Skip to contentSkip to conversation

drm/doc: add rfc section for small BAR uapi

PATCH Superseded Matthew Auld 7 participants 29 messages Apr 20, 2022 – May 3, 2022

Subsystems (3)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 drm/doc: add rfc section for small BAR uapi

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
---
 Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++
 Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
 Documentation/gpu/rfc/index.rst          |   4 +
 3 files changed, 252 insertions(+)
 create mode 100644 Documentation/gpu/rfc/i915_small_bar.h
 create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst
Documentation/gpu/rfc/i915_small_bar.h new file +190 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {
+	/** @region: The class:instance pair encoding */
+	struct drm_i915_gem_memory_class_instance region;
+
+	/** @rsvd0: MBZ */
+	__u32 rsvd0;
+
+	/** @probed_size: Memory probed by the driver (-1 = unknown) */
+	__u64 probed_size;
+
+	/** @unallocated_size: Estimate of memory remaining (-1 = unknown) */
+	__u64 unallocated_size;
+
+	union {
+		/** @rsvd1: MBZ */
+		__u64 rsvd1[8];
+		struct {
+			/**
+			 * @probed_cpu_visible_size: Memory probed by the driver
+			 * that is CPU accessible. (-1 = unknown).
+			 *
+			 * This will be always be <= @probed_size, and the
+			 * remainder(if there is any) will not be CPU
+			 * accessible.
+			 */
+			__u64 probed_cpu_visible_size;
+		};
+	};
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+	/**
+	 * @size: Requested size for the object.
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 *
+	 * Note that for some devices we have might have further minimum
+	 * page-size restrictions(larger than 4K), like for device local-memory.
+	 * However in general the final size here should always reflect any
+	 * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+	 * extension to place the object in device local-memory.
+	 */
+	__u64 size;
+	/**
+	 * @handle: Returned handle for the object.
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+	/**
+	 * @flags: Optional flags.
+	 *
+	 * Supported values:
+	 *
+	 * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+	 * the object will need to be accessed via the CPU.
+	 *
+	 * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+	 * only strictly required on platforms where only some of the device
+	 * memory is directly visible or mappable through the CPU, like on DG2+.
+	 *
+	 * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+	 * ensure we can always spill the allocation to system memory, if we
+	 * can't place the object in the mappable part of
+	 * I915_MEMORY_CLASS_DEVICE.
+	 *
+	 * Note that since the kernel only supports flat-CCS on objects that can
+	 * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+	 * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+	 * flat-CCS.
+	 *
+	 * Without this hint, the kernel will assume that non-mappable
+	 * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+	 * kernel can still migrate the object to the mappable part, as a last
+	 * resort, if userspace ever CPU faults this object, but this might be
+	 * expensive, and so ideally should be avoided.
+	 */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+	__u32 flags;
+	/**
+	 * @extensions: The chain of extensions to apply to this object.
+	 *
+	 * This will be useful in the future when we need to support several
+	 * different extensions, and we need to apply more than one when
+	 * creating the object. See struct i915_user_extension.
+	 *
+	 * If we don't supply any extensions then we get the same old gem_create
+	 * behaviour.
+	 *
+	 * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+	 * struct drm_i915_gem_create_ext_memory_regions.
+	 *
+	 * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+	 * struct drm_i915_gem_create_ext_protected_content.
+	 */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+	__u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO	5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *	struct drm_i915_query_vma_info info = {};
+ *	struct drm_i915_query_item item = {
+ *		.data_ptr = (uintptr_t)&info,
+ *		.query_id = DRM_I915_QUERY_VMA_INFO,
+ *	};
+ *	struct drm_i915_query query = {
+ *		.num_items = 1,
+ *		.items_ptr = (uintptr_t)&item,
+ *	};
+ *	int err;
+ *
+ *	// Unlike some other types of queries, there is no need to first query
+ *	// the size of the data_ptr blob here, since we already know ahead of
+ *	// time how big this needs to be.
+ *	item.length = sizeof(info);
+ *
+ *	// Next we fill in the vm_id and ppGTT address of the vma we wish
+ *	// to query, before then firing off the query.
+ *	info.vm_id = vm_id;
+ *	info.offset = gtt_address;
+ *	err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *	if (err || item.length < 0) ...
+ *
+ *	// If all went well we can now inspect the returned attributes.
+ *	if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+	/**
+	 * @vm_id: The given vm id that contains the vma. The id is the value
+	 * returned by the DRM_I915_GEM_VM_CREATE. See struct
+	 * drm_i915_gem_vm_control.vm_id.
+	 */
+	__u32 vm_id;
+	/** @pad: MBZ. */
+	__u32 pad;
+	/**
+	 * @offset: The corresponding ppGTT address of the vma which the kernel
+	 * will use to perform the lookup.
+	 */
+	__u64 offset;
+	/**
+	 * @attributes: The returned attributes for the given vma.
+	 *
+	 * Possible values:
+	 *
+	 * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+	 * vma are currently CPU accessible. If this is not set then the vma is
+	 * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+	 * cannot directly access(this is only possible on discrete devices with
+	 * a small BAR). Attempting to MMAP and fault such an object will
+	 * require the kernel first synchronising any GPU work tied to the
+	 * object, before then migrating the pages, either to the CPU accessible
+	 * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+	 * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+	 *
+	 * Note that this is inherently racy.
+	 */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+	__u64 attributes;
+	/** @rsvd: MBZ */
+	__u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +58 -0

58 lines not on this page · raw message

Documentation/gpu/rfc/index.rst +4 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:
 .. toctree::
 
     i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
Signature · 1 line
2.34.1
#1 lore raw

[Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/doc: add rfc section for small BAR uapi

== Series Details ==

== Series Details ==

Series: drm/doc: add rfc section for small BAR uapi
URL : https://patchwork.freedesktop.org/series/102875/
State : warning

== Summary ==

Error: dim checkpatch failed
778440992ee2 drm/doc: add rfc section for small BAR uapi
-:28: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating? #28:
new file mode 100644

-:33: WARNING:SPDX_LICENSE_TAG: Missing or malformed SPDX-License-Identifier tag in line 1 #33: FILE: Documentation/gpu/rfc/i915_small_bar.h:1:
+/**

-:218: CHECK:SPACING: spaces preferred around that '<<' (ctx:VxV) #218: FILE: Documentation/gpu/rfc/i915_small_bar.h:186:
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)

                                               ^

-:229: WARNING:SPDX_LICENSE_TAG: Missing or malformed SPDX-License-Identifier tag in line 1 #229: FILE: Documentation/gpu/rfc/i915_small_bar.rst:1:
+==========================

total: 0 errors, 3 warnings, 1 checks, 255 lines checked

#2 lore raw

[Intel-gfx] ✗ Fi.CI.BAT: failure for drm/doc: add rfc section for small BAR uapi

In reply to Matthew Auld
== Series Details ==

== Series Details ==

Series: drm/doc: add rfc section for small BAR uapi
URL : https://patchwork.freedesktop.org/series/102875/
State : failure

== Summary ==

CI Bug Log - changes from CI_DRM_11528 -> Patchwork_102875v1
====================================================

Summary
-------

**FAILURE**

Serious unknown changes coming with Patchwork_102875v1 absolutely need to be verified manually.

If you think the reported changes have nothing to do with the changes introduced in Patchwork_102875v1, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.

External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/index.html

Participating hosts (46 -> 46)
------------------------------

Additional (2): fi-cml-u2 fi-icl-u2
Missing (2): fi-bsw-cyan bat-jsl-1

Possible new issues
-------------------

Here are the unknown changes that may have been introduced in Patchwork_102875v1:

### IGT changes ###

#### Possible regressions ####

* igt@kms_pipe_crc_basic@suspend-read-crc-pipe-a:
- fi-cml-u2: NOTRUN -> [INCOMPLETE][1]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]

Known issues
------------

Here are the changes found in Patchwork_102875v1 that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@gem_exec_fence@basic-busy@bcs0:
- fi-cml-u2: NOTRUN -> [SKIP][2] ([i915#1208]) +1 similar issue
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@gem_exec_fence@[email protected]

* igt@gem_huc_copy@huc-copy:
- fi-cml-u2: NOTRUN -> [SKIP][3] ([i915#2190])
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]
- fi-rkl-11600: NOTRUN -> [SKIP][4] ([i915#2190])
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]
- fi-icl-u2: NOTRUN -> [SKIP][5] ([i915#2190])
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@gem_lmem_swapping@basic:
- fi-rkl-11600: NOTRUN -> [SKIP][6] ([i915#4613]) +3 similar issues
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]

* igt@gem_lmem_swapping@parallel-random-engines:
- fi-icl-u2: NOTRUN -> [SKIP][7] ([i915#4613]) +3 similar issues
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@gem_tiled_pread_basic:
- fi-rkl-11600: NOTRUN -> [SKIP][8] ([i915#3282])
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@gem_tiled_pread_basic.html

* igt@i915_module_load@reload:
- fi-kbl-soraka: [PASS][9] -> [DMESG-WARN][10] ([i915#1982])
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/fi-kbl-soraka/igt@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-kbl-soraka/igt@[email protected]

* igt@i915_pm_backlight@basic-brightness:
- fi-rkl-11600: NOTRUN -> [SKIP][11] ([i915#3012])
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]

* igt@i915_selftest@live@hangcheck:
- bat-dg1-6: [PASS][12] -> [DMESG-FAIL][13] ([i915#4494] / [i915#4957])
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/bat-dg1-6/igt@i915_selftest@[email protected]
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/bat-dg1-6/igt@i915_selftest@[email protected]

* igt@kms_chamelium@dp-crc-fast:
- fi-rkl-11600: NOTRUN -> [SKIP][14] ([fdo#111827]) +8 similar issues
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]

* igt@kms_chamelium@dp-hpd-fast:
- fi-cml-u2: NOTRUN -> [SKIP][15] ([fdo#109284] / [fdo#111827]) +8 similar issues
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]

* igt@kms_chamelium@hdmi-hpd-fast:
- fi-icl-u2: NOTRUN -> [SKIP][16] ([fdo#111827]) +8 similar issues
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@kms_cursor_legacy@basic-busy-flip-before-cursor-atomic:
- fi-rkl-11600: NOTRUN -> [SKIP][17] ([i915#4070] / [i915#4103]) +1 similar issue
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]
- fi-cml-u2: NOTRUN -> [SKIP][18] ([fdo#109278]) +1 similar issue
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]

* igt@kms_cursor_legacy@basic-busy-flip-before-cursor-legacy:
- fi-icl-u2: NOTRUN -> [SKIP][19] ([fdo#109278]) +2 similar issues
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@kms_flip@basic-plain-flip@a-edp1:
- fi-tgl-u2: [PASS][20] -> [DMESG-WARN][21] ([i915#402]) +1 similar issue
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/fi-tgl-u2/igt@kms_flip@[email protected]
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-tgl-u2/igt@kms_flip@[email protected]

* igt@kms_force_connector_basic@force-load-detect:
- fi-cml-u2: NOTRUN -> [SKIP][22] ([fdo#109285])
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]
- fi-rkl-11600: NOTRUN -> [SKIP][23] ([fdo#109285] / [i915#4098])
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]
- fi-icl-u2: NOTRUN -> [SKIP][24] ([fdo#109285])
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@kms_frontbuffer_tracking@basic:
- fi-cml-u2: NOTRUN -> [DMESG-WARN][25] ([i915#4269])
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]

* igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-d:
- fi-rkl-11600: NOTRUN -> [SKIP][26] ([i915#4070] / [i915#533])
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]
- fi-cml-u2: NOTRUN -> [SKIP][27] ([fdo#109278] / [i915#533])
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-cml-u2/igt@[email protected]

* igt@kms_psr@primary_mmap_gtt:
- fi-rkl-11600: NOTRUN -> [SKIP][28] ([i915#1072]) +3 similar issues
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@kms_psr@primary_mmap_gtt.html

* igt@kms_setmode@basic-clone-single-crtc:
- fi-icl-u2: NOTRUN -> [SKIP][29] ([i915#3555])
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]
- fi-rkl-11600: NOTRUN -> [SKIP][30] ([i915#3555] / [i915#4098])
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]

* igt@prime_vgem@basic-userptr:
- fi-rkl-11600: NOTRUN -> [SKIP][31] ([i915#3301] / [i915#3708])
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]
- fi-icl-u2: NOTRUN -> [SKIP][32] ([i915#3301])
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-icl-u2/igt@[email protected]

* igt@prime_vgem@basic-write:
- fi-rkl-11600: NOTRUN -> [SKIP][33] ([i915#3291] / [i915#3708]) +2 similar issues
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@[email protected]

#### Possible fixes ####

* igt@gem_exec_suspend@basic-s3@smem:
- fi-rkl-11600: [INCOMPLETE][34] ([i915#5127]) -> [PASS][35]
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/fi-rkl-11600/igt@gem_exec_suspend@[email protected]
[35]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-11600/igt@gem_exec_suspend@[email protected]

* igt@i915_selftest@live@gt_mocs:
- fi-rkl-guc: [DMESG-WARN][36] -> [PASS][37]
[36]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/fi-rkl-guc/igt@i915_selftest@live@gt_mocs.html
[37]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/fi-rkl-guc/igt@i915_selftest@live@gt_mocs.html

* igt@i915_selftest@live@reset:
- {bat-adlp-6}: [DMESG-FAIL][38] ([i915#4983]) -> [PASS][39]
[38]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/bat-adlp-6/igt@i915_selftest@[email protected]
[39]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/bat-adlp-6/igt@i915_selftest@[email protected]

#### Warnings ####

* igt@i915_selftest@live@hangcheck:
- bat-dg1-5: [DMESG-FAIL][40] ([i915#4494] / [i915#4957]) -> [INCOMPLETE][41] ([i915#5757])
[40]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11528/bat-dg1-5/igt@i915_selftest@[email protected]
[41]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/bat-dg1-5/igt@i915_selftest@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

