File:  [NetBSD Developer Wiki] / wikisrc / tutorials / bus_space_tutorial.mdwn
Revision 1.5: download - view: text, annotated - select for diffs
Sun Jun 23 14:41:01 2013 UTC (8 years, 5 months ago) by mspo
Branches: MAIN
CVS tags: HEAD
more cleanup at the end

    1: [[!toc ]]
    2: 
    3: ## Introduction
    4: 
    5: ### Why was this tutorial created?
    6: 
    7: -   Introductory-level documentation is scarce
    8: -   Writing device drivers is often considered black magic
    9: -   Reading the man pages won’t give you the big picture
   10: -   BSD systems are always in need of new drivers
   11: -   Device drivers are fun
   12: 
   13: ### What won’t be covered here?
   14: 
   15: We don’t have much time, so several ~~advanced~~ topics were omitted:
   16: 
   17: -   Interrupt handling
   18: -   Direct Memory Access and the bus\_dma framework
   19: -   Power management
   20: -   Driver detachment
   21: -   Drivers as kernel modules
   22: -   Examples for buses other than PCI
   23: -   Pretty much everything else...
   24: 
   25: However, once you finish this tutorial, you should be able to pursue
   26: this knowledge yourself.
   27: 
   28: ### What is a driver anyway?
   29: 
   30: -   The interface between user space and hardware, implemented as a part
   31:     of the kernel
   32: -   The NetBSD drivers are written mostly in C
   33: -   Sometimes they have machine dependent assembler parts, but this is a
   34:     rare case
   35: 
   36: ### What do you need to write a driver?
   37: 
   38: -   C programming skills
   39: -   Hardware documentation (or the ability to reverse engineer the
   40:     hardware)
   41: -   A reference driver implementation will help but is not essential
   42: -   A NetBSD installation and kernel source, or a cross-build
   43:     environment (the latter is usually preferred for development of
   44:     drivers)
   45: -   A lot of time, coffee and patience
   46: 
   47: ### Why is writing the device drivers considered difficult?
   48: 
   49: -   It’s not as difficult as you may expect, in fact during this
   50:     tutorial we’ll prove that it’s quite easy
   51: -   You need to think on a very low level
   52:     -   Good understanding of computer architecture is a must
   53: -   Often documentation is the main problem – writing the driver is not
   54:     possible if you don’t understand how the device works
   55:     -   No access to documentation (uncooperative hardware vendors,
   56:         vendors out of business)
   57:     -   Documentation is incomplete or plain wrong
   58:     -   Reverse engineering can solve these problems but it’s a very
   59:         time consuming process
   60: 
   61: ## The NetBSD driver model
   62: 
   63: ### The NetBSD kernel basics
   64: 
   65: -   NetBSD has a classic monolithic UNIX-like kernel - all drivers are
   66:     running in the same address space
   67: -   Thanks to the above, communication between drivers and other kernel
   68:     layers is simple
   69: -   However, it also means that one badly written driver can affect the
   70:     whole kernel
   71: -   Numerous in-kernel frameworks standardise the way drivers are
   72:     written (bus\_space, autoconf, etc.)
   73: 
   74: ### The NetBSD source directory structure
   75: 
   76: -   We’ll only cover parts interesting for a device driver programmer
   77: -   src/sys/
   78:     - kernel source directory
   79: -   src/sys/dev/
   80:     - machine-independent device drivers
   81: -   src/sys/arch/
   82:     - port-specific or architecture-specific parts (such as the
   83:     low-level system initialisation procedures or machine-dependent
   84:     drivers)
   85: -   src/sys/arch/\$PORTNAME/conf/
   86:     - kernel configuration files for a given port
   87: 
   88: ### Kernel autoconfiguration framework - autoconf(9)
   89: 
   90: -   Autoconfiguration is the process of matching hardware devices with
   91:     an appropriate device driver
   92: -   The kernel message buffer (dmesg) contains information about
   93:     autoconfiguration of devices
   94: -   driver0 at bus0: Foo hardware
   95:     -   Instance 0 of the driver has attached to instance 0 of the
   96:         particular bus
   97:     -   Such messages often carry additional bus-specific information
   98:         about the exact location of the device (like the device and
   99:         function number on the PCI bus)
  100: -   driver0: some message
  101:     -   Additional information about the driver state or device
  102:         configuration
  103: 
  104: ### Autoconfiguration as seen in the dmesg
  105: 
  106:     NetBSD 6.99.12 (GENERIC) #7: Fri Oct  5 18:43:21 CEST 2012
  107:             rkujawa@saiko.local:/Users/rkujawa/netbsd-eurobsdcon2012/src/sys/arch/cobalt/compile/obj/GENERIC
  108:     Cobalt Qube 2
  109:     total memory = 32768 KB
  110:     avail memory = 27380 KB
  111:     mainbus0 (root)
  112:     com0 at mainbus0 addr 0x1c800000 level 3: ns16550a, working fifo
  113:     com0: console
  114:     cpu0 at mainbus0: QED RM5200 CPU (0x28a0) Rev. 10.0 with built-in FPU Rev. 1.0
  115:     cpu0: 48 TLB entries, 256MB max page size
  116:     cpu0: 32KB/32B 2-way set-associative L1 instruction cache
  117:     cpu0: 32KB/32B 2-way set-associative write-back L1 data cache
  118:     mcclock0 at mainbus0 addr 0x10000070: mc146818 compatible time-of-day clock
  119:     panel0 at mainbus0 addr 0x1f000000
  120:     gt0 at mainbus0 addr 0x14000000
  121:     pci0 at gt0
  122:     pchb0 at pci0 dev 0 function 0: Galileo GT-64011 System Controller, rev 1
  123:     pcib0 at pci0 dev 9 function 0
  124:     pcib0: VIA Technologies VT82C586 PCI-ISA Bridge, rev 57
  125:     viaide0 at pci0 dev 9 function 1
  126:     viaide0: VIA Technologies VT82C586 (Apollo VP) ATA33 controller
  127:     viaide0: primary channel interrupting at irq 14
  128:     atabus0 at viaide0 channel 0
  129:     viaide0: secondary channel interrupting at irq 15
  130:     atabus1 at viaide0 channel 1
  131:     wd0 at atabus0 drive 0
  132:     wd0: <netbsd-cobalt.img>
  133:     wd0: 750 MB, 1524 cyl, 16 head, 63 sec, 512 bytes/sect x 1536192 sectors
  134: 
  135: ### Autoconfiguration as seen in the dmesg
  136: 
  137: ![image](img_cobaltdevices.png)
  138: 
  139: ### The bus\_space(9) framework
  140: 
  141: -   “The goal of the bus\_space functions is to allow a single driver
  142:     source file to manipulate a set of devices on different system
  143:     architectures, and to allow a single driver object file to
  144:     manipulate a set of devices on multiple bus types on a single
  145:     architecture.”
