From: Vladimir Murzin <vladimir.murzin@arm.com>
To: linux-arm-kernel@lists.infradead.org
Cc: mark.rutland@arm.com, maz@kernel.org, will@kernel.org,
catalin.marinas@arm.com, ruanjinjie@huawei.com,
liaochang1@huawei.com
Subject: [RFC PATCH v2 39/45] arm64: smp: Fall back to IRQ when IPI NMI request fails
Date: Mon, 27 Jul 2026 17:34:47 +0100 [thread overview]
Message-ID: <20260727163453.7969-40-vladimir.murzin@arm.com> (raw)
In-Reply-To: <20260727163453.7969-1-vladimir.murzin@arm.com>
Existing IPI setup logic relies on ipi_should_be_nmi() to decide
whether to request an NMI (if supported) or an ordinary IRQ. If NMI is
not supported, using an ordinary IRQ is acceptable.
FEAT_NMI requires complementary support in the interrupt
controller. If FEAT_NMI is present but interrupt controller support is
absent, request_percpu_nmi() can fail.
One way to prevent this would be to extend ipi_should_be_nmi() to also
check for interrupt controller support. However, since falling back to
an ordinary IRQ is acceptable, treat ipi_should_be_nmi() as a hint
instead. Whether an IPI is actually registered as an NMI is determined
by the success or failure of request_percpu_nmi(). If the request
fails, fall back to request_irq().
Track which IPIs were actually registered as NMIs and use that state
in the per-CPU enable and disable paths. This avoids calling the
percpu NMI helpers for an IPI that fell back to regular IRQ handling.
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com>
---
arch/arm64/kernel/smp.c | 31 +++++++++++++++++++++++--------
1 file changed, 23 insertions(+), 8 deletions(-)
diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c
index 3dd4bc02caed..9bacdbb2aa44 100644
--- a/arch/arm64/kernel/smp.c
+++ b/arch/arm64/kernel/smp.c
@@ -69,11 +69,13 @@ static int nr_ipi __ro_after_init = NR_IPI;
struct ipi_descs {
struct irq_desc *descs[MAX_IPI];
+ DECLARE_BITMAP(nmi_bitmap, MAX_IPI);
};
static DEFINE_PER_CPU_READ_MOSTLY(struct ipi_descs, pcpu_ipi_desc);
#define get_ipi_desc(__cpu, __ipi) (per_cpu_ptr(&pcpu_ipi_desc, __cpu)->descs[__ipi])
+#define get_ipi_nmi_bitmap(__cpu) (per_cpu_ptr(&pcpu_ipi_desc, __cpu)->nmi_bitmap)
struct ipi_irq_ops {
void (*setup)(int ipi, int ncpus);
@@ -1051,31 +1053,44 @@ static bool ipi_should_be_nmi(enum ipi_msg_type ipi)
static void ipi_sgi_setup(int ipi, int ncpus)
{
+ bool is_nmi = ipi_should_be_nmi(ipi);
int err, irq, cpu;
if (WARN_ON_ONCE(ncpus))
return;
-
irq = ipi_irq_base + ipi;
- if (ipi_should_be_nmi(ipi)) {
- err = request_percpu_nmi(irq, ipi_handler, "IPI", NULL, &irq_stat);
- WARN(err, "Could not request IRQ %d as NMI, err=%d\n", irq, err);
- } else {
+ /*
+ * Try to request NMI if it should be
+ */
+ if (is_nmi && request_percpu_nmi(irq, ipi_handler, "IPI", NULL, &irq_stat)) {
+ /*
+ * Fallback to regular IRQ
+ */
+ is_nmi = false;
+ }
+
+ /*
+ * Otherwise regular IRQ
+ */
+ if (!is_nmi) {
err = request_percpu_irq(irq, ipi_handler, "IPI", &irq_stat);
WARN(err, "Could not request IRQ %d as IRQ, err=%d\n", irq, err);
}
- for_each_possible_cpu(cpu)
+ for_each_possible_cpu(cpu) {
get_ipi_desc(cpu, ipi) = irq_to_desc(irq);
+ if (is_nmi)
+ set_bit(ipi, get_ipi_nmi_bitmap(cpu));
+ }
irq_set_status_flags(irq, IRQ_HIDDEN);
}
static void ipi_sgi_enable(int cpu, int ipi)
{
- if (ipi_should_be_nmi(ipi)) {
+ if (test_bit(ipi, get_ipi_nmi_bitmap(cpu))) {
prepare_percpu_nmi(ipi_irq_base + ipi);
enable_percpu_nmi(ipi_irq_base + ipi, 0);
} else {
@@ -1085,7 +1100,7 @@ static void ipi_sgi_enable(int cpu, int ipi)
static void ipi_sgi_disable(int cpu, int ipi)
{
- if (ipi_should_be_nmi(ipi)) {
+ if (test_bit(ipi, get_ipi_nmi_bitmap(cpu))) {
disable_percpu_nmi(ipi_irq_base + ipi);
teardown_percpu_nmi(ipi_irq_base + ipi);
} else {
--
2.34.1
next prev parent reply other threads:[~2026-07-27 17:21 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-27 16:34 [RFC PATCH v2 00/45] arm64: Add support for FEAT_NMI Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 01/45] arm64: ptrace: Remove INIT_PSTATE_EL2 Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 02/45] arm64: debug: don't mask DAIF for mdscr_write() Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 03/45] arm64: hibernate: mask DAIF before restoring hibernated kernel Vladimir Murzin
2026-07-28 1:17 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 04/45] arm64: hibernate: Restore DAIF state on error Vladimir Murzin
2026-07-28 1:16 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 05/45] arm64: suspend: rely on daif helpers to handle PMR Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 06/45] arm64: suspend: Initialize PMR on resume Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 07/45] arm64: entry: mask DAIF before returning from C EL1 handlers Vladimir Murzin
2026-07-28 2:00 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 08/45] irqchip/gic-v3: make the unmasking of pseudo-NMIs explicit when handling IRQs Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 09/45] arm64: entry: Avoid unnecessary local_irq_disable() on kernel exit Vladimir Murzin
2026-07-28 3:18 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 10/45] arm64: irqflags: Introduce arm64-specific irqflags type Vladimir Murzin
2026-07-28 2:42 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 11/45] arm64: irqflags: save and use both DAIF and PMR Vladimir Murzin
2026-07-28 3:46 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 12/45] arm64: interrupts: Add common exception state helpers Vladimir Murzin
2026-07-28 8:20 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 13/45] arm64: process: Use helper to check exception state Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 14/45] arm64: entry: Introduce entry specific exception masking helpers Vladimir Murzin
2026-07-28 8:48 ` Jinjie Ruan
2026-07-28 9:18 ` Jinjie Ruan
2026-07-28 9:29 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 15/45] arm64: entry: replace DAIF helpers with entry helpers Vladimir Murzin
2026-07-28 9:53 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 16/45] arm64: interrupts: Introduce exception masking save/restore helpers Vladimir Murzin
2026-07-28 11:56 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 17/45] arm64: interrupts: introduce a helper for GIC priority initialization Vladimir Murzin
2026-07-28 11:21 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 18/45] arm64: replace local_daif helpers Vladimir Murzin
2026-07-28 12:04 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 19/45] arm64: cpuidle: use new helpers to bypass interrupt priority masking Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 20/45] arm64: remove daifflags.h Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 21/45] arm64: gicv3: remove GIC_PRIO_PSR_I_SET Vladimir Murzin
2026-07-28 12:08 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 22/45] arm64: cpufeature: Remove system_has_prio_mask_debugging() Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 23/45] arm64: irqflags: Switch to CONFIG_DEBUG_IRQFLAGS Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 24/45] arm64: Kconfig: Remove CONFIG_ARM64_DEBUG_PRIORITY_MASKING Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 25/45] efi/runtime-wrappers: Permit architectures to override IRQ flags checks Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 26/45] arm64/efi: Implement override for " Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 27/45] arm64: booting: Document boot requirements for FEAT_NMI Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 28/45] arm64: sysreg: Add definitions for immediate versions of MSR ALLINT Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 29/45] arm64: ptrace: Add PSR_ALLINT_BIT Vladimir Murzin
2026-07-28 3:58 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 30/45] arm64: idreg: Add an override for FEAT_NMI Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 31/45] arm64: cpufeature: Detect PE support " Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 32/45] arm64: nmi: Manage masking for superpriority interrupts Vladimir Murzin
2026-07-28 12:14 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 33/45] arm64: irq: Report FEAT_NMI masking local IRQs Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 34/45] arm64: nmi: Add handling of superpriority interrupts as NMIs Vladimir Murzin
2026-07-28 12:25 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 35/45] arm64: suspend: Always initialise PSTATE.ALLINT Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 36/45] arm64/efi: Add ALLINT to IRQ flags checks Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 37/45] arm64: kprobes: Disable NMIs Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 38/45] arm64: smp: Abstract SGI and LPI operations Vladimir Murzin
2026-07-27 16:34 ` Vladimir Murzin [this message]
2026-07-27 16:34 ` [RFC PATCH v2 40/45] arm64: nmi: Add Kconfig for NMI Vladimir Murzin
2026-07-28 6:24 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 41/45] irqchip/gic-v3: Prepare for FEAT_GICv3_NMI support Vladimir Murzin
2026-07-28 12:29 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 42/45] irqchip/gic-v3: Implement " Vladimir Murzin
2026-07-28 12:31 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 43/45] arm64: smp: Add NMI support for LPI-backed IPIs Vladimir Murzin
2026-07-28 12:34 ` Jinjie Ruan
2026-07-27 16:34 ` [RFC PATCH v2 44/45] irqchip/gic-v5: Add NMI support for PPIs, SPIs and LPIs Vladimir Murzin
2026-07-27 16:34 ` [RFC PATCH v2 45/45] irqchip/gic-v5: Add NMI support for IPIs Vladimir Murzin
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=20260727163453.7969-40-vladimir.murzin@arm.com \
--to=vladimir.murzin@arm.com \
--cc=catalin.marinas@arm.com \
--cc=liaochang1@huawei.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=mark.rutland@arm.com \
--cc=maz@kernel.org \
--cc=ruanjinjie@huawei.com \
--cc=will@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