[fdo#109278]: https://bugs.freedesktop.org/show_bug.cgi?id=109278
[fdo#109284]: https://bugs.freedesktop.org/show_bug.cgi?id=109284
[fdo#109285]: https://bugs.freedesktop.org/show_bug.cgi?id=109285
[fdo#109308]: https://bugs.freedesktop.org/show_bug.cgi?id=109308
[fdo#111825]: https://bugs.freedesktop.org/show_bug.cgi?id=111825
[fdo#111827]: https://bugs.freedesktop.org/show_bug.cgi?id=111827
[i915#1072]: https://gitlab.freedesktop.org/drm/intel/issues/1072
[i915#1155]: https://gitlab.freedesktop.org/drm/intel/issues/1155
[i915#1208]: https://gitlab.freedesktop.org/drm/intel/issues/1208
[i915#1845]: https://gitlab.freedesktop.org/drm/intel/issues/1845
[i915#1849]: https://gitlab.freedesktop.org/drm/intel/issues/1849
[i915#1982]: https://gitlab.freedesktop.org/drm/intel/issues/1982
[i915#2190]: https://gitlab.freedesktop.org/drm/intel/issues/2190
[i915#2582]: https://gitlab.freedesktop.org/drm/intel/issues/2582
[i915#3012]: https://gitlab.freedesktop.org/drm/intel/issues/3012
[i915#3282]: https://gitlab.freedesktop.org/drm/intel/issues/3282
[i915#3291]: https://gitlab.freedesktop.org/drm/intel/issues/3291
[i915#3301]: https://gitlab.freedesktop.org/drm/intel/issues/3301
[i915#3555]: https://gitlab.freedesktop.org/drm/intel/issues/3555
[i915#3576]: https://gitlab.freedesktop.org/drm/intel/issues/3576
[i915#3637]: https://gitlab.freedesktop.org/drm/intel/issues/3637
[i915#3708]: https://gitlab.freedesktop.org/drm/intel/issues/3708
[i915#402]: https://gitlab.freedesktop.org/drm/intel/issues/402
[i915#4070]: https://gitlab.freedesktop.org/drm/intel/issues/4070
[i915#4098]: https://gitlab.freedesktop.org/drm/intel/issues/4098
[i915#4103]: https://gitlab.freedesktop.org/drm/intel/issues/4103
[i915#4269]: https://gitlab.freedesktop.org/drm/intel/issues/4269
[i915#4312]: https://gitlab.freedesktop.org/drm/intel/issues/4312
[i915#4391]: https://gitlab.freedesktop.org/drm/intel/issues/4391
[i915#4494]: https://gitlab.freedesktop.org/drm/intel/issues/4494
[i915#4613]: https://gitlab.freedesktop.org/drm/intel/issues/4613
[i915#4897]: https://gitlab.freedesktop.org/drm/intel/issues/4897
[i915#4957]: https://gitlab.freedesktop.org/drm/intel/issues/4957
[i915#4983]: https://gitlab.freedesktop.org/drm/intel/issues/4983
[i915#5127]: https://gitlab.freedesktop.org/drm/intel/issues/5127
[i915#533]: https://gitlab.freedesktop.org/drm/intel/issues/533
[i915#5341]: https://gitlab.freedesktop.org/drm/intel/issues/5341
[i915#5757]: https://gitlab.freedesktop.org/drm/intel/issues/5757

Build changes
-------------

* Linux: CI_DRM_11528 -> Patchwork_102875v1

CI-20190529: 20190529
CI_DRM_11528: bd638cf6c04abbc39e46649f820253a303131df5 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_6441: a44d1d4c9e7198a2d59d7dcaae38e340f7cadcf9 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
Patchwork_102875v1: bd638cf6c04abbc39e46649f820253a303131df5 @ git://anongit.freedesktop.org/gfx-ci/linux

== Linux commits ==

d0658ae6e664 drm/doc: add rfc section for small BAR uapi

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_102875v1/index.html

#3 lore raw
In reply to Matthew Auld

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 297 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
  Documentation/gpu/rfc/index.rst          |   4 +
  3 files changed, 252 insertions(+)

create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +190 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {
+	/** @region: The class:instance pair encoding */
+	struct drm_i915_gem_memory_class_instance region;
+
+	/** @rsvd0: MBZ */
+	__u32 rsvd0;
+
+	/** @probed_size: Memory probed by the driver (-1 = unknown) */
+	__u64 probed_size;
+
+	/** @unallocated_size: Estimate of memory remaining (-1 = unknown) */
+	__u64 unallocated_size;
+
+	union {
+		/** @rsvd1: MBZ */
+		__u64 rsvd1[8];
+		struct {
+			/**
+			 * @probed_cpu_visible_size: Memory probed by the driver
+			 * that is CPU accessible. (-1 = unknown).
+			 *
+			 * This will be always be <= @probed_size, and the
+			 * remainder(if there is any) will not be CPU
+			 * accessible.
+			 */
+			__u64 probed_cpu_visible_size;
+		};
+	};
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+	/**
+	 * @size: Requested size for the object.
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 *
+	 * Note that for some devices we have might have further minimum
+	 * page-size restrictions(larger than 4K), like for device local-memory.
+	 * However in general the final size here should always reflect any
+	 * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+	 * extension to place the object in device local-memory.
+	 */
+	__u64 size;
+	/**
+	 * @handle: Returned handle for the object.
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+	/**
+	 * @flags: Optional flags.
+	 *
+	 * Supported values:
+	 *
+	 * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+	 * the object will need to be accessed via the CPU.
+	 *
+	 * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+	 * only strictly required on platforms where only some of the device
+	 * memory is directly visible or mappable through the CPU, like on DG2+.
+	 *
+	 * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+	 * ensure we can always spill the allocation to system memory, if we
+	 * can't place the object in the mappable part of
+	 * I915_MEMORY_CLASS_DEVICE.
+	 *
+	 * Note that since the kernel only supports flat-CCS on objects that can
+	 * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+	 * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+	 * flat-CCS.
+	 *
+	 * Without this hint, the kernel will assume that non-mappable
+	 * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+	 * kernel can still migrate the object to the mappable part, as a last
+	 * resort, if userspace ever CPU faults this object, but this might be
+	 * expensive, and so ideally should be avoided.
+	 */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+	__u32 flags;
+	/**
+	 * @extensions: The chain of extensions to apply to this object.
+	 *
+	 * This will be useful in the future when we need to support several
+	 * different extensions, and we need to apply more than one when
+	 * creating the object. See struct i915_user_extension.
+	 *
+	 * If we don't supply any extensions then we get the same old gem_create
+	 * behaviour.
+	 *
+	 * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+	 * struct drm_i915_gem_create_ext_memory_regions.
+	 *
+	 * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+	 * struct drm_i915_gem_create_ext_protected_content.
+	 */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+	__u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO	5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *	struct drm_i915_query_vma_info info = {};
+ *	struct drm_i915_query_item item = {
+ *		.data_ptr = (uintptr_t)&info,
+ *		.query_id = DRM_I915_QUERY_VMA_INFO,
+ *	};
+ *	struct drm_i915_query query = {
+ *		.num_items = 1,
+ *		.items_ptr = (uintptr_t)&item,
+ *	};
+ *	int err;
+ *
+ *	// Unlike some other types of queries, there is no need to first query
+ *	// the size of the data_ptr blob here, since we already know ahead of
+ *	// time how big this needs to be.
+ *	item.length = sizeof(info);
+ *
+ *	// Next we fill in the vm_id and ppGTT address of the vma we wish
+ *	// to query, before then firing off the query.
+ *	info.vm_id = vm_id;
+ *	info.offset = gtt_address;
+ *	err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *	if (err || item.length < 0) ...
+ *
+ *	// If all went well we can now inspect the returned attributes.
+ *	if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+	/**
+	 * @vm_id: The given vm id that contains the vma. The id is the value
+	 * returned by the DRM_I915_GEM_VM_CREATE. See struct
+	 * drm_i915_gem_vm_control.vm_id.
+	 */
+	__u32 vm_id;
+	/** @pad: MBZ. */
+	__u32 pad;
+	/**
+	 * @offset: The corresponding ppGTT address of the vma which the kernel
+	 * will use to perform the lookup.
+	 */
+	__u64 offset;
+	/**
+	 * @attributes: The returned attributes for the given vma.
+	 *
+	 * Possible values:
+	 *
+	 * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+	 * vma are currently CPU accessible. If this is not set then the vma is
+	 * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+	 * cannot directly access(this is only possible on discrete devices with
+	 * a small BAR). Attempting to MMAP and fault such an object will
+	 * require the kernel first synchronising any GPU work tied to the
+	 * object, before then migrating the pages, either to the CPU accessible
+	 * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+	 * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+	 *
+	 * Note that this is inherently racy.
+	 */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+	__u64 attributes;
+	/** @rsvd: MBZ */
+	__u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +58 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total size of the
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.
Documentation/gpu/rfc/index.rst +4 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:
  .. toctree::
 
      i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
#4 lore raw

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Thanks,

-Lionel

Quoted 361 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 348 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#5 lore raw

Well usually we increment the drm minor version when adding some new flags on amdgpu.

Additional to that just one comment from our experience with that: You don't just need one flag, but two. The first one is a hint which says "CPU access needed" and the second is a promise which says "CPU access never needed".

The background is that on a whole bunch of buffers you can 100% certain say that you will never ever need CPU access.

Then at least we have a whole bunch of buffers where we might need CPU access, but can't tell for sure.

And last we have stuff like transfer buffers you can be 100% sure that you need CPU access.

Separating it like this helped a lot with performance on small BAR systems.

Regards,
Christian.

Am 27.04.22 um 08:48 schrieb Lionel Landwerlin:

Quoted 378 lines

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Thanks,

-Lionel

Quoted 366 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 353 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#6 lore raw
Quoted 11 lines · Christian König

Well usually we increment the drm minor version when adding some new flags on amdgpu.

Additional to that just one comment from our experience with that: You don't just need one flag, but two. The first one is a hint which says "CPU access needed" and the second is a promise which says "CPU access never needed".

The background is that on a whole bunch of buffers you can 100% certain say that you will never ever need CPU access.

Then at least we have a whole bunch of buffers where we might need CPU access, but can't tell for sure.

And last we have stuff like transfer buffers you can be 100% sure that you need CPU access.

Separating it like this helped a lot with performance on small BAR systems.

Thanks for the comments. For the "CPU access never needed" flag, what extra stuff does that do on the kernel side vs not specifying any flag/hint? I assume it still prioritizes using the non-CPU visible portion first? What else does it do?

Quoted 384 lines

Regards,
Christian.

Am 27.04.22 um 08:48 schrieb Lionel Landwerlin:

Quoted 378 lines

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Thanks,

-Lionel

Quoted 366 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 353 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#7 lore raw

Am 27.04.22 um 17:02 schrieb Matthew Auld:

Quoted 7 lines · Christian König
Quoted 6 lines

Well usually we increment the drm minor version when adding some new flags on amdgpu.

Additional to that just one comment from our experience with that: You don't just need one flag, but two. The first one is a hint which says "CPU access needed" and the second is a promise which says "CPU access never needed".

The background is that on a whole bunch of buffers you can 100% certain say that you will never ever need CPU access.

Quoted 7 lines
Quoted 5 lines

Then at least we have a whole bunch of buffers where we might need CPU access, but can't tell for sure.

And last we have stuff like transfer buffers you can be 100% sure that you need CPU access.

Separating it like this helped a lot with performance on small BAR systems.

Thanks for the comments. For the "CPU access never needed" flag, what extra stuff does that do on the kernel side vs not specifying any flag/hint? I assume it still prioritizes using the non-CPU visible portion first? What else does it do?

It's used as a hint when you need to pin BOs for scanout for example.

In general we try to allocate BOs which are marked "CPU access needed" in the CPU visible window if possible, but fallback to any memory if that won't fit.

Christian.

Quoted 394 lines
Quoted 394 lines

Regards,
Christian.

Am 27.04.22 um 08:48 schrieb Lionel Landwerlin:

Quoted 387 lines

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Thanks,

-Lionel

Quoted 375 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 362 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will
+     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find 

the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#8 lore raw
In reply to Christian König
Quoted 10 lines · Christian König

Well usually we increment the drm minor version when adding some new flags on amdgpu.

Additional to that just one comment from our experience with that: You don't just need one flag, but two. The first one is a hint which says "CPU access needed" and the second is a promise which says "CPU access never needed".

The background is that on a whole bunch of buffers you can 100% certain say that you will never ever need CPU access.

Quoted 7 lines

Then at least we have a whole bunch of buffers where we might need CPU access, but can't tell for sure.

And last we have stuff like transfer buffers you can be 100% sure that you need CPU access.

Separating it like this helped a lot with performance on small BAR systems.

So my assumption was that for transfer buffers you'd fill them with the cpu first anyway, so no need for the extra flag.

I guess this if for transfer buffers for gpu -> cpu transfers, where it would result in costly bo move and stalls and it's better to make sure it's cpu accessible from the start? At least on current gpu we have where there's no coherent interconnect, those buffers have to be in system memory or your cpu access will be a disaster, so again they're naturally cpu accessible.

What's the use-case for the "cpu access required" flag where "cpu access before gpu access" isn't a good enough hint already to get the same perf benefits?

Also for scanout my idea at least is that we just fail mmap when you haven't set the flag and the scanout is pinned to unmappable, for two reasons:
- 4k buffers are big, if we force them all into mappable things are non-pretty.
- You need mesa anyway to access tiled buffers, and mesa knows how to use a transfer buffer. That should work even when you do desktop switching and fastboot and stuff like that with the getfb2 ioctl should all work (and without getfb2 it's doomed to garbage anyway).

So only dumb kms buffers (which are linear) would ever get the NEEDS_CPU_ACCESS flag, and only those we'd ever pin into cpu accessible range for scanout. Is there a hole in that plan?

Cheers, Daniel

Quoted 393 lines

Regards,
Christian.

Am 27.04.22 um 08:48 schrieb Lionel Landwerlin:

Quoted 387 lines

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel?
I assume older kernels are going to reject object creation if we use this flag?

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address?
You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Thanks,

-Lionel

Quoted 369 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 356 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region

as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 =

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma,

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will

require CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the

total size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate

VkMemoryHeap with the +VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we

only support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if

the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
#9 lore raw
In reply to Lionel Landwerlin
Quoted 2 lines · Lionel Landwerlin

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

From some offline discussion with Lionel, the plan here is to just do a dummy gem_create_ext to check if the kernel throws an error with the new flag or not.

Quoted 3 lines

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Since mesa at this time doesn't currently have a use for this one, then I guess we should maybe just drop this part of the uapi, in this version at least, if no objections.

Quoted 368 lines

Thanks,

-Lionel

Quoted 361 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 348 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#10 lore raw
Quoted 11 lines · Matthew Auld
Quoted 2 lines · Lionel Landwerlin

One question though, how do we detect that this flag (I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) is accepted on a given kernel? I assume older kernels are going to reject object creation if we use this flag?

From some offline discussion with Lionel, the plan here is to just do a dummy gem_create_ext to check if the kernel throws an error with the new flag or not.

Quoted 3 lines

I didn't plan to use __drm_i915_query_vma_info, but isn't it inconsistent to select the placement on the GEM object and then query whether it's mappable by address? You made a comment stating this is racy, wouldn't querying on the GEM object prevent this?

Since mesa at this time doesn't currently have a use for this one, then I guess we should maybe just drop this part of the uapi, in this version at least, if no objections.