  146: -   Provides a set of functions implementing common operations on the
  147:     bus like mapping, reading, writing, copying, etc.
  148: -   The bus\_space(9) is implemented at the machine-dependent level
  149:     (typically it’s a part of architecture-specific code), but all
  150:     implementations present the same interface[^1]
  151: 
  152: ### Machine independent drivers
  153: 
  154: -   If possible drivers should work on any hardware platform
  155: -   High quality, machine-independent (MI) drivers are an important
  156:     factor that adds to NetBSD portability
  157: -   Some drivers are completely MI, some have MD or bus dependent
  158:     attachments and some are completely MD
  159:     -   A driver for a typical PCI card will be completely MI
  160:     -   A driver for the components of a SoC will usually be completely
  161:         MD
  162: -   The bus\_space abstraction helps to achieve portability,
  163:     transparently handling endianness issues and hiding bus
  164:     implementation details from the device driver
  165: -   Even if we have MI drivers, writing the drivers is always
  166:     significant part of effort needed to port NetBSD to new hardware
  167: 
  168: ## Example driver from scratch
  169: 
  170: ### Development environment
  171: 
  172: -   Out of scope of this course, but very well documented
  173: -   Cross compiling is an easy task with the build.sh script
  174: -   Described in [Part V of the NetBSD
  175:     Guide](http://www.netbsd.org/docs/guide/en/part-compile.html)
  176: -   Check out the NetBSD sources
  177: -   \$ build.sh -m cobalt tools
  178:     will build compiler, assembler, linker, etc. for cobalt port
  179: -   \$ build.sh -m cobalt kernel=GENERIC
  180:     will build the GENERIC kernel for cobalt
  181: -   Call build.sh with a -u parameter to update (won’t rebuilding
  182:     everything)
  183: -   build.sh
  184:     is calling nbconfig and nbmake tools, no magic involved
  185: 
  186: ### Quick introduction to GXemul
  187: 
  188: -   A framework for full-system computer architecture emulation,
  189:     excellent for educational purposes
  190: -   Capable of emulating several real machines supported by NetBSD
  191: -   We’ll emulate a [Cobalt](http://en.wikipedia.org/wiki/Cobalt_Qube),
  192:     MIPS-based micro server with PCI bus
  193: -   I’ve modified GXemul and implemented an emulation of an additional
  194:     PCI device
  195: -   It will be used to show (almost) a real-life example of the driver
  196:     development process
  197: 
  198: ### Our hardware - functional description
  199: 
  200: -   Business applications often use arithmetic operations like addition
  201: -   Fake Cards Inc. responded to market needs and created a new product,
  202:     Advanced Addition Accelerator
  203: -   Pointy Haired Bosses will certainly buy it to accelerate their
  204:     business applications, so let’s create a driver for NetBSD!
  205: 
  206: ### Our hardware - technical details
  207: 
  208: -   Overview
  209:     -   Implemented as a PCI device
  210:     -   Arithmetic unit capable of addition of two numbers
  211:     -   Four[^2] registers in the PCI memory space
  212: -   PCI configuration space
  213:     -   Identified by the PCI vendor ID 0xfabc and product ID 0x0001
  214:     -   Base Address Register 0x10 used to configure the engine address
  215:     -   4 x 32-bit registers = 16 bytes
  216:     -   Other configuration registers irrelevant
  217: 
  218: ### Our hardware - technical details (memory mapped register set)
  219: 
  220: -   Advanced Addition Acceleration registers
  221: 
  222: <table>
  223: <tr><th>Register Name</th><th>Offset</th><th>Description</th></tr>
  224: <tr><td>COMMAND</td><td>0x4</td><td>Register used to issue commands to the engine</td></tr>
  225: <tr><td>DATA</td><td>0x8</td><td>Register used to load data to internal engine registers</td></tr>
  226: <tr><td>RESULT</td><td>0xC</td><td>Register used to store the result of arithmetic operation</td></tr>
  227: </table>
  228: 
  229: -   COMMAND register
  230: 
  231: <table>
  232: <tr><th>Bit</th><th>R/W</th><th>Description</th></tr>
  233: <tr><td>0</td><td>W</td><td>Execute ADD operation on values loaded into internal register A and B</td></tr>
  234: <tr><td>1</td><td>R/W</td><td>Select internal register A for access through DATA register</td></tr>
  235: <tr><td>2</td><td>R/W</td><td>Select internal register B for access through DATA register</td></tr>
  236: </table>
  237: 
  238: -   Selecting internal register A and B at the same time will lead to
  239:     undefined behaviour
  240: 
  241: ### Our hardware - technical details (memory mapped register set)
  242: 
  243: -   DATA register
  244: 
  245: <table>
  246: <tr><th>Bit</th><th>R/W</th><th>Description</th></tr>
  247: <tr><td>0:31</td><td>R/W</td><td>Read/write the value in internal engine register</td></tr>
  248: </table>
  249: 
  250: -   RESULT register
  251: 
  252: <table>
  253: <tr><th>Bit</th><th>R/W</th><th>Description</th></tr>
  254: <tr><td>0:31</td><td>R</td><td>Holds the result of last ADD operation</td></tr>
  255: </table>
  256: 
  257: ### Our hardware - technical details (operation algorithm)
  258: 
  259: -   Select the internal register A for access (write 0x2 into COMMAND
  260:     register)
  261: -   Write the first number into DATA register
  262: -   Select the internal register B for access (write 0x4 into COMMAND
  263:     register)
  264: -   Write the second number into DATA register
  265: -   Issue the ADD operation (write 0x1 into COMMAND register)
  266: -   Read the result from RESULT register
  267: 
  268: ### Adding a new driver to the NetBSD kernel
  269: 
  270: -   We’ll discuss the steps needed to add a new MI PCI device driver to
  271:     the NetBSD kernel
  272:     -   Add the vendor and device ID to the database of PCI IDs
  273:     -   Create a set of the driver source files in
  274:         src/sys/dev/\$BUSNAME/
  275:     -   Add the new driver to src/sys/dev/\$BUSNAME/\$BUSNAME.files file
  276:     -   Add the new driver to DEVNAMES[^3] file
  277: 
  278: ### Modifying the PCI device database
  279: 
