1irq_domain interrupt number mapping library 2 3The current design of the Linux kernel uses a single large number 4space where each separate IRQ source is assigned a different number. 5This is simple when there is only one interrupt controller, but in 6systems with multiple interrupt controllers the kernel must ensure 7that each one gets assigned non-overlapping allocations of Linux 8IRQ numbers. 9 10The number of interrupt controllers registered as unique irqchips 11show a rising tendency: for example subdrivers of different kinds 12such as GPIO controllers avoid reimplementing identical callback 13mechanisms as the IRQ core system by modelling their interrupt 14handlers as irqchips, i.e. in effect cascading interrupt controllers. 15 16Here the interrupt number loose all kind of correspondence to 17hardware interrupt numbers: whereas in the past, IRQ numbers could 18be chosen so they matched the hardware IRQ line into the root 19interrupt controller (i.e. the component actually fireing the 20interrupt line to the CPU) nowadays this number is just a number. 21 22For this reason we need a mechanism to separate controller-local 23interrupt numbers, called hardware irq's, from Linux IRQ numbers. 24 25The irq_alloc_desc*() and irq_free_desc*() APIs provide allocation of 26irq numbers, but they don't provide any support for reverse mapping of 27the controller-local IRQ (hwirq) number into the Linux IRQ number 28space. 29 30The irq_domain library adds mapping between hwirq and IRQ numbers on 31top of the irq_alloc_desc*() API. An irq_domain to manage mapping is 32preferred over interrupt controller drivers open coding their own 33reverse mapping scheme. 34 35irq_domain also implements translation from Device Tree interrupt 36specifiers to hwirq numbers, and can be easily extended to support 37other IRQ topology data sources. 38 39=== irq_domain usage === 40An interrupt controller driver creates and registers an irq_domain by 41calling one of the irq_domain_add_*() functions (each mapping method 42has a different allocator function, more on that later). The function 43will return a pointer to the irq_domain on success. The caller must 44provide the allocator function with an irq_domain_ops structure. 45 46In most cases, the irq_domain will begin empty without any mappings 47between hwirq and IRQ numbers. Mappings are added to the irq_domain 48by calling irq_create_mapping() which accepts the irq_domain and a 49hwirq number as arguments. If a mapping for the hwirq doesn't already 50exist then it will allocate a new Linux irq_desc, associate it with 51the hwirq, and call the .map() callback so the driver can perform any 52required hardware setup. 53 54When an interrupt is received, irq_find_mapping() function should 55be used to find the Linux IRQ number from the hwirq number. 56 57The irq_create_mapping() function must be called *atleast once* 58before any call to irq_find_mapping(), lest the descriptor will not 59be allocated. 60 61If the driver has the Linux IRQ number or the irq_data pointer, and 62needs to know the associated hwirq number (such as in the irq_chip 63callbacks) then it can be directly obtained from irq_data->hwirq. 64 65=== Types of irq_domain mappings === 66There are several mechanisms available for reverse mapping from hwirq 67to Linux irq, and each mechanism uses a different allocation function. 68Which reverse map type should be used depends on the use case. Each 69of the reverse map types are described below: 70 71==== Linear ==== 72irq_domain_add_linear() 73 74The linear reverse map maintains a fixed size table indexed by the 75hwirq number. When a hwirq is mapped, an irq_desc is allocated for 76the hwirq, and the IRQ number is stored in the table. 77 78The Linear map is a good choice when the maximum number of hwirqs is 79fixed and a relatively small number (~ < 256). The advantages of this 80map are fixed time lookup for IRQ numbers, and irq_descs are only 81allocated for in-use IRQs. The disadvantage is that the table must be 82as large as the largest possible hwirq number. 83 84The majority of drivers should use the linear map. 85 86==== Tree ==== 87irq_domain_add_tree() 88 89The irq_domain maintains a radix tree map from hwirq numbers to Linux 90IRQs. When an hwirq is mapped, an irq_desc is allocated and the 91hwirq is used as the lookup key for the radix tree. 92 93The tree map is a good choice if the hwirq number can be very large 94since it doesn't need to allocate a table as large as the largest 95hwirq number. The disadvantage is that hwirq to IRQ number lookup is 96dependent on how many entries are in the table. 97 98Very few drivers should need this mapping. 99 100==== No Map ===- 101irq_domain_add_nomap() 102 103The No Map mapping is to be used when the hwirq number is 104programmable in the hardware. In this case it is best to program the 105Linux IRQ number into the hardware itself so that no mapping is 106required. Calling irq_create_direct_mapping() will allocate a Linux 107IRQ number and call the .map() callback so that driver can program the 108Linux IRQ number into the hardware. 109 110Most drivers cannot use this mapping. 