Linux-ARM-Kernel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Vincent Donnefort <vdonnefort@google.com>
To: maz@kernel.org, oupton@kernel.org, kvmarm@lists.linux.dev,
	 linux-arm-kernel@lists.infradead.org
Cc: joey.gouly@arm.com, seiden@linux.ibm.com, suzuki.poulose@arm.com,
	 yuzenghui@huawei.com, catalin.marinas@arm.com, will@kernel.org,
	 kernel-team@android.com, fuad.tabba@linux.dev,
	qperret@google.com,  weilin.chang@arm.com,
	Vincent Donnefort <vdonnefort@google.com>
Subject: [PATCH v4 18/23] KVM: arm64: Introduce kvm_pgtable_stage2_table_install()
Date: Tue, 17 Nov 2026 17:43:57 +0000	[thread overview]
Message-ID: <20261117174403.58762-19-vdonnefort@google.com> (raw)
In-Reply-To: <20261117174403.58762-1-vdonnefort@google.com>

The existing kvm_pgtable_stage2_split() is unfit for pKVM: it relies on
kvm_mmu_memory_cache which pKVM doesn't use and has no easy way to
handle pKVM-specific annotations.

Add the pgtable helper kvm_pgtable_stage2_table_install() so pKVM can
handle block splitting in mem_protect.c. That helper can take a table
either populated with kvm_pgtable_stage2_create_unlinked() or one filled
with annotations.

Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
---
 arch/arm64/include/asm/kvm_pgtable.h | 13 +++++
 arch/arm64/kvm/hyp/pgtable.c         | 86 ++++++++++++++++++++++++----
 2 files changed, 88 insertions(+), 11 deletions(-)

diff --git a/arch/arm64/include/asm/kvm_pgtable.h b/arch/arm64/include/asm/kvm_pgtable.h
index 1c9c57cd6ae4..5d5d2165dbec 100644
--- a/arch/arm64/include/asm/kvm_pgtable.h
+++ b/arch/arm64/include/asm/kvm_pgtable.h
@@ -835,6 +835,19 @@ int kvm_pgtable_stage2_flush(struct kvm_pgtable *pgt, u64 addr, u64 size);
 int kvm_pgtable_stage2_split(struct kvm_pgtable *pgt, u64 addr, u64 size,
 			     struct kvm_mmu_memory_cache *mc);
 
+/**
+ * kvm_pgtable_stage2_table_install() - Replace a stage-2 leaf entry with a table.
+ * @pgt:	Page-table structure initialised by kvm_pgtable_stage2_init*().
+ * @table:	Unlinked stage-2 paging structure to install, such as one
+ *		created by kvm_pgtable_stage2_create_unlinked().
+ * @addr:	Address identifying the leaf entry to replace.
+ * @level:	Level of the leaf entry to replace.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int kvm_pgtable_stage2_table_install(struct kvm_pgtable *pgt, kvm_pte_t *table,
+				     u64 addr, s8 level);
+
 /**
  * kvm_pgtable_walk() - Walk a page-table.
  * @pgt:	Page-table structure initialised by kvm_pgtable_*_init().
diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
index abd32e94f6ff..1ad4f854c442 100644
--- a/arch/arm64/kvm/hyp/pgtable.c
+++ b/arch/arm64/kvm/hyp/pgtable.c
@@ -1537,18 +1537,38 @@ static int stage2_block_get_nr_page_tables(s8 level)
 	};
 }
 
+static int stage2_table_install(const struct kvm_pgtable_visit_ctx *ctx,
+				struct kvm_s2_mmu *mmu,
+				kvm_pte_t *table)
+{
+	kvm_pte_t new;
+
+	if (!stage2_try_break_pte(ctx, mmu))
+		return -EAGAIN;
+
+	/*
+	 * Note, the contents of the page table are guaranteed to be made
+	 * visible before the new PTE is assigned because stage2_make_pte()
+	 * writes the PTE using smp_store_release().
+	 */
+	new = kvm_init_table_pte(table, ctx->mm_ops);
+	stage2_make_pte(ctx, new);
+	return 0;
+}
+
 static int stage2_split_walker(const struct kvm_pgtable_visit_ctx *ctx,
 			       enum kvm_pgtable_walk_flags visit)
 {
 	struct kvm_pgtable_mm_ops *mm_ops = ctx->mm_ops;
 	struct kvm_mmu_memory_cache *mc = ctx->arg;
-	struct kvm_s2_mmu *mmu;
-	kvm_pte_t pte = ctx->old, new, *childp;
+	kvm_pte_t pte = ctx->old, *childp;
 	enum kvm_pgtable_prot prot;
+	struct kvm_s2_mmu *mmu;
 	s8 level = ctx->level;
 	bool force_pte;
 	int nr_pages;
 	u64 phys;
+	int ret;
 
 	/* No huge-pages exist at the last level */
 	if (level == KVM_PGTABLE_LAST_LEVEL)
@@ -1590,18 +1610,12 @@ static int stage2_split_walker(const struct kvm_pgtable_visit_ctx *ctx,
 	if (IS_ERR(childp))
 		return PTR_ERR(childp);
 
-	if (!stage2_try_break_pte(ctx, mmu)) {
+	ret = stage2_table_install(ctx, mmu, childp);
+	if (ret) {
 		kvm_pgtable_stage2_free_unlinked(mm_ops, childp, level);
-		return -EAGAIN;
+		return ret;
 	}
 
