.TL Regenerating System Software .AU Charles B. Haley Dennis. M. Ritchie .AI .MH .PP This document has been updated to include the modifications to .UX version seven by the Unix Systems Engineering group at DIGITAL. .sp 3 .ce 1 Unix/v7m Release 2.1 .sp 2 .ce 3 Fred Canter .AI .MK .sp 2 .PP Digital Equipment Corporation assumes no responsibilities for this software and makes no warranties or guaranties as to its suitability or completeness. .bp .SH Introduction .PP This document discusses how to assemble or compile various parts of the unix system software. This may be necessary because a command or library is accidentally deleted or otherwise destroyed; also, it may be desirable to install a modified version of some command or library routine. A few commands depend to some degree on the current configuration of the system; thus in any new system modifications to some commands are advisable. Most of the likely modifications relate to the standard disk devices contained in the system. For example, the df(1) (`disk free') command has built into it the names of the standardly present disk storage drives (e.g. `/dev/rf0', `/dev/rp0'). Df(1) takes an argument to indicate which disk to examine, but it is convenient if its default argument is adjusted to reflect the ordinarily present devices. The companion document `Setting up UNIX' discusses which commands are likely to require changes. Several of the command sources `include' the file <sys/param.h>, it may be necessary to recompile some or all of these commands if the system tunable parameters are changed, see System Tuning below. .bp .SH Where Commands and Subroutines Live .PP The source files for commands and subroutines reside in several subdirectories of the directory /usr/src. These subdirectories, and a general description of their contents, are .IP cmd 12 Source files for commands. .IP libc/stdio 12 Source files making up the `standard i/o package'. .IP libc/sys 12 Source files for the C system call interfaces. .IP libc/gen 12 Source files for most of the remaining routines described in section 3 of the manual. .IP libc/crt 12 Source files making up the C runtime support package, as in call save-return and long arithmetic. .IP libc/csu 12 Source for the C startup routines. .IP games 12 Source for (some of) the games. No great care has been taken to try to make it obvious how to compile these; treat it as a game. .IP libF77 12 Source for the Fortran 77 runtime library, exclusive of IO. .IP libI77 12 Source for the Fortran 77 IO runtime routines. .IP libdbm 12 Source for the `data-base manager' package .I dbm (3). .IP libfpsim 12 Source for the floating-point simulator routine. .IP libm 12 Source for the mathematical library. .IP libplot 12 Source for plotting routines. .bp .SH Commands .PP The regeneration of most commands is straightforward. The `cmd' directory will contain either a source file for the command or a subdirectory containing the set of files that make up the command. If it is a single file the command .DS cd /usr/src/cmd cmake cmd_name .DE suffices. (Cmd_name is the name of the command you are playing with.) The result of the cmake command will be an executable version. If you type .DS cmake \-cp cmd_name .DE the result will be copied to /bin (or perhaps /etc or other places if appropriate). .PP The cmake command has been modified to make the alternate version of certian commands required for operation of unix version seven on the PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60 processors. These commands are compiled by .DS cmake cmd_name40 .DE and are named as follows: .DS dcheck40 dump40 dumpdir40 icheck40 ncheck40 restor40 .DE .bp .PP Prior to making a command the correct version of the parameter file must be copied to /usr/include/sys/param.h, use /sys/h/param_ov.h for non separate I and D space CPUs and /sys/h/param_id.h for separate I and D space CPUs. In order to simplify this process a `makefile' has been provided, which automatically selects the correct parameter file, recompiles all necessary commands, and installs them in /bin or /etc as cmd_name40 or cmd_name70. For non separate I and D space CPUs use: .DS cd /usr/src/cmd make cmd40 .DE For separate I and D space CPUs use: .DS cd /usr/src/cmd make cmd70 .DE The command `make all' can be used to make both cmd40 and cmd70. The `makefile' in /bin can then be used to copy desired version of each command to its normal name, i.e., dump40 or dump70 to dump, etc. .PP If the source files are in a subdirectory there will be a `makefile' (see make(1)) to control the regeneration. After changing to the proper directory (cd(1)) you type one of the following: .IP "make all" 12 The program is compiled and loaded; the executable is left in the current directory. .IP "make cp" 12 The program is compiled and loaded, and the executable is installed. Everything is cleaned up afterwards; for example .o files are deleted. .IP "make cmp" 12 The program is compiled and loaded, and the executable is compared against the one in /bin. .PP Some of the makefiles have other options. Print (cat(1)) the ones you are interested in to find out. .PP The makefile for the tar command has been updated