1i386 Micro Channel Architecture Support
   4MCA support is enabled using the CONFIG_MCA define.  A machine with a MCA
   5bus will have the kernel variable MCA_bus set, assuming the BIOS feature
   6bits are set properly (see arch/i386/boot/setup.S for information on
   7how this detection is done).
   9Adapter Detection
  12The ideal MCA adapter detection is done through the use of the
  13Programmable Option Select registers.  Generic functions for doing
  14this have been added in include/linux/mca.h and arch/i386/kernel/mca.c.
  15Everything needed to detect adapters and read (and write) configuration
  16information is there.  A number of MCA-specific drivers already use
  17this.  The typical probe code looks like the following:
  19        #include <linux/mca.h>
  21        unsigned char pos2, pos3, pos4, pos5;
  22        struct net_device* dev;
  23        int slot;
  25        if( MCA_bus ) {
  26                slot = mca_find_adapter( ADAPTER_ID, 0 );
  27                if( slot == MCA_NOTFOUND ) {
  28                        return -ENODEV;
  29                }
  30                /* optional - see below */
  31                mca_set_adapter_name( slot, "adapter name & description" );
  32                mca_set_adapter_procfn( slot, dev_getinfo, dev );
  34                /* read the POS registers.  Most devices only use 2 and 3 */
  35                pos2 = mca_read_stored_pos( slot, 2 );
  36                pos3 = mca_read_stored_pos( slot, 3 );
  37                pos4 = mca_read_stored_pos( slot, 4 );
  38                pos5 = mca_read_stored_pos( slot, 5 );
  39        } else {
  40                return -ENODEV;
  41        }
  43        /* extract configuration from pos[2345] and set everything up */
  45Loadable modules should modify this to test that the specified IRQ and
  46IO ports (plus whatever other stuff) match.  See 3c523.c for example
  47code (actually, smc-mca.c has a slightly more complex example that can
  48handle a list of adapter ids).
  50Keep in mind that devices should never directly access the POS registers
  51(via inb(), outb(), etc).  While it's generally safe, there is a small
  52potential for blowing up hardware when it's done at the wrong time.
  53Furthermore, accessing a POS register disables a device temporarily.
  54This is usually okay during startup, but do _you_ want to rely on it?
  55During initial configuration, mca_init() reads all the POS registers
  56into memory.  mca_read_stored_pos() accesses that data.  mca_read_pos()
  57and mca_write_pos() are also available for (safer) direct POS access,
  58but their use is _highly_ discouraged.  mca_write_pos() is particularly
  59dangerous, as it is possible for adapters to be put in inconsistent
  60states (i.e. sharing IO address, etc) and may result in crashes, toasted
  61hardware, and blindness.
  63User level drivers (such as the AGX X server) can use /proc/mca/pos to
  64find adapters (see below).
  66Some MCA adapters can also be detected via the usual ISA-style device
  67probing (many SCSI adapters, for example).  This sort of thing is highly
  68discouraged.  Perfectly good information is available telling you what's
  69there, so there's no excuse for messing with random IO ports.  However,
  70we MCA people still appreciate any ISA-style driver that will work with
  71our hardware.  You take what you can get...
  73Level-Triggered Interrupts
  76Because MCA uses level-triggered interrupts, a few problems arise with
  77what might best be described as the ISA mindset and its effects on
  78drivers.  These sorts of problems are expected to become less common as
  79more people use shared IRQs on PCI machines.
  81In general, an interrupt must be acknowledged not only at the ICU (which
  82is done automagically by the kernel), but at the device level.  In
  83particular, IRQ 0 must be reset after a timer interrupt (now done in
  84arch/i386/kernel/time.c) or the first timer interrupt hangs the system.
  85There were also problems with the 1.3.x floppy drivers, but that seems
  86to have been fixed.
  88IRQs are also shareable, and most MCA-specific devices should be coded
  89with shared IRQs in mind.
  94/proc/mca is a directory containing various files for adapters and
  95other stuff.
  97        /proc/mca/pos           Straight listing of POS registers
  98        /proc/mca/slot[1-8]     Information on adapter in specific slot
  99        /proc/mca/video         Same for integrated video
 100        /proc/mca/scsi          Same for integrated SCSI
 101        /proc/mca/machine       Machine information
 103See Appendix A for a sample.
