1=========================== 2InfiniBand Midlayer Locking 3=========================== 4 5 This guide is an attempt to make explicit the locking assumptions 6 made by the InfiniBand midlayer. It describes the requirements on 7 both low-level drivers that sit below the midlayer and upper level 8 protocols that use the midlayer. 9 10Sleeping and interrupt context 11============================== 12 13 With the following exceptions, a low-level driver implementation of 14 all of the methods in struct ib_device may sleep. The exceptions 15 are any methods from the list: 16 17 - create_ah 18 - modify_ah 19 - query_ah 20 - destroy_ah 21 - post_send 22 - post_recv 23 - poll_cq 24 - req_notify_cq 25 26 which may not sleep and must be callable from any context. 27 28 The corresponding functions exported to upper level protocol 29 consumers: 30 31 - rdma_create_ah 32 - rdma_modify_ah 33 - rdma_query_ah 34 - rdma_destroy_ah 35 - ib_post_send 36 - ib_post_recv 37 - ib_req_notify_cq 38 39 are therefore safe to call from any context. 40 41 In addition, the function 42 43 - ib_dispatch_event 44 45 used by low-level drivers to dispatch asynchronous events through 46 the midlayer is also safe to call from any context. 47 48Reentrancy 49---------- 50 51 All of the methods in struct ib_device exported by a low-level 52 driver must be fully reentrant. The low-level driver is required to 53 perform all synchronization necessary to maintain consistency, even 54 if multiple function calls using the same object are run 55 simultaneously. 56 57 The IB midlayer does not perform any serialization of function calls. 58 59 Because low-level drivers are reentrant, upper level protocol 60 consumers are not required to perform any serialization. However, 61 some serialization may be required to get sensible results. For 62 example, a consumer may safely call ib_poll_cq() on multiple CPUs 63 simultaneously. However, the ordering of the work completion 64 information between different calls of ib_poll_cq() is not defined. 65 66Callbacks 67--------- 68 69 A low-level driver must not perform a callback directly from the 70 same callchain as an ib_device method call. For example, it is not 71 allowed for a low-level driver to call a consumer's completion event 72 handler directly from its post_send method. Instead, the low-level 73 driver should defer this callback by, for example, scheduling a 74 tasklet to perform the callback. 75 76 The low-level driver is responsible for ensuring that multiple 77 completion event handlers for the same CQ are not called 78 simultaneously. The driver must guarantee that only one CQ event 79 handler for a given CQ is running at a time. In other words, the 80 following situation is not allowed:: 81 82 CPU1 CPU2 83 84 low-level driver -> 85 consumer CQ event callback: 86 /* ... */ 87 ib_req_notify_cq(cq, ...); 88 low-level driver -> 89 /* ... */ consumer CQ event callback: 90 /* ... */ 91 return from CQ event handler 92 93 The context in which completion event and asynchronous event 94 callbacks run is not defined. Depending on the low-level driver, it 95 may be process context, softirq context, or interrupt context. 96 Upper level protocol consumers may not sleep in a callback. 97 98Hot-plug 99-------- 100 101 A low-level driver announces that a device is ready for use by 102 consumers when it calls ib_register_device(), all initialization 103 must be complete before this call. The device must remain usable 104 until the driver's call to ib_unregister_device() has returned. 105 106 A low-level driver must call ib_register_device() and 107 ib_unregister_device() from process context. It must not hold any 108 semaphores that could cause deadlock if a consumer calls back into 109 the driver across these calls. 110 111 An upper level protocol consumer may begin using an IB device as 112 soon as the add method of its struct ib_client is called for that 113 device. A consumer must finish all cleanup and free all resources 114 relating to a device before returning from the remove method. 115 116 A consumer is permitted to sleep in its add and remove methods. 117