From: Dave Airlie <airlied@gmail.com>
To: dri-devel@lists.freedesktop.org
Cc: nouveau@lists.freedesktop.org
Subject: [PATCH 2/3] nouveau/vmm: increase size of vmm pte tracker struct to u32 (v2)
Date: Wed, 4 Feb 2026 13:00:06 +1000 [thread overview]
Message-ID: <20260204030208.2313241-3-airlied@gmail.com> (raw)
In-Reply-To: <20260204030208.2313241-1-airlied@gmail.com>
From: Dave Airlie <airlied@redhat.com>
We need to tracker large counts of spte than previously due to unref
getting delayed sometimes.
This doesn't fix LPT tracking yet, it just creates space for it.
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
v2: fix memset32 wrong length
---
drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c | 6 +++---
drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h | 9 +++++----
2 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c
index efc334f6104c..44daeec0aa6d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c
@@ -387,7 +387,7 @@ nvkm_vmm_sparse_ptes(const struct nvkm_vmm_desc *desc,
} else
if (desc->type == LPT) {
union nvkm_pte_tracker sparse = { .s.sparse = 1 };
- memset(&pgt->pte[ptei].u, sparse.u, ptes);
+ memset32(&pgt->pte[ptei].u, sparse.u, ptes);
}
}
@@ -399,7 +399,7 @@ nvkm_vmm_sparse_unref_ptes(struct nvkm_vmm_iter *it, bool pfn, u32 ptei, u32 pte
memset(&pt->pde[ptei], 0x00, sizeof(pt->pde[0]) * ptes);
else
if (it->desc->type == LPT)
- memset(&pt->pte[ptei].u, 0x00, sizeof(pt->pte[0]) * ptes);
+ memset32(&pt->pte[ptei].u, 0x00, ptes);
return nvkm_vmm_unref_ptes(it, pfn, ptei, ptes);
}
@@ -458,7 +458,7 @@ nvkm_vmm_ref_hwpt(struct nvkm_vmm_iter *it, struct nvkm_vmm_pt *pgd, u32 pdei)
desc->func->sparse(vmm, pt, pteb, ptes);
else
desc->func->invalid(vmm, pt, pteb, ptes);
- memset(&pgt->pte[pteb], 0x00, ptes);
+ memset32(&pgt->pte[pteb].u, 0x00, ptes);
} else {
desc->func->unmap(vmm, pt, pteb, ptes);
while (ptes--)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h
index a6312a0e6b84..a8b08126e8dc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h
@@ -5,11 +5,12 @@
enum nvkm_memory_target;
union nvkm_pte_tracker {
- u8 u;
+ u32 u;
struct {
- u8 sparse:1;
- u8 spte_valid:1;
- u8 sptes:6;
+ u32 sparse:1;
+ u32 spte_valid:1;
+ u32 padding:14;
+ u32 sptes:16;
} s;
};
--
2.52.0
next prev parent reply other threads:[~2026-02-04 3:02 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-02-04 3:00 [PATCH 0/3] nouveau/vmm: fix switching between small and large PTEs (series v2) Dave Airlie
2026-02-04 3:00 ` [PATCH 1/3] nouveau/vmm: rewrite pte tracker using a struct and bitfields Dave Airlie
2026-02-04 3:00 ` Dave Airlie [this message]
2026-02-04 3:00 ` [PATCH 3/3] nouveau/vmm: start tracking if the LPT PTE is valid. (v6) Dave Airlie
2026-02-04 12:43 ` [PATCH 0/3] nouveau/vmm: fix switching between small and large PTEs (series v2) Mary Guillemard
2026-02-04 16:40 ` M Henning
2026-02-04 16:40 ` M Henning
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=20260204030208.2313241-3-airlied@gmail.com \
--to=airlied@gmail.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=nouveau@lists.freedesktop.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.