to make tar40, which is requried for non separate I & D space processors. .PP There are now six versions of `adb', in order to deal with the various types of unix kernels and the absence of floating point hardware. Refer to the `README' file in `/usr/src/cmd/adb' for information on how to select the appropriate version of adb and how to generate it. .bp .SH The Assembler .PP The assembler consists of two executable files: /bin/as and /lib/as2. The first is the 0-th pass: it reads the source program, converts it to an intermediate form in a temporary file `/tmp/atm0?', and estimates the final locations of symbols. It also makes two or three other temporary files which contain the ordinary symbol table, a table of temporary symbols (like 1:) and possibly an overflow intermediate file. The program /lib/as2 acts as an ordinary multiple pass assembler with input taken from the files produced by /bin/as. .PP The source files for /bin/as are named `/usr/src/cmd/as/as1?.s' (there are 9 of them); /lib/as2 is produced from the source files `/usr/src/cmd/as/as2?.s'; they likewise are 9 in number. Considerable care should be exercised in replacing either component of the assembler. Remember that if the assembler is lost, the only recourse is to replace it from some backup storage; a broken assembler cannot assemble itself. .PP There is now a second assembler `ovas', which is used by the overlay C compiler to generate the overlay text unix kernel. The `makefile' in `/usr/src/cmd/as' has been modified to make the overlay assembler. The file `/usr/src/cmd/as/README' contains more information on the overlay assembler. .SH The C Compiler .PP There is now an overlay C compiler, used to generate the overlay text unix kernel. The overlay C compiler and its make procedure have been integrated with the standard C compiler. The file `/usr/src/cmd/c/README' further explains the overlay C compiler and its generation. .PP The C compiler consists of seven routines: `/bin/cc', which calls the phases of the compiler proper, the compiler control line expander `/lib/cpp', the assembler (`as'), and the loader (`ld'). The phases of the C compiler are `/lib/c0' or `/lib/ovc0', which is the first phase of the compiler; `/lib/c1', which is the second phase of the compiler; and `/lib/c2', which is the optional third phase optimizer. The loss of the C compiler is as serious as that of the assembler. .bp .PP The source for /bin/cc resides in `/usr/src/cmd/cc.c'. Its loss alone (or that of c2) is not fatal. If needed, prog.c can be compiled by .DS /lib/cpp prog.c >temp0 /lib/c0 temp0 temp1 temp2 /lib/c1 temp1 temp2 temp3 as \- temp3 ld \-n /lib/crt0.o a.out \-lc .DE .PP The source for the compiler proper is in the directory /usr/src/cmd/c. The first phase (/lib/c0) or (/lib/ovc0) is generated from the files c00.c, ..., c05.c, which must be compiled by the C compiler. There is also c0.h, a header file .I included by the C programs of the first phase. To make a new /lib/c0 and /lib/ovc0 use .DS make -f mfnov c0 make -f mfov ovc0 .DE Before installing the new c0s, it is prudent to save the old ones someplace. .PP The second phase of C (/lib/c1) is generated from the source files c10.c, ..., c13.c, the include-file c1.h, and a set of object-code tables combined into table.o. To generate a new second phase use .DS make -f mfov c1 .DE It is likewise prudent to save c1 before installing a new version. In fact in general it is wise to save the object files for the C compiler so that if disaster strikes C can be reconstituted without a working version of the compiler. .PP In a similar manner, the third phase of the C compiler (/lib/c2) is made up from the files c20.c and c21.c together with c2.h, and is compiled by the command: .DS make -f mfov c2 .DE Its loss is not critical since it is completely optional. .PP The set of tables mentioned above is generated from the file table.s. This `.s' file is not in fact assembler source; it must be converted by use of the .I cvopt program, whose source and object are located in the C directory. Normally this is taken care of by make(1). You might want to look at the makefile to see what it does. .bp .SH UNIX .PP The source and object programs for UNIX are kept in the subdirectories of .I /sys. In the subdirectory .I h there are several files ending in `.h'; these are header files which are picked up (via `#include ...') as required by each system module. The file param.h contains the parameters for unix, there are two versions of this file, param_ov.h for the PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60 and param_id.h for the PDP11/44, PDP11/45, PDP11/55, and PDP11/70. The subdirectory .I dev consists mostly of the device drivers together with a few other things. The subdirectory .I sys is the rest of the system. The files LIB1_id in sys and LIB2_id in dev are archives (ar(1)) which contain the object versions of the routines in the directory, for the separate I & D space processors. The overlay text kernel object modules are in ovdev and the overlay system objects and a partial archive (LIB1_ov) are in ovsys. .PP Subdirectory .I conf contains the files which control device configuration of the system. .I L.s specifies the contents of the interrupt vectors; .I c.c contains the tables which relate device numbers to handler routines. A third file, mch_ov.s or mch_id.s , contains all the machine-language code in the system. A fourth file, .I mch0.s , is generated by mkconf(1) and contains flags indicating what sort of tape drive is available for taking crash dumps. It also specifies the device address of the tape controller used for crash dumps. The mch0.s file also contains a parameter for controlling the inclusion of floating point support in the machine language code. .bp .PP The first step in the unix system generation process is to select the unix kernel that is most appropriate for your type of processor, there are three. The separate I & D space kernel `unix_id' is used with the PDP11/44, PDP11/45, PDP11/55, and PDP11/70 processors. The overlay text kernel `unix_ov' is used with the PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60 processors. The `unix_i' kernel is only used for the preconfigured unix systems needed to initially load unix onto the system disk from the distribution tape. After the type of kernel has been chosen, use the following procedure to make unix: .IP 1. Examine the appropriate parameter file (param_ov.h or param_id.h), you will find several `#define ...' statements used to control the inclusion of various features in the kernel. The features are ACCT, FP, SEP_ID, DH, MX, UBUSMAP, PARITY, and LCKPHYS, their meanings are explained by the comments in the parameter file. Features are excluded by commenting out the define statements in the parameter file. The system tuning parameters, NBUF, NPROC and the like, are also in the parameter file. It is not advisable to modify these for the first system generation, the best thing to do is make unix for your configuration and test it, then experiment with the tuning parameters. .IP 2. The device drivers contain statements defining the CSR address and the number of units to be supported. Check the drivers for the devices in your configuration to insure that these values are correct, edit the drivers if necessary. .bp .IP 3. You must insure that the `sys' and `dev' archives are up to date, the archives supplied with the system are current. If no changes to the drivers or the parameter file were made in steps 1 and 2, then no action is required. There are two methods of updating the archives (LIB1 and LIB2). The first is to recompile all the source files and recreate the archives as follows: .DS cd /sys/conf make all?? .DE where ?? is the CPU type, 23, 24, 34, 40, 44, 45, 55, 60, or 70. This would normally not be necessary unless the system tuning parameters in param.h are changed. The second method is to recompile only the source files that were chenged and rearchive them as follows: .DS cd /sys/conf mksys_id file1 file2 ... file6 or mkdev_id file1 file2 ... file6 .DE for unix_id or .DS cd /sys/conf mksys_ov file1 file2 ... file6 or mkdev_ov file1 file2 ... file6 .DE for unix_ov. As many as six source files may be recompiled at once, only the filename is typed, not the`.c'. These `mk' files automatically select the appropriate parameter file and copy it to param.h. For example if the hp and dz drivers were changed the commands would be: .DS cd /sys/conf mkdev_id hp dz .DE those drivers will be recompiled and replaced in the LIB2_id archive. If the parameter file was changed in step 1, the comments in that file indicate which source files must be recompiled. .bp .IP 4. Prepare a configuration file, named `unixconf' or something like that, which describes your system configuration. Use mkconf(1m) and the many existing `conf' files in /sys/conf as a guide. If the overlay text kernel is to be used, the `conf' file must contain the `ov' declaration. .IP 5. Run the mkconf program with the `conf' file as input: .DS mkconf <unixconf .DE mkconf will print a list of the configured devices and their vectors. .IP 6. Examine the core dump tape CSR address, in mch0.s to verify that is matches your hardware. You may need to edit the low core vector file `l.s' to correct the device interrupt vectors, in any case examine `l.s' to insure that the vectors match your configuration. It is wise to print the configuration tables, in the file `c.c', and verify that the correct devices are entered in the bdevsw and cdevsw tables. You will need a copy of `c.c' later on anyway. .IP 7. Use the `makefile', in /sys/conf to make unix as follows: .DS make unix?? .DE where ?? is the CPU type, 23, 24, 34, 40, 44, 45, 55, 60, 70. .IP 8. When the make is done, the new system is present in the current directory as `unix_ov' or `unix_id'. It should be tested before destroying the currently running `/unix', this is best done by doing something like .DS mv /unix /ounix mv unix_ov /unix .DE or .DS mv unix_id /unix .DE You must be super-user to move unix to the root. If the new system doesn't work, you can still boot `ounix' and come up (see boot(8)). When you have satisfied yourself that the new system works, remove /ounix. .bp .SH Installing new devices .PP Refer to mkconf(1m) and the `Unix/v7m Software Description' for information on what devices are supported