 105Device drivers can easily add their own information function for
 106specific slots (including integrated ones) via the
 107mca_set_adapter_procfn() call.  Drivers that support this are ESDI, IBM
 108SCSI, and 3c523.  If a device is also a module, make sure that the proc
 109function is removed in the module cleanup.  This will require storing
 110the slot information in a private structure somewhere.  See the 3c523
 111driver for details.
 113Your typical proc function will look something like this:
 115        static int
 116        dev_getinfo( char* buf, int slot, void* d ) {
 117                struct net_device* dev = (struct net_device*) d;
 118                int len = 0;
 120                len += sprintf( buf+len, "Device: %s\n", dev->name );
 121                len += sprintf( buf+len, "IRQ: %d\n", dev->irq );
 122                len += sprintf( buf+len, "IO Port: %#lx-%#lx\n", ... );
 123                ...
 125                return len;
 126        }
 128Some of the standard MCA information will already be printed, so don't
 129bother repeating it.  Don't try putting in more than 3K of information.
 131Enable this function with:
 132        mca_set_adapter_procfn( slot, dev_getinfo, dev );
 134Disable it with:
 135        mca_set_adapter_procfn( slot, NULL, NULL );
 137It is also recommended that, even if you don't write a proc function, to
 138set the name of the adapter (i.e. "PS/2 ESDI Controller") via
 139mca_set_adapter_name( int slot, char* name ).
 141MCA Device Drivers
 144Currently, there are a number of MCA-specific device drivers.
 1461) PS/2 SCSI
 147        drivers/scsi/ibmmca.c
 148        drivers/scsi/ibmmca.h
 149   The driver for the IBM SCSI subsystem.  Includes both integrated
 150   controllers and adapter cards.  May require command-line arg
 151   "ibmmcascsi=io_port" to force detection of an adapter.  If you have a
 152   machine with a front-panel display (i.e. model 95), you can use
 153   "ibmmcascsi=display" to enable a drive activity indicator.
 1552) 3c523
 156        drivers/net/3c523.c
 157        drivers/net/3c523.h
 158   3Com 3c523 Etherlink/MC ethernet driver.
 1603) SMC Ultra/MCA and IBM Adapter/A
 161        drivers/net/smc-mca.c
 162        drivers/net/smc-mca.h
 163        Driver for the MCA version of the SMC Ultra and various other
 164        OEM'ed and work-alike cards (Elite, Adapter/A, etc).
 1664) NE/2
 167        driver/net/ne2.c
 168        driver/net/ne2.h
 169        The NE/2 is the MCA version of the NE2000.  This may not work
 170        with clones that have a different adapter id than the original
 171        NE/2.
 1735) Future Domain MCS-600/700, OEM'd IBM Fast SCSI Adapter/A and
 174   Reply Sound Blaster/SCSI (SCSI part)
 175        Better support for these cards than the driver for ISA.
 176   Supports multiple cards with IRQ sharing.
 178Also added boot time option of scsi-probe, which can do reordering of
 179SCSI host adapters. This will direct the kernel on the order which
 180SCSI adapter should be detected. Example:
 181  scsi-probe=ibmmca,fd_mcs,adaptec1542,buslogic
 183The serial drivers were modified to support the extended IO port range
 184of the typical MCA system (also #ifdef CONFIG_MCA).
 186The following devices work with existing drivers:
 1871) Token-ring
 1882) Future Domain SCSI (MCS-600, MCS-700, not MCS-350, OEM'ed IBM SCSI)
 1893) Adaptec 1640 SCSI (using the aha1542 driver)
 1904) Bustek/Buslogic SCSI (various)
 1915) Probably all Arcnet cards.
 1926) Some, possibly all, MCA IDE controllers.
 1937) 3Com 3c529 (MCA version of 3c509) (patched)
 1958) Intel EtherExpressMC  (patched version)
 196   You need to have CONFIG_MCA defined to have EtherExpressMC support.
 1979) Reply Sound Blaster/SCSI (SB part) (patched version)
 199Bugs & Other Weirdness
 202NMIs tend to occur with MCA machines because of various hardware
 203weirdness, bus timeouts, and many other non-critical things.  Some basic
 204code to handle them (inspired by the NetBSD MCA code) has been added to
 205detect the guilty device, but it's pretty incomplete.  If NMIs are a
 206persistent problem (on some model 70 or 80s, they occur every couple
 207shell commands), the CONFIG_IGNORE_NMI flag will take care of that.