Just repeating what we discussed (maybe I missed some other discussion and that's why I was confused) :

The way I was planning to use this is to have 3 heaps in Vulkan :

    - heap0: local only, no cpu visible

    - heap1: system, cpu visible

    - heap2: local & cpu visible

With heap2 having the reported probed_cpu_visible_size size.

It is an error for the application to map from heap0 [1].

With that said, it means if we created a GEM BO without I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS, we'll never mmap it.

So why the query?

I guess it would be useful when we import a buffer from another application. But in that case, why not have the query on the BO?

-Lionel

[1] : https://www.khronos.org/registry/vulkan/specs/1.3-extensions/man/html/vkMapMemory.html (VUID-vkMapMemory-memory-00682)

Quoted 380 lines
Quoted 380 lines

Thanks,

-Lionel

Quoted 372 lines · Lionel Landwerlin

Hi Matt,

The proposal looks good to me.

Looking forward to try it on drm-tip.

-Lionel

Quoted 359 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;

Documentation/gpu/rfc/i915_small_bar.h new file +20 -0
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver
+             * that is CPU accessible. (-1 = unknown).
+             *
+             * This will be always be <= @probed_size, and the
+             * remainder(if there is any) will not be CPU
+             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };
+    };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#11 lore raw
In reply to Matthew Auld
Quoted 55 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
  Documentation/gpu/rfc/index.rst          |   4 +
  3 files changed, 252 insertions(+)

create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +24 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {
+	/** @region: The class:instance pair encoding */
+	struct drm_i915_gem_memory_class_instance region;
+
+	/** @rsvd0: MBZ */
+	__u32 rsvd0;
+
+	/** @probed_size: Memory probed by the driver (-1 = unknown) */
+	__u64 probed_size;
+
+	/** @unallocated_size: Estimate of memory remaining (-1 = unknown) */
+	__u64 unallocated_size;
+
+	union {
+		/** @rsvd1: MBZ */
+		__u64 rsvd1[8];

Documentation/gpu/rfc/i915_small_bar.h line 34

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+		struct {
+			/**
+			 * @probed_cpu_visible_size: Memory probed by the driver
+			 * that is CPU accessible. (-1 = unknown).
+			 *
+			 * This will be always be <= @probed_size, and the
+			 * remainder(if there is any) will not be CPU
+			 * accessible.
+			 */
+			__u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Quoted 51 lines
Documentation/gpu/rfc/i915_small_bar.h new file +51 -0
+		};
+	};
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+	/**
+	 * @size: Requested size for the object.
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 *
+	 * Note that for some devices we have might have further minimum
+	 * page-size restrictions(larger than 4K), like for device local-memory.
+	 * However in general the final size here should always reflect any
+	 * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+	 * extension to place the object in device local-memory.
+	 */
+	__u64 size;
+	/**
+	 * @handle: Returned handle for the object.
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+	/**
+	 * @flags: Optional flags.
+	 *
+	 * Supported values:
+	 *
+	 * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+	 * the object will need to be accessed via the CPU.
+	 *
+	 * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+	 * only strictly required on platforms where only some of the device
+	 * memory is directly visible or mappable through the CPU, like on DG2+.
+	 *
+	 * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+	 * ensure we can always spill the allocation to system memory, if we
+	 * can't place the object in the mappable part of
+	 * I915_MEMORY_CLASS_DEVICE.
+	 *
+	 * Note that since the kernel only supports flat-CCS on objects that can

Documentation/gpu/rfc/i915_small_bar.h line 95

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+	 * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+	 * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+	 * flat-CCS.
+	 *
+	 * Without this hint, the kernel will assume that non-mappable
+	 * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+	 * kernel can still migrate the object to the mappable part, as a last
+	 * resort, if userspace ever CPU faults this object, but this might be
+	 * expensive, and so ideally should be avoided.
+	 */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Is this correct?

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Quoted 134 lines
Documentation/gpu/rfc/i915_small_bar.h new file +95 -0
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+	__u32 flags;
+	/**
+	 * @extensions: The chain of extensions to apply to this object.
+	 *
+	 * This will be useful in the future when we need to support several
+	 * different extensions, and we need to apply more than one when
+	 * creating the object. See struct i915_user_extension.
+	 *
+	 * If we don't supply any extensions then we get the same old gem_create
+	 * behaviour.
+	 *
+	 * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+	 * struct drm_i915_gem_create_ext_memory_regions.
+	 *
+	 * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+	 * struct drm_i915_gem_create_ext_protected_content.
+	 */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+	__u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO	5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *	struct drm_i915_query_vma_info info = {};
+ *	struct drm_i915_query_item item = {
+ *		.data_ptr = (uintptr_t)&info,
+ *		.query_id = DRM_I915_QUERY_VMA_INFO,
+ *	};
+ *	struct drm_i915_query query = {
+ *		.num_items = 1,
+ *		.items_ptr = (uintptr_t)&item,
+ *	};
+ *	int err;
+ *
+ *	// Unlike some other types of queries, there is no need to first query
+ *	// the size of the data_ptr blob here, since we already know ahead of
+ *	// time how big this needs to be.
+ *	item.length = sizeof(info);
+ *
+ *	// Next we fill in the vm_id and ppGTT address of the vma we wish
+ *	// to query, before then firing off the query.
+ *	info.vm_id = vm_id;
+ *	info.offset = gtt_address;
+ *	err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *	if (err || item.length < 0) ...
+ *
+ *	// If all went well we can now inspect the returned attributes.
+ *	if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+	/**
+	 * @vm_id: The given vm id that contains the vma. The id is the value
+	 * returned by the DRM_I915_GEM_VM_CREATE. See struct
+	 * drm_i915_gem_vm_control.vm_id.
+	 */
+	__u32 vm_id;
+	/** @pad: MBZ. */
+	__u32 pad;
+	/**
+	 * @offset: The corresponding ppGTT address of the vma which the kernel
+	 * will use to perform the lookup.
+	 */
+	__u64 offset;
+	/**
+	 * @attributes: The returned attributes for the given vma.
+	 *
+	 * Possible values:
+	 *
+	 * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+	 * vma are currently CPU accessible. If this is not set then the vma is
+	 * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+	 * cannot directly access(this is only possible on discrete devices with
+	 * a small BAR). Attempting to MMAP and fault such an object will
+	 * require the kernel first synchronising any GPU work tied to the
+	 * object, before then migrating the pages, either to the CPU accessible
+	 * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+	 * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+	 *
+	 * Note that this is inherently racy.
+	 */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+	__u64 attributes;
+	/** @rsvd: MBZ */
+	__u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +33 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total size of the
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info

Documentation/gpu/rfc/i915_small_bar.rst line 43

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Regards,

Tvrtko

Quoted 27 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +15 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.
Documentation/gpu/rfc/index.rst +4 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:
  .. toctree::
 
      i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
#12 lore raw
Quoted 65 lines · Tvrtko Ursulin
Quoted 63 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver +             * that is CPU accessible. (-1 = unknown).

Quoted 8 lines
Quoted 6 lines

+             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Make sense. But I don't think unallocated_size has actually been properly wired up yet. It still just gives the same value as probed_size. IIRC for unallocated_size we still need a real user/usecase/umd, before wiring that up for real with the existing avail tracking. Once we have that we can also add unallocated_cpu_visible_size.

Quoted 2 lines

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Note sure. But just in case we do add it for real at some point, why the added restriction?

Quoted 79 lines
Quoted 68 lines

+        };
+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), likefor device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device +     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation tosystem memory, if we +     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable

Quoted 8 lines

+     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

Quoted 4 lines

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

Yup.

Quoted 5 lines

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Here it would only be: device_cpu_mappable, system. We would completely ignore "device" in this case.

Quoted 4 lines

Is this correct?

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Yeah, the object will only be placed somewhere that is also CPU mappable, with the flag set.

Quoted 166 lines
Quoted 166 lines

+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several +     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usagesee
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is noneed to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for thegiven vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set ifthe pages backing the
+     * vma are currently CPU accessible. If this isnot set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICEmemory, which the CPU
+     * cannot directly access(this is only possibleon discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 

256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU 

access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should 
Quoted 8 lines
Quoted 6 lines

always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Yeah, I'm wondering now if we can just drop this part of the uapi, for now at least, and focus on landing the new flag stuff first.

Quoted 2 lines

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Yeah, just using the object handle or so I guess would also work. Thanks for the comments.

Quoted 35 lines

Regards,

Tvrtko

Quoted 30 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
      i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#13 lore raw
Quoted 75 lines · Matthew Auld
Quoted 65 lines · Tvrtko Ursulin
Quoted 63 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver +             * that is CPU accessible. (-1 = unknown).

Quoted 8 lines
Quoted 6 lines

+             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Quoted 1 line

Make sense. But I don't think unallocated_size has actually been properly wired up yet. It still just gives the same value as probed_size. IIRC for unallocated_size we still need a real user/usecase/umd, before wiring that up for real with the existing avail tracking. Once we have that we can also add unallocated_cpu_visible_size.

So this does nothing at the moment:

info.unallocated_size = mr->avail;

Right, it is set to "mem->avail = mem->total;" at region init time and I indeed can't find it ever getting modified. Okay.

Quoted 3 lines
Quoted 1 line

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Note sure. But just in case we do add it for real at some point, why the added restriction?

To avoid a side channel, albeit perhaps a very weak one. For engine utilization we require CAP_SYS_PERFMON, but that is implied by the perf core API. It's open for discussion. I guess it may make sense to limit it also because it is questionable the field(s) are even useful.

Quoted 94 lines
Quoted 81 lines
Quoted 70 lines

+        };
+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), likefor device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation tosystem memory, if we +     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable

Quoted 8 lines

+     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

Quoted 4 lines

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

Yup.

Quoted 5 lines

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Quoted 8 lines

Here it would only be: device_cpu_mappable, system. We would completely ignore "device" in this case.

Quoted 4 lines

Is this correct?

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Yeah, the object will only be placed somewhere that is also CPU mappable, with the flag set.

Hm, wouldn't it be more efficient to be able to migrate it over to non-mappable in cases when mappable is over-subscribed?

Quoted 175 lines
Quoted 166 lines
Quoted 166 lines

+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usagesee
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is noneed to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for thegiven vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set ifthe pages backing the
+     * vma are currently CPU accessible. If this isnot set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICEmemory, which the CPU
+     * cannot directly access(this is only possibleon discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 
Quoted 8 lines
Quoted 6 lines

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Quoted 6 lines

Yeah, I'm wondering now if we can just drop this part of the uapi, for now at least, and focus on landing the new flag stuff first.

Quoted 2 lines

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Yeah, just using the object handle or so I guess would also work. Thanks for the comments.

I saw other folks have said the same so omitting for now sounds good to me indeed.

Regards,

Tvrtko

#14 lore raw
Quoted 193 lines · Tvrtko Ursulin
Quoted 76 lines · Matthew Auld
Quoted 66 lines · Tvrtko Ursulin
Quoted 64 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver +             * that is CPU accessible. (-1 = unknown).

Quoted 8 lines
Quoted 6 lines

+             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Quoted 1 line

Make sense. But I don't think unallocated_size has actually been properly wired up yet. It still just gives the same value as probed_size. IIRC for unallocated_size we still need a real user/usecase/umd, before wiring that up for real with the existing avail tracking. Once we have that we can also add unallocated_cpu_visible_size.

So this does nothing at the moment:

 info.unallocated_size = mr->avail;

Right, it is set to "mem->avail = mem->total;" at region init time and I indeed can't find it ever getting modified. Okay.

Quoted 3 lines
Quoted 1 line

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Note sure. But just in case we do add it for real at some point, why the added restriction?

To avoid a side channel, albeit perhaps a very weak one. For engine utilization we require CAP_SYS_PERFMON, but that is implied by the perf core API. It's open for discussion. I guess it may make sense to limit it also because it is questionable the field(s) are even useful.

Quoted 101 lines
Quoted 85 lines
Quoted 74 lines

+        };
+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), likefor device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation tosystem memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable

Quoted 8 lines

+     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

Quoted 4 lines

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

Yup.

Quoted 5 lines

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Here it would only be: device_cpu_mappable, system. We would completely ignore "device" in this case.

Quoted 1 line
Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

Is this correct?

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Yeah, the object will only be placed somewhere that is also CPU mappable, with the flag set.

Hm, wouldn't it be more efficient to be able to migrate it over to non-mappable in cases when mappable is over-subscribed?

Not sure. As an alternative strategy, I guess that might be interesting, and if userspace wants something like that we can always add a new flag I guess? It's a toss up whether just using system memory is better/worse than incurring an extra move at fault time?

Quoted 189 lines
Quoted 177 lines
Quoted 168 lines
Quoted 168 lines

+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usagesee
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is noneed to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for thegiven vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set ifthe pages backing the
+     * vma are currently CPU accessible. If this isnot set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICEmemory, which the CPU
+     * cannot directly access(this is only possibleon discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 
Quoted 8 lines
Quoted 6 lines

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Quoted 6 lines

Yeah, I'm wondering now if we can just drop this part of the uapi, for now at least, and focus on landing the new flag stuff first.

Quoted 2 lines

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Yeah, just using the object handle or so I guess would also work. Thanks for the comments.

I saw other folks have said the same so omitting for now sounds good to me indeed.

Regards,

Tvrtko

#15 lore raw
Quoted 198 lines · Matthew Auld
Quoted 197 lines · Tvrtko Ursulin
Quoted 76 lines · Matthew Auld
Quoted 66 lines · Tvrtko Ursulin
Quoted 64 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver +             * that is CPU accessible. (-1 = unknown).

Quoted 8 lines
Quoted 6 lines

+             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Quoted 1 line

Make sense. But I don't think unallocated_size has actually been properly wired up yet. It still just gives the same value as probed_size. IIRC for unallocated_size we still need a real user/usecase/umd, before wiring that up for real with the existing avail tracking. Once we have that we can also add unallocated_cpu_visible_size.

So this does nothing at the moment:

  info.unallocated_size = mr->avail;

Right, it is set to "mem->avail = mem->total;" at region init time and I indeed can't find it ever getting modified. Okay.

Quoted 3 lines
Quoted 1 line

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Note sure. But just in case we do add it for real at some point, why the added restriction?

To avoid a side channel, albeit perhaps a very weak one. For engine utilization we require CAP_SYS_PERFMON, but that is implied by the perf core API. It's open for discussion. I guess it may make sense to limit it also because it is questionable the field(s) are even useful.

Quoted 105 lines
Quoted 87 lines
Quoted 76 lines

+        };
+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), likefor device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation tosystem memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable

Quoted 8 lines

+     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

Quoted 4 lines

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

Yup.

