  0:50     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 13/26] drm/xe/pf: Remove GuC migration data save/restore from GT debugfs Michał Winiarski
2025-10-13 11:36   ` Michal Wajdeczko
2025-10-11 19:38 ` [PATCH 14/26] drm/xe/pf: Don't save GuC VF migration data on pause Michał Winiarski
2025-10-13 11:42   ` Michal Wajdeczko
2025-10-11 19:38 ` [PATCH 15/26] drm/xe/pf: Switch VF migration GuC save/restore to struct migration data Michał Winiarski
2025-10-11 19:38 ` [PATCH 16/26] drm/xe/pf: Handle GuC migration data as part of PF control Michał Winiarski
2025-10-13 11:56   ` Michal Wajdeczko
2025-10-21  0:52     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 17/26] drm/xe/pf: Add helpers for VF GGTT migration data handling Michał Winiarski
2025-10-13 12:17   ` Michal Wajdeczko [this message]
2025-10-21  1:00     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 18/26] drm/xe/pf: Handle GGTT migration data as part of PF control Michał Winiarski
2025-10-13 12:36   ` Michal Wajdeczko
2025-10-21  1:16     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 19/26] drm/xe/pf: Add helpers for VF MMIO migration data handling Michał Winiarski
2025-10-13 13:28   ` Michal Wajdeczko
2025-10-11 19:38 ` [PATCH 20/26] drm/xe/pf: Handle MMIO migration data as part of PF control Michał Winiarski
2025-10-11 19:38 ` [PATCH 21/26] drm/xe/pf: Add helper to retrieve VF's LMEM object Michał Winiarski
2025-10-11 19:38 ` [PATCH 22/26] drm/xe/migrate: Add function for raw copy of VRAM and CCS Michał Winiarski
2025-10-12 18:54   ` Matthew Brost
2025-10-11 19:38 ` [PATCH 23/26] drm/xe/pf: Handle VRAM migration data as part of PF control Michał Winiarski
2025-10-11 19:38 ` [PATCH 24/26] drm/xe/pf: Add wait helper for VF FLR Michał Winiarski
2025-10-13 13:49   ` Michal Wajdeczko
2025-10-11 19:38 ` [PATCH 25/26] drm/xe/pf: Export helpers for VFIO Michał Winiarski
2025-10-12 18:32   ` Matthew Brost
2025-10-21  1:38     ` Michał Winiarski
2025-10-13 14:02   ` Michal Wajdeczko
2025-10-21  1:49     ` Michał Winiarski
2025-10-11 19:38 ` [PATCH 26/26] vfio/xe: Add vendor-specific vfio_pci driver for Intel graphics Michał Winiarski
2025-10-13 19:00   ` Rodrigo Vivi
2025-10-21 23:03   ` Jason Gunthorpe
2025-10-21 23:14     ` Matthew Brost
2025-10-21 23:38       ` Jason Gunthorpe
2025-10-22  1:15         ` Matthew Brost
2025-10-22 13:02           ` Jason Gunthorpe
2025-10-22  9:05     ` Michał Winiarski
2025-10-27  7:02       ` Tian, Kevin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=83f83d86-a89d-4315-aa01-9c48e782bd60@intel.com \
    --to=michal.wajdeczko@intel.com \
    --cc=airlied@gmail.com \
    --cc=alex.williamson@redhat.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=jani.nikula@linux.intel.com \
    --cc=jgg@ziepe.ca \
    --cc=joonas.lahtinen@linux.intel.com \
    --cc=kevin.tian@intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lucas.demarchi@intel.com \
    --cc=lukasz.laguna@intel.com \
    --cc=matthew.brost@intel.com \
    --cc=michal.winiarski@intel.com \
    --cc=rodrigo.vivi@intel.com \
    --cc=shameerali.kolothum.thodi@huawei.com \
    --cc=simona@ffwll.ch \
    --cc=thomas.hellstrom@linux.intel.com \
    --cc=tursulin@ursulin.net \
    --cc=yishaih@nvidia.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox