1================================ 2POWER9 XIVE interrupt controller 3================================ 4 5The POWER9 processor comes with a new interrupt controller 6architecture, called XIVE as "eXternal Interrupt Virtualization 7Engine". 8 9Compared to the previous architecture, the main characteristics of 10XIVE are to support a larger number of interrupt sources and to 11deliver interrupts directly to virtual processors without hypervisor 12assistance. This removes the context switches required for the 13delivery process. 14 15 16XIVE architecture 17================= 18 19The XIVE IC is composed of three sub-engines, each taking care of a 20processing layer of external interrupts: 21 22- Interrupt Virtualization Source Engine (IVSE), or Source Controller 23 (SC). These are found in PCI PHBs, in the Processor Service 24 Interface (PSI) host bridge Controller, but also inside the main 25 controller for the core IPIs and other sub-chips (NX, CAP, NPU) of 26 the chip/processor. They are configured to feed the IVRE with 27 events. 28- Interrupt Virtualization Routing Engine (IVRE) or Virtualization 29 Controller (VC). It handles event coalescing and perform interrupt 30 routing by matching an event source number with an Event 31 Notification Descriptor (END). 32- Interrupt Virtualization Presentation Engine (IVPE) or Presentation 33 Controller (PC). It maintains the interrupt context state of each 34 thread and handles the delivery of the external interrupt to the 35 thread. 36 37:: 38 39 XIVE Interrupt Controller 40 +------------------------------------+ IPIs 41 | +---------+ +---------+ +--------+ | +-------+ 42 | |IVRE | |Common Q | |IVPE |----> | CORES | 43 | | esb | | | | |----> | | 44 | | eas | | Bridge | | tctx |----> | | 45 | |SC end | | | | nvt | | | | 46 +------+ | +---------+ +----+----+ +--------+ | +-+-+-+-+ 47 | RAM | +------------------|-----------------+ | | | 48 | | | | | | 49 | | | | | | 50 | | +--------------------v------------------------v-v-v--+ other 51 | <--+ Power Bus +--> chips 52 | esb | +---------+-----------------------+------------------+ 53 | eas | | | 54 | end | +--|------+ | 55 | nvt | +----+----+ | +----+----+ 56 +------+ |IVSE | | |IVSE | 57 | | | | | 58 | PQ-bits | | | PQ-bits | 59 | local |-+ | in VC | 60 +---------+ +---------+ 61 PCIe NX,NPU,CAPI 62 63 64 PQ-bits: 2 bits source state machine (P:pending Q:queued) 65 esb: Event State Buffer (Array of PQ bits in an IVSE) 66 eas: Event Assignment Structure 67 end: Event Notification Descriptor 68 nvt: Notification Virtual Target 69 tctx: Thread interrupt Context registers 70 71 72 73XIVE internal tables 74-------------------- 75 76Each of the sub-engines uses a set of tables to redirect interrupts 77from event sources to CPU threads. 78 79:: 80 81 +-------+ 82 User or O/S | EQ | 83 or +------>|entries| 84 Hypervisor | | .. | 85 Memory | +-------+ 86 | ^ 87 | | 88 +-------------------------------------------------+ 89 | | 90 Hypervisor +------+ +---+--+ +---+--+ +------+ 91 Memory | ESB | | EAT | | ENDT | | NVTT | 92 (skiboot) +----+-+ +----+-+ +----+-+ +------+ 93 ^ | ^ | ^ | ^ 94 | | | | | | | 95 +-------------------------------------------------+ 96 | | | | | | | 97 | | | | | | | 98 +----|--|--------|--|--------|--|-+ +-|-----+ +------+ 99 | | | | | | | | | | tctx| |Thread| 100 IPI or ---+ + v + v + v |---| + .. |-----> | 101 HW events | | | | | | 102 | IVRE | | IVPE | +------+ 103 +---------------------------------+ +-------+ 104 105 106The IVSE have a 2-bits state machine, P for pending and Q for queued, 107for each source that allows events to be triggered. They are stored in 108an Event State Buffer (ESB) array and can be controlled by MMIOs. 109 110If the event is let through, the IVRE looks up in the Event Assignment 111Structure (EAS) table for an Event Notification Descriptor (END) 112configured for the source. Each Event Notification Descriptor defines 113a notification path to a CPU and an in-memory Event Queue, in which 114will be enqueued an EQ data for the O/S to pull. 115 116The IVPE determines if a Notification Virtual Target (NVT) can handle 117the event by scanning the thread contexts of the VCPUs dispatched on 118the processor HW threads. It maintains the interrupt context state of 119each thread in a NVT table. 