264 lines · plain
1===========2NTB Drivers3===========4 5NTB (Non-Transparent Bridge) is a type of PCI-Express bridge chip that connects6the separate memory systems of two or more computers to the same PCI-Express7fabric. Existing NTB hardware supports a common feature set: doorbell8registers and memory translation windows, as well as non common features like9scratchpad and message registers. Scratchpad registers are read-and-writable10registers that are accessible from either side of the device, so that peers can11exchange a small amount of information at a fixed address. Message registers can12be utilized for the same purpose. Additionally they are provided with13special status bits to make sure the information isn't rewritten by another14peer. Doorbell registers provide a way for peers to send interrupt events.15Memory windows allow translated read and write access to the peer memory.16 17NTB Core Driver (ntb)18=====================19 20The NTB core driver defines an api wrapping the common feature set, and allows21clients interested in NTB features to discover NTB the devices supported by22hardware drivers. The term "client" is used here to mean an upper layer23component making use of the NTB api. The term "driver," or "hardware driver,"24is used here to mean a driver for a specific vendor and model of NTB hardware.25 26NTB Client Drivers27==================28 29NTB client drivers should register with the NTB core driver. After30registering, the client probe and remove functions will be called appropriately31as ntb hardware, or hardware drivers, are inserted and removed. The32registration uses the Linux Device framework, so it should feel familiar to33anyone who has written a pci driver.34 35NTB Typical client driver implementation36----------------------------------------37 38Primary purpose of NTB is to share some peace of memory between at least two39systems. So the NTB device features like Scratchpad/Message registers are40mainly used to perform the proper memory window initialization. Typically41there are two types of memory window interfaces supported by the NTB API:42inbound translation configured on the local ntb port and outbound translation43configured by the peer, on the peer ntb port. The first type is44depicted on the next figure::45 46 Inbound translation:47 48 Memory: Local NTB Port: Peer NTB Port: Peer MMIO:49 ____________50 | dma-mapped |-ntb_mw_set_trans(addr) |51 | memory | _v____________ | ______________52 | (addr) |<======| MW xlat addr |<====| MW base addr |<== memory-mapped IO53 |------------| |--------------| | |--------------|54 55So typical scenario of the first type memory window initialization looks:561) allocate a memory region, 2) put translated address to NTB config,573) somehow notify a peer device of performed initialization, 4) peer device58maps corresponding outbound memory window so to have access to the shared59memory region.60 61The second type of interface, that implies the shared windows being62initialized by a peer device, is depicted on the figure::63 64 Outbound translation:65 66 Memory: Local NTB Port: Peer NTB Port: Peer MMIO:67 ____________ ______________68 | dma-mapped | | | MW base addr |<== memory-mapped IO69 | memory | | |--------------|70 | (addr) |<===================| MW xlat addr |<-ntb_peer_mw_set_trans(addr)71 |------------| | |--------------|72 73Typical scenario of the second type interface initialization would be:741) allocate a memory region, 2) somehow deliver a translated address to a peer75device, 3) peer puts the translated address to NTB config, 4) peer device maps76outbound memory window so to have access to the shared memory region.77 78As one can see the described scenarios can be combined in one portable79algorithm.80 81 Local device:82 1) Allocate memory for a shared window83 2) Initialize memory window by translated address of the allocated region84 (it may fail if local memory window initialization is unsupported)85 3) Send the translated address and memory window index to a peer device86 87 Peer device:88 1) Initialize memory window with retrieved address of the allocated89 by another device memory region (it may fail if peer memory window90 initialization is unsupported)91 2) Map outbound memory window92 93In accordance with this scenario, the NTB Memory Window API can be used as94follows:95 96 Local device:97 1) ntb_mw_count(pidx) - retrieve number of memory ranges, which can98 be allocated for memory windows between local device and peer device99 of port with specified index.100 2) ntb_get_align(pidx, midx) - retrieve parameters restricting the101 shared memory region alignment and size. Then memory can be properly102 allocated.103 3) Allocate physically contiguous memory region in compliance with104 restrictions retrieved in 2).105 4) ntb_mw_set_trans(pidx, midx) - try to set translation address of106 the memory window with specified index for the defined peer device107 (it may fail if local translated address setting is not supported)108 5) Send translated base address (usually together with memory window109 number) to the peer device using, for instance, scratchpad or message110 registers.111 112 Peer device:113 1) ntb_peer_mw_set_trans(pidx, midx) - try to set received from other114 device (related to pidx) translated address for specified memory115 window. It may fail if retrieved address, for instance, exceeds116 maximum possible address or isn't properly aligned.117 2) ntb_peer_mw_get_addr(widx) - retrieve MMIO address to map the memory118 window so to have an access to the shared memory.119 120Also it is worth to note, that method ntb_mw_count(pidx) should return the121same value as ntb_peer_mw_count() on the peer with port index - pidx.122 123NTB Transport Client (ntb\_transport) and NTB Netdev (ntb\_netdev)124------------------------------------------------------------------125 126The primary client for NTB is the Transport client, used in tandem with NTB127Netdev. These drivers function together to create a logical link to the peer,128across the ntb, to exchange packets of network data. The Transport client129establishes a logical link to the peer, and creates queue pairs to exchange130messages and data. The NTB Netdev then creates an ethernet device using a131Transport queue pair. Network data is copied between socket buffers and the132Transport queue pair buffer. The Transport client may be used for other things133besides Netdev, however no other applications have yet been written.134 135NTB Ping Pong Test Client (ntb\_pingpong)136-----------------------------------------137 138The Ping Pong test client serves as a demonstration to exercise the doorbell139and scratchpad registers of NTB hardware, and as an example simple NTB client.140Ping Pong enables the link when started, waits for the NTB link to come up, and141then proceeds to read and write the doorbell scratchpad registers of the NTB.142The peers interrupt each other using a bit mask of doorbell bits, which is143shifted by one in each round, to test the behavior of multiple doorbell bits144and interrupt vectors. The Ping Pong driver also reads the first local145scratchpad, and writes the value plus one to the first peer scratchpad, each146round before writing the peer doorbell register.147 148Module Parameters:149 150* unsafe - Some hardware has known issues with scratchpad and doorbell151 registers. By default, Ping Pong will not attempt to exercise such152 hardware. You may override this behavior at your own risk by setting153 unsafe=1.154* delay\_ms - Specify the delay between receiving a doorbell155 interrupt event and setting the peer doorbell register for the next156 round.157* init\_db - Specify the doorbell bits to start new series of rounds. A new158 series begins once all the doorbell bits have been shifted out of159 range.160* dyndbg - It is suggested to specify dyndbg=+p when loading this module, and161 then to observe debugging output on the console.162 163NTB Tool Test Client (ntb\_tool)164--------------------------------165 166The Tool test client serves for debugging, primarily, ntb hardware and drivers.167The Tool provides access through debugfs for reading, setting, and clearing the168NTB doorbell, and reading and writing scratchpads.169 170The Tool does not currently have any module parameters.171 172Debugfs Files:173 174* *debugfs*/ntb\_tool/*hw*/175 A directory in debugfs will be created for each176 NTB device probed by the tool. This directory is shortened to *hw*177 below.178* *hw*/db179 This file is used to read, set, and clear the local doorbell. Not180 all operations may be supported by all hardware. To read the doorbell,181 read the file. To set the doorbell, write `s` followed by the bits to182 set (eg: `echo 's 0x0101' > db`). To clear the doorbell, write `c`183 followed by the bits to clear.184* *hw*/mask185 This file is used to read, set, and clear the local doorbell mask.186 See *db* for details.187* *hw*/peer\_db188 This file is used to read, set, and clear the peer doorbell.189 See *db* for details.190* *hw*/peer\_mask191 This file is used to read, set, and clear the peer doorbell192 mask. See *db* for details.193* *hw*/spad194 This file is used to read and write local scratchpads. To read195 the values of all scratchpads, read the file. To write values, write a196 series of pairs of scratchpad number and value197 (eg: `echo '4 0x123 7 0xabc' > spad`198 # to set scratchpads `4` and `7` to `0x123` and `0xabc`, respectively).199* *hw*/peer\_spad200 This file is used to read and write peer scratchpads. See201 *spad* for details.202 203NTB MSI Test Client (ntb\_msi\_test)204------------------------------------205 206The MSI test client serves to test and debug the MSI library which207allows for passing MSI interrupts across NTB memory windows. The208test client is interacted with through the debugfs filesystem:209 210* *debugfs*/ntb\_msi\_test/*hw*/211 A directory in debugfs will be created for each212 NTB device probed by the msi test. This directory is shortened to *hw*213 below.214* *hw*/port215 This file describes the local port number216* *hw*/irq*_occurrences217 One occurrences file exists for each interrupt and, when read,218 returns the number of times the interrupt has been triggered.219* *hw*/peer*/port220 This file describes the port number for each peer221* *hw*/peer*/count222 This file describes the number of interrupts that can be223 triggered on each peer224* *hw*/peer*/trigger225 Writing an interrupt number (any number less than the value226 specified in count) will trigger the interrupt on the227 specified peer. That peer's interrupt's occurrence file228 should be incremented.229 230NTB Hardware Drivers231====================232 233NTB hardware drivers should register devices with the NTB core driver. After234registering, clients probe and remove functions will be called.235 236NTB Intel Hardware Driver (ntb\_hw\_intel)237------------------------------------------238 239The Intel hardware driver supports NTB on Xeon and Atom CPUs.240 241Module Parameters:242 243* b2b\_mw\_idx244 If the peer ntb is to be accessed via a memory window, then use245 this memory window to access the peer ntb. A value of zero or positive246 starts from the first mw idx, and a negative value starts from the last247 mw idx. Both sides MUST set the same value here! The default value is248 `-1`.249* b2b\_mw\_share250 If the peer ntb is to be accessed via a memory window, and if251 the memory window is large enough, still allow the client to use the252 second half of the memory window for address translation to the peer.253* xeon\_b2b\_usd\_bar2\_addr64254 If using B2B topology on Xeon hardware, use255 this 64 bit address on the bus between the NTB devices for the window256 at BAR2, on the upstream side of the link.257* xeon\_b2b\_usd\_bar4\_addr64 - See *xeon\_b2b\_bar2\_addr64*.258* xeon\_b2b\_usd\_bar4\_addr32 - See *xeon\_b2b\_bar2\_addr64*.259* xeon\_b2b\_usd\_bar5\_addr32 - See *xeon\_b2b\_bar2\_addr64*.260* xeon\_b2b\_dsd\_bar2\_addr64 - See *xeon\_b2b\_bar2\_addr64*.261* xeon\_b2b\_dsd\_bar4\_addr64 - See *xeon\_b2b\_bar2\_addr64*.262* xeon\_b2b\_dsd\_bar4\_addr32 - See *xeon\_b2b\_bar2\_addr64*.263* xeon\_b2b\_dsd\_bar5\_addr32 - See *xeon\_b2b\_bar2\_addr64*.264