  280:     unmatched vendor 0xfabc product 0x0001 (Co-processor 
  281:     processor, revision 0x01) at pci0 dev 12 function 0 
  282:     not configured
  283: 
  284: -   The kernel does not know anything about this vendor and device
  285: -   Add it to the PCI device database - src/sys/dev/pci/pcidevs
  286: -   vendor VENDORNAME 0xVENDORID Long Vendor Name
  287: -   product VENDORNAME PRODUCTNAME 0xPRODUCTID Long Product Name
  288: -   To regenerate pcidevs\*.h run awk -f devlist2h.awk pcidevs or
  289:     Makefile.pcidevs if you’re on NetBSD
  290: 
  291: ### Modifying the PCI device database - example
  292: 
  293:     --- pcidevs 29 Sep 2012 10:26:14 -0000  1.1139
  294:     +++ pcidevs 5 Oct 2012 08:52:59 -0000
  295:     @@ -669,6 +669,7 @@
  296:      vendor CHRYSALIS   0xcafe  Chrysalis-ITS
  297:      vendor MIDDLE_DIGITAL  0xdeaf  Middle Digital
  298:      vendor ARC     0xedd8  ARC Logic
  299:     +vendor FAKECARDS   0xfabc  Fake Cards
  300:      vendor INVALID     0xffff  INVALID VENDOR ID
  301:      
  302:      /*
  303:     @@ -2120,6 +2121,9 @@
  304:      /* Eumitcom products */
  305:      product EUMITCOM WL11000P  0x1100  WL11000P PCI WaveLAN/IEEE 802.11
  306:      
  307:     +/* FakeCards products */
  308:     +product FAKECARDS AAA      0x0001  Advanced Addition Accelerator
  309:     +
  310:      /* O2 Micro */
  311:      product O2MICRO 00F7       0x00f7  Integrated OHCI IEEE 1394 Host Controller
  312:      product O2MICRO OZ6729     0x6729  OZ6729 PCI-PCMCIA Bridge
  313: 
  314: ### Modifying the PCI device database - example
  315: 
  316:     Fake Cards Advanced Addition Accelerator (Co-processor 
  317:     processor, revision 0x01) at pci0 dev 12 function 0 
  318:     not configured
  319: 
  320: -   Now the kernel knows the vendor and product ID
  321: -   But there’s still no driver for this device
  322: 
  323: ### Adding the new PCI driver
  324: 
  325: -   Choose a name - short, easy to remember, avoid numbers
  326:     -   faa looks like a good name, but you can choose any name you like
  327: -   Create a set of new files in src/sys/dev/pci
  328:     -   faa.c
  329:         - main driver code
  330:     -   faareg.h
  331:         - register definitions[^4]
  332:     -   faavar.h
  333:         - driver structures and functions used in other parts of the
  334:         kernel[^5]
  335: -   Modify driver definitions
  336:     -   src/sys/dev/pci/files.pci
  337:     -   src/sys/dev/DEVNAMES
  338: -   Configure the kernel to use the newly added driver -
  339:     src/sys/arch/\$PORTNAME/conf/GENERIC
  340: 
  341: ### Adding the new PCI driver - main driver
  342: 
  343: -   Kernel includes are at the beginning, followed by machine-specific
  344:     and bus-specific includes
  345: -   Should also include faareg.h and faavar.h files
  346: -   A minimal driver needs just two functions
  347:     -   faa\_match
  348:         (or faa\_probe for some buses)
  349:     -   faa\_attach
  350: -   The CFATTACH\_DECL\_NEW macro plugs the above functions into
  351:     autoconf(9) mechanism
  352: 
  353: ### Adding the new PCI driver - main driver
  354: 
  355: -   static int faa\_match(device\_t parent, cfdata\_t match, void
  356:     \*aux);
  357:     -   Check if the driver should attach to a given device (for example
  358:         in case of PCI bus, it will be used to check vendor and product
  359:         ID)
  360:     -   parent
  361:         - pointer to parent’s driver device structure
  362:     -   match
  363:         - pointer to autoconf(9) details structure
  364:     -   aux
  365:         - despite the name the most important argument, usually contains
  366:         bus-specific structure describing device details
  367: 
  368: -   static void faa\_attach(device\_t parent, device\_t self, void
  369:     \*aux);
  370:     -   Attach the driver to a given device
  371:     -   parent
  372:         - same as with match function
  373:     -   self
  374:         - pointer to driver’s device structure
  375:     -   aux
  376:         - same as with match function
  377: -   See definitions of these functions in the
  378:     [driver(9)](http://netbsd.gw.com/cgi-bin/man-cgi?driver+9+NetBSD-current)
  379:     man page.
  380: 
  381: ### Adding the new PCI driver - main driver cont’d
  382: 
  383: -   CFATTACH\_DECL\_NEW(faa, sizeof(struct faa\_softc), faa\_match,
  384:     faa\_attach, NULL, NULL);
  385:     -   driver name
  386:     -   size of softc structure containing state of driver’s instance
  387:     -   match/probe function
  388:     -   attach function
  389:     -   detach function
  390:     -   activate function
  391: -   The “\_NEW” name is unfortunate
  392: -   Pass NULL for unimplemented functions
  393: -   We won’t cover detach and activate now, as they are not needed for a
  394:     simple driver
  395: 
  396: ### Adding the new PCI driver - main driver example
  397: 
  398: -   src/sys/dev/pci/faa.c
  399: 
  400: <!-- -->
  401: 
  402:     #include <sys/cdefs.h>
  403:     __KERNEL_RCSID(0, "$NetBSD: bus_space_tutorial.mdwn,v 1.5 2013/06/23 14:41:01 mspo Exp $");
  404:     #include <sys/param.h>
  405:     #include <sys/device.h>
  406:     #include <dev/pci/pcivar.h>
  407:     #include <dev/pci/pcidevs.h>
  408:     #include <dev/pci/faareg.h>
  409:     #include <dev/pci/faavar.h>
  410: 
  411:     static int      faa_match(device_t, cfdata_t, void *);
  412:     static void     faa_attach(device_t, device_t, void *);
  413: 
  414:     CFATTACH_DECL_NEW(faa, sizeof(struct faa_softc),
  415:         faa_match, faa_attach, NULL, NULL);
  416: 
  417:     static int
  418:     faa_match(device_t parent, cfdata_t match, void *aux)
  419:     {
  420:             return 0;
  421:     }
  422: 
  423:     static void
  424:     faa_attach(device_t parent, device_t self, void *aux)
  425:     { 
  426:     }
  427: 
  428: ### Adding the new PCI driver - auxiliary includes
  429: 
  430: -   src/sys/dev/pci/faareg.h
  431: 
  432: <!-- -->
  433: 
  434:     #ifndef FAAREG_H
  435:     #define FAAREG_H
  436:     /* 
  437:      * Registers are defined using preprocessor:
  438:      * #define FAA_REGNAME  0x0
  439:      * We'll add them later, let's leave it empty for now.
