From: Matthew Brost <matthew.brost@intel.com>
To: <intel-xe@lists.freedesktop.org>
Cc: <thomas.hellstrom@linux.intel.com>,
Matthew Brost <matthew.brost@intel.com>,
Paulo Zanoni <paulo.r.zanoni@intel.com>
Subject: [PATCH v2 3/3] drm/xe: Return 2MB page size for compact 64k PTEs
Date: Mon, 19 Feb 2024 13:19:42 -0800 [thread overview]
Message-ID: <20240219211942.3633795-4-matthew.brost@intel.com> (raw)
In-Reply-To: <20240219211942.3633795-1-matthew.brost@intel.com>
Compact 64k PTEs are only intended to be used within a single VMA which
covers the entire 2MB range of the compact 64k PTEs. Add
XE_VMA_PTE_COMPACT VMA flag to indicate compact 64k PTEs are used and
update xe_vma_max_pte_size to return at least 2MB if set.
v2: Include missing changes
Link: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/758
Fixes: 8f33b4f054fc ("drm/xe: Avoid doing rebinds")
Fixes: c47794bdd63d ("drm/xe: Set max pte size when skipping rebinds")
Reported-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
drivers/gpu/drm/xe/xe_pt.c | 5 ++++-
drivers/gpu/drm/xe/xe_vm.c | 2 +-
drivers/gpu/drm/xe/xe_vm_types.h | 1 +
3 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index 46cea37bdac5..7f54bc3e389d 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -547,13 +547,16 @@ xe_pt_stage_bind_entry(struct xe_ptw *parent, pgoff_t offset,
*child = &xe_child->base;
/*
- * Prefer the compact pagetable layout for L0 if possible.
+ * Prefer the compact pagetable layout for L0 if possible. Only
+ * possible if VMA covers entire 2MB region as compact 64k and
+ * 4k pages cannot be mixed within a 2MB region.
* TODO: Suballocate the pt bo to avoid wasting a lot of
* memory.
*/
if (GRAPHICS_VERx100(tile_to_xe(xe_walk->tile)) >= 1250 && level == 1 &&
covers && xe_pt_scan_64K(addr, next, xe_walk)) {
walk->shifts = xe_compact_pt_shifts;
+ xe_walk->vma->gpuva.flags |= XE_VMA_PTE_COMPACT;
flags |= XE_PDE_64K;
xe_child->is_compact = true;
}
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 7781290d2b99..23a44ef85aa4 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -2190,7 +2190,7 @@ static u64 xe_vma_max_pte_size(struct xe_vma *vma)
{
if (vma->gpuva.flags & XE_VMA_PTE_1G)
return SZ_1G;
- else if (vma->gpuva.flags & XE_VMA_PTE_2M)
+ else if (vma->gpuva.flags & (XE_VMA_PTE_2M | XE_VMA_PTE_COMPACT))
return SZ_2M;
else if (vma->gpuva.flags & XE_VMA_PTE_64K)
return SZ_64K;
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index 91800ce70845..a603cc2eb56b 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -30,6 +30,7 @@ struct xe_vm;
#define XE_VMA_PTE_2M (DRM_GPUVA_USERBITS << 6)
#define XE_VMA_PTE_1G (DRM_GPUVA_USERBITS << 7)
#define XE_VMA_PTE_64K (DRM_GPUVA_USERBITS << 8)
+#define XE_VMA_PTE_COMPACT (DRM_GPUVA_USERBITS << 9)
/** struct xe_userptr - User pointer */
struct xe_userptr {
--
2.34.1
next prev parent reply other threads:[~2024-02-19 21:18 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-19 21:19 [PATCH v2 0/3] Skip remap fixes Matthew Brost
2024-02-19 21:19 ` [PATCH v2 1/3] drm/xe: Fix xe_vma_set_pte_size Matthew Brost
2024-02-20 8:43 ` Thomas Hellström
2024-02-19 21:19 ` [PATCH v2 2/3] drm/xe: Add XE_VMA_PTE_64K VMA flag Matthew Brost
2024-02-20 8:43 ` Thomas Hellström
2024-02-19 21:19 ` Matthew Brost [this message]
2024-02-20 8:44 ` [PATCH v2 3/3] drm/xe: Return 2MB page size for compact 64k PTEs Thomas Hellström
2024-02-19 22:10 ` ✓ CI.Patch_applied: success for Skip remap fixes (rev2) Patchwork
2024-02-19 22:11 ` ✗ CI.checkpatch: warning " Patchwork
2024-02-19 22:11 ` ✓ CI.KUnit: success " Patchwork
2024-02-19 22:22 ` ✓ CI.Build: " Patchwork
2024-02-19 22:23 ` ✓ CI.Hooks: " Patchwork
2024-02-19 22:25 ` ✓ CI.checksparse: " Patchwork
2024-02-19 22:45 ` ✓ CI.BAT: " Patchwork
2024-02-20 8:42 ` [PATCH v2 0/3] Skip remap fixes Thomas Hellström
2024-02-20 19:51 ` Zanoni, Paulo R
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240219211942.3633795-4-matthew.brost@intel.com \
--to=matthew.brost@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=paulo.r.zanoni@intel.com \
--cc=thomas.hellstrom@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox