Intel-XE Archive on lore.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>,
	intel-xe@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	kvm@vger.kernel.org, "Matthew Brost" <matthew.brost@intel.com>
Cc: <dri-devel@lists.freedesktop.org>,
	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 v2 10/26] drm/xe: Allow the caller to pass guc_buf_cache size
Date: Thu, 23 Oct 2025 01:13:26 +0200	[thread overview]
Message-ID: <dc361e6e-a5bd-429f-8988-038723450a03@intel.com> (raw)
In-Reply-To: <20251021224133.577765-11-michal.winiarski@intel.com>



On 10/22/2025 12:41 AM, Michał Winiarski wrote:
> An upcoming change will use GuC buffer cache as a place where GuC
> migration data will be stored, and the memory requirement for that is
> larger than indirect data.
> Allow the caller to pass the size based on the intended usecase.
> 
> Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
> ---
>  drivers/gpu/drm/xe/tests/xe_guc_buf_kunit.c | 2 +-
>  drivers/gpu/drm/xe/xe_guc.c                 | 4 ++--
>  drivers/gpu/drm/xe/xe_guc_buf.c             | 6 +++---
>  drivers/gpu/drm/xe/xe_guc_buf.h             | 4 +++-
>  4 files changed, 9 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/tests/xe_guc_buf_kunit.c b/drivers/gpu/drm/xe/tests/xe_guc_buf_kunit.c
> index d266882adc0e0..485e7a70e6bb7 100644
> --- a/drivers/gpu/drm/xe/tests/xe_guc_buf_kunit.c
> +++ b/drivers/gpu/drm/xe/tests/xe_guc_buf_kunit.c
> @@ -72,7 +72,7 @@ static int guc_buf_test_init(struct kunit *test)
>  	kunit_activate_static_stub(test, xe_managed_bo_create_pin_map,
>  				   replacement_xe_managed_bo_create_pin_map);
>  
> -	KUNIT_ASSERT_EQ(test, 0, xe_guc_buf_cache_init(&guc->buf));
> +	KUNIT_ASSERT_EQ(test, 0, xe_guc_buf_cache_init(&guc->buf, XE_GUC_BUF_CACHE_DEFAULT_SIZE));
>  
>  	test->priv = &guc->buf;
>  	return 0;
> diff --git a/drivers/gpu/drm/xe/xe_guc.c b/drivers/gpu/drm/xe/xe_guc.c
> index ecc3e091b89e6..7c65528859ecb 100644
> --- a/drivers/gpu/drm/xe/xe_guc.c
> +++ b/drivers/gpu/drm/xe/xe_guc.c
> @@ -812,7 +812,7 @@ static int vf_guc_init_post_hwconfig(struct xe_guc *guc)
>  	if (err)
>  		return err;
>  
> -	err = xe_guc_buf_cache_init(&guc->buf);
> +	err = xe_guc_buf_cache_init(&guc->buf, XE_GUC_BUF_CACHE_DEFAULT_SIZE);
>  	if (err)
>  		return err;
>  
> @@ -860,7 +860,7 @@ int xe_guc_init_post_hwconfig(struct xe_guc *guc)
>  	if (ret)
>  		return ret;
>  
> -	ret = xe_guc_buf_cache_init(&guc->buf);
> +	ret = xe_guc_buf_cache_init(&guc->buf, XE_GUC_BUF_CACHE_DEFAULT_SIZE);
>  	if (ret)
>  		return ret;
>  
> diff --git a/drivers/gpu/drm/xe/xe_guc_buf.c b/drivers/gpu/drm/xe/xe_guc_buf.c
> index 4d8a4712309f4..ed096a0331244 100644
> --- a/drivers/gpu/drm/xe/xe_guc_buf.c
> +++ b/drivers/gpu/drm/xe/xe_guc_buf.c
> @@ -28,16 +28,16 @@ static struct xe_gt *cache_to_gt(struct xe_guc_buf_cache *cache)
>   * @cache: the &xe_guc_buf_cache to initialize
>   *
>   * The Buffer Cache allows to obtain a reusable buffer that can be used to pass
> - * indirect H2G data to GuC without a need to create a ad-hoc allocation.
> + * data to GuC or read data from GuC without a need to create a ad-hoc allocation.
>   *
>   * Return: 0 on success or a negative error code on failure.
>   */
> -int xe_guc_buf_cache_init(struct xe_guc_buf_cache *cache)
> +int xe_guc_buf_cache_init(struct xe_guc_buf_cache *cache, u32 size)
>  {
>  	struct xe_gt *gt = cache_to_gt(cache);
>  	struct xe_sa_manager *sam;
>  
> -	sam = __xe_sa_bo_manager_init(gt_to_tile(gt), SZ_8K, 0, sizeof(u32));
> +	sam = __xe_sa_bo_manager_init(gt_to_tile(gt), size, 0, sizeof(u32));
>  	if (IS_ERR(sam))
>  		return PTR_ERR(sam);
>  	cache->sam = sam;
> diff --git a/drivers/gpu/drm/xe/xe_guc_buf.h b/drivers/gpu/drm/xe/xe_guc_buf.h
> index c5e0f1fd24d74..5210703309e81 100644
> --- a/drivers/gpu/drm/xe/xe_guc_buf.h
> +++ b/drivers/gpu/drm/xe/xe_guc_buf.h
> @@ -11,7 +11,9 @@
>  
>  #include "xe_guc_buf_types.h"
>  
> -int xe_guc_buf_cache_init(struct xe_guc_buf_cache *cache);
> +#define XE_GUC_BUF_CACHE_DEFAULT_SIZE SZ_8K
> +
> +int xe_guc_buf_cache_init(struct xe_guc_buf_cache *cache, u32 size);