-	/*
-	 * Note, the contents of the page table are guaranteed to be made
-	 * visible before the new PTE is assigned because stage2_make_pte()
-	 * writes the PTE using smp_store_release().
-	 */
-	new = kvm_init_table_pte(childp, mm_ops);
-	stage2_make_pte(ctx, new);
 	return 0;
 }
 
@@ -1620,6 +1634,56 @@ int kvm_pgtable_stage2_split(struct kvm_pgtable *pgt, u64 addr, u64 size,
 	return ret;
 }
 
+struct stage2_table_install_data {
+	struct kvm_s2_mmu	*mmu;
+	kvm_pte_t		*table;
+	s8			level;
+};
+
+static int stage2_table_install_walker(const struct kvm_pgtable_visit_ctx *ctx,
+				       enum kvm_pgtable_walk_flags visit)
+{
+	struct stage2_table_install_data *data = ctx->arg;
+
+	if (ctx->level != data->level)
+		return -EINVAL;
+
+	/*
+	 * On success, return a positive value to prevent descending into the
+	 * new table.
+	 */
+	return stage2_table_install(ctx, data->mmu, data->table) ?: 1;
+}
+
+int kvm_pgtable_stage2_table_install(struct kvm_pgtable *pgt, kvm_pte_t *table,
+				     u64 addr, s8 level)
+{
+	struct stage2_table_install_data data = {
+		.mmu	= pgt->mmu,
+		.table	= table,
+		.level	= level,
+	};
+	struct kvm_pgtable_walker walker = {
+		.cb	= stage2_table_install_walker,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+		.arg	= &data,
+	};
+	int ret;
+
+	if (level < pgt->start_level || level >= KVM_PGTABLE_LAST_LEVEL)
+		return -EINVAL;
+
+	if (!IS_ALIGNED(addr, kvm_granule_size(level)))
+		return -EINVAL;
+
+	ret = kvm_pgtable_walk(pgt, addr, kvm_granule_size(level), &walker);
+	if (ret < 0)
+		return ret;
+
+	dsb(ishst);
+	return 0;
+}
+
 int __kvm_pgtable_stage2_init(struct kvm_pgtable *pgt, struct kvm_s2_mmu *mmu,
 			      struct kvm_pgtable_mm_ops *mm_ops,
 			      enum kvm_pgtable_stage2_flags flags,
-- 
2.56.1.708.g77570a6a1d-goog



  parent reply	other threads:[~2026-11-17 17:45 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-17 17:43 [PATCH v4 00/23] Huge mapping support for protected VMs Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 01/23] KVM: arm64: Prefault host stage-2 entries on block split Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 02/23] KVM: arm64: Propagate host stage-2 annotated " Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 03/23] KVM: arm64: Allow block-level stage-2 annotation Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 04/23] KVM: arm64: Use block-level annotations when setting up the host stage-2 Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 05/23] KVM: arm64: Make pKVM ownership selftest an HVC Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 06/23] KVM: arm64: Reject zero-sized ranges in pfn_range_is_valid() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 07/23] KVM: arm64: Add a range to __pkvm_host_share/unshare_hyp() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 08/23] KVM: arm64: Add a range to __pkvm_host_donate_guest() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 09/23] KVM: arm64: Add a range to hyp_poison_page() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 10/23] KVM: arm64: Add a range to __pkvm_host_reclaim_guest() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 11/23] KVM: arm64: Add a range to __pkvm_guest_share_host() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 12/23] KVM: arm64: Add a range to __pkvm_guest_unshare_host() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 13/23] KVM: arm64: Handle huge mappings in __pkvm_host_force_reclaim_page_guest() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 14/23] KVM: arm64: Handle huge mappings in __pkvm_vcpu_in_poison_fault() Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 15/23] KVM: arm64: Add a range to pKVM ownership selftest Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 16/23] KVM: arm64: Warn on pKVM guest stage-2 block collapse Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 17/23] KVM: arm64: Add vcpu conduit for pkvm_hyp_req Vincent Donnefort
2026-11-17 17:43 ` Vincent Donnefort [this message]
2026-11-17 17:43 ` [PATCH v4 19/23] KVM: arm64: Add __pkvm_host_split_guest HVC Vincent Donnefort
2026-11-17 17:43 ` [PATCH v4 20/23] KVM: arm64: Extend pKVM page ownership selftests to cover guest block split Vincent Donnefort
2026-11-17 17:44 ` [PATCH v4 21/23] KVM: arm64: Add PKVM_HYP_REQ_SPLIT Vincent Donnefort
2026-11-17 17:44 ` [PATCH v4 22/23] KVM: arm64: Raise PKVM_HYP_REQ_SPLIT on guest to host sharing Vincent Donnefort
2026-11-17 17:44 ` [PATCH v4 23/23] KVM: arm64: Stage-2 huge mappings for protected VMs Vincent Donnefort

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=20261117174403.58762-19-vdonnefort@google.com \
    --to=vdonnefort@google.com \
    --cc=catalin.marinas@arm.com \
    --cc=fuad.tabba@linux.dev \
    --cc=joey.gouly@arm.com \
    --cc=kernel-team@android.com \
    --cc=kvmarm@lists.linux.dev \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=maz@kernel.org \
    --cc=oupton@kernel.org \
    --cc=qperret@google.com \
    --cc=seiden@linux.ibm.com \
    --cc=suzuki.poulose@arm.com \
    --cc=weilin.chang@arm.com \
    --cc=will@kernel.org \
    --cc=yuzenghui@huawei.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