 209Various Pentium machines have had serious problems with the FPU test in
 210bugs.h.  Basically, the machine hangs after the HLT test.  This occurs,
 211as far as we know, on the Pentium-equipped 85s, 95s, and some PC Servers.
 212The PCI/MCA PC 750s are fine as far as I can tell.  The ``mca-pentium''
 213boot-prompt flag will disable the FPU bug check if this is a problem
 214with your machine.
 216The model 80 has a raft of problems that are just too weird and unique
 217to get into here.  Some people have no trouble while others have nothing
 218but problems.  I'd suspect some problems are related to the age of the
 219average 80 and accompanying hardware deterioration, although others
 220are definitely design problems with the hardware.  Among the problems
 221include SCSI controller problems, ESDI controller problems, and serious
 222screw-ups in the floppy controller.  Oh, and the parallel port is also
 223pretty flaky.  There were about 5 or 6 different model 80 motherboards
 224produced to fix various obscure problems.  As far as I know, it's pretty
 225much impossible to tell which bugs a particular model 80 has (other than
 226triggering them, that is).
 228Drivers are required for some MCA memory adapters.  If you're suddenly
 229short a few megs of RAM, this might be the reason.  The (I think) Enhanced
 230Memory Adapter commonly found on the model 70 is one.  There's a very
 231alpha driver floating around, but it's pretty ugly (disassembled from
 232the DOS driver, actually).  See the MCA Linux web page (URL below)
 233for more current memory info.
 235The Thinkpad 700 and 720 will work, but various components are either
 236non-functional, flaky, or we don't know anything about them.  The
 237graphics controller is supposed to be some WD, but we can't get things
 238working properly.  The PCMCIA slots don't seem to work.  Ditto for APM.
 239The serial ports work, but detection seems to be flaky.
 243A whole pile of people have contributed to the MCA code.  I'd include
 244their names here, but I don't have a list handy.  Check the MCA Linux
 245home page (URL below) for a perpetually out-of-date list.
 248MCA Linux Home Page:
 250Christophe Beauregard
 255Appendix A: Sample /proc/mca
 257This is from my model 8595.  Slot 1 contains the standard IBM SCSI
 258adapter, slot 3 is an Adaptec AHA-1640, slot 5 is a XGA-1 video adapter,
 259and slot 7 is the 3c523 Etherlink/MC.
 262Model Id: 0xf8
 263Submodel Id: 0x14
 264BIOS Revision: 0x5
 267Slot 1: ff 8e f1 fc a0 ff ff ff  IBM SCSI Adapter w/Cache
 268Slot 2: ff ff ff ff ff ff ff ff  
 269Slot 3: 1f 0f 81 3b bf b6 ff ff  
 270Slot 4: ff ff ff ff ff ff ff ff  
 271Slot 5: db 8f 1d 5e fd c0 00 00  
 272Slot 6: ff ff ff ff ff ff ff ff  
 273Slot 7: 42 60 ff 08 ff ff ff ff  3Com 3c523 Etherlink/MC
 274Slot 8: ff ff ff ff ff ff ff ff  
 275Video : ff ff ff ff ff ff ff ff  
 276SCSI  : ff ff ff ff ff ff ff ff  
 279Slot: 1
 280Adapter Name: IBM SCSI Adapter w/Cache
 281Id: 8eff
 282Enabled: Yes
 283POS: ff 8e f1 fc a0 ff ff ff 
 284Subsystem PUN: 7
 285Detected at boot: Yes
 288Slot: 3
 289Adapter Name: Unknown
 290Id: 0f1f
 291Enabled: Yes
 292POS: 1f 0f 81 3b bf b6 ff ff 
 295Slot: 5
 296Adapter Name: Unknown
 297Id: 8fdb
 298Enabled: Yes
 299POS: db 8f 1d 5e fd c0 00 00 
 302Slot: 7
 303Adapter Name: 3Com 3c523 Etherlink/MC
 304Id: 6042
 305Enabled: Yes
 306POS: 42 60 ff 08 ff ff ff ff 
 307Revision: 0xe
 308IRQ: 9
 309IO Address: 0x3300-0x3308
 310Memory: 0xd8000-0xdbfff
 311Transceiver: External
 312Device: eth0
 313Hardware Address: 02 60 8c 45 c4 2a