irq_desc not found for my interrupt number during request_irq (radix tree of irq_desc has only 64 entries..)

Chan Kim ckim at etri.re.kr
Tue Apr 5 22:50:45 EDT 2022


This is the boot message of the virtual machine related to IRQ.
[    0.000000] NR_IRQS: 64, nr_irqs: 64, preallocated irqs: 0
[    0.000000] GICv3: 224 SPIs implemented
[    0.000000] GICv3: 0 Extended SPIs implemented
[    0.000000] GICv3: Distributor has no Range Selector support
[    0.000000] GICv3: 16 PPIs implemented
[    0.000000] GICv3: no VLPI support, no direct LPI support
[    0.000000] GICv3: CPU0: found redistributor 0 region
0:0x00000000080a0000
Somehow the NR_IRQS is 64.
So I tried using SPI 15 which was not assigned. (32 internal IRQs + 15 < 64)
and now request_irq returns 0.
So this irq_desc problem was solved(?) for now. (I don't know why this is so
small when CONFIG_SPARSE_IRQS=y).
And still the handler is not called even when I set qemu_set_irq.
		_____________________________________________
		From: Chan Kim <ckim at etri.re.kr> 
		Sent: Wednesday, April 6, 2022 10:46 AM
		To: 'qemu-discuss' <qemu-discuss at nongnu.org>;
kernelnewbies at kernelnewbies.org
		Subject: irq_desc not found for my interrupt number during
request_irq (radix tree of irq_desc has only 64 entries..)
		
		
		Hello all, 

		I'm doing an linux driver and application test on a qemu
arm64 virtual machine.
		My virtual machine is slightly modified version of arm
'virt' machine and it has our device model included.
		When I do request_irq for INTID 208 (SPI 176) in linux on my
arm64 virtual machine, it returns -EINVAL.
		I poked into that function and found the desc is returning
NULL in function request_threaded_irq.

			int request_threaded_irq(unsigned int irq,
irq_handler_t handler,
			             irq_handler_t thread_fn, unsigned long
irqflags,
			             const char *devname, void *dev_id)
			{
			    struct irqaction *action;
			    struct irq_desc *desc;
			    int retval;
			...
			desc = irq_to_desc(irq);
			    if (!desc) 
			        return -EINVAL;
			...
			}
		
		and irq_to_desc is supposed to return an irq_desc from the
irq_radix_tree as shown below.

			struct irq_desc *irq_to_desc(unsigned int irq)
			{
			    return radix_tree_lookup(&irq_desc_tree, irq);
			}

			void *radix_tree_lookup(const struct radix_tree_root
*root, unsigned long index)
			{
			    return __radix_tree_lookup(root, index, NULL,
NULL);
			}

			void *__radix_tree_lookup(const struct
radix_tree_root *root,
			              unsigned long index, struct
radix_tree_node **nodep,
			              void __rcu ***slotp)
			{
			    struct radix_tree_node *node, *parent;
			    unsigned long maxindex;
			    void __rcu **slot;

			 restart:
			    parent = NULL;
			    slot = (void __rcu **)&root->xa_head;
			    radix_tree_load_root(root, &node, &maxindex);
			    if (index > maxindex)
			        return NULL;
			...
			}

		and I found my request index is 208 but the maxindex of the
radix tree is 63. 
		So it looks like the radix tree of irq_desc has been set to
have maximum 64 irq_descs. 
		Where and how can I set this radix tree size?
		I looked at qemu's hw/arm/ab21q-build-acpi.c for MADT or
DSDT table parts but couldn't find what to add.
		I have my devices address range and irq added in the DSDT
and MADT.

		static void
		build_dsdt(GArray *table_data, BIOSLinker *linker,
Ab21qMachineState *vms)
		{
		...
		acpi_dsdt_add_axpu(scope, &memmap[AB21Q_AXPU],
		                       (irqmap[AB21Q_AXPU] + ARM_SPI_BASE));
		...
		}
		static void acpi_dsdt_add_axpu(Aml *scope, const MemMapEntry
*axpu_memmap,
		                                           uint32_t irq)
		{
		    Aml *dev = aml_device("AXPU");
		    aml_append(dev, aml_name_decl("_HID",
aml_string("AXPU0011")));
		    aml_append(dev, aml_name_decl("_UID", aml_int(0)));

		    Aml *crs = aml_resource_template();
		    aml_append(crs, aml_memory32_fixed(axpu_memmap->base,
		                                       axpu_memmap->size,
AML_READ_WRITE));
		    aml_append(crs,
		               aml_interrupt(AML_CONSUMER, AML_LEVEL,
AML_ACTIVE_HIGH,
		                             AML_EXCLUSIVE, &irq, 1));
		    aml_append(dev, aml_name_decl("_CRS", crs));

		    aml_append(scope, dev);
		}

		I'm not sure if I can just use _HID value as "AXPU0011". 
		If you see anything wrong and have a suggestion, please tell
me.

		Thank you!
		Chan Kim
		
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20220406/71f538df/attachment-0001.html>


More information about the Kernelnewbies mailing list