Quoted 5 lines

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Here it would only be: device_cpu_mappable, system. We would completely ignore "device" in this case.

Quoted 3 lines

Is this correct?

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
Quoted 1 line

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Yeah, the object will only be placed somewhere that is also CPU mappable, with the flag set.

Hm, wouldn't it be more efficient to be able to migrate it over to non-mappable in cases when mappable is over-subscribed?

Not sure. As an alternative strategy, I guess that might be interesting, and if userspace wants something like that we can always add a new flag I guess? It's a toss up whether just using system memory is better/worse than incurring an extra move at fault time?

For me it doesn't make sense to allow bos wo/ cpu mappable flag to use the mappable section (albeit as 2nd priority) and not allowing the cpu mappable ones to temporarily go anywhere in lmem.

From i915 side it needs to know the allowed regions (for this argument I speak of mappable / non-mappable as separate regions, even if uapi does not expose them as separate memory regions). i915 does not know in advance the exact usage pattern.

In case of multiple clients, one might touch a buffer from the CPU once and then render with GPU many times. Another client might touch from CPU a lot more. With mappable space contention this would cause the buffer from the first client to constantly get migrated between smem and lmem, while in reality it could have been migrated to non mappable lmem and used by the GPU without any problems.

As some sort of diagram:

Client A BO Client B BO
-------------- --------------
CPU access
CPU access (Client A BO "evicted" to smem)
GPU access GPU access
GPU access CPU access
GPU access GPU access

If we assume there is only space for one BO in mappable, allowing non-mappable placement allows Client A to be unaffected by Client B activity. While with the current proposal it needlessly takes a hit on every, or every other GPU access.

So I think, if there isn't a fundamental reason to disallow it which I am missing, not limiting the implied placement when cpu access flag is given is beneficial to flexibility of migration decisions i915 can make.

Regards,

Tvrtko

Quoted 194 lines
Quoted 194 lines
Quoted 182 lines
Quoted 173 lines
Quoted 173 lines

+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usagesee
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is noneed to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for thegiven vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set ifthe pages backing the
+     * vma are currently CPU accessible. If this isnot set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICEmemory, which the CPU
+     * cannot directly access(this is only possibleon discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This

Quoted 8 lines
Quoted 6 lines

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Quoted 6 lines

Yeah, I'm wondering now if we can just drop this part of the uapi, for now at least, and focus on landing the new flag stuff first.

Quoted 2 lines

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Yeah, just using the object handle or so I guess would also work. Thanks for the comments.

I saw other folks have said the same so omitting for now sounds good to me indeed.

Regards,

Tvrtko

#16 lore raw
Quoted 225 lines · Tvrtko Ursulin
Quoted 198 lines · Matthew Auld
Quoted 197 lines · Tvrtko Ursulin
Quoted 76 lines · Matthew Auld
Quoted 66 lines · Tvrtko Ursulin
Quoted 64 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver +             * that is CPU accessible. (-1 = unknown).

Quoted 8 lines
Quoted 6 lines

+             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;

Would unallocated_cpu_visible_size be useful, to follow the total unallocated_size?

Quoted 1 line

Make sense. But I don't think unallocated_size has actually been properly wired up yet. It still just gives the same value as probed_size. IIRC for unallocated_size we still need a real user/usecase/umd, before wiring that up for real with the existing avail tracking. Once we have that we can also add unallocated_cpu_visible_size.

So this does nothing at the moment:

  info.unallocated_size = mr->avail;

Right, it is set to "mem->avail = mem->total;" at region init time and I indeed can't find it ever getting modified. Okay.

Quoted 3 lines
Quoted 1 line

Btw, have we ever considered whether unallocated_size should require CAP_SYS_ADMIN/PERFMON or something?

Note sure. But just in case we do add it for real at some point, why the added restriction?

To avoid a side channel, albeit perhaps a very weak one. For engine utilization we require CAP_SYS_PERFMON, but that is implied by the perf core API. It's open for discussion. I guess it may make sense to limit it also because it is questionable the field(s) are even useful.

Quoted 105 lines
Quoted 87 lines
Quoted 76 lines

+        };
+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), likefor device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation tosystem memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable

Quoted 8 lines

+     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */

So "needs_cpu_access" flag could almost be viewed as a sub-region placement priority? What I mean is this:

Quoted 4 lines

1)
placements=device,system flags=

This results in placement priorities: device, device_cpu_mappable, system.

Yup.

Quoted 5 lines

2)
placements=device,system flags=needs_cpu_access

This results in placement priorities: device_cpu_mappable, device, system.

Here it would only be: device_cpu_mappable, system. We would completely ignore "device" in this case.

Quoted 3 lines

Is this correct?

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
Quoted 1 line

The benefit of the flag is that i915 can place the object to the right place from the start instead of on the first CPU access? Is that worth it or is there more to it?

Yeah, the object will only be placed somewhere that is also CPU mappable, with the flag set.

Hm, wouldn't it be more efficient to be able to migrate it over to non-mappable in cases when mappable is over-subscribed?

Not sure. As an alternative strategy, I guess that might be interesting, and if userspace wants something like that we can always add a new flag I guess? It's a toss up whether just using system memory is better/worse than incurring an extra move at fault time?

For me it doesn't make sense to allow bos wo/ cpu mappable flag to use the mappable section (albeit as 2nd priority) and not allowing the cpu mappable ones to temporarily go anywhere in lmem.

From i915 side it needs to know the allowed regions (for this argument
I speak of mappable / non-mappable as separate regions, even if uapi does not expose them as separate memory regions). i915 does not know in advance the exact usage pattern.

In case of multiple clients, one might touch a buffer from the CPU once and then render with GPU many times. Another client might touch from CPU a lot more. With mappable space contention this would cause the buffer from the first client to constantly get migrated between smem and lmem, while in reality it could have been migrated to non mappable lmem and used by the GPU without any problems.

As some sort of diagram:

    Client A BO    Client B BO
    --------------    --------------
    CPU access
            CPU access (Client A BO "evicted" to smem) GPU access    GPU access
    GPU access    CPU access
    GPU access    GPU access

Quoted 4 lines

If we assume there is only space for one BO in mappable, allowing non-mappable placement allows Client A to be unaffected by Client B activity. While with the current proposal it needlessly takes a hit on every, or every other GPU access.

So I think, if there isn't a fundamental reason to disallow it which I am missing, not limiting the implied placement when cpu access flag is given is beneficial to flexibility of migration decisions i915 can make.

Jon/Daniel, any thoughts/inputs here?

Quoted 201 lines

Regards,

Tvrtko

Quoted 197 lines
Quoted 197 lines
Quoted 185 lines
Quoted 176 lines
Quoted 176 lines

+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usagesee
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is noneed to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for thegiven vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set ifthe pages backing the
+     * vma are currently CPU accessible. If this isnot set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICEmemory, which the CPU
+     * cannot directly access(this is only possibleon discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will
+     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find 

the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This

Quoted 8 lines
Quoted 6 lines

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.

I think there should be justification for the new query documented as well. (Why on top of what.)

Quoted 6 lines

Yeah, I'm wondering now if we can just drop this part of the uapi, for now at least, and focus on landing the new flag stuff first.

Quoted 2 lines

Without it personally I can't immediately understand why the disconnect between the object based and VMA based API. Userspace has to do some intervening operations like either execbuf, or vm bind in the future, to make this query usable after object creation. So question is why wouldn't it know already which placements it allowed and so would i915 auto-migrate or not for this particular object. No? Or in other words why this wouldn't be an object based query since the question it is answering is about the object backing store and not the VMA.

Yeah, just using the object handle or so I guess would also work. Thanks for the comments.

I saw other folks have said the same so omitting for now sounds good to me indeed.

Regards,

Tvrtko

#17 lore raw
In reply to Matthew Auld
Quoted 56 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
  Documentation/gpu/rfc/index.rst          |   4 +
  3 files changed, 252 insertions(+)

create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +25 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {
+	/** @region: The class:instance pair encoding */
+	struct drm_i915_gem_memory_class_instance region;
+
+	/** @rsvd0: MBZ */
+	__u32 rsvd0;
+
+	/** @probed_size: Memory probed by the driver (-1 = unknown) */
+	__u64 probed_size;
+
+	/** @unallocated_size: Estimate of memory remaining (-1 = unknown) */
+	__u64 unallocated_size;
+
+	union {
+		/** @rsvd1: MBZ */
+		__u64 rsvd1[8];
+		struct {

Documentation/gpu/rfc/i915_small_bar.h line 35

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+			/**
+			 * @probed_cpu_visible_size: Memory probed by the driver
+			 * that is CPU accessible. (-1 = unknown).
+			 *
+			 * This will be always be <= @probed_size, and the
+			 * remainder(if there is any) will not be CPU
+			 * accessible.
+			 */
+			__u64 probed_cpu_visible_size;
+		};

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Quoted 231 lines
Documentation/gpu/rfc/i915_small_bar.h new file +155 -0
+	};
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+	/**
+	 * @size: Requested size for the object.
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 *
+	 * Note that for some devices we have might have further minimum
+	 * page-size restrictions(larger than 4K), like for device local-memory.
+	 * However in general the final size here should always reflect any
+	 * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+	 * extension to place the object in device local-memory.
+	 */
+	__u64 size;
+	/**
+	 * @handle: Returned handle for the object.
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+	/**
+	 * @flags: Optional flags.
+	 *
+	 * Supported values:
+	 *
+	 * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+	 * the object will need to be accessed via the CPU.
+	 *
+	 * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+	 * only strictly required on platforms where only some of the device
+	 * memory is directly visible or mappable through the CPU, like on DG2+.
+	 *
+	 * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+	 * ensure we can always spill the allocation to system memory, if we
+	 * can't place the object in the mappable part of
+	 * I915_MEMORY_CLASS_DEVICE.
+	 *
+	 * Note that since the kernel only supports flat-CCS on objects that can
+	 * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+	 * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+	 * flat-CCS.
+	 *
+	 * Without this hint, the kernel will assume that non-mappable
+	 * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+	 * kernel can still migrate the object to the mappable part, as a last
+	 * resort, if userspace ever CPU faults this object, but this might be
+	 * expensive, and so ideally should be avoided.
+	 */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+	__u32 flags;
+	/**
+	 * @extensions: The chain of extensions to apply to this object.
+	 *
+	 * This will be useful in the future when we need to support several
+	 * different extensions, and we need to apply more than one when
+	 * creating the object. See struct i915_user_extension.
+	 *
+	 * If we don't supply any extensions then we get the same old gem_create
+	 * behaviour.
+	 *
+	 * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+	 * struct drm_i915_gem_create_ext_memory_regions.
+	 *
+	 * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+	 * struct drm_i915_gem_create_ext_protected_content.
+	 */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+	__u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO	5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *	struct drm_i915_query_vma_info info = {};
+ *	struct drm_i915_query_item item = {
+ *		.data_ptr = (uintptr_t)&info,
+ *		.query_id = DRM_I915_QUERY_VMA_INFO,
+ *	};
+ *	struct drm_i915_query query = {
+ *		.num_items = 1,
+ *		.items_ptr = (uintptr_t)&item,
+ *	};
+ *	int err;
+ *
+ *	// Unlike some other types of queries, there is no need to first query
+ *	// the size of the data_ptr blob here, since we already know ahead of
+ *	// time how big this needs to be.
+ *	item.length = sizeof(info);
+ *
+ *	// Next we fill in the vm_id and ppGTT address of the vma we wish
+ *	// to query, before then firing off the query.
+ *	info.vm_id = vm_id;
+ *	info.offset = gtt_address;
+ *	err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *	if (err || item.length < 0) ...
+ *
+ *	// If all went well we can now inspect the returned attributes.
+ *	if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+	/**
+	 * @vm_id: The given vm id that contains the vma. The id is the value
+	 * returned by the DRM_I915_GEM_VM_CREATE. See struct
+	 * drm_i915_gem_vm_control.vm_id.
+	 */
+	__u32 vm_id;
+	/** @pad: MBZ. */
+	__u32 pad;
+	/**
+	 * @offset: The corresponding ppGTT address of the vma which the kernel
+	 * will use to perform the lookup.
+	 */
+	__u64 offset;
+	/**
+	 * @attributes: The returned attributes for the given vma.
+	 *
+	 * Possible values:
+	 *
+	 * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+	 * vma are currently CPU accessible. If this is not set then the vma is
+	 * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+	 * cannot directly access(this is only possible on discrete devices with
+	 * a small BAR). Attempting to MMAP and fault such an object will
+	 * require the kernel first synchronising any GPU work tied to the
+	 * object, before then migrating the pages, either to the CPU accessible
+	 * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+	 * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+	 *
+	 * Note that this is inherently racy.
+	 */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+	__u64 attributes;
+	/** @rsvd: MBZ */
+	__u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +58 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total size of the
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.
Documentation/gpu/rfc/index.rst +4 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:
  .. toctree::
 
      i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
#18 lore raw
Quoted 75 lines · Lionel Landwerlin
Quoted 62 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

Quoted 8 lines

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

-Lionel

Quoted 282 lines
Quoted 280 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#19 lore raw
Quoted 80 lines · Lionel Landwerlin
Quoted 79 lines · Lionel Landwerlin
Quoted 62 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

Quoted 7 lines
Quoted 4 lines

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

unallocated_size should always give the same value as probed_size. We have the avail tracking, but we don't currently expose that through unallocated_size, due to lack of real userspace/user etc.

Quoted 2 lines

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Yeah, if we ever properly expose the unallocated_size, then we could also just add unallocated_cpu_visible_size.

Quoted 2 lines

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

I don't think it's out of the question...

I guess user-space should be able to get the current flag behaviour just by specifying: device, system. And it does give more flexibly to allow something like: device, device-nm, smem.

We can also drop the probed_cpu_visible_size, which would now just be the probed_size with device/device-nm. And if we lack device-nm, then the entire thing must be CPU mappable.

One of the downsides though, is that we can no longer easily mix object pages from both device + device-nm, which we could previously do when we didn't specify the flag. At least according to the current design/behaviour for @regions that would not be allowed. I guess some kind of new flag like ALLOC_MIXED or so? Although currently that is only possible with device + device-nm in ttm/i915.

Quoted 288 lines

-Lionel

Quoted 282 lines
Quoted 280 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#20 lore raw
Quoted 100 lines · Matthew Auld
Quoted 81 lines · Lionel Landwerlin
Quoted 80 lines · Lionel Landwerlin
Quoted 63 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

Quoted 7 lines
Quoted 4 lines

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

unallocated_size should always give the same value as probed_size. We have the avail tracking, but we don't currently expose that through unallocated_size, due to lack of real userspace/user etc.