  440:      */
  441:     #endif /* FAAREG_H */
  442: 
  443: -   src/sys/dev/pci/faavar.h
  444: 
  445: <!-- -->
  446: 
  447:     #ifndef FAAVAR_H
  448:     #define FAAVAR_H
  449: 
  450:     /* sc_dev is an absolute minimum, we'll add more later */
  451:     struct faa_softc {
  452:             device_t sc_dev;
  453:     };
  454:     #endif /* FAAVAR_H */
  455: 
  456: ### Adding the new PCI driver - registering the driver (courtesy)
  457: 
  458: -   src/sys/dev/DEVNAMES
  459: 
  460: <!-- -->
  461: 
  462:     --- DEVNAMES    1 Sep 2012 11:19:58 -0000   1.279
  463:     +++ DEVNAMES    6 Oct 2012 19:59:06 -0000
  464:     @@ -436,6 +436,7 @@
  465:      ex         MI
  466:      exphy          MI
  467:      ezload         MI      Attribute
  468:     +faa            MI
  469:      fb         luna68k
  470:      fb         news68k
  471:      fb         newsmips
  472: 
  473: ### Adding the new PCI driver - registering the driver
  474: 
  475: -   See config(5)
  476: -   src/sys/dev/pci/files.pci
  477: 
  478: <!-- -->
  479: 
  480:     --- pci/files.pci   2 Aug 2012 00:17:44 -0000   1.360
  481:     +++ pci/files.pci   6 Oct 2012 19:59:10 -0000
  482:     @@ -1122,3 +1122,9 @@
  483:      device tdvfb: wsemuldisplaydev, rasops8, vcons, videomode
  484:      attach tdvfb at pci
  485:      file   dev/pci/tdvfb.c     tdvfb   
  486:     +
  487:     +# FakeCards Advanced Addition Accelerator
  488:     +device faa
  489:     +attach faa at pci
  490:     +file   dev/pci/faa.c       faa 
  491:     +
  492: 
  493: ### Adding the new PCI driver to the kernel configuration
  494: 
  495: -   src/sys/arch/cobalt/conf/GENERIC
  496: 
  497: <!-- -->
  498: 
  499:     --- GENERIC 10 Mar 2012 21:51:50 -0000  1.134
  500:     +++ GENERIC 6 Oct 2012 20:12:37 -0000
  501:     @@ -302,6 +302,9 @@
  502:      #fms*      at pci? dev ? function ?    # Forte Media FM801
  503:      #sv*       at pci? dev ? function ?    # S3 SonicVibes
  504:      
  505:     +# Fake Cards Advanced Addition Accelerator
  506:     +faa*       at pci? dev ? function ?
  507:     +
  508:      # Audio support
  509:      #audio*        at audiobus?
  510: 
  511: -   The above definition means that an instance of faa may be attached
  512:     to any PCI bus, any device, any function
  513: -   The exact position of the rule in the configuration file is not
  514:     important in this case
  515: -   See
  516:     [config(5)](http://netbsd.gw.com/cgi-bin/man-cgi?config+5+NetBSD-current)
  517:     for a description of the device definition language
  518: 
  519: ### Adding the new PCI driver - example
  520: 
  521: -   The driver should compile now
  522: -   The driver’s match function will check if the driver is able to work
  523:     with a given device
  524: -   Since it is not implemented, the kernel will not attach the driver
  525: 
  526: ### Matching the PCI device
  527: 
  528: -   Modify the faa\_match function to match the specified PCI device
  529: -   Use PCI\_VENDOR and PCI\_PRODUCT macros to obtain the IDs
  530: 
  531: <!-- -->
  532: 
  533:     static int
  534:     faa_match(device_t parent, cfdata_t match, void *aux)
  535:     {
  536:             const struct pci_attach_args *pa = (const struct pci_attach_args *)aux;
  537: 
  538:             if ((PCI_VENDOR(pa->pa_id) == PCI_VENDOR_FAKECARDS) 
  539:                 && (PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_FAKECARDS_AAA))
  540:                     return 1;
  541: 
  542:             return 0;
  543:     }
  544: 
  545: ### Attaching to the PCI device
  546: 
  547:     faa0 at pci0 dev 12 function 0
  548: 
  549: -   The driver has successfully matched and attached to the PCI device
  550:     but still is not doing anything useful
  551: -   Let’s fill the attach function and actually program the hardware
  552: 
  553: ### Variable types used with bus\_space
  554: 
  555: -   bus\_space\_tag\_t
  556: 
  557:     – type used to describe a particular bus, usually passed to the
  558:     driver from MI bus structures
  559: 
  560: -   bus\_space\_handle\_t
  561: 
  562:     – used to describe a mapped range of bus space, usually created with
  563:     the bus\_space\_map() function
  564: 
  565: -   bus\_addr\_t
  566: 
  567:     – address on the bus
  568: 
  569: -   bus\_size\_t
  570: 
  571:     – an amount of space on the bus
  572: 
  573: -   Contents of these types are MD, so avoid modifying from within the
  574:     driver[^6]
  575: 
  576: ### Why do we need to “map” the resources?
