From: Marc Zyngier <maz@kernel.org>
To: kvmarm@lists.linux.dev, linux-arm-kernel@lists.infradead.org,
kvm@vger.kernel.org
Cc: Joey Gouly <joey.gouly@arm.com>,
Suzuki K Poulose <suzuki.poulose@arm.com>,
Oliver Upton <oupton@kernel.org>,
Zenghui Yu <yuzenghui@huawei.com>,
Christoffer Dall <christoffer.dall@arm.com>,
Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>,
Yao Yuan <yaoyuan@linux.alibaba.com>
Subject: [PATCH v2 32/45] KVM: arm64: GICv3: Avoid broadcast kick on CPUs lacking TDIR
Date: Sun, 9 Nov 2025 17:16:06 +0000 [thread overview]
Message-ID: <20251109171619.1507205-33-maz@kernel.org> (raw)
In-Reply-To: <20251109171619.1507205-1-maz@kernel.org>
CPUs lacking TDIR always trap ICV_DIR_EL1, no matter what, since
we have ICH_HCR_EL2.TC set permanently. For these CPUs, it is
useless to use a broadcast kick on SPI injection, as the sole
purpose of this is to set TDIR.
We can therefore skip this on these CPUs, which are challenged
enough not to be burdened by extra IPIs. As a consequence,
permanently set the TDIR bit in the shadow state to notify the
fast-path emulation code of the exit reason.
Signed-off-by: Marc Zyngier <maz@kernel.org>
---
arch/arm64/kvm/vgic/vgic-v3.c | 6 +++++-
arch/arm64/kvm/vgic/vgic.c | 13 ++++++++-----
2 files changed, 13 insertions(+), 6 deletions(-)
diff --git a/arch/arm64/kvm/vgic/vgic-v3.c b/arch/arm64/kvm/vgic/vgic-v3.c
index bf4dde158d516..32654ea51418b 100644
--- a/arch/arm64/kvm/vgic/vgic-v3.c
+++ b/arch/arm64/kvm/vgic/vgic-v3.c
@@ -50,10 +50,14 @@ void vgic_v3_configure_hcr(struct kvm_vcpu *vcpu,
* need to deal with SPIs that can be deactivated on another
* CPU.
*
+ * On systems that do not implement TDIR, force the bit in the
+ * shadow state anyway to avoid IPI-ing on these poor sods.
+ *
* Note that we set the trap irrespective of EOIMode, as that
* can change behind our back without any warning...
*/
- if (irqs_active_outside_lrs(als) ||
+ if (!cpus_have_final_cap(ARM64_HAS_ICH_HCR_EL2_TDIR) ||
+ irqs_active_outside_lrs(als) ||
atomic_read(&vcpu->kvm->arch.vgic.active_spis))
cpuif->vgic_hcr |= ICH_HCR_EL2_TDIR;
}
diff --git a/arch/arm64/kvm/vgic/vgic.c b/arch/arm64/kvm/vgic/vgic.c
index 60bdddf3472e7..a4c42884f1e16 100644
--- a/arch/arm64/kvm/vgic/vgic.c
+++ b/arch/arm64/kvm/vgic/vgic.c
@@ -370,12 +370,15 @@ static bool vgic_validate_injection(struct vgic_irq *irq, bool level, void *owne
static bool vgic_model_needs_bcst_kick(struct kvm *kvm)
{
/*
- * A GICv3 (or GICv3-like) system exposing a GICv3 to the
- * guest needs a broadcast kick to set TDIR globally, even if
- * the bit doesn't really exist (we still need to check for
- * the shadow bit in the DIR emulation fast-path).
+ * A GICv3 (or GICv3-like) system exposing a GICv3 to the guest
+ * needs a broadcast kick to set TDIR globally.
+ *
+ * For systems that do not have TDIR (ARM's own v8.0 CPUs), the
+ * shadow TDIR bit is always set, and so is the register's TC bit,
+ * so no need to kick the CPUs.
