linux/arch/x86/kernel/apic
Dongli Zhang a6c11c0a52 genirq/cpuhotplug, x86/vector: Prevent vector leak during CPU offline
The absence of IRQD_MOVE_PCNTXT prevents immediate effectiveness of
interrupt affinity reconfiguration via procfs. Instead, the change is
deferred until the next instance of the interrupt being triggered on the
original CPU.

When the interrupt next triggers on the original CPU, the new affinity is
enforced within __irq_move_irq(). A vector is allocated from the new CPU,
but the old vector on the original CPU remains and is not immediately
reclaimed. Instead, apicd->move_in_progress is flagged, and the reclaiming
process is delayed until the next trigger of the interrupt on the new CPU.

Upon the subsequent triggering of the interrupt on the new CPU,
irq_complete_move() adds a task to the old CPU's vector_cleanup list if it
remains online. Subsequently, the timer on the old CPU iterates over its
vector_cleanup list, reclaiming old vectors.

However, a rare scenario arises if the old CPU is outgoing before the
interrupt triggers again on the new CPU.

In that case irq_force_complete_move() is not invoked on the outgoing CPU
to reclaim the old apicd->prev_vector because the interrupt isn't currently
affine to the outgoing CPU, and irq_needs_fixup() returns false. Even
though __vector_schedule_cleanup() is later called on the new CPU, it
doesn't reclaim apicd->prev_vector; instead, it simply resets both
apicd->move_in_progress and apicd->prev_vector to 0.

As a result, the vector remains unreclaimed in vector_matrix, leading to a
CPU vector leak.

To address this issue, move the invocation of irq_force_complete_move()
before the irq_needs_fixup() call to reclaim apicd->prev_vector, if the
interrupt is currently or used to be affine to the outgoing CPU.

Additionally, reclaim the vector in __vector_schedule_cleanup() as well,
following a warning message, although theoretically it should never see
apicd->move_in_progress with apicd->prev_cpu pointing to an offline CPU.

Fixes: f0383c24b4 ("genirq/cpuhotplug: Add support for cleaning up move in progress")
Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20240522220218.162423-1-dongli.zhang@oracle.com
2024-05-23 21:51:50 +02:00
..
apic_common.c x86/apic: Remove yet another dubious callback 2024-02-15 22:07:41 +01:00
apic_flat_64.c x86/apic: Remove yet another dubious callback 2024-02-15 22:07:41 +01:00
apic_noop.c x86/apic: Remove check_apicid_used() and ioapic_phys_id_map() 2024-02-15 22:07:39 +01:00
apic_numachip.c x86/apic: Remove the pointless writeback of boot_cpu_physical_apicid 2024-02-15 22:07:41 +01:00
apic.c Improve data types to fix Coccinelle division warnings 2024-05-14 09:24:14 -07:00
bigsmp_32.c x86/apic: Remove the pointless writeback of boot_cpu_physical_apicid 2024-02-15 22:07:41 +01:00
hw_nmi.c * Rework apic callbacks, getting rid of unnecessary ones and 2023-08-30 10:44:46 -07:00
init.c x86/apic: Provide static call infrastructure for APIC callbacks 2023-08-09 12:00:55 -07:00
io_apic.c Rework of APIC enumeration and topology evaluation: 2024-03-11 15:45:55 -07:00
ipi.c Major microcode loader restructuring, cleanup and improvements by Thomas 2023-11-04 08:46:37 -10:00
local.h x86/apic: Remove yet another dubious callback 2024-02-15 22:07:41 +01:00
Makefile arch/x86: Fix typos 2024-01-03 11:46:22 +01:00
msi.c Revert "x86/apic/msi: Enable PCI/IMS" 2024-05-15 17:02:04 -05:00
probe_32.c x86/apic: Remove yet another dubious callback 2024-02-15 22:07:41 +01:00
probe_64.c x86/apic: Provide common init infrastructure 2023-08-09 11:58:34 -07:00
vector.c genirq/cpuhotplug, x86/vector: Prevent vector leak during CPU offline 2024-05-23 21:51:50 +02:00
x2apic_cluster.c x86/cpu: Improve readability of per-CPU cpumask initialization code 2024-04-10 07:02:33 +02:00
x2apic_phys.c x86/apic: Remove the pointless writeback of boot_cpu_physical_apicid 2024-02-15 22:07:41 +01:00
x2apic_uv_x.c x86/apic: Remove the pointless writeback of boot_cpu_physical_apicid 2024-02-15 22:07:41 +01:00