  577: 
  578: -   “The bus space must be mapped before it can be used, and should be
  579:     unmapped when it is no longer needed”
  580: -   It’s a machine-dependent process but it’s also conveniently hidden
  581:     from the programmer by the bus\_space framework
  582: 
  583: ### Mapping the hardware resources
  584: 
  585: -   The generic bus\_space(9) way to map space
  586: 
  587: <!-- -->
  588: 
  589:     bus_space_map(bus_space_tag_t space, bus_addr_t address, 
  590:     bus_size_t size, int flags, bus_space_handle_t  *handlep);
  591: 
  592: -   bus\_space\_map
  593: 
  594:     creates a mapping from the physical address to a kernel virtual
  595:     address
  596: 
  597: -   space
  598: 
  599:     – represents the bus on which the mapping will be created
  600: 
  601: -   address
  602: 
  603:     – typically represents the physical address for which a mapping will
  604:     be created
  605: 
  606: -   size
  607: 
  608:     – describes the amount of bus space to be mapped
  609: 
  610: -   handlep
  611: 
  612:     – pointer to mapped space (filled after successful mapping)
  613: 
  614: -   Separate space and address
  615: 
  616: ### Mapping the hardware resources
  617: 
  618: -   The PCI-specific way to map space
  619: 
  620: <!-- -->
  621: 
  622:     pci_mapreg_map(const struct pci_attach_args *pa, int reg, pcireg_t type, 
  623:     int busflags, bus_space_tag_t *tagp, bus_space_handle_t *handlep, 
  624:     bus_addr_t *basep, bus_size_t *sizep);
  625: 
  626: -   pci\_mapreg\_map
  627: 
  628:     creates mapping from physical address present in specified BAR
  629:     register to kernel virtual address
  630: 
  631: -   pa
  632: 
  633:     – struct describing PCI attachment details (passed through aux)
  634: 
  635: -   reg
  636: 
  637:     – BAR register number
  638: 
  639: -   type
  640: 
  641:     – Select mapping type (I/O, memory)
  642: 
  643: -   busflags
  644: 
  645:     – Passed to bus\_space\_map flags argument
  646: 
  647: -   tagp
  648: 
  649:     – pointer to bus\_space\_tag
  650: 
  651: -   handlep
  652: 
  653:     – pointer to a mapped space
  654: 
  655: -   basep
  656: 
  657:     – address of a mapped space
  658: 
  659: -   sizep
  660: 
  661:     – size of mapped space (equivalent to BAR size)
  662: 
  663: -   The last four parameters are filled after successful mapping
  664: 
  665: ### Mapping the registers using BAR - adding auxiliary includes
  666: 
  667: -   src/sys/dev/pci/faareg.h
  668: 
  669: <!-- -->
  670: 
  671:     #define FAA_MMREG_BAR   0x10
  672: 
  673: -   src/sys/dev/pci/faavar.h
  674: 
  675: <!-- -->
  676: 
  677:     struct faa_softc {
  678:             device_t sc_dev;
  679: 
  680:             bus_space_tag_t sc_regt;
  681:             bus_space_handle_t sc_regh;
  682:             bus_addr_t sc_reg_pa;
  683: 
  684:     };
  685: 
  686: ### Mapping the registers using BAR - main driver code
  687: 
  688: -   src/sys/dev/pci/faa.c
  689: 
  690: <!-- -->
  691: 
  692:     static void
  693:     faa_attach(device_t parent, device_t self, void *aux)
  694:     {
  695:             struct faa_softc *sc = device_private(self);
  696:             const struct pci_attach_args *pa = aux;
  697: 
  698:             sc->sc_dev = self;
  699: 
  700:             pci_aprint_devinfo(pa, NULL);
  701: 
  702:             if (pci_mapreg_map(pa, FAA_MMREG_BAR, PCI_MAPREG_TYPE_MEM, 0, 
  703:                 &sc->sc_regt, &sc->sc_regh, &sc->sc_reg_pa, 0) != 0 ) {
  704:                 aprint_error_dev(sc->sc_dev, "can't map the BAR\n");
  705:                 return;
  706:             }
  707: 
  708:             aprint_normal_dev(sc->sc_dev, "regs at 0x%08x\n", (uint32_t) sc->sc_reg_pa);
  709:     }
  710: 
  711: ### Accessing the hardware registers
  712: 
  713: -   The bus\_space\_read\_ and bus\_space\_write\_ functions are basic
  714:     methods of reading and writing the hardware registers
  715: 
  716: -   uintX\_t bus\_space\_read\_X(bus\_space\_tag\_t space,
  717:     bus\_space\_handle\_t handle, bus\_size\_t offset);
  718: 
  719: -   void bus\_space\_write\_X(bus\_space\_tag\_t space,
  720:     bus\_space\_handle\_t handle, bus\_size\_t offset, uintX\_t value);
  721: 
  722:     -   space
  723: 
  724:         - tag describing the bus
  725: 
  726:     -   handle
  727: 
  728:         - describes the exact location on the bus where read/write
  729:         should occur, this handle is obtained by bus\_space\_map
  730: 
  731:     -   offset
  732: 
  733:         - offset from handle location
  734: 
  735:     -   The read function returns the data read from the specified
  736:         location, while write has an argument value which should be
  737:         filled with data to be written
  738: 
  739: ### Variants of bus\_space\_read and bus\_space\_write
  740: 
  741:     Data       Read function         Write function
  742:   -------- --------------------- ----------------------
  743:    8-bit    bus\_space\_read\_1   bus\_space\_write\_1
  744:    16-bit   bus\_space\_read\_2   bus\_space\_write\_2
  745:    32-bit   bus\_space\_read\_4   bus\_space\_write\_4
  746:    64-bit   bus\_space\_read\_8   bus\_space\_write\_8
  747: 
  748: -   There are many more variants of read and write functions and they
  749:     are useful in certain situations, see the
  750:     [bus\_space(9)](http://netbsd.gw.com/cgi-bin/man-cgi?bus_space++NetBSD-current)
  751:     man page
  752: 
  753: ### Accessing the hardware registers - example
  754: 
  755: -   Create a function that will write a value into the DATA register of
  756:     our device, then read it back and check if the value is the same as
  757:     written
  758: 
  759: -   Define the DATA register in the driver
  760: 
  761: -   src/sys/dev/pci/faareg.h
  762: 
  763: <!-- -->
  764: 
  765:     #define FAA_DATA                0x8
  766:     #define FAA_COMMAND             0x4
  767:     #define FAA_COMMAND_STORE_A         __BIT(1)
  768: 
  769: -   Define the new function in main driver code
  770: 
  771: -   static bool faa\_check(struct faa\_softc \*sc);
  772: 
  773: ### Accessing the hardware registers - example
  774: 
  775: -   src/sys/dev/pci/faa.c
  776: 
  777: <!-- -->
  778: 
  779:     static void
  780:     faa_attach(device_t parent, device_t self, void *aux)
  781:     {
  782:        /* ... */
  783:        if (!faa_check(sc)) {
  784:             aprint_error_dev(sc->sc_dev, "hardware not responding\n");
  785:             return;
  786:        }
  787:     }
  788: 
  789:     static bool
  790:     faa_check(struct faa_softc *sc)
  791:     {
  792:             uint32_t testval = 0xff11ee22; 
  793:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_COMMAND, FAA_COMMAND_STORE_A);
  794:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_DATA, testval);
  795:             if (bus_space_read_4(sc->sc_regt, sc->sc_regh, FAA_DATA) == testval)
  796:                     return true;
  797: 
  798:             return false;
  799:     }
  800: 
  801: ### Accessing the hardware registers - running the example
  802: 
  803: -   Update the kernel binary and run it again
  804: 
  805: -   Check the GXemul log
  806: 
  807: <!-- -->
  808: 
  809:     [ faa: COMMAND register (0x4) WRITE value 0x2 ]
  810:     [ faa: DATA register (0x8) WRITE value 0xff11ee22 ]
  811:     [ faa: DATA register (0x8) READ value 0xff11ee22 ]
  812: 
  813: -   GXemul will conveniently display all accesses to our device
  814: 
  815: -   The faa driver still does attach without error, which means that the
  816:     check function is working properly
  817: 
  818: <!-- -->
  819: 
  820:     faa0 at pci0 dev 12 function 0: Fake Cards Advanced Addition Accelerator (rev. 0x01)
  821:     faa0: registers at 0x10110000
  822: 
  823: ### Implementing addition using the hardware
  824: 
  825: -   The basic principle of device operation should be laid out in the
  826:     data sheet
  827: 
  828: -   We need to implement an algorithm based on this description
  829: 
  830: -   Writing such an algorithm is often not needed, since the NetBSD
  831:     kernel already has frameworks for common device types (such as
  832:     atabus/wd for IDE and SATA hard disk controllers, wsdisplay/wscons
  833:     for frame buffers, etc.)
