Merge tag 'irq-msi-2024-03-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull MSI updates from Thomas Gleixner:
 "Updates for the MSI interrupt subsystem and initial RISC-V MSI
  support.

  The core changes have been adopted from previous work which converted
  ARM[64] to the new per device MSI domain model, which was merged to
  support multiple MSI domain per device. The ARM[64] changes are being
  worked on too, but have not been ready yet. The core and platform-MSI
  changes have been split out to not hold up RISC-V and to avoid that
  RISC-V builds on the scheduled for removal interfaces.

  The core support provides new interfaces to handle wire to MSI bridges
  in a straight forward way and introduces new platform-MSI interfaces
  which are built on top of the per device MSI domain model.

  Once ARM[64] is converted over the old platform-MSI interfaces and the
  related ugliness in the MSI core code will be removed.

  The actual MSI parts for RISC-V were finalized late and have been
  post-poned for the next merge window.

  Drivers:

   - Add a new driver for the Andes hart-level interrupt controller

   - Rework the SiFive PLIC driver to prepare for MSI suport

   - Expand the RISC-V INTC driver to support the new RISC-V AIA
     controller which provides the basis for MSI on RISC-V

   - A few fixup for the fallout of the core changes"

* tag 'irq-msi-2024-03-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (29 commits)
  irqchip/riscv-intc: Fix low-level interrupt handler setup for AIA
  x86/apic/msi: Use DOMAIN_BUS_GENERIC_MSI for HPET/IO-APIC domain search
  genirq/matrix: Dynamic bitmap allocation
  irqchip/riscv-intc: Add support for RISC-V AIA
  irqchip/sifive-plic: Improve locking safety by using irqsave/irqrestore
  irqchip/sifive-plic: Parse number of interrupts and contexts early in plic_probe()
  irqchip/sifive-plic: Cleanup PLIC contexts upon irqdomain creation failure
  irqchip/sifive-plic: Use riscv_get_intc_hwnode() to get parent fwnode
  irqchip/sifive-plic: Use devm_xyz() for managed allocation
  irqchip/sifive-plic: Use dev_xyz() in-place of pr_xyz()
  irqchip/sifive-plic: Convert PLIC driver into a platform driver
  irqchip/riscv-intc: Introduce Andes hart-level interrupt controller
  irqchip/riscv-intc: Allow large non-standard interrupt number
  genirq/irqdomain: Don't call ops->select for DOMAIN_BUS_ANY tokens
  irqchip/imx-intmux: Handle pure domain searches correctly
  genirq/msi: Provide MSI_FLAG_PARENT_PM_DEV
  genirq/irqdomain: Reroute device MSI create_mapping
  genirq/msi: Provide allocation/free functions for "wired" MSI interrupts
  genirq/msi: Optionally use dev->fwnode for device domain
  genirq/msi: Provide DOMAIN_BUS_WIRED_TO_MSI
  ...

1 files changed, 17 insertions, 11 deletions

diff --git a/kernel/irq/matrix.c b/kernel/irq/matrix.c
index 75d0ae490e29..8f222d1cccec 100644
--- a/kernel/irq/matrix.c
+++ b/kernel/irq/matrix.c
@@ -8,8 +8,6 @@
 #include <linux/cpu.h>
 #include <linux/irq.h>
 
-#define IRQ_MATRIX_SIZE	(BITS_TO_LONGS(IRQ_MATRIX_BITS))
-
 struct cpumap {
 	unsigned int		available;
 	unsigned int		allocated;
@@ -17,8 +15,8 @@ struct cpumap {
 	unsigned int		managed_allocated;
 	bool			initialized;
 	bool			online;
-	unsigned long		alloc_map[IRQ_MATRIX_SIZE];
-	unsigned long		managed_map[IRQ_MATRIX_SIZE];
+	unsigned long		*managed_map;
+	unsigned long		alloc_map[];
 };
 
 struct irq_matrix {
@@ -32,8 +30,8 @@ struct irq_matrix {
 	unsigned int		total_allocated;
 	unsigned int		online_maps;
 	struct cpumap __percpu	*maps;
-	unsigned long		scratch_map[IRQ_MATRIX_SIZE];
-	unsigned long		system_map[IRQ_MATRIX_SIZE];
+	unsigned long		*system_map;
+	unsigned long		scratch_map[];
 };
 
 #define CREATE_TRACE_POINTS
@@ -50,24 +48,32 @@ __init struct irq_matrix *irq_alloc_matrix(unsigned int matrix_bits,
 					   unsigned int alloc_start,
 					   unsigned int alloc_end)
 {
+	unsigned int cpu, matrix_size = BITS_TO_LONGS(matrix_bits);
 	struct irq_matrix *m;
 
-	if (matrix_bits > IRQ_MATRIX_BITS)
-		return NULL;
-
-	m = kzalloc(sizeof(*m), GFP_KERNEL);
+	m = kzalloc(struct_size(m, scratch_map, matrix_size * 2), GFP_KERNEL);
 	if (!m)
 		return NULL;
 
+	m->system_map = &m->scratch_map[matrix_size];
+
 	m->matrix_bits = matrix_bits;
 	m->alloc_start = alloc_start;
 	m->alloc_end = alloc_end;
 	m->alloc_size = alloc_end - alloc_start;
-	m->maps = alloc_percpu(*m->maps);
+	m->maps = __alloc_percpu(struct_size(m->maps, alloc_map, matrix_size * 2),
+				 __alignof__(*m->maps));
 	if (!m->maps) {
 		kfree(m);
 		return NULL;
 	}
+
+	for_each_possible_cpu(cpu) {
+		struct cpumap *cm = per_cpu_ptr(m->maps, cpu);
+
+		cm->managed_map = &cm->alloc_map[matrix_size];
+	}
+
 	return m;
 }