Quoted 2 lines

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Yeah, if we ever properly expose the unallocated_size, then we could also just add unallocated_cpu_visible_size.

Quoted 2 lines

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

Quoted 7 lines

I don't think it's out of the question...

I guess user-space should be able to get the current flag behaviour just by specifying: device, system. And it does give more flexibly to allow something like: device, device-nm, smem.

We can also drop the probed_cpu_visible_size, which would now just be the probed_size with device/device-nm. And if we lack device-nm, then the entire thing must be CPU mappable.

One of the downsides though, is that we can no longer easily mix object pages from both device + device-nm, which we could previously do when we didn't specify the flag. At least according to the current design/behaviour for @regions that would not be allowed. I guess some kind of new flag like ALLOC_MIXED or so? Although currently that is only possible with device + device-nm in ttm/i915.

Thanks, I wasn't aware of the restrictions.

Adding unallocated_cpu_visible_size would be great.

-Lionel

Quoted 299 lines
Quoted 299 lines

-Lionel

Quoted 292 lines
Quoted 290 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#21 lore raw
Quoted 105 lines · Lionel Landwerlin
Quoted 104 lines · Matthew Auld
Quoted 81 lines · Lionel Landwerlin
Quoted 80 lines · Lionel Landwerlin
Quoted 63 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

Quoted 7 lines
Quoted 4 lines

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

unallocated_size should always give the same value as probed_size. We have the avail tracking, but we don't currently expose that through unallocated_size, due to lack of real userspace/user etc.

Quoted 2 lines

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Yeah, if we ever properly expose the unallocated_size, then we could also just add unallocated_cpu_visible_size.

Quoted 2 lines

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

I don't think it's out of the question...

I guess user-space should be able to get the current flag behaviour just by specifying: device, system. And it does give more flexibly to allow something like: device, device-nm, smem.

Quoted 8 lines
Quoted 3 lines

We can also drop the probed_cpu_visible_size, which would now just be the probed_size with device/device-nm. And if we lack device-nm, then the entire thing must be CPU mappable.

One of the downsides though, is that we can no longer easily mix object pages from both device + device-nm, which we could previously do when we didn't specify the flag. At least according to the current design/behaviour for @regions that would not be allowed. I guess some kind of new flag like ALLOC_MIXED or so? Although currently that is only possible with device + device-nm in ttm/i915.

Thanks, I wasn't aware of the restrictions.

Adding unallocated_cpu_visible_size would be great.

So do we want this in the next version? i.e we already have a current real use case in mind for unallocated_size where probed_size is not good enough?

Quoted 305 lines

-Lionel

Quoted 299 lines
Quoted 299 lines

-Lionel

Quoted 292 lines
Quoted 290 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#22 lore raw
Quoted 109 lines · Matthew Auld
Quoted 108 lines · Lionel Landwerlin
Quoted 107 lines · Matthew Auld
Quoted 82 lines · Lionel Landwerlin
Quoted 81 lines · Lionel Landwerlin
Quoted 64 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region 

as known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

Quoted 7 lines
Quoted 4 lines

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

unallocated_size should always give the same value as probed_size. We have the avail tracking, but we don't currently expose that through unallocated_size, due to lack of real userspace/user etc.

Quoted 2 lines

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Yeah, if we ever properly expose the unallocated_size, then we could also just add unallocated_cpu_visible_size.

Quoted 2 lines

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

I don't think it's out of the question...

I guess user-space should be able to get the current flag behaviour just by specifying: device, system. And it does give more flexibly to allow something like: device, device-nm, smem.

We can also drop the probed_cpu_visible_size, which would now just be the probed_size with device/device-nm. And if we lack device-nm, then the entire thing must be CPU mappable.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

One of the downsides though, is that we can no longer easily mix object pages from both device + device-nm, which we could previously do when we didn't specify the flag. At least according to the current design/behaviour for @regions that would not be allowed. I guess some kind of new flag like ALLOC_MIXED or so? Although currently that is only possible with device + device-nm in ttm/i915.

Thanks, I wasn't aware of the restrictions.

Adding unallocated_cpu_visible_size would be great.

So do we want this in the next version? i.e we already have a current real use case in mind for unallocated_size where probed_size is not good enough?

Yeah in the  next iteration.

We're using unallocated_size to implement VK_EXT_memory_budget and since I'm going to expose lmem mappable/unmappable as 2 different heaps on Vulkan, I would use that there too.

-Lionel

Quoted 313 lines
Quoted 313 lines

-Lionel

Quoted 306 lines
Quoted 306 lines

-Lionel

Quoted 299 lines
Quoted 297 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum
+     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU. +     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to
+     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with
+     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object.
+     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when
+     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see +     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content. +     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE)
...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct +     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will
+     * require the kernel first synchronising any GPU work tied to the
+     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will 

require CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the
+CPU accessible portion, for the particular region. This should only be

Documentation/gpu/rfc/i915_small_bar.rst new file +3 -0
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate 

VkMemoryHeap with the +VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, 

find the
+respective vma, and return its set of attributes. For now we only support
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#23 lore raw
In reply to Matthew Auld
Quoted 95 lines · Matthew Auld
Quoted 81 lines · Lionel Landwerlin
Quoted 80 lines · Lionel Landwerlin
Quoted 63 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190
+++++++++++++++++++++++
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I'm guessing before there is support for this value it'll be 0 (MBZ).

Quoted 7 lines
Quoted 4 lines

After after it should either be the entire lmem or something smaller.

-Lionel

Other pain point of this new uAPI, previously we could query the unallocated size for each heap.

unallocated_size should always give the same value as probed_size. We have the avail tracking, but we don't currently expose that through unallocated_size, due to lack of real userspace/user etc.

Quoted 2 lines

Now lmem is effectively divided into 2 heaps, but unallocated_size is tracking allocation from both parts of lmem.

Quoted 8 lines

Yeah, if we ever properly expose the unallocated_size, then we could also just add unallocated_cpu_visible_size.

Quoted 2 lines

Is adding new I915_MEMORY_CLASS_DEVICE_NON_MAPPABLE out of question?

I don't think it's out of the question...

I guess user-space should be able to get the current flag behaviour just by specifying: device, system. And it does give more flexibly to allow something like: device, device-nm, smem.

I was also thinking about that option, albeit both regions under the existing class just separate instances with "capability" flags differing.

Downsides I thought were a) it does not really match the underlying resource, which is one and not two from the backing storage POV, and b) it allows userspace to do potentially do too much of restrictive regions=device-mappable,system (even if only an innocent mistake); disallowing i915 to manage the space better in cases where multiple clients happen to fight over it.

The last part is going back to what I was commenting earlier, where I though migrating objects which had cpu-access set to any device should be allowed. (When mappable region is over subscribed.)

Regards,

Tvrtko

Quoted 296 lines

We can also drop the probed_cpu_visible_size, which would now just be the probed_size with device/device-nm. And if we lack device-nm, then the entire thing must be CPU mappable.

One of the downsides though, is that we can no longer easily mix object pages from both device + device-nm, which we could previously do when we didn't specify the flag. At least according to the current design/behaviour for @regions that would not be allowed. I guess some kind of new flag like ALLOC_MIXED or so? Although currently that is only possible with device + device-nm in ttm/i915.

Quoted 292 lines

-Lionel

Quoted 286 lines
Quoted 284 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create 

behaviour, with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any
+     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and
+     * only strictly required on platforms where only some of the device
+     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we
+     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't
+     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several
+     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the 

first 256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the 

total size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This 

should always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
        i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#24 lore raw
In reply to Lionel Landwerlin
Quoted 93 lines

-----Original Message-----
From: Landwerlin, Lionel G <[email protected]>
Sent: Monday, May 2, 2022 12:55 AM
To: Auld, Matthew <[email protected]>; [email protected]
Cc: [email protected]; Thomas Hellström <[email protected]>; Bloomfield, Jon <[email protected]>; Daniel Vetter <[email protected]>; Justen, Jordan L <[email protected]>; Kenneth Graunke <[email protected]>; Abodunrin, Akeem G <[email protected]>; [email protected]
Subject: Re: [PATCH v2] drm/doc: add rfc section for small BAR uapi

Quoted 17 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
Quoted 2 lines
---
  Documentation/gpu/rfc/i915_small_bar.h   | 190

+++++++++++++++++++++++

Quoted 13 lines
  Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
  Documentation/gpu/rfc/index.rst          |   4 +
  3 files changed, 252 insertions(+)

create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +0 -0
new file mode 100644
index 000000000000..7bfd0cf44d35

Documentation/gpu/rfc/i915_small_bar.h line 6

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as
+known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct

drm_i915_query.

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+ * For this new query we are adding the new query id
+DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+	/** @region: The class:instance pair encoding */
+	struct drm_i915_gem_memory_class_instance region;
+
+	/** @rsvd0: MBZ */
+	__u32 rsvd0;
+
+	/** @probed_size: Memory probed by the driver (-1 = unknown) */
+	__u64 probed_size;
+
+	/** @unallocated_size: Estimate of memory remaining (-1 = unknown)

*/

Documentation/gpu/rfc/i915_small_bar.h

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+	__u64 unallocated_size;
+
+	union {
+		/** @rsvd1: MBZ */
+		__u64 rsvd1[8];
+		struct {
+			/**
+			 * @probed_cpu_visible_size: Memory probed by the

driver

Documentation/gpu/rfc/i915_small_bar.h line 35

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+			 * that is CPU accessible. (-1 = unknown).
+			 *
+			 * This will be always be <= @probed_size, and the
+			 * remainder(if there is any) will not be CPU
+			 * accessible.
+			 */
+			__u64 probed_cpu_visible_size;
+		};

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

Quoted 1 line

When is it set to -1?

I believe it is set to -1 if it is unknown, and/or not cpu accessible...

Cheers!
~Akeem

Quoted 265 lines

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Quoted 18 lines
Documentation/gpu/rfc/i915_small_bar.h new file +18 -0
+	};
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour,
+with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the
+stuff that
+ * is immutable. Previously we would have two ioctls, one to create
+the object
+ * with gem_create, and another to apply various parameters, however
+this
+ * creates some ambiguity for the params which are considered
+immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+struct __drm_i915_gem_create_ext {
+	/**
+	 * @size: Requested size for the object.
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 *
+	 * Note that for some devices we have might have further minimum
+	 * page-size restrictions(larger than 4K), like for device local-memory.
+	 * However in general the final size here should always reflect any
+	 * rounding up, if for example using the

I915_GEM_CREATE_EXT_MEMORY_REGIONS

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+	 * extension to place the object in device local-memory.
+	 */
+	__u64 size;
+	/**
+	 * @handle: Returned handle for the object.

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+	/**
+	 * @flags: Optional flags.
+	 *
+	 * Supported values:
+	 *
+	 * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the

kernel that

Documentation/gpu/rfc/i915_small_bar.h

Quoted 3 lines
Documentation/gpu/rfc/i915_small_bar.h new file +3 -0
+	 * the object will need to be accessed via the CPU.
+	 *
+	 * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE,

and

Documentation/gpu/rfc/i915_small_bar.h

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+	 * only strictly required on platforms where only some of the device
+	 * memory is directly visible or mappable through the CPU, like on DG2+.
+	 *
+	 * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM,

to

Documentation/gpu/rfc/i915_small_bar.h

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
+	 * ensure we can always spill the allocation to system memory, if we
+	 * can't place the object in the mappable part of
+	 * I915_MEMORY_CLASS_DEVICE.
+	 *
+	 * Note that since the kernel only supports flat-CCS on objects that can
+	 * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore

don't

Documentation/gpu/rfc/i915_small_bar.h line 94

Quoted 1 line
Documentation/gpu/rfc/i915_small_bar.h new file +1 -0
+	 * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS

together with

Documentation/gpu/rfc/i915_small_bar.h

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+	 * flat-CCS.
+	 *
+	 * Without this hint, the kernel will assume that non-mappable
+	 * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that

the

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+	 * kernel can still migrate the object to the mappable part, as a last
+	 * resort, if userspace ever CPU faults this object, but this might be
+	 * expensive, and so ideally should be avoided.
+	 */

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+	__u32 flags;
+	/**
+	 * @extensions: The chain of extensions to apply to this object.
+	 *
+	 * This will be useful in the future when we need to support several
+	 * different extensions, and we need to apply more than one when
+	 * creating the object. See struct i915_user_extension.
+	 *
+	 * If we don't supply any extensions then we get the same old

gem_create

Quoted 51 lines
Documentation/gpu/rfc/i915_small_bar.h new file +51 -0
+	 * behaviour.
+	 *
+	 * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+	 * struct drm_i915_gem_create_ext_memory_regions.
+	 *
+	 * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+	 * struct drm_i915_gem_create_ext_protected_content.
+	 */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0 #define
+I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+	__u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO	5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma,
+returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *	struct drm_i915_query_vma_info info = {};
+ *	struct drm_i915_query_item item = {
+ *		.data_ptr = (uintptr_t)&info,
+ *		.query_id = DRM_I915_QUERY_VMA_INFO,
+ *	};
+ *	struct drm_i915_query query = {
+ *		.num_items = 1,
+ *		.items_ptr = (uintptr_t)&item,
+ *	};
+ *	int err;
+ *
+ *	// Unlike some other types of queries, there is no need to first query
+ *	// the size of the data_ptr blob here, since we already know ahead of
+ *	// time how big this needs to be.
+ *	item.length = sizeof(info);
+ *
+ *	// Next we fill in the vm_id and ppGTT address of the vma we wish
+ *	// to query, before then firing off the query.
+ *	info.vm_id = vm_id;
+ *	info.offset = gtt_address;
+ *	err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *	if (err || item.length < 0) ...
+ *
+ *	// If all went well we can now inspect the returned attributes.
+ *	if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+	/**
+	 * @vm_id: The given vm id that contains the vma. The id is the value
+	 * returned by the DRM_I915_GEM_VM_CREATE. See struct
+	 * drm_i915_gem_vm_control.vm_id.
+	 */
+	__u32 vm_id;
+	/** @pad: MBZ. */
+	__u32 pad;
+	/**
+	 * @offset: The corresponding ppGTT address of the vma which the

kernel

Documentation/gpu/rfc/i915_small_bar.h

Quoted 9 lines
Documentation/gpu/rfc/i915_small_bar.h new file +9 -0
+	 * will use to perform the lookup.
+	 */
+	__u64 offset;
+	/**
+	 * @attributes: The returned attributes for the given vma.
+	 *
+	 * Possible values:
+	 *
+	 * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages

backing the

Documentation/gpu/rfc/i915_small_bar.h line 184

Quoted 2 lines
Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
+	 * vma are currently CPU accessible. If this is not set then the vma is
+	 * currently backed by I915_MEMORY_CLASS_DEVICE memory, which

the CPU

Documentation/gpu/rfc/i915_small_bar.h

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+	 * cannot directly access(this is only possible on discrete devices with
+	 * a small BAR). Attempting to MMAP and fault such an object will
+	 * require the kernel first synchronising any GPU work tied to the
+	 * object, before then migrating the pages, either to the CPU accessible
+	 * part of I915_MEMORY_CLASS_DEVICE, or

I915_MEMORY_CLASS_SYSTEM, if the

Documentation/gpu/rfc/i915_small_bar.h line 190

Quoted 1 line
Documentation/gpu/rfc/i915_small_bar.h new file +1 -0
+	 * placements permit it. See

I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.