*/
- return (kvm->arch.vgic.vgic_model == KVM_DEV_TYPE_ARM_VGIC_V3);
+ return (cpus_have_final_cap(ARM64_HAS_ICH_HCR_EL2_TDIR) &&
+ kvm->arch.vgic.vgic_model == KVM_DEV_TYPE_ARM_VGIC_V3);
}
/*
--
2.47.3
next prev parent reply other threads:[~2025-11-09 17:17 UTC|newest]
Thread overview: 83+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-09 17:15 [PATCH v2 00/45] KVM: arm64: Add LR overflow infrastructure Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 01/45] irqchip/gic: Add missing GICH_HCR control bits Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 02/45] irqchip/gic: Expose CPU interface VA to KVM Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 03/45] irqchip/apple-aic: Spit out ICH_MISR_EL2 value on spurious vGIC MI Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 04/45] KVM: arm64: Turn vgic-v3 errata traps into a patched-in constant Marc Zyngier
2025-11-10 10:40 ` Suzuki K Poulose
2025-11-10 11:47 ` Marc Zyngier
2025-11-11 23:53 ` Oliver Upton
2025-11-13 9:52 ` Marek Szyprowski
2025-11-13 10:56 ` Marc Zyngier
2025-11-13 11:04 ` Marek Szyprowski
2025-11-13 11:23 ` Joey Gouly
2025-11-13 11:42 ` Marc Zyngier
2025-11-13 10:59 ` Marc Zyngier
2025-11-13 11:20 ` Marek Szyprowski
2025-11-13 18:01 ` Mark Brown
2025-11-14 9:37 ` Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 05/45] KVM: arm64: GICv3: Detect and work around the lack of ICV_DIR_EL1 trapping Marc Zyngier
2025-11-13 14:33 ` Mark Brown
2025-11-13 18:15 ` Marc Zyngier
2025-11-13 19:06 ` Mark Brown
2025-11-13 20:10 ` Marc Zyngier
2025-11-13 21:59 ` Oliver Upton
2025-11-09 17:15 ` [PATCH v2 06/45] KVM: arm64: Repack struct vgic_irq fields Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 07/45] KVM: arm64: Add tracking of vgic_irq being present in a LR Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 08/45] KVM: arm64: Add LR overflow handling documentation Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 09/45] KVM: arm64: GICv3: Drop LPI active state when folding LRs Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 10/45] KVM: arm64: GICv3: Preserve EOIcount on exit Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 11/45] KVM: arm64: GICv3: Decouple ICH_HCR_EL2 programming from LRs Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 12/45] KVM: arm64: GICv3: Extract LR folding primitive Marc Zyngier
2025-11-10 9:01 ` Yao Yuan
2025-11-10 9:18 ` Marc Zyngier
2025-11-10 9:48 ` Yao Yuan
2025-11-09 17:15 ` [PATCH v2 13/45] KVM: arm64: GICv3: Extract LR computing primitive Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 14/45] KVM: arm64: GICv2: Preserve EOIcount on exit Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 15/45] KVM: arm64: GICv2: Decouple GICH_HCR programming from LRs being loaded Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 16/45] KVM: arm64: GICv2: Extract LR folding primitive Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 17/45] KVM: arm64: GICv2: Extract LR computing primitive Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 18/45] KVM: arm64: Compute vgic state irrespective of the number of interrupts Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 19/45] KVM: arm64: Eagerly save VMCR on exit Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 20/45] KVM: arm64: Revamp vgic maintenance interrupt configuration Marc Zyngier
2025-11-12 0:08 ` Oliver Upton
2025-11-12 8:33 ` Marc Zyngier
2025-11-12 8:45 ` Oliver Upton
2025-11-12 9:56 ` Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 21/45] KVM: arm64: Turn kvm_vgic_vcpu_enable() into kvm_vgic_vcpu_reset() Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 22/45] KVM: arm64: Make vgic_target_oracle() globally available Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 23/45] KVM: arm64: Invert ap_list sorting to push active interrupts out Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 24/45] KVM: arm64: Move undeliverable interrupts to the end of ap_list Marc Zyngier