alternatively, to minimize code changes, we can have:

int xe_guc_buf_cache_init_size(struct xe_guc_buf_cache *cache, u32 size);

static inline int xe_guc_buf_cache_init(struct xe_guc_buf_cache *cache)
{
	return xe_guc_buf_cache_init_size(cache, XE_GUC_BUF_CACHE_DEFAULT_SIZE);
}

but up to you,

Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com>

>  u32 xe_guc_buf_cache_dwords(struct xe_guc_buf_cache *cache);
>  struct xe_guc_buf xe_guc_buf_reserve(struct xe_guc_buf_cache *cache, u32 dwords);
>  struct xe_guc_buf xe_guc_buf_from_data(struct xe_guc_buf_cache *cache,


  reply	other threads:[~2025-10-22 23:13 UTC|newest]

Thread overview: 75+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-21 22:41 [PATCH v2 00/26] vfio/xe: Add driver variant for Xe VF migration Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 01/26] drm/xe/pf: Remove GuC version check for migration support Michał Winiarski
2025-10-28  2:33   ` Tian, Kevin
2025-10-28  8:06     ` Winiarski, Michal
2025-10-21 22:41 ` [PATCH v2 02/26] drm/xe: Move migration support to device-level struct Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 03/26] drm/xe/pf: Add save/restore control state stubs and connect to debugfs Michał Winiarski
2025-10-22 22:31   ` Michal Wajdeczko
2025-10-27 12:02     ` Michał Winiarski
2025-10-28  3:06   ` Tian, Kevin
2025-10-28  8:02     ` Michal Wajdeczko
2025-10-21 22:41 ` [PATCH v2 04/26] drm/xe/pf: Add data structures and handlers for migration rings Michał Winiarski
2025-10-22 22:06   ` Michal Wajdeczko
2025-10-27 12:33     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 05/26] drm/xe/pf: Add helpers for migration data allocation / free Michał Winiarski
2025-10-22 22:18   ` Michal Wajdeczko
2025-10-27 12:47     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 06/26] drm/xe/pf: Add support for encap/decap of bitstream to/from packet Michał Winiarski
2025-10-22 22:34   ` Michal Wajdeczko
2025-10-27 13:27     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 07/26] drm/xe/pf: Add minimalistic migration descriptor Michał Winiarski
2025-10-22 22:49   ` Michal Wajdeczko
2025-10-27 14:52     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 08/26] drm/xe/pf: Expose VF migration data size over debugfs Michał Winiarski
2025-10-22 23:02   ` Michal Wajdeczko
2025-10-21 22:41 ` [PATCH v2 09/26] drm/xe: Add sa/guc_buf_cache sync interface Michał Winiarski
2025-10-22 23:05   ` Michal Wajdeczko
2025-10-21 22:41 ` [PATCH v2 10/26] drm/xe: Allow the caller to pass guc_buf_cache size Michał Winiarski
2025-10-22 23:13   ` Michal Wajdeczko [this message]
2025-10-21 22:41 ` [PATCH v2 11/26] drm/xe/pf: Increase PF GuC Buffer Cache size and use it for VF migration Michał Winiarski
2025-10-23 17:37   ` Michal Wajdeczko