Quoted 15 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+	 *
+	 * Note that this is inherently racy.
+	 */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+	__u64 attributes;
+	/** @rsvd: MBZ */
+	__u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +0 -0
new file mode 100644
index 000000000000..be3d9bcdd86d

Documentation/gpu/rfc/i915_small_bar.rst line 8

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device
+local-memory(i.e I915_MEMORY_CLASS_DEVICE), but in some cases the
+final BAR size might still be smaller than the total probed_size. In
+such cases, only some subset of I915_MEMORY_CLASS_DEVICE will be CPU
+accessible(for example the first 256M), while the remainder is only accessible

via the GPU.

Documentation/gpu/rfc/i915_small_bar.rst line 17

Quoted 9 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU access.
+This becomes important when placing an object in
+I915_MEMORY_CLASS_DEVICE, where underneath the device has a small
+BAR, meaning only some portion of it is CPU accessible. Without this
+flag the kernel will assume that CPU access is not required, and
+prioritize using the non-CPU visible portion of

I915_MEMORY_CLASS_DEVICE.

Quoted 16 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +16 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total
+size of the CPU accessible portion, for the particular region. This
+should only be applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap
+with the VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the
+CPU visible portion, where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info

Documentation/gpu/rfc/i915_small_bar.rst

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only
+support DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the
+object/vma is currently placed in memory that is accessible by the
+CPU. This should always be set on devices where the CPU
+probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE matches the
+probed_size. If this is not set then CPU faulting the object will likely first

require migrating the pages.

Quoted 28 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +15 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.
(quoted diff) +0 -0

b/Documentation/gpu/rfc/index.rst index 91e93a705230..5a3bd3924ba6 100644

Documentation/gpu/rfc/index.rst +4 -0
@@ -23,3 +23,7 @@ host such documentation:
  .. toctree::
 
      i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
#25 lore raw
Quoted 88 lines · Akeem G
Quoted 88 lines

-----Original Message-----
From: Landwerlin, Lionel G <[email protected]>
Sent: Monday, May 2, 2022 12:55 AM
To: Auld, Matthew <[email protected]>; [email protected]
Cc: [email protected]; Thomas Hellström <[email protected]>; Bloomfield, Jon <[email protected]>; Daniel Vetter <[email protected]>; Justen, Jordan L <[email protected]>; Kenneth Graunke <[email protected]>; Abodunrin, Akeem G <[email protected]>; [email protected]
Subject: Re: [PATCH v2] drm/doc: add rfc section for small BAR uapi

Quoted 17 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
- Some spelling fixes and other small tweaks. (Akeem & Thomas)
- Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
- Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
Quoted 2 lines
---
   Documentation/gpu/rfc/i915_small_bar.h   | 190

+++++++++++++++++++++++

Quoted 13 lines
   Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++
   Documentation/gpu/rfc/index.rst          |   4 +
   3 files changed, 252 insertions(+)

create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +0 -0
new file mode 100644
index 000000000000..7bfd0cf44d35

Documentation/gpu/rfc/i915_small_bar.h line 6

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as
+known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct

drm_i915_query.

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+ * For this new query we are adding the new query id
+DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+   /** @region: The class:instance pair encoding */
+   struct drm_i915_gem_memory_class_instance region;
+
+   /** @rsvd0: MBZ */
+   __u32 rsvd0;
+
+   /** @probed_size: Memory probed by the driver (-1 = unknown) */
+   __u64 probed_size;
+
+   /** @unallocated_size: Estimate of memory remaining (-1 = unknown)

*/

Documentation/gpu/rfc/i915_small_bar.h

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+   __u64 unallocated_size;
+
+   union {
+           /** @rsvd1: MBZ */
+           __u64 rsvd1[8];
+           struct {
+                   /**
+                    * @probed_cpu_visible_size: Memory probed by the

driver

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+                    * that is CPU accessible. (-1 = unknown).
+                    *
+                    * This will be always be <= @probed_size, and the
+                    * remainder(if there is any) will not be CPU
+                    * accessible.
+                    */
+                   __u64 probed_cpu_visible_size;
+           };
Quoted 8 lines
Quoted 4 lines

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I believe it is set to -1 if it is unknown, and/or not cpu accessible...

Cheers!
~Akeem

So what should I expect on system memory?

What value is returned when all of probed_size is CPU visible on local memory?

Thanks,

-Lionel

Quoted 263 lines
Quoted 263 lines

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Quoted 18 lines
Documentation/gpu/rfc/i915_small_bar.h new file +18 -0
+   };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour,
+with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the
+stuff that
+ * is immutable. Previously we would have two ioctls, one to create
+the object
+ * with gem_create, and another to apply various parameters, however
+this
+ * creates some ambiguity for the params which are considered
+immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+struct __drm_i915_gem_create_ext {
+   /**
+    * @size: Requested size for the object.
+    *
+    * The (page-aligned) allocated size for the object will be returned.
+    *
+    * Note that for some devices we have might have further minimum
+    * page-size restrictions(larger than 4K), like for device local-memory.
+    * However in general the final size here should always reflect any
+    * rounding up, if for example using the

I915_GEM_CREATE_EXT_MEMORY_REGIONS

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+    * extension to place the object in device local-memory.
+    */
+   __u64 size;
+   /**
+    * @handle: Returned handle for the object.

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+    *
+    * Object handles are nonzero.
+    */
+   __u32 handle;
+   /**
+    * @flags: Optional flags.
+    *
+    * Supported values:
+    *
+    * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the

kernel that

Documentation/gpu/rfc/i915_small_bar.h

Quoted 3 lines
Documentation/gpu/rfc/i915_small_bar.h new file +3 -0
+    * the object will need to be accessed via the CPU.
+    *
+    * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE,

and

Documentation/gpu/rfc/i915_small_bar.h

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+    * only strictly required on platforms where only some of the device
+    * memory is directly visible or mappable through the CPU, like on DG2+.
+    *
+    * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM,

to

Documentation/gpu/rfc/i915_small_bar.h

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
+    * ensure we can always spill the allocation to system memory, if we
+    * can't place the object in the mappable part of
+    * I915_MEMORY_CLASS_DEVICE.
+    *
+    * Note that since the kernel only supports flat-CCS on objects that can
+    * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore

don't

Documentation/gpu/rfc/i915_small_bar.h line 94

Quoted 1 line
Documentation/gpu/rfc/i915_small_bar.h new file +1 -0
+    * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS

together with

Documentation/gpu/rfc/i915_small_bar.h

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+    * flat-CCS.
+    *
+    * Without this hint, the kernel will assume that non-mappable
+    * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that

the

Quoted 4 lines
Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+    * kernel can still migrate the object to the mappable part, as a last
+    * resort, if userspace ever CPU faults this object, but this might be
+    * expensive, and so ideally should be avoided.
+    */

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+   __u32 flags;
+   /**
+    * @extensions: The chain of extensions to apply to this object.
+    *
+    * This will be useful in the future when we need to support several
+    * different extensions, and we need to apply more than one when
+    * creating the object. See struct i915_user_extension.
+    *
+    * If we don't supply any extensions then we get the same old

gem_create

Quoted 51 lines
Documentation/gpu/rfc/i915_small_bar.h new file +51 -0
+    * behaviour.
+    *
+    * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+    * struct drm_i915_gem_create_ext_memory_regions.
+    *
+    * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+    * struct drm_i915_gem_create_ext_protected_content.
+    */
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0 #define
+I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+   __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma,
+returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ * struct drm_i915_query_vma_info info = {};
+ * struct drm_i915_query_item item = {
+ *         .data_ptr = (uintptr_t)&info,
+ *         .query_id = DRM_I915_QUERY_VMA_INFO,
+ * };
+ * struct drm_i915_query query = {
+ *         .num_items = 1,
+ *         .items_ptr = (uintptr_t)&item,
+ * };
+ * int err;
+ *
+ * // Unlike some other types of queries, there is no need to first query
+ * // the size of the data_ptr blob here, since we already know ahead of
+ * // time how big this needs to be.
+ * item.length = sizeof(info);
+ *
+ * // Next we fill in the vm_id and ppGTT address of the vma we wish
+ * // to query, before then firing off the query.
+ * info.vm_id = vm_id;
+ * info.offset = gtt_address;
+ * err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ * if (err || item.length < 0) ...
+ *
+ * // If all went well we can now inspect the returned attributes.
+ * if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+   /**
+    * @vm_id: The given vm id that contains the vma. The id is the value
+    * returned by the DRM_I915_GEM_VM_CREATE. See struct
+    * drm_i915_gem_vm_control.vm_id.
+    */
+   __u32 vm_id;
+   /** @pad: MBZ. */
+   __u32 pad;
+   /**
+    * @offset: The corresponding ppGTT address of the vma which the

kernel

Documentation/gpu/rfc/i915_small_bar.h

Quoted 9 lines
Documentation/gpu/rfc/i915_small_bar.h new file +9 -0
+    * will use to perform the lookup.
+    */
+   __u64 offset;
+   /**
+    * @attributes: The returned attributes for the given vma.
+    *
+    * Possible values:
+    *
+    * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages

backing the

Documentation/gpu/rfc/i915_small_bar.h line 184

Quoted 2 lines
Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
+    * vma are currently CPU accessible. If this is not set then the vma is
+    * currently backed by I915_MEMORY_CLASS_DEVICE memory, which

the CPU

Documentation/gpu/rfc/i915_small_bar.h

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+    * cannot directly access(this is only possible on discrete devices with
+    * a small BAR). Attempting to MMAP and fault such an object will
+    * require the kernel first synchronising any GPU work tied to the
+    * object, before then migrating the pages, either to the CPU accessible
+    * part of I915_MEMORY_CLASS_DEVICE, or

I915_MEMORY_CLASS_SYSTEM, if the

Documentation/gpu/rfc/i915_small_bar.h line 190

Quoted 1 line
Documentation/gpu/rfc/i915_small_bar.h new file +1 -0
+    * placements permit it. See

I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.

Quoted 15 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+    *
+    * Note that this is inherently racy.
+    */
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+   __u64 attributes;
+   /** @rsvd: MBZ */
+   __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +0 -0
new file mode 100644
index 000000000000..be3d9bcdd86d

Documentation/gpu/rfc/i915_small_bar.rst line 8

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device
+local-memory(i.e I915_MEMORY_CLASS_DEVICE), but in some cases the
+final BAR size might still be smaller than the total probed_size. In
+such cases, only some subset of I915_MEMORY_CLASS_DEVICE will be CPU
+accessible(for example the first 256M), while the remainder is only accessible

via the GPU.

Documentation/gpu/rfc/i915_small_bar.rst line 17

Quoted 9 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU access.
+This becomes important when placing an object in
+I915_MEMORY_CLASS_DEVICE, where underneath the device has a small
+BAR, meaning only some portion of it is CPU accessible. Without this
+flag the kernel will assume that CPU access is not required, and
+prioritize using the non-CPU visible portion of

I915_MEMORY_CLASS_DEVICE.

Quoted 16 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +16 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total
+size of the CPU accessible portion, for the particular region. This
+should only be applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap
+with the VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the
+CPU visible portion, where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info

Documentation/gpu/rfc/i915_small_bar.rst

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only
+support DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the
+object/vma is currently placed in memory that is accessible by the
+CPU. This should always be set on devices where the CPU
+probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE matches the
+probed_size. If this is not set then CPU faulting the object will likely first

require migrating the pages.

Quoted 28 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +15 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.
(quoted diff) +0 -0

b/Documentation/gpu/rfc/index.rst index 91e93a705230..5a3bd3924ba6 100644

Documentation/gpu/rfc/index.rst +4 -0
@@ -23,3 +23,7 @@ host such documentation:
   .. toctree::
 
       i915_scheduler.rst
+
+.. toctree::
+
+    i915_small_bar.rst
#26 lore raw
Quoted 92 lines · Lionel Landwerlin
Quoted 91 lines · Akeem G
Quoted 91 lines

-----Original Message-----
From: Landwerlin, Lionel G <[email protected]>
Sent: Monday, May 2, 2022 12:55 AM
To: Auld, Matthew <[email protected]>; [email protected]
Cc: [email protected]; Thomas Hellström <[email protected]>; Bloomfield, Jon <[email protected]>; Daniel Vetter <[email protected]>; Justen, Jordan L <[email protected]>; Kenneth Graunke <[email protected]>; Abodunrin, Akeem G <[email protected]>; [email protected]
Subject: Re: [PATCH v2] drm/doc: add rfc section for small BAR uapi

Quoted 17 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
    - Some spelling fixes and other small tweaks. (Akeem & Thomas)
    - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
    - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
Quoted 2 lines

---
   Documentation/gpu/rfc/i915_small_bar.h   | 190

+++++++++++++++++++++++

Quoted 13 lines

   Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

   3 files changed, 252 insertions(+)

   create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +0 -0
new file mode 100644
index 000000000000..7bfd0cf44d35

Documentation/gpu/rfc/i915_small_bar.h line 6

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as
+known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct

drm_i915_query.

Quoted 5 lines
Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+ * For this new query we are adding the new query id
+DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {

Documentation/gpu/rfc/i915_small_bar.h

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.h new file +10 -0
+   /** @region: The class:instance pair encoding */
+   struct drm_i915_gem_memory_class_instance region;
+
+   /** @rsvd0: MBZ */
+   __u32 rsvd0;
+
+   /** @probed_size: Memory probed by the driver (-1 = unknown) */
+   __u64 probed_size;
+
+   /** @unallocated_size: Estimate of memory remaining (-1 = unknown)

*/

Documentation/gpu/rfc/i915_small_bar.h

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+   __u64 unallocated_size;
+
+   union {
+           /** @rsvd1: MBZ */
+           __u64 rsvd1[8];
+           struct {
+                   /**
+                    * @probed_cpu_visible_size: Memory probed by the

driver

Documentation/gpu/rfc/i915_small_bar.h line 35

Quoted 8 lines
Documentation/gpu/rfc/i915_small_bar.h new file +8 -0
+                    * that is CPU accessible. (-1 = unknown).
+                    *
+                    * This will be always be <= @probed_size, and the
+                    * remainder(if there is any) will not be CPU
+                    * accessible.
+                    */
+                   __u64 probed_cpu_visible_size;
+           };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

When is it set to -1?