2025-11-09 17:15 ` [PATCH v2 25/45] KVM: arm64: Use MI to detect groups being enabled/disabled Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 26/45] KVM: arm64: GICv3: Handle LR overflow when EOImode==0 Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 27/45] KVM: arm64: GICv3: Handle deactivation via ICV_DIR_EL1 traps Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 28/45] KVM: arm64: GICv3: Add GICv2 SGI handling to deactivation primitive Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 29/45] KVM: arm64: GICv3: Set ICH_HCR_EL2.TDIR when interrupts overflow LR capacity Marc Zyngier
2025-11-14 14:20 ` Fuad Tabba
2025-11-14 15:02 ` Marc Zyngier
2025-11-14 15:53 ` Fuad Tabba
2025-11-14 17:41 ` Marc Zyngier
2025-11-17 8:22 ` Fuad Tabba
2025-11-17 11:56 ` Marc Zyngier
2025-11-24 11:52 ` Mark Brown
2025-11-24 13:06 ` Marc Zyngier
2025-11-24 13:23 ` Mark Brown
2025-11-24 13:40 ` Marc Zyngier
2025-11-24 14:12 ` Marc Zyngier
2025-11-24 15:06 ` Mark Brown
2025-11-09 17:16 ` [PATCH v2 30/45] KVM: arm64: GICv3: Add SPI tracking to handle asymmetric deactivation Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 31/45] KVM: arm64: GICv3: Handle in-LR deactivation when possible Marc Zyngier
2025-11-09 17:16 ` Marc Zyngier [this message]
2025-11-09 17:16 ` [PATCH v2 33/45] KVM: arm64: GICv2: Handle LR overflow when EOImode==0 Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 34/45] KVM: arm64: GICv2: Handle deactivation via GICV_DIR traps Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 35/45] KVM: arm64: GICv2: Always trap GICV_DIR register Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 36/45] KVM: arm64: selftests: gic_v3: Add irq group setting helper Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 37/45] KVM: arm64: selftests: gic_v3: Disable Group-0 interrupts by default Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 38/45] KVM: arm64: selftests: vgic_irq: Fix GUEST_ASSERT_IAR_EMPTY() helper Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 39/45] KVM: arm64: selftests: vgic_irq: Change configuration before enabling interrupt Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 40/45] KVM: arm64: selftests: vgic_irq: Exclude timer-controlled interrupts Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 41/45] KVM: arm64: selftests: vgic_irq: Remove LR-bound limitation Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 42/45] KVM: arm64: selftests: vgic_irq: Perform EOImode==1 deactivation in ack order Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 43/45] KVM: arm64: selftests: vgic_irq: Add asymmetric SPI deaectivation test Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 44/45] KVM: arm64: selftests: vgic_irq: Add Group-0 enable test Marc Zyngier
2025-11-09 17:16 ` [PATCH v2 45/45] KVM: arm64: selftests: vgic_irq: Add timer deactivation test Marc Zyngier
2025-11-12 9:13 ` [PATCH v2 00/45] KVM: arm64: Add LR overflow infrastructure Oliver Upton
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=20251109171619.1507205-33-maz@kernel.org \
--to=maz@kernel.org \
--cc=Volodymyr_Babchuk@epam.com \
--cc=christoffer.dall@arm.com \
--cc=joey.gouly@arm.com \
--cc=kvm@vger.kernel.org \
--cc=kvmarm@lists.linux.dev \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=oupton@kernel.org \
--cc=suzuki.poulose@arm.com \
--cc=yaoyuan@linux.alibaba.com \
--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;
as well as URLs for NNTP newsgroup(s).