111 112==== Legacy ==== 113irq_domain_add_simple() 114irq_domain_add_legacy() 115irq_domain_add_legacy_isa() 116 117The Legacy mapping is a special case for drivers that already have a 118range of irq_descs allocated for the hwirqs. It is used when the 119driver cannot be immediately converted to use the linear mapping. For 120example, many embedded system board support files use a set of #defines 121for IRQ numbers that are passed to struct device registrations. In that 122case the Linux IRQ numbers cannot be dynamically assigned and the legacy 123mapping should be used. 124 125The legacy map assumes a contiguous range of IRQ numbers has already 126been allocated for the controller and that the IRQ number can be 127calculated by adding a fixed offset to the hwirq number, and 128visa-versa. The disadvantage is that it requires the interrupt 129controller to manage IRQ allocations and it requires an irq_desc to be 130allocated for every hwirq, even if it is unused. 131 132The legacy map should only be used if fixed IRQ mappings must be 133supported. For example, ISA controllers would use the legacy map for 134mapping Linux IRQs 0-15 so that existing ISA drivers get the correct IRQ 135numbers. 136 137Most users of legacy mappings should use irq_domain_add_simple() which 138will use a legacy domain only if an IRQ range is supplied by the 139system and will otherwise use a linear domain mapping. The semantics 140of this call are such that if an IRQ range is specified then 141descriptors will be allocated on-the-fly for it, and if no range is 142specified it will fall through to irq_domain_add_linear() which means 143*no* irq descriptors will be allocated. 144 145A typical use case for simple domains is where an irqchip provider 146is supporting both dynamic and static IRQ assignments. 147 148In order to avoid ending up in a situation where a linear domain is 149used and no descriptor gets allocated it is very important to make sure 150that the driver using the simple domain call irq_create_mapping() 151before any irq_find_mapping() since the latter will actually work 152for the static IRQ assignment case. 153 154==== Hierarchy IRQ domain ==== 155On some architectures, there may be multiple interrupt controllers 156involved in delivering an interrupt from the device to the target CPU. 157Let's look at a typical interrupt delivering path on x86 platforms: 158 159Device --> IOAPIC -> Interrupt remapping Controller -> Local APIC -> CPU 160 161There are three interrupt controllers involved: 1621) IOAPIC controller 1632) Interrupt remapping controller 1643) Local APIC controller 165 166To support such a hardware topology and make software architecture match 167hardware architecture, an irq_domain data structure is built for each 168interrupt controller and those irq_domains are organized into hierarchy. 169When building irq_domain hierarchy, the irq_domain near to the device is 170child and the irq_domain near to CPU is parent. So a hierarchy structure 171as below will be built for the example above. 172 CPU Vector irq_domain (root irq_domain to manage CPU vectors) 173 ^ 174 | 175 Interrupt Remapping irq_domain (manage irq_remapping entries) 176 ^ 177 | 178 IOAPIC irq_domain (manage IOAPIC delivery entries/pins) 179 180There are four major interfaces to use hierarchy irq_domain: 1811) irq_domain_alloc_irqs(): allocate IRQ descriptors and interrupt 182 controller related resources to deliver these interrupts. 1832) irq_domain_free_irqs(): free IRQ descriptors and interrupt controller 184 related resources associated with these interrupts. 1853) irq_domain_activate_irq(): activate interrupt controller hardware to 186 deliver the interrupt. 1873) irq_domain_deactivate_irq(): deactivate interrupt controller hardware 188 to stop delivering the interrupt. 189 190Following changes are needed to support hierarchy irq_domain. 1911) a new field 'parent' is added to struct irq_domain; it's used to 192 maintain irq_domain hierarchy information. 1932) a new field 'parent_data' is added to struct irq_data; it's used to 194 build hierarchy irq_data to match hierarchy irq_domains. The irq_data 195 is used to store irq_domain pointer and hardware irq number. 1963) new callbacks are added to struct irq_domain_ops to support hierarchy 197 irq_domain operations. 198 199With support of hierarchy irq_domain and hierarchy irq_data ready, an 200irq_domain structure is built for each interrupt controller, and an 201irq_data structure is allocated for each irq_domain associated with an 202IRQ. Now we could go one step further to support stacked(hierarchy) 203irq_chip. That is, an irq_chip is associated with each irq_data along 204the hierarchy. A child irq_chip may implement a required action by 205itself or by cooperating with its parent irq_chip. 206 207With stacked irq_chip, interrupt controller driver only needs to deal 208with the hardware managed by itself and may ask for services from its 209parent irq_chip when needed. So we could achieve a much cleaner 210software architecture. 211 212For an interrupt controller driver to support hierarchy irq_domain, it 213needs to: 2141) Implement irq_domain_ops.alloc and irq_domain_ops.free 2152) Optionally implement irq_domain_ops.activate and 216 irq_domain_ops.deactivate. 2173) Optionally implement an irq_chip to manage the interrupt controller 218 hardware. 2194) No need to implement irq_domain_ops.map and irq_domain_ops.unmap, 220 they are unused with hierarchy irq_domain. 221 222Hierarchy irq_domain may also be used to support other architectures, 223such as ARM, ARM64 etc. 224