From: Radu Rendec <radu@rendec.net>
To: Marc Zyngier <maz@kernel.org>,
linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org
Cc: Thomas Gleixner <tglx@kernel.org>,
Sascha Bischoff <sascha.bischoff@arm.com>
Subject: Re: [PATCH] irqchip/gic: Simplify Maintenance Interrupt ACPI parsing
Date: Sat, 01 Aug 2026 14:22:50 -0400 [thread overview]
Message-ID: <d04a7313b44a3d671024c2d15ad25acbce2ebc2e.camel@rendec.net> (raw)
In-Reply-To: <20260725155941.114322-1-maz@kernel.org>
On Sat, 2026-07-25 at 16:59 +0100, Marc Zyngier wrote:
> The ACPI spec describes the vgic Maintenance Interrupt with attributes
> indicating whether the signalling is Level or Edge. While this
> superficially looks like a sensible thing to do, it doesn't actually
> make much sense.
>
> By construction, this interrupt is Level, and cannot be anything
> else, as the hypervisor actively needs to interact with the vgic
> for the level to drop. This indicates a state, and not an event.
>
> Get rid of this nonsense, always register the corresponding GSI as
> Level, and let the user know that they run on quality FW should
> the interrupt be advertised as Edge.
>
> Signed-off-by: Marc Zyngier <maz@kernel.org>
> ---
> drivers/irqchip/irq-gic-v3.c | 10 +++-------
> drivers/irqchip/irq-gic.c | 8 ++++----
> 2 files changed, 7 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/irqchip/irq-gic-v3.c b/drivers/irqchip/irq-gic-v3.c
> index 99444a1b2ffa2..22ecd28a491a8 100644
> --- a/drivers/irqchip/irq-gic-v3.c
> +++ b/drivers/irqchip/irq-gic-v3.c
> @@ -2279,7 +2279,6 @@ static struct
> bool single_redist;
> int enabled_rdists;
> u32 maint_irq;
> - int maint_irq_mode;
> phys_addr_t vcpu_base;
> } acpi_data __initdata;
>
> @@ -2457,21 +2456,19 @@ static int __init gic_acpi_parse_virt_madt_gicc(union acpi_subtable_headers *hea
> {
> struct acpi_madt_generic_interrupt *gicc =
> (struct acpi_madt_generic_interrupt *)header;
> - int maint_irq_mode;
> static int first_madt = true;
>
> if (!(gicc->flags &
> (ACPI_MADT_ENABLED | ACPI_MADT_GICC_ONLINE_CAPABLE)))
> return 0;
>
> - maint_irq_mode = (gicc->flags & ACPI_MADT_VGIC_IRQ_MODE) ?
> - ACPI_EDGE_SENSITIVE : ACPI_LEVEL_SENSITIVE;
> + if (gicc->flags & ACPI_MADT_VGIC_IRQ_MODE)
> + pr_warn_once(FW_BUG "MI wrongly advertised as Edge-triggered\n");
>
> if (first_madt) {
> first_madt = false;
>
> acpi_data.maint_irq = gicc->vgic_interrupt;
> - acpi_data.maint_irq_mode = maint_irq_mode;
> acpi_data.vcpu_base = gicc->gicv_base_address;
>
> return 0;
> @@ -2481,7 +2478,6 @@ static int __init gic_acpi_parse_virt_madt_gicc(union acpi_subtable_headers *hea
> * The maintenance interrupt and GICV should be the same for every CPU
> */
> if ((acpi_data.maint_irq != gicc->vgic_interrupt) ||
> - (acpi_data.maint_irq_mode != maint_irq_mode) ||
> (acpi_data.vcpu_base != gicc->gicv_base_address))
> return -EINVAL;
>
> @@ -2514,7 +2510,7 @@ static void __init gic_acpi_setup_kvm_info(void)
> gic_v3_kvm_info.type = GIC_V3;
>
> irq = acpi_register_gsi(NULL, acpi_data.maint_irq,
> - acpi_data.maint_irq_mode,
> + ACPI_LEVEL_SENSITIVE,
> ACPI_ACTIVE_HIGH);
> if (irq <= 0)
> return;
> diff --git a/drivers/irqchip/irq-gic.c b/drivers/irqchip/irq-gic.c
> index ec70c84e9f91d..74bda5e246345 100644
> --- a/drivers/irqchip/irq-gic.c
> +++ b/drivers/irqchip/irq-gic.c
> @@ -1527,7 +1527,6 @@ static struct
> {
> phys_addr_t cpu_phys_base;
> u32 maint_irq;
> - int maint_irq_mode;
> phys_addr_t vctrl_base;
> phys_addr_t vcpu_base;
> } acpi_data __initdata;
> @@ -1553,10 +1552,11 @@ gic_acpi_parse_madt_cpu(union acpi_subtable_headers *header,
> if (cpu_base_assigned && gic_cpu_base != acpi_data.cpu_phys_base)
> return -EINVAL;
>
> + if (processor->flags & ACPI_MADT_VGIC_IRQ_MODE)
> + pr_warn_once(FW_BUG "MI wrongly advertised as Edge-triggered\n");
> +
> acpi_data.cpu_phys_base = gic_cpu_base;
> acpi_data.maint_irq = processor->vgic_interrupt;
> - acpi_data.maint_irq_mode = (processor->flags & ACPI_MADT_VGIC_IRQ_MODE) ?
> - ACPI_EDGE_SENSITIVE : ACPI_LEVEL_SENSITIVE;
> acpi_data.vctrl_base = processor->gich_base_address;
> acpi_data.vcpu_base = processor->gicv_base_address;
>
> @@ -1616,7 +1616,7 @@ static void __init gic_acpi_setup_kvm_info(void)
> vcpu_res->end = vcpu_res->start + ACPI_GICV2_VCPU_MEM_SIZE - 1;
>
> irq = acpi_register_gsi(NULL, acpi_data.maint_irq,
> - acpi_data.maint_irq_mode,
> + ACPI_LEVEL_SENSITIVE,
> ACPI_ACTIVE_HIGH);
> if (irq <= 0)
> return;
Reviewed-by: Radu Rendec <radu@rendec.net>
prev parent reply other threads:[~2026-08-01 18:23 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-25 15:59 Marc Zyngier
2026-07-27 10:50 ` Zenghui Yu
2026-08-01 18:22 ` Radu Rendec [this message]
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=d04a7313b44a3d671024c2d15ad25acbce2ebc2e.camel@rendec.net \
--to=radu@rendec.net \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=maz@kernel.org \
--cc=sascha.bischoff@arm.com \
--cc=tglx@kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®