2025-10-28 10:46     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 12/26] drm/xe/pf: Remove GuC migration data save/restore from GT debugfs Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 13/26] drm/xe/pf: Don't save GuC VF migration data on pause Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 14/26] drm/xe/pf: Switch VF migration GuC save/restore to struct migration data Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 15/26] drm/xe/pf: Handle GuC migration data as part of PF control Michał Winiarski
2025-10-23 20:39   ` Michal Wajdeczko
2025-10-28 13:04     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 16/26] drm/xe/pf: Add helpers for VF GGTT migration data handling Michał Winiarski
2025-10-23 21:50   ` Michal Wajdeczko
2025-10-28 17:03     ` Michał Winiarski
2025-10-28  3:22   ` Tian, Kevin
2025-10-28  7:38     ` Michal Wajdeczko
2025-10-21 22:41 ` [PATCH v2 17/26] drm/xe/pf: Handle GGTT migration data as part of PF control Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 18/26] drm/xe/pf: Add helpers for VF MMIO migration data handling Michał Winiarski
2025-10-23 22:10   ` Michal Wajdeczko
2025-10-28 23:37     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 19/26] drm/xe/pf: Handle MMIO migration data as part of PF control Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 20/26] drm/xe/pf: Add helper to retrieve VF's LMEM object Michał Winiarski
2025-10-23 20:25   ` Michal Wajdeczko
2025-10-28 23:40     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 21/26] drm/xe/migrate: Add function to copy of VRAM data in chunks Michał Winiarski
2025-10-23 19:29   ` Michal Wajdeczko
2025-10-30  6:07     ` Laguna, Lukasz
2025-10-21 22:41 ` [PATCH v2 22/26] drm/xe/pf: Handle VRAM migration data as part of PF control Michał Winiarski
2025-10-23 11:44   ` kernel test robot
2025-10-23 19:54   ` Michal Wajdeczko
2025-10-29  8:54     ` Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 23/26] drm/xe/pf: Add wait helper for VF FLR Michał Winiarski
2025-10-21 22:41 ` [PATCH v2 24/26] drm/xe/pf: Enable SR-IOV VF migration for PTL and BMG Michał Winiarski
2025-10-23 20:15   ` Michal Wajdeczko
2025-10-21 22:41 ` [PATCH v2 25/26] drm/xe/pf: Export helpers for VFIO Michał Winiarski
2025-10-28  3:28   ` Tian, Kevin
2025-10-21 22:41 ` [PATCH v2 26/26] vfio/xe: Add vendor-specific vfio_pci driver for Intel graphics Michał Winiarski
2025-10-22  7:12   ` Christoph Hellwig
2025-10-22  8:52     ` Michał Winiarski
2025-10-22  8:54       ` Christoph Hellwig
2025-10-22  9:12         ` Michał Winiarski
     [not found]           ` <20251022113355.GC21554@ziepe.ca>
2025-10-22 13:27             ` Michał Winiarski
2025-10-27  7:24   ` Tian, Kevin
2025-10-29 20:46     ` Winiarski, Michal
2025-10-27  7:26   ` Tian, Kevin
2025-10-21 22:50 ` ✗ CI.checkpatch: warning for vfio/xe: Add driver variant for Xe VF migration (rev2) Patchwork
2025-10-21 22:52 ` ✓ CI.KUnit: success " Patchwork
2025-10-21 23:31 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-22  2:54 ` ✗ Xe.CI.Full: failure " Patchwork

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=dc361e6e-a5bd-429f-8988-038723450a03@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=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