120 121XIVE thread interrupt context 122----------------------------- 123 124The XIVE presenter can generate four different exceptions to its 125HW threads: 126 127- hypervisor exception 128- O/S exception 129- Event-Based Branch (user level) 130- msgsnd (doorbell) 131 132Each exception has a state independent from the others called a Thread 133Interrupt Management context. This context is a set of registers which 134lets the thread handle priority management and interrupt 135acknowledgment among other things. The most important ones being : 136 137- Interrupt Priority Register (PIPR) 138- Interrupt Pending Buffer (IPB) 139- Current Processor Priority (CPPR) 140- Notification Source Register (NSR) 141 142TIMA 143~~~~ 144 145The Thread Interrupt Management registers are accessible through a 146specific MMIO region, called the Thread Interrupt Management Area 147(TIMA), four aligned pages, each exposing a different view of the 148registers. First page (page address ending in ``0b00``) gives access 149to the entire context and is reserved for the ring 0 view for the 150physical thread context. The second (page address ending in ``0b01``) 151is for the hypervisor, ring 1 view. The third (page address ending in 152``0b10``) is for the operating system, ring 2 view. The fourth (page 153address ending in ``0b11``) is for user level, ring 3 view. 154 155Interrupt flow from an O/S perspective 156~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 157 158After an event data has been enqueued in the O/S Event Queue, the IVPE 159raises the bit corresponding to the priority of the pending interrupt 160in the register IBP (Interrupt Pending Buffer) to indicate that an 161event is pending in one of the 8 priority queues. The Pending 162Interrupt Priority Register (PIPR) is also updated using the IPB. This 163register represent the priority of the most favored pending 164notification. 165 166The PIPR is then compared to the Current Processor Priority 167Register (CPPR). If it is more favored (numerically less than), the 168CPU interrupt line is raised and the EO bit of the Notification Source 169Register (NSR) is updated to notify the presence of an exception for 170the O/S. The O/S acknowledges the interrupt with a special load in the 171Thread Interrupt Management Area. 172 173The O/S handles the interrupt and when done, performs an EOI using a 174MMIO operation on the ESB management page of the associate source. 175 176Overview of the QEMU models for XIVE 177==================================== 178 179The XiveSource models the IVSE in general, internal and external. It 180handles the source ESBs and the MMIO interface to control them. 181 182The XiveNotifier is a small helper interface interconnecting the 183XiveSource to the XiveRouter. 184 185The XiveRouter is an abstract model acting as a combined IVRE and 186IVPE. It routes event notifications using the EAS and END tables to 187the IVPE sub-engine which does a CAM scan to find a CPU to deliver the 188exception. Storage should be provided by the inheriting classes. 189 190XiveEnDSource is a special source object. It exposes the END ESB MMIOs 191of the Event Queues which are used for coalescing event notifications 192and for escalation. Not used on the field, only to sync the EQ cache 193in OPAL. 194 195Finally, the XiveTCTX contains the interrupt state context of a thread, 196four sets of registers, one for each exception that can be delivered 197to a CPU. These contexts are scanned by the IVPE to find a matching VP 198when a notification is triggered. It also models the Thread Interrupt 199Management Area (TIMA), which exposes the thread context registers to 200the CPU for interrupt management. 201