  834: 
  835: ### Implementing addition using the hardware
  836: 
  837: -   Define all registers
  838: 
  839: -   src/sys/dev/pci/faareg.h
  840: 
  841: <!-- -->
  842: 
  843:     #define FAA_STATUS              0x0
  844:     #define FAA_COMMAND             0x4
  845:     #define FAA_COMMAND_ADD             __BIT(0)        
  846:     #define FAA_COMMAND_STORE_A         __BIT(1)
  847:     #define FAA_COMMAND_STORE_B         __BIT(2)
  848:     #define FAA_DATA                0x8
  849:     #define FAA_RESULT              0xC
  850: 
  851: ### Implementing addition using the hardware
  852: 
  853: -   Add a new function to the main driver code
  854: 
  855: -   src/sys/dev/pci/faa.c
  856: 
  857: <!-- -->
  858: 
  859:     static void
  860:     faa_attach(device_t parent, device_t self, void *aux)
  861:     {
  862:             /* ... */
  863:             aprint_normal_dev(sc->sc_dev, "just checking: 1 + 2 = %d\n", faa_add(sc, 1, 2));
  864:     }
  865: 
  866:     static uint32_t
  867:     faa_add(struct faa_softc *sc, uint32_t a, uint32_t b)
  868:     {
  869:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_COMMAND, FAA_COMMAND_STORE_A);
  870:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_DATA, a);
  871:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_COMMAND, FAA_COMMAND_STORE_B);
  872:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_DATA, b);
  873:             bus_space_write_4(sc->sc_regt, sc->sc_regh, FAA_COMMAND, FAA_COMMAND_ADD);
  874:             return bus_space_read_4(sc->sc_regt, sc->sc_regh, FAA_RESULT);
  875:     }
  876: 
  877: ### Implementing addition using the hardware - running the example
  878: 
  879: -   Update the kernel binary and run it again
  880: 
  881: -   Check GXemul log
  882: 
  883: <!-- -->
  884: 
  885:     [ faa: COMMAND register (0x4) WRITE value 0x2 ]
  886:     [ faa: DATA register (0x8) WRITE value 0x1 ]
  887:     [ faa: COMMAND register (0x4) WRITE value 0x4 ]
  888:     [ faa: DATA register (0x8) WRITE value 0x2 ]
  889:     [ faa: COMMAND register (0x4) WRITE value 0x1 ]
  890:     [ faa: RESULT register (0xC) READ value 0x3 ]
  891: 
  892: -   Looks like it worked!
  893: 
  894: <!-- -->
  895: 
  896:     faa0 at pci0 dev 12 function 0: Fake Cards Advanced Addition Accelerator (rev. 0x01)
  897:     faa0: registers at 0x10110000
  898:     faa0: just checking: 1 + 2 = 3
  899: 
  900: ## Interacting with userspace
  901: 
  902: ### The kernel-user space interface
  903: 
  904: -   Now that the core functionality of the kernel driver is working, it
  905:     should be exposed to user space
  906: -   The interface between kernel driver and userspace can be designed in
  907:     many different ways
  908: -   The classic UNIX way of interfacing between the kernel and user
  909:     space is a device file
  910: -   Even when using device files there is no single interfacing method
  911:     that fits all use cases
  912: -   It’s up to the programmer to define the communication protocol
  913: 
  914: ### Device files
  915: 
  916: -   crw-r—– 1 root wheel 101, 1 Aug 12 21:53 /dev/file
  917: -   The kernel identifies which driver should service the request to
  918:     this file by using major and minor numbers (101 and 1 in the example
  919:     above)
  920: -   The major number identifies the driver
  921: -   The minor number usually identifies the driver instance, although
  922:     the driver is free to use it in any other way
  923: -   In NetBSD device files are created statically
  924:     -   By the MAKEDEV script during installation or boot
  925:     -   Manually by using the mknod utility
  926: 
  927: ### Operations on device files
  928: 
  929: -   [open(2)](http://netbsd.gw.com/cgi-bin/man-cgi?read++NetBSD-current)
  930:     and
  931:     [close(2)](http://netbsd.gw.com/cgi-bin/man-cgi?read++NetBSD-current)
  932: -   [read(2)](http://netbsd.gw.com/cgi-bin/man-cgi?read++NetBSD-current)
  933:     and
  934:     [write(2)](http://netbsd.gw.com/cgi-bin/man-cgi?write++NetBSD-current)
  935: -   [ioctl(2)](http://netbsd.gw.com/cgi-bin/man-cgi?write++NetBSD-current)
  936: -   [poll(2)](http://netbsd.gw.com/cgi-bin/man-cgi?poll++NetBSD-current)
  937: -   [mmap(2)](http://netbsd.gw.com/cgi-bin/man-cgi?write++NetBSD-current)
  938: -   and more…
  939: -   Any mix of the above system calls might be used to interface between
  940:     the kernel and user space
  941: -   We’ll implement an ioctl(2)-based communication mechanism
  942: 
  943: ### Adding cdevsw
  944: 
  945: -   cdevsw
  946: 
  947:     is used to decide which operation on the character device file calls
  948:     which driver function
  949: 
  950: -   Not all calls have to be implemented, although some device layers
  951:     define a set of calls that a driver must implement
  952: 
  953: -   For example disk drivers must implement open, close, read, write and
  954:     ioctl
  955: 
  956: -   src/sys/dev/pci/faa.c
  957: 
  958: <!-- -->
  959: 
  960:     dev_type_open(faaopen);
  961:     dev_type_close(faaclose);
  962:     dev_type_ioctl(faaioctl);
  963: 
  964:     const struct cdevsw faa_cdevsw = {
  965:             faaopen, faaclose, noread, nowrite, faaioctl,
  966:             nostop, notty, nopoll, nommap, nokqfilter, D_OTHER
  967:     };
  968: 
  969: ### Prototyping the cdevsw operations
  970: 
  971: -   The dev\_type\* macros are used to prototype the functions passed to
  972:     cdevsw
  973: -   Pass no followed by a function name to the appropriate cdevsw field
  974:     if it is not implemented