I believe it is set to -1 if it is unknown, and/or not cpu accessible...

Cheers!
~Akeem

So what should I expect on system memory?

I guess just probed_cpu_visible_size == probed_size. Or maybe we can just use -1 here?

Quoted 2 lines

What value is returned when all of probed_size is CPU visible on local memory?

probed_size == probed_cpu_visible_size.

Quoted 296 lines

Thanks,

-Lionel

Quoted 286 lines
Quoted 286 lines

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Quoted 23 lines
Documentation/gpu/rfc/i915_small_bar.h new file +23 -0
+   };
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour,
+with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the
+stuff that
+ * is immutable. Previously we would have two ioctls, one to create
+the object
+ * with gem_create, and another to apply various parameters, however
+this
+ * creates some ambiguity for the params which are considered
+immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+   /**
+    * @size: Requested size for the object.
+    *
+    * The (page-aligned) allocated size for the object will be 
Quoted 7 lines

returned.
+    *
+    * Note that for some devices we have might have further minimum +    * page-size restrictions(larger than 4K), like for device local-memory.
+    * However in general the final size here should always reflect any +    * rounding up, if for example using the

I915_GEM_CREATE_EXT_MEMORY_REGIONS

Quoted 7 lines

+    * extension to place the object in device local-memory. +    */
+   __u64 size;
+   /**
+    * @handle: Returned handle for the object.
+    *
+    * Object handles are nonzero.

Quoted 8 lines

+    */
+   __u32 handle;
+   /**
+    * @flags: Optional flags.
+    *
+    * Supported values:
+    *
+    * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the

kernel that

Quoted 3 lines

+    * the object will need to be accessed via the CPU. +    *
+    * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE,

and

Quoted 6 lines

+    * only strictly required on platforms where only some of the device
+    * memory is directly visible or mappable through the CPU, like on DG2+.
+    *
+    * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM,

to

Quoted 8 lines

+    * ensure we can always spill the allocation to system memory, if we
+    * can't place the object in the mappable part of
+    * I915_MEMORY_CLASS_DEVICE.
+    *
+    * Note that since the kernel only supports flat-CCS on objects that can
+    * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore

don't

Quoted 1 line

+    * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS

together with

Quoted 4 lines

+    * flat-CCS.
+    *
+    * Without this hint, the kernel will assume that non-mappable +    * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that

the

Quoted 9 lines

+    * kernel can still migrate the object to the mappable part, as a last
+    * resort, if userspace ever CPU faults this object, but this might be
+    * expensive, and so ideally should be avoided.
+    */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +   __u32 flags;
+   /**

Quoted 8 lines

+    * @extensions: The chain of extensions to apply to this object. +    *
+    * This will be useful in the future when we need to support several
+    * different extensions, and we need to apply more than one when +    * creating the object. See struct i915_user_extension. +    *
+    * If we don't supply any extensions then we get the same old

gem_create

Quoted 56 lines

+    * behaviour.
+    *
+    * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see +    * struct drm_i915_gem_create_ext_memory_regions. +    *
+    * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +    * struct drm_i915_gem_create_ext_protected_content. +    */

Documentation/gpu/rfc/i915_small_bar.h new file +28 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0 #define
+I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+   __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma,
+returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ * struct drm_i915_query_vma_info info = {};
+ * struct drm_i915_query_item item = {
+ *         .data_ptr = (uintptr_t)&info,
+ *         .query_id = DRM_I915_QUERY_VMA_INFO,
+ * };
+ * struct drm_i915_query query = {
+ *         .num_items = 1,
+ *         .items_ptr = (uintptr_t)&item,
+ * };
+ * int err;
+ *
+ * // Unlike some other types of queries, there is no need to first 

query
+ * // the size of the data_ptr blob here, since we already know ahead of
+ * // time how big this needs to be.
+ * item.length = sizeof(info);
+ *
+ * // Next we fill in the vm_id and ppGTT address of the vma we wish
+ * // to query, before then firing off the query.
+ * info.vm_id = vm_id;
+ * info.offset = gtt_address;
+ * err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ * if (err || item.length < 0) ...
+ *
+ * // If all went well we can now inspect the returned attributes.
+ * if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+   /**
+    * @vm_id: The given vm id that contains the vma. The id is the value

Quoted 8 lines

+    * returned by the DRM_I915_GEM_VM_CREATE. See struct +    * drm_i915_gem_vm_control.vm_id.
+    */
+   __u32 vm_id;
+   /** @pad: MBZ. */
+   __u32 pad;
+   /**
+    * @offset: The corresponding ppGTT address of the vma which the

kernel

Quoted 9 lines

+    * will use to perform the lookup.
+    */
+   __u64 offset;
+   /**
+    * @attributes: The returned attributes for the given vma. +    *
+    * Possible values:
+    *
+    * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages

backing the

Quoted 3 lines

+    * vma are currently CPU accessible. If this is not set then the vma is
+    * currently backed by I915_MEMORY_CLASS_DEVICE memory, which

the CPU

Quoted 7 lines

+    * cannot directly access(this is only possible on discrete devices with
+    * a small BAR). Attempting to MMAP and fault such an object will +    * require the kernel first synchronising any GPU work tied to the +    * object, before then migrating the pages, either to the CPU accessible
+    * part of I915_MEMORY_CLASS_DEVICE, or

I915_MEMORY_CLASS_SYSTEM, if the

Quoted 1 line

+    * placements permit it. See

I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.

Quoted 23 lines

+    *
+    * Note that this is inherently racy.
+    */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+   __u64 attributes;
+   /** @rsvd: MBZ */
+   __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device
+local-memory(i.e I915_MEMORY_CLASS_DEVICE), but in some cases the
+final BAR size might still be smaller than the total probed_size. In
+such cases, only some subset of I915_MEMORY_CLASS_DEVICE will be CPU
+accessible(for example the first 256M), while the remainder is only 
Quoted 1 line

accessible

via the GPU.

Documentation/gpu/rfc/i915_small_bar.rst

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+This becomes important when placing an object in
+I915_MEMORY_CLASS_DEVICE, where underneath the device has a small
+BAR, meaning only some portion of it is CPU accessible. Without this
+flag the kernel will assume that CPU access is not required, and
+prioritize using the non-CPU visible portion of

I915_MEMORY_CLASS_DEVICE.

Quoted 27 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +13 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total
+size of the CPU accessible portion, for the particular region. This
+should only be applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap
+with the VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the
+CPU visible portion, where the total size of the heap needs to be 

known.

Documentation/gpu/rfc/i915_small_bar.rst new file +13 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only
+support DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the
+object/vma is currently placed in memory that is accessible by the
+CPU. This should always be set on devices where the CPU
+probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE matches the
+probed_size. If this is not set then CPU faulting the object will 
Quoted 1 line

likely first

require migrating the pages.

Quoted 32 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

(quoted diff) +0 -0

b/Documentation/gpu/rfc/index.rst index 91e93a705230..5a3bd3924ba6 100644

Documentation/gpu/rfc/index.rst +0 -0
@@ -23,3 +23,7 @@ host such documentation:

   .. toctree::

       i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#27 lore raw
Quoted 102 lines · Matthew Auld
Quoted 101 lines · Lionel Landwerlin
Quoted 98 lines · Akeem G
Quoted 90 lines

-----Original Message-----
From: Landwerlin, Lionel G <[email protected]>
Sent: Monday, May 2, 2022 12:55 AM
To: Auld, Matthew <[email protected]>; [email protected]
Cc: [email protected]; Thomas Hellström <[email protected]>; Bloomfield, Jon <[email protected]>; Daniel Vetter <[email protected]>; Justen, Jordan L <[email protected]>; Kenneth Graunke <[email protected]>; Abodunrin, Akeem G <[email protected]>; [email protected]
Subject: Re: [PATCH v2] drm/doc: add rfc section for small BAR uapi

Quoted 17 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
    - Some spelling fixes and other small tweaks. (Akeem & Thomas)
    - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
    - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>
Quoted 2 lines

---
   Documentation/gpu/rfc/i915_small_bar.h   | 190

+++++++++++++++++++++++

Quoted 13 lines

Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

   3 files changed, 252 insertions(+)

   create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +0 -0
new file mode 100644
index 000000000000..7bfd0cf44d35

Documentation/gpu/rfc/i915_small_bar.h line 6

Quoted 6 lines
Documentation/gpu/rfc/i915_small_bar.h new file +6 -0
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as
+known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct

drm_i915_query.

Quoted 15 lines
Documentation/gpu/rfc/i915_small_bar.h new file +15 -0
+ * For this new query we are adding the new query id
+DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */
+struct __drm_i915_memory_region_info {
+   /** @region: The class:instance pair encoding */
+   struct drm_i915_gem_memory_class_instance region;
+
+   /** @rsvd0: MBZ */
+   __u32 rsvd0;
+
+   /** @probed_size: Memory probed by the driver (-1 = unknown) */
+   __u64 probed_size;
+
+   /** @unallocated_size: Estimate of memory remaining (-1 = 
Quoted 1 line

unknown)

*/

Quoted 8 lines

+   __u64 unallocated_size;
+
+   union {
+           /** @rsvd1: MBZ */
+           __u64 rsvd1[8];
+           struct {
+                   /**
+                    * @probed_cpu_visible_size: Memory probed by the

driver

Quoted 9 lines

+                    * that is CPU accessible. (-1 = unknown). +                    *
+                    * This will be always be <= @probed_size, and the
+                    * remainder(if there is any) will not be CPU +                    * accessible.
+                    */
+                   __u64 probed_cpu_visible_size;
+           };

Quoted 4 lines

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

When is it set to -1?

I believe it is set to -1 if it is unknown, and/or not cpu accessible...

Cheers!
~Akeem

Quoted 8 lines
Quoted 1 line

So what should I expect on system memory?

I guess just probed_cpu_visible_size == probed_size. Or maybe we can just use -1 here?

Quoted 2 lines

What value is returned when all of probed_size is CPU visible on local memory?

probed_size == probed_cpu_visible_size.

Thanks, looks good to me.

Then maybe we should update the comment to say that.

Looks like there are no cases where we'll get -1.

-Lionel

Quoted 301 lines
Quoted 301 lines

Thanks,

-Lionel

Quoted 290 lines
Quoted 290 lines

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

-Lionel

Quoted 23 lines

+   };

Documentation/gpu/rfc/i915_small_bar.h new file +22 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour,
+with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the
+stuff that
+ * is immutable. Previously we would have two ioctls, one to create
+the object
+ * with gem_create, and another to apply various parameters, however
+this
+ * creates some ambiguity for the params which are considered
+immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+   /**
+    * @size: Requested size for the object.
+    *
+    * The (page-aligned) allocated size for the object will be 
Quoted 8 lines

returned.
+    *
+    * Note that for some devices we have might have further minimum +    * page-size restrictions(larger than 4K), like for device local-memory.
+    * However in general the final size here should always reflect any
+    * rounding up, if for example using the

I915_GEM_CREATE_EXT_MEMORY_REGIONS

Quoted 7 lines

+    * extension to place the object in device local-memory. +    */
+   __u64 size;
+   /**
+    * @handle: Returned handle for the object.
+    *
+    * Object handles are nonzero.

Quoted 8 lines

+    */
+   __u32 handle;
+   /**
+    * @flags: Optional flags.
+    *
+    * Supported values:
+    *
+    * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the

kernel that

Quoted 3 lines

+    * the object will need to be accessed via the CPU. +    *
+    * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE,

and

Quoted 6 lines

+    * only strictly required on platforms where only some of the device
+    * memory is directly visible or mappable through the CPU, like on DG2+.
+    *
+    * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM,

to

Quoted 8 lines

+    * ensure we can always spill the allocation to system memory, if we
+    * can't place the object in the mappable part of
+    * I915_MEMORY_CLASS_DEVICE.
+    *
+    * Note that since the kernel only supports flat-CCS on objects that can
+    * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore

don't

Quoted 1 line

+    * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS

together with

Quoted 5 lines

+    * flat-CCS.
+    *
+    * Without this hint, the kernel will assume that non-mappable +    * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that

the

Quoted 9 lines

+    * kernel can still migrate the object to the mappable part, as a last
+    * resort, if userspace ever CPU faults this object, but this might be
+    * expensive, and so ideally should be avoided.
+    */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0) +   __u32 flags;
+   /**

Quoted 8 lines

+    * @extensions: The chain of extensions to apply to this object. +    *
+    * This will be useful in the future when we need to support several
+    * different extensions, and we need to apply more than one when +    * creating the object. See struct i915_user_extension. +    *
+    * If we don't supply any extensions then we get the same old

gem_create

Quoted 56 lines

+    * behaviour.
+    *
+    * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see +    * struct drm_i915_gem_create_ext_memory_regions. +    *
+    * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see +    * struct drm_i915_gem_create_ext_protected_content. +    */

Documentation/gpu/rfc/i915_small_bar.h new file +28 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0 #define
+I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+   __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma,
+returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ * struct drm_i915_query_vma_info info = {};
+ * struct drm_i915_query_item item = {
+ *         .data_ptr = (uintptr_t)&info,
+ *         .query_id = DRM_I915_QUERY_VMA_INFO,
+ * };
+ * struct drm_i915_query query = {
+ *         .num_items = 1,
+ *         .items_ptr = (uintptr_t)&item,
+ * };
+ * int err;
+ *
+ * // Unlike some other types of queries, there is no need to 

first query
+ * // the size of the data_ptr blob here, since we already know ahead of
+ * // time how big this needs to be.
+ * item.length = sizeof(info);
+ *
+ * // Next we fill in the vm_id and ppGTT address of the vma we wish
+ * // to query, before then firing off the query.
+ * info.vm_id = vm_id;
+ * info.offset = gtt_address;
+ * err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ * if (err || item.length < 0) ...
+ *
+ * // If all went well we can now inspect the returned attributes.
+ * if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+   /**
+    * @vm_id: The given vm id that contains the vma. The id is the value

Quoted 8 lines

+    * returned by the DRM_I915_GEM_VM_CREATE. See struct +    * drm_i915_gem_vm_control.vm_id.
+    */
+   __u32 vm_id;
+   /** @pad: MBZ. */
+   __u32 pad;
+   /**
+    * @offset: The corresponding ppGTT address of the vma which the

kernel

Quoted 9 lines

+    * will use to perform the lookup.
+    */
+   __u64 offset;
+   /**
+    * @attributes: The returned attributes for the given vma. +    *
+    * Possible values:
+    *
+    * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages

backing the

Quoted 3 lines

+    * vma are currently CPU accessible. If this is not set then the vma is
+    * currently backed by I915_MEMORY_CLASS_DEVICE memory, which

the CPU

Quoted 8 lines

+    * cannot directly access(this is only possible on discrete devices with
+    * a small BAR). Attempting to MMAP and fault such an object will +    * require the kernel first synchronising any GPU work tied to the
+    * object, before then migrating the pages, either to the CPU accessible
+    * part of I915_MEMORY_CLASS_DEVICE, or

I915_MEMORY_CLASS_SYSTEM, if the

Quoted 1 line

+    * placements permit it. See

I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.

