Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
	dri-devel@lists.freedesktop.org,
	"Kenneth Graunke" <kenneth@whitecape.org>,
	"Daniel Vetter" <daniel.vetter@ffwll.ch>
Subject: [Intel-gfx] [PATCH v3 2/5] drm/i915/uapi: convert drm_i915_gem_caching to kernel doc
Date: Mon,  5 Jul 2021 14:53:07 +0100	[thread overview]
Message-ID: <20210705135310.1502437-2-matthew.auld@intel.com> (raw)
In-Reply-To: <20210705135310.1502437-1-matthew.auld@intel.com>

Convert all the drm_i915_gem_caching bits to proper kernel doc.

Suggested-by: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Jason Ekstrand <jason@jlekstrand.net>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Ramalingam C <ramalingam.c@intel.com>
---
 include/uapi/drm/i915_drm.h | 70 +++++++++++++++++++++++--------------
 1 file changed, 43 insertions(+), 27 deletions(-)

diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
index 2f70c48567c0..d13c6c5fad04 100644
--- a/include/uapi/drm/i915_drm.h
+++ b/include/uapi/drm/i915_drm.h
@@ -1363,43 +1363,59 @@ struct drm_i915_gem_busy {
 };
 
 /**
- * I915_CACHING_NONE
+ * struct drm_i915_gem_caching - Set or get the caching for given object
+ * handle.
  *
- * GPU access is not coherent with cpu caches. Default for machines without an
- * LLC.
- */
-#define I915_CACHING_NONE		0
-/**
- * I915_CACHING_CACHED
+ * Allow userspace to control the GTT caching bits for a given object when the
+ * object is later mapped through the ppGTT(or GGTT on older platforms lacking
+ * ppGTT support, or if the object is used for scanout). Note that this might
+ * require unbinding the object from the GTT first, if its current caching value
+ * doesn't match.
  *
- * GPU access is coherent with cpu caches and furthermore the data is cached in
- * last-level caches shared between cpu cores and the gpu GT. Default on
- * machines with HAS_LLC.
- */
-#define I915_CACHING_CACHED		1
-/**
- * I915_CACHING_DISPLAY
  *
- * Special GPU caching mode which is coherent with the scanout engines.
- * Transparently falls back to I915_CACHING_NONE on platforms where no special
- * cache mode (like write-through or gfdt flushing) is available. The kernel
- * automatically sets this mode when using a buffer as a scanout target.
- * Userspace can manually set this mode to avoid a costly stall and clflush in
- * the hotpath of drawing the first frame.
  */
-#define I915_CACHING_DISPLAY		2
-
 struct drm_i915_gem_caching {
 	/**
-	 * Handle of the buffer to set/get the caching level of. */
+	 * @handle: Handle of the buffer to set/get the caching level.
+	 */
 	__u32 handle;
 
 	/**
-	 * Cacheing level to apply or return value
+	 * @caching: The GTT caching level to apply or possible return value.
+	 *
+	 * The supported @caching values:
 	 *
-	 * bits0-15 are for generic caching control (i.e. the above defined
-	 * values). bits16-31 are reserved for platform-specific variations
-	 * (e.g. l3$ caching on gen7). */
+	 * I915_CACHING_NONE:
+	 *
+	 * GPU access is not coherent with CPU caches.  Default for machines
+	 * without an LLC. This means we need to manually clflush, if we want
+	 * GPU access to be coherent.
+	 *
+	 * I915_CACHING_CACHED:
+	 *
+	 * GPU access is coherent with CPU caches and furthermore the data is
+	 * cached in last-level caches shared between CPU cores and the GPU GT.
+	 * Default on machines with HAS_LLC. In general the fast shared
+	 * last-level cache(HAS_LLC) is considered much faster then platforms
+	 * which only support snooping(HAS_SNOOP), hence by default
+	 *
+	 * I915_CACHING_DISPLAY:
+	 *
+	 * Special GPU caching mode which is coherent with the scanout engines.
+	 * Transparently falls back to I915_CACHING_NONE on platforms where no
+	 * special cache mode (like write-through or gfdt flushing) is
+	 * available. The kernel automatically sets this mode when using a
+	 * buffer as a scanout target.  Userspace can manually set this mode to
+	 * avoid a costly stall and clflush in the hotpath of drawing the first
+	 * frame.
+	 *
+	 * Side note: On gen8+ this no longer does much since we lost the GGTT
+	 * caching bits. Although setting this is harmless, since it still
+	 * effectively falls back to I915_CACHING_NONE.
+	 */
+#define I915_CACHING_NONE		0
+#define I915_CACHING_CACHED		1
+#define I915_CACHING_DISPLAY		2
 	__u32 caching;
 };
 
-- 
2.26.3

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2021-07-05 13:53 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-05 13:53 [Intel-gfx] [PATCH v3 1/5] drm/i915: use consistent CPU mappings for pin_map users Matthew Auld
2021-07-05 13:53 ` Matthew Auld [this message]
2021-07-07  0:29   ` [Intel-gfx] [PATCH v3 2/5] drm/i915/uapi: convert drm_i915_gem_caching to kernel doc Ramalingam C
2021-07-05 13:53 ` [Intel-gfx] [PATCH v3 3/5] drm/i915/uapi: reject caching ioctls for discrete Matthew Auld
2021-07-07  0:38   ` Ramalingam C
2021-07-13 18:03   ` Kenneth Graunke
2021-07-05 13:53 ` [Intel-gfx] [PATCH v3 4/5] drm/i915/uapi: convert drm_i915_gem_set_domain to kernel doc Matthew Auld
2021-07-07  0:41   ` Ramalingam C
2021-07-05 13:53 ` [Intel-gfx] [PATCH v3 5/5] drm/i915/uapi: reject set_domain for discrete Matthew Auld
2021-07-07  0:49   ` Ramalingam C
2021-07-05 15:15 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [v3,1/5] drm/i915: use consistent CPU mappings for pin_map users Patchwork
2021-07-05 15:46 ` [Intel-gfx] ✗ Fi.CI.BAT: failure " Patchwork
2021-07-07  0:12 ` [Intel-gfx] [PATCH v3 1/5] " Ramalingam C
2021-07-07 11:46 ` Daniel Vetter

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=20210705135310.1502437-2-matthew.auld@intel.com \
    --to=matthew.auld@intel.com \
    --cc=daniel.vetter@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=kenneth@whitecape.org \
    --cc=thomas.hellstrom@linux.intel.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