by Unix/v7m. The information in this section is of general interest, however, the steps described below are only necessary if you need to add a new device that is not presently supported by mkconf(1m). .PP To install a new driver, compile it and put it into its library. The best way to put it into the library is to edit its name into the `mkdev' files in `/sys/conf' and the `mklib' files in /sys/dev, and then use `mkdev' to recompile and archive it. There is no LIB2 device driver library for the overlay kernel, `unix_ov'. .PP Next, the device's interrupt vector must be entered in l.s. This is probably already done by the routine mkconf(1), but if the device is esoteric or nonstandard you will have to massage l.s by hand. This involves placing a pointer to a callout routine and the device's priority level in the vector. Use some other device (like the console) as a guide. Notice that the entries in l.s must be in order as the assembler does not permit moving the location counter `.' backwards. The assembler also does not permit assignation of an absolute number to `.', which is the reason for the `. = ZERO+100' subterfuge. If a constant smaller than 16(10) is added to the priority level, this number will be available as the first argument of the interrupt routine. This stratagem is used when several similar devices share the same interrupt routine (as in dl11's). .PP If you have to massage l.s, be sure to add the code to actually transfer to the interrupt routine. Again use the console as a guide. The apparent strangeness of this code is due to running the kernel in separate I&D space. The .I call routine saves registers as required and prepares a C-style call on the actual interrupt routine named after the `jmp' instruction. When the routine returns, .I call restores the registers and performs an rti instruction. As an aside, note that external names in C programs have an underscore (`_') prepended to them. .bp .PP The second step which must be performed to add a device unknown to mkconf is to add it to the configuration table /sys/conf/c.c. This file contains two subtables, one for block-type devices, and one for character-type devices. Block devices include disks, DECtape, and magtape. All other devices are character devices. A line in each of these tables gives all the information the system needs to know about the device handler; the ordinal position of the line in the table implies its major device number, starting at 0. .PP There are four subentries per line in the block device table, which give its open routine, close routine, strategy routine, and device table. The open and close routines may be nonexistent, in which case the name `nulldev' is given; this routine merely returns. The strategy routine is called to do any I/O, and the device table contains status information for the device. .PP For character devices, each line in the table specifies a routine for open, close, read, and write, and one which sets and returns device-specific status (used, for example, for stty and gtty on typewriters). If there is no open or close routine, `nulldev' may be given; if there is no read, write, or status routine, `nodev' may be given. Nodev sets an error flag and returns. .PP The final step which must be taken to install a device is to make a special file for it. This is done by mknod(1), to which you must specify the device class (block or character), major device number (relative line in the configuration table) and minor device number (which is made available to the driver at appropriate times). .PP The documents `Setting up Unix' and `The Unix IO system' may aid in comprehending these steps. .bp .SH The Library libc.a .PP The library /lib/libc.a is where most of the subroutines described in sections 2 and 3 of the manual are kept. This library can be remade using the following commands: .DS cd /usr/src/libc sh compall sh mklib mv libc.a /lib/libc.a .DE If single routines need to be recompiled and replaced, use .DS cc \-c \-O x.c ar vr /lib/libc.a x.o rm x.o .DE The above can also be used to put new items into the library. See ar(1), lorder(1), and tsort(1). .PP The routines in /usr/src/cmd/libc/csu (C start up) are not in libc.a. These are separately assembled and put into /lib. The commands to do this are .DS cd /usr/src/libc/csu as \- x.s mv a.out /lib/x .DE where x is the routine you want. .SH Other Libraries .PP Likewise, the directories containing the source for the other libraries have files compall (that recompiles everything) and mklib (that recreates the library). .bp .SH System Tuning .PP There are several tunable parameters in the system. These set the size of various tables and limits. They are found in the file /sys/h/param.h as manifests (`#define's), remember that there are two versions of this file, param_ov.h and param_id.h. Their values are rather generous in the system as distributed. Our typical maximum number of users is about 20, but there are many daemon processes. The values of the parameters in the param_ov.h file are set for about 10 users. .PP When any parameter is changed, it is prudent to recompile the entire system, as discussed above. A brief discussion of each follows: .IP NBUF 12 This sets the size of the disk buffer cache. Each buffer is 512 bytes. This number should be around 25 plus NMOUNT, or as big as can be if the above number of buffers cause the system to not fit in memory. .IP NFILE 12 This sets the maximum number of open files. An entry is made in this table every time a file is `opened' (see open(2), creat(2)). Processes share these table entries across forks (fork(2)). This number should be about the same size as NINODE below. (It can be a bit smaller.) .IP NMOUNT 12 This indicates the maximum number of mounted file systems. Make it big enough that you don't run out at inconvenient times. .IP MAXMEM 12 This sets an administrative limit on the amount of memory a process may have. It is set automatically if the amount of physical memory is small, and thus should not need to be changed. .IP MAXUPRC 12 This sets the maximum number of processes that any one user can be running at any one time. This should be set just large enough that people can get work done but not so large that a user can hog all the processes available (usually by accident!). .IP NPROC 12 This sets the maximum number of processes that can be active. It depends on the demand pattern of the typical user; we seem to need about 8 times the number of terminals. .DE .IP NINODE 12 This sets the size of the inode table. There is one entry in the inode table for every open device, current working directory, sticky text segment, open file, and mounted device. Note that if two users have a file open there is still only one entry in the inode table. A reasonable rule of thumb for the size of this table is .DS NPROC + NMOUNT + (number of terminals) .DE .IP SSIZE 12 The initial size of a process stack. This may be made bigger if commonly run processes have large data areas on the stack. .IP SINCR 12 The size of the stack growth increment. .IP NOFILE 12 This sets the maximum number of files that any one process can have open. 20 is plenty. .IP CANBSIZ 12 This is the size of the typewriter canonicalization buffer. It is in this buffer that erase and kill processing is done. Thus this is the maximum size of an input typewriter line. 256 is usually plenty. .IP CMAPSIZ 12 The number of fragments that memory can be broken into. This should be big enough that it never runs out. This parameter automatically grows as NPROC is increased. .IP SMAPSIZ 12 Same as CMAPSIZ except for secondary (swap) memory. .IP NCALL 12 This is the size of the callout table. Callouts are entered in this table when some sort of internal system timing must be done, as in carriage return delays for terminals. The number must be big enough to handle all such requests. .IP NTEXT 12 The maximum number of simultaneously executing pure programs. This should be big enough so as to not run out of space under heavy load. A reasonable rule of thumb is about .br .nf .sp (number of terminals) + (number of sticky programs) .br .fi .IP NCLIST 12 The number of clist segments. A clist segment is 6 characters. NCLIST should be big enough so that the list doesn't become exhausted when the machine is busy. The characters that have arrived from a terminal and are waiting to be given to a process live here. Thus enough space should be left so that every terminal can have at least one average line pending (about 30 or 40 characters). .IP TIMEZONE 12 The number of minutes westward from Greenwich. See `Setting Up UNIX'. .IP DSTFLAG 12 See `Setting Up UNIX' section on time conversion. .IP MSGBUFS 12 The maximum number of characters of system error messages saved. This is used as a circular buffer. .IP NCARGS 12 The maximum number of characters in an exec(2) arglist. This number controls how many arguments can be passed into a process. 5120 is practically infinite. .IP HZ 12 Set to the frequency of the system clock (e.g., 50 for a 50 Hz. clock). .bp .SH System tuning on non separate I & D space CPUs .PP The overlay text unix kernel is used for the non separate I & D space processors, PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60. The system tuning parameters are set for about 10 users, as follows: .DS NBUF = 14 NINODE = 100 NFILE = 80 NPROC = 70 NTEXT = 25 NCLIST = 125 .DE .PP The following table can be used as an aid when tuning unix version seven on the non separate I & D space CPUs. It lists the name of the paramter, the size increase in bytes of incrementing the parameter by one, and the source files which must be recompiled if the parameter is changed. .DS PARAMETER SIZE FILES NBUF 542 c.c, bio.c , main.c NINODE 74 c.c, alloc.c, iget.c, sys3.c NFILE 8 c.c, mx2.c, fio.c, iget.c NMOUNT 6 c.c, alloc.c, iget.c, nami.c, sys3.c (one system buffer per NMOUNT) MAXUPRC 0 c.c, sys1.c NOFILE 0 c.c, fio.c, slp.c, sys1.c, sys3.c CMAPSIZ 4 c.c, (any file that `includes' map.h) SMAPSIZ 4 c.c, (any file that `includes' map.h) NCALL 6 c.c, clock.c NPROC 28 c.c NTEXT 12 c.c, text.c NCLIST 8 c.c, prim.c (clists are larger on I & D space CPUs) .DE .PP The size of the overlay text kernel can be reduced by deselecting unneeded features in the param_ov.h, rebuilding the sys and dev archives, and remaking unix, as described in the section on generating UNIX above.