  975: -   There’s also bdevsw for block devices, but we won’t use it in this
  976:     example
  977: -   The last member of the cdevsw structure defines the device flags,
  978:     originally it was used to define the device type (still used for
  979:     disks, tape drives and ttys, for other devices pass D\_OTHER)
  980: 
  981: ### Implemeting the cdevsw operations - open / close
  982: 
  983: -   src/sys/dev/pci/faa.c
  984: 
  985: <!-- -->
  986: 
  987:     int
  988:     faaopen(dev_t dev, int flags, int mode, struct lwp *l)
  989:     {
  990:             struct faa_softc *sc;
  991:             sc = device_lookup_private(&faa_cd, minor(dev));
  992: 
  993:             if (sc == NULL)
  994:                     return ENXIO;
  995:             if (sc->sc_flags & FAA_OPEN)
  996:                     return EBUSY;
  997: 
  998:             sc->sc_flags |= FAA_OPEN;
  999:             return 0;
 1000:     }
 1001:     int
 1002:     faaclose(dev_t dev, int flag, int mode, struct lwp *l)
 1003:     {
 1004:             struct faa_softc *sc;
 1005:             sc = device_lookup_private(&faa_cd, minor(dev));
 1006: 
 1007:             if (sc->sc_flags & FAA_OPEN)
 1008:                     sc->sc_flags =~ FAA_OPEN;
 1009: 
 1010:             return 0;
 1011:     }
 1012: 
 1013: ### Defining the ioctls
 1014: 
 1015: -   ioctl(2)
 1016: 
 1017:     can be used to call kernel-level functions and exchange data between
 1018:     the kernel and user space
 1019: 
 1020: -   The classic way of passing data is by using structures, their
 1021:     definitions are shared between the kernel and user space code
 1022: -   The driver might support more than one ioctl, the \_IO\* macros are
 1023:     used to define the operation and associated structure used to
 1024:     exchange data
 1025: 
 1026:     -   \_IO
 1027: 
 1028:         - just a kernel function call, no data exchange
 1029: 
 1030:     -   \_IOR
 1031: 
 1032:         - kernel function call and data pass from kernel to user space
 1033: 
 1034:     -   \_IOW
 1035: 
 1036:         - kernel function call and data pass from user space to kernel
 1037: 
 1038:     -   \_IOWR
 1039: 
 1040:         - kernel function call and data exchange in both directions
 1041: 
 1042:     -   \#define DRIVERIO\_IOCTLNAME \_IOXXX(group, ioctl\_number, data
 1043:         structure)
 1044: 
 1045: ## Using ioctls
 1046: 
 1047: ### Defining the ioctls
 1048: 
 1049: -   src/sys/dev/pci/faaio.h
 1050: 
 1051: <!-- -->
 1052: 
 1053:     #include <sys/ioccom.h>
 1054:                          
 1055:     #define FAAIO_ADD   _IOWR(0, 1, struct faaio_add)
 1056: 
 1057:     struct faaio_add {
 1058:         uint32_t a;
 1059:         uint32_t b;
 1060:         uint32_t *result;
 1061:     };
 1062: 
 1063: -   In the above example the ioctl group is not defined (0), but a
 1064:     single letter identifier could appear as first argument to \_IOWR
 1065: 
 1066: ### Implemeting the cdevsw operations - ioctl
 1067: 
 1068: -   src/sys/dev/pci/faa.c
 1069: 
 1070: <!-- -->
 1071: 
 1072:     int
 1073:     faaioctl(dev_t dev, u_long cmd, void *data, int flag, struct lwp *l)
 1074:     {
 1075:             struct faa_softc *sc = device_lookup_private(&faa_cd, minor(dev));
 1076:             int err;
 1077: 
 1078:             switch (cmd) {
 1079:             case FAAIO_ADD:
 1080:                     err = faaioctl_add(sc, (struct faaio_add *) data);
 1081:                     break;
 1082:             default:
 1083:                     err = EINVAL;
 1084:                     break;
 1085:             }
 1086:             return(err);
 1087:     }
 1088:     static int
 1089:     faaioctl_add(struct faa_softc *sc, struct faaio_add *data)
 1090:     {
 1091:             uint32_t result; int err;
 1092: 
 1093:             aprint_normal_dev(sc->sc_dev, "got ioctl with a %d, b %d\n",
 1094:                 data->a, data->b);
 1095: 
 1096:             result = faa_add(sc, data->a, data->b);
 1097:             err = copyout(&result, data->result, sizeof(uint32_t));
 1098:             return err;
 1099:     }
 1100: 
 1101: ### Using copyout to pass data to userspace
 1102: 
 1103: -   The copy(9) functions are used to copy kernel space data from/to
 1104:     user space
 1105: -   copyout(kernel\_address, user space\_address, size);
 1106: -   Actually on Cobalt we could just do data-\>result = faa\_add();
 1107:     instead of calling the copyout function, but that is a bad idea
 1108: -   Some architectures (such as sparc64) have totally separate kernel
 1109:     and user address spaces $ \implies $ user space addresses are
 1110:     meaningless in the kernel
 1111: 
 1112: ### Defining device major number
 1113: 
 1114: -   Device major numbers for hardware drivers are usually defined in a
 1115:     per-port manner[^7]
 1116: -   src/sys/arch/\$PORTNAME/conf/majors.\$PORTNAME
 1117: -   src/sys/arch/cobalt/conf/majors.cobalt
 1118: -   The following defines a new character device file called /dev/faa\*
 1119:     with major number 101, but only if the faa driver is included in the
 1120:     kernel (last argument)
 1121: -   device-major faa char 101 faa
 1122: 
 1123: ### Creating the device node
 1124: 
 1125: -   The mknod utility can be used to create the device file manually
 1126: -   The driver name can be specified instead of the major number - it
 1127:     will be automatically resolved into the correct major number
 1128: -   mknod name [b | c] [major | driver] minor
 1129: -   mknod /dev/faa0 c faa 0
 1130: -   Created successfully
 1131: -   crw-r–r– 1 root wheel 101, 0 Oct 8 2012 /dev/faa0
 1132: 