Quoted 23 lines

+    *
+    * Note that this is inherently racy.
+    */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+   __u64 attributes;
+   /** @rsvd: MBZ */
+   __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +8 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device
+local-memory(i.e I915_MEMORY_CLASS_DEVICE), but in some cases the
+final BAR size might still be smaller than the total probed_size. In
+such cases, only some subset of I915_MEMORY_CLASS_DEVICE will be CPU
+accessible(for example the first 256M), while the remainder is 
Quoted 1 line

only accessible

via the GPU.

Documentation/gpu/rfc/i915_small_bar.rst

Quoted 10 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require 

CPU access.

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+This becomes important when placing an object in
+I915_MEMORY_CLASS_DEVICE, where underneath the device has a small
+BAR, meaning only some portion of it is CPU accessible. Without this
+flag the kernel will assume that CPU access is not required, and
+prioritize using the non-CPU visible portion of

I915_MEMORY_CLASS_DEVICE.

Quoted 28 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +13 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total
+size of the CPU accessible portion, for the particular region. This
+should only be applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap
+with the VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the
+CPU visible portion, where the total size of the heap needs to be 

known.

Documentation/gpu/rfc/i915_small_bar.rst new file +7 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find 

the

Documentation/gpu/rfc/i915_small_bar.rst new file +6 -0
+respective vma, and return its set of attributes. For now we only
+support DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the
+object/vma is currently placed in memory that is accessible by the
+CPU. This should always be set on devices where the CPU
+probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE matches the
+probed_size. If this is not set then CPU faulting the object will 
Quoted 1 line

likely first

require migrating the pages.

Quoted 32 lines
Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

(quoted diff) +0 -0

b/Documentation/gpu/rfc/index.rst index 91e93a705230..5a3bd3924ba6 100644

Documentation/gpu/rfc/index.rst +0 -0
@@ -23,3 +23,7 @@ host such documentation:

   .. toctree::

       i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst
#28 lore raw
In reply to Lionel Landwerlin
Quoted 75 lines · Lionel Landwerlin
Quoted 62 lines · Matthew Auld

Add an entry for the new uapi needed for small BAR on DG2+.

v2:
   - Some spelling fixes and other small tweaks. (Akeem & Thomas)
   - Rework error capture interactions, including no longer needing NEEDS_CPU_ACCESS for objects marked for capture. (Thomas)
   - Add probed_cpu_visible_size. (Lionel)

Cc Thomas Hellström <[email protected]>
Cc Lionel Landwerlin <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jordan Justen <[email protected]>
Cc Kenneth Graunke <[email protected]>
Cc Akeem G Abodunrin <[email protected]>

---
  Documentation/gpu/rfc/i915_small_bar.h   | 190 +++++++++++++++++++++++ Documentation/gpu/rfc/i915_small_bar.rst |  58 +++++++ Documentation/gpu/rfc/index.rst          |   4 +

  3 files changed, 252 insertions(+)

  create mode 100644 Documentation/gpu/rfc/i915_small_bar.h create mode 100644 Documentation/gpu/rfc/i915_small_bar.rst

Documentation/gpu/rfc/i915_small_bar.h new file +2 -0
new file mode 100644
index 000000000000..7bfd0cf44d35
@@ -0,0 +1,190 @@
+/**
+ * struct __drm_i915_memory_region_info - Describes one region as 

known to the
+ * driver.
+ *
+ * Note this is using both struct drm_i915_query_item and struct drm_i915_query.
+ * For this new query we are adding the new query id DRM_I915_QUERY_MEMORY_REGIONS
+ * at &drm_i915_query_item.query_id.
+ */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+struct __drm_i915_memory_region_info {
+    /** @region: The class:instance pair encoding */
+    struct drm_i915_gem_memory_class_instance region;
+
+    /** @rsvd0: MBZ */
+    __u32 rsvd0;
+
+    /** @probed_size: Memory probed by the driver (-1 = unknown) */
+    __u64 probed_size;
+
+    /** @unallocated_size: Estimate of memory remaining (-1 = 

unknown) */
+    __u64 unallocated_size;
+
+    union {
+        /** @rsvd1: MBZ */
+        __u64 rsvd1[8];
+        struct {
+            /**
+             * @probed_cpu_visible_size: Memory probed by the driver

Quoted 8 lines

+             * that is CPU accessible. (-1 = unknown). +             *
+             * This will be always be <= @probed_size, and the +             * remainder(if there is any) will not be CPU +             * accessible.
+             */
+            __u64 probed_cpu_visible_size;
+        };

Trying to implement userspace support in Vulkan for this, I have an additional question about the value of probed_cpu_visible_size.

Quoted 1 line

When is it set to -1?

I don't anything is currently using -1, for any of these fields.

Quoted 4 lines

I'm guessing before there is support for this value it'll be 0 (MBZ).

After after it should either be the entire lmem or something smaller.

Yup.

Quoted 282 lines

-Lionel

Quoted 275 lines

+    };

Documentation/gpu/rfc/i915_small_bar.h new file +4 -0
+};
+
+/**
+ * struct __drm_i915_gem_create_ext - Existing gem_create behaviour, 

with added
+ * extension support using struct i915_user_extension.
+ *
+ * Note that new buffer flags should be added here, at least for the stuff that
+ * is immutable. Previously we would have two ioctls, one to create the object
+ * with gem_create, and another to apply various parameters, however this
+ * creates some ambiguity for the params which are considered immutable. Also in
+ * general we're phasing out the various SET/GET ioctls.
+ */
+struct __drm_i915_gem_create_ext {
+    /**
+     * @size: Requested size for the object.
+     *
+     * The (page-aligned) allocated size for the object will be returned.
+     *
+     * Note that for some devices we have might have further minimum +     * page-size restrictions(larger than 4K), like for device local-memory.
+     * However in general the final size here should always reflect any +     * rounding up, if for example using the I915_GEM_CREATE_EXT_MEMORY_REGIONS
+     * extension to place the object in device local-memory. +     */
+    __u64 size;
+    /**
+     * @handle: Returned handle for the object.
+     *
+     * Object handles are nonzero.
+     */
+    __u32 handle;
+    /**
+     * @flags: Optional flags.
+     *
+     * Supported values:
+     *
+     * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS - Signal to the kernel that
+     * the object will need to be accessed via the CPU.
+     *
+     * Only valid when placing objects in I915_MEMORY_CLASS_DEVICE, and +     * only strictly required on platforms where only some of the device +     * memory is directly visible or mappable through the CPU, like on DG2+.
+     *
+     * One of the placements MUST also be I915_MEMORY_CLASS_SYSTEM, to +     * ensure we can always spill the allocation to system memory, if we +     * can't place the object in the mappable part of
+     * I915_MEMORY_CLASS_DEVICE.
+     *
+     * Note that since the kernel only supports flat-CCS on objects that can
+     * *only* be placed in I915_MEMORY_CLASS_DEVICE, we therefore don't +     * support I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS together with +     * flat-CCS.
+     *
+     * Without this hint, the kernel will assume that non-mappable +     * I915_MEMORY_CLASS_DEVICE is preferred for this object. Note that the
+     * kernel can still migrate the object to the mappable part, as a last
+     * resort, if userspace ever CPU faults this object, but this might be
+     * expensive, and so ideally should be avoided.
+     */
+#define I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS (1 << 0)
+    __u32 flags;
+    /**
+     * @extensions: The chain of extensions to apply to this object. +     *
+     * This will be useful in the future when we need to support several +     * different extensions, and we need to apply more than one when +     * creating the object. See struct i915_user_extension. +     *
+     * If we don't supply any extensions then we get the same old gem_create
+     * behaviour.
+     *
+     * For I915_GEM_CREATE_EXT_MEMORY_REGIONS usage see
+     * struct drm_i915_gem_create_ext_memory_regions.
+     *
+     * For I915_GEM_CREATE_EXT_PROTECTED_CONTENT usage see
+     * struct drm_i915_gem_create_ext_protected_content.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +11 -0
+#define I915_GEM_CREATE_EXT_MEMORY_REGIONS 0
+#define I915_GEM_CREATE_EXT_PROTECTED_CONTENT 1
+    __u64 extensions;
+};
+
+#define DRM_I915_QUERY_VMA_INFO    5
+
+/**
+ * struct __drm_i915_query_vma_info
+ *
+ * Given a vm and GTT address, lookup the corresponding vma, 

returning its set
+ * of attributes.
+ *
+ * .. code-block:: C
+ *
+ *    struct drm_i915_query_vma_info info = {};
+ *    struct drm_i915_query_item item = {
+ *        .data_ptr = (uintptr_t)&info,
+ *        .query_id = DRM_I915_QUERY_VMA_INFO,
+ *    };
+ *    struct drm_i915_query query = {
+ *        .num_items = 1,
+ *        .items_ptr = (uintptr_t)&item,
+ *    };
+ *    int err;
+ *
+ *    // Unlike some other types of queries, there is no need to first query
+ *    // the size of the data_ptr blob here, since we already know ahead of
+ *    // time how big this needs to be.
+ *    item.length = sizeof(info);
+ *
+ *    // Next we fill in the vm_id and ppGTT address of the vma we wish
+ *    // to query, before then firing off the query.
+ *    info.vm_id = vm_id;
+ *    info.offset = gtt_address;
+ *    err = ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+ *    if (err || item.length < 0) ...
+ *
+ *    // If all went well we can now inspect the returned attributes.
+ *    if (info.attributes & DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE) ...
+ */
+struct __drm_i915_query_vma_info {
+    /**
+     * @vm_id: The given vm id that contains the vma. The id is the value
+     * returned by the DRM_I915_GEM_VM_CREATE. See struct
+     * drm_i915_gem_vm_control.vm_id.
+     */
+    __u32 vm_id;
+    /** @pad: MBZ. */
+    __u32 pad;
+    /**
+     * @offset: The corresponding ppGTT address of the vma which the kernel
+     * will use to perform the lookup.
+     */
+    __u64 offset;
+    /**
+     * @attributes: The returned attributes for the given vma. +     *
+     * Possible values:
+     *
+     * DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE - Set if the pages backing the
+     * vma are currently CPU accessible. If this is not set then the vma is
+     * currently backed by I915_MEMORY_CLASS_DEVICE memory, which the CPU
+     * cannot directly access(this is only possible on discrete devices with
+     * a small BAR). Attempting to MMAP and fault such an object will +     * require the kernel first synchronising any GPU work tied to the +     * object, before then migrating the pages, either to the CPU accessible
+     * part of I915_MEMORY_CLASS_DEVICE, or I915_MEMORY_CLASS_SYSTEM, if the
+     * placements permit it. See I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS.
+     *
+     * Note that this is inherently racy.
+     */

Documentation/gpu/rfc/i915_small_bar.h new file +5 -0
+#define DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE (1<<0)
+    __u64 attributes;
+    /** @rsvd: MBZ */
+    __u32 rsvd[4];
+};
Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
new file mode 100644
index 000000000000..be3d9bcdd86d
@@ -0,0 +1,58 @@
+==========================
+I915 Small BAR RFC Section
+==========================
+Starting from DG2 we will have resizable BAR support for device 

local-memory(i.e
+I915_MEMORY_CLASS_DEVICE), but in some cases the final BAR size might still be

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+smaller than the total probed_size. In such cases, only some subset of
+I915_MEMORY_CLASS_DEVICE will be CPU accessible(for example the first 

256M),

Documentation/gpu/rfc/i915_small_bar.rst new file +5 -0
+while the remainder is only accessible via the GPU.
+
+I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS flag
+----------------------------------------------
+New gem_create_ext flag to tell the kernel that a BO will require CPU 

access.
+This becomes important when placing an object in I915_MEMORY_CLASS_DEVICE, where
+underneath the device has a small BAR, meaning only some portion of it is CPU
+accessible. Without this flag the kernel will assume that CPU access is not

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+required, and prioritize using the non-CPU visible portion of
+I915_MEMORY_CLASS_DEVICE.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_gem_create_ext
+
+probed_cpu_visible_size attribute
+---------------------------------
+New struct__drm_i915_memory_region attribute which returns the total 

size of the

Documentation/gpu/rfc/i915_small_bar.rst new file +4 -0
+CPU accessible portion, for the particular region. This should only be
+applicable for I915_MEMORY_CLASS_DEVICE.
+
+Vulkan will need this as part of creating a separate VkMemoryHeap 

with the
+VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT set, to represent the CPU visible portion,

Documentation/gpu/rfc/i915_small_bar.rst new file +9 -0
+where the total size of the heap needs to be known.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_memory_region_info
+
+DRM_I915_QUERY_VMA_INFO query
+-----------------------------
+Query the attributes of some vma. Given a vm and GTT offset, find the
+respective vma, and return its set of attributes. For now we only 

support

Documentation/gpu/rfc/i915_small_bar.rst new file +2 -0
+DRM_I915_QUERY_VMA_INFO_CPU_VISIBLE, which is set if the object/vma is
+currently placed in memory that is accessible by the CPU. This should 

always be
+set on devices where the CPU probed_cpu_visible_size of I915_MEMORY_CLASS_DEVICE
+matches the probed_size. If this is not set then CPU faulting the object will

Documentation/gpu/rfc/i915_small_bar.rst new file +10 -0
+likely first require migrating the pages.
+
+.. kernel-doc:: Documentation/gpu/rfc/i915_small_bar.h
+   :functions: __drm_i915_query_vma_info
+
+Error Capture restrictions
+--------------------------
+With error capture we have two new restrictions:
+
+    1) Error capture is best effort on small BAR systems; if the 

pages are not
+    CPU accessible, at the time of capture, then the kernel is free to skip
+    trying to capture them.
+
+    2) On discrete we now reject error capture on recoverable contexts. In the
+    future the kernel may want to blit during error capture, when for example
+    something is not currently CPU accessible.

Documentation/gpu/rfc/index.rst +0 -0
index 91e93a705230..5a3bd3924ba6 100644
@@ -23,3 +23,7 @@ host such documentation:

  .. toctree::
      i915_scheduler.rst

Documentation/gpu/rfc/index.rst +4 -0
+
+.. toctree::
+
+    i915_small_bar.rst