 1133: ### An example user space program
 1134: 
 1135: -   The example program will open the device file and call ioctl(2) on
 1136:     it
 1137: -   As simple as possible, just to show how communication is done
 1138: -   Using ioctls from the user space
 1139:     -   Open the device file with O\_RDWR
 1140:     -   Call ioctl(2) with the operation number and structure as
 1141:         parameters
 1142: 
 1143: ### An example user space program - source
 1144: 
 1145:     void add(int, uint32_t, uint32_t);
 1146: 
 1147:     static const char* faa_device = "/dev/faa0";
 1148: 
 1149:     int
 1150:     main(int argc, char *argv[])
 1151:     {
 1152:             int devfd;
 1153: 
 1154:             if (argc != 3) {
 1155:                     printf("usage: %s a b\n", argv[0]);
 1156:                     return 1;
 1157:             }
 1158:             if ( (devfd = open(faa_device, O_RDWR)) == -1) {
 1159:                     perror("can't open device file");
 1160:                     return 1;
 1161:             }
 1162: 
 1163:             add(devfd, atoi(argv[1]), atoi(argv[2]));
 1164: 
 1165:             close(devfd);
 1166:             return 0;
 1167:     }
 1168: 
 1169: ### An example user space program - source
 1170: 
 1171:     void
 1172:     add(int devfd, uint32_t a, uint32_t b)
 1173:     {
 1174:             struct faaio_add faaio;
 1175:             uint32_t result = 0;
 1176: 
 1177:             faaio.result = &result;
 1178:             faaio.a = a;
 1179:             faaio.b = b;
 1180: 
 1181:             if (ioctl(devfd, FAAIO_ADD, &faaio) == -1) {
 1182:                     perror("ioctl failed");
 1183:             }
 1184:             printf("%d\n", result);
 1185:     }
 1186: 
 1187: ### An example user space program - running it
 1188: 
 1189:     # make
 1190:     cc -o aaa_add aaa_add.c
 1191:     # ./aaa_add 3 7
 1192:     faa0: got ioctl with a 3, b 7
 1193:     10
 1194: 
 1195: -   The program is successfully accessing the faa driver through the
 1196:     ioctl
 1197: -   The faa0:... line is a kernel message, normally only seen on the
 1198:     console terminal
 1199: 
 1200: ## A few tips
 1201: 
 1202: ### Avoiding common pitfalls
 1203: 
 1204: -   Always free resources allocated in the match or probe functions
 1205: -   Always use bus\_space methods, don’t access the hardware using a
 1206:     pointer dereference
 1207: -   If possible test on more than one hardware architecture, some bugs
 1208:     may surface
 1209: -   Don’t reinvent the wheel, try to use existing kernel frameworks as
 1210:     much as possible
 1211: -   Use copy(9) (or uiomove(9) or store(9)/fetch(9)) to move data
 1212:     between the kernel and user space
 1213: 
 1214: ### Basic driver debugging
 1215: 
 1216: -   Use aprint\_debug to print debug-level messages on console and log
 1217:     them (enabled by passing AB\_DEBUG from the boot loader)
 1218: -   Use the built-in DDB debugger
 1219:     -   Enabled by the kernel option DDB
 1220:     -   A kernel panic will start DDB if the DDB\_ONPANIC=1 kernel
 1221:         option is specified or the ddb.onpanic sysctl is set to 1.
 1222:     -   Run \# sysctl -w kern.panic\_now=1 to trigger a panic manually
 1223:         (DIAGNOSTIC option)
 1224: -   Remote debugging is possible on some ports
 1225:     -   With KGDB through the serial port
 1226:     -   With IPKDB through the network
 1227: 
 1228: ## Summary
 1229: 
 1230: ### Further reading
 1231: 
 1232: -   Documentation, articles:
 1233:     -   [A Machine-Independent DMA Framework for NetBSD, Jason R.
 1234:         Thorpe](http://www.netbsd.org/docs/kernel/bus_dma.pdf)
 1235:     -   [Writing Drivers for NetBSD, Jochen
 1236:         Kunz](ftp://ftp.netbsd.org/pub/NetBSD/misc/ddwg/NetBSD-driver_writing-1.0.1e.pdf)
 1237:     -   [NetBSD Documentation: Writing a pseudo
 1238:         device](http://www.netbsd.org/docs/kernel/pseudo/)
 1239:     -   [autoconf(9)](http://netbsd.gw.com/cgi-bin/man-cgi?autoconf+9+NetBSD-current),
 1240:         [bus\_space(9)](http://netbsd.gw.com/cgi-bin/man-cgi?bus_space+9+NetBSD-current)
 1241:         [bus\_dma(9)](http://netbsd.gw.com/cgi-bin/man-cgi?bus_dma+9+NetBSD-current)
 1242:         [driver(9)](http://netbsd.gw.com/cgi-bin/man-cgi?driver+9+NetBSD-current),
 1243:         [pci(9)](http://netbsd.gw.com/cgi-bin/man-cgi?pci+9+NetBSD-current)
 1244:         man pages, etc.
 1245: 
 1246: -   Example source code of drivers:
 1247:     -   tdvfb
 1248: 
 1249:         voodoofb are fairly good frame buffer driver examples with
 1250:         documentation publicly available.
 1251: 
 1252:     -   etsec
 1253: 
 1254:         is a nice example of a more complicated network interface driver
 1255: 
 1256: ### Get the source code
 1257: 
 1258: -   Download the source code and materials for this tutorial
 1259: -   <https://github.com/rkujawa/busspace-tutorial>
 1260: -   <https://github.com/rkujawa/gxemul-tutorial>
 1261: 
 1262: ### The End…
 1263: 
 1264: [^1]: At least they should, some functions are missing on less popular ports
 1265: 
 1266: [^2]: Only three of these registers are of any importance for us at this moment
 1267: 
 1268: [^3]: Required if you are NetBSD developer, optional otherwise.
 1269: 
 1270: [^4]: Might not exist if the driver is only a simple passthrough from a specific bus to another MI driver.
 1271: 
 1272: [^5]: Omitted if not needed.
 1273: 
 1274: [^6]: although you’ll often have to use bus\_size\_t
 1275: 
 1276: [^7]: It’s also possible to define a major in a machine-independent way in src/sys/conf/majors

CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb