|
|
1.1 root 1: .sp
2: .ce 100
3: A Package to Support VAX Compatability Mode on UNIX-32V
4: .sp
5: Arthur W. Wetzel
6: .br
7: 735 LIS Bldg
8: .br
9: Interdisciplinary Department of Information Science
10: .br
11: University of Pittsburgh
12: .br
13: Pittsburgh, Pa. 15260
14: (412)-624-5203
15: .ce 0
16: .sp 2
17: This is a brief description of a
18: package to support the execution of PDP-11 programs on
19: VAX UNIX-32V
20: or Berkeley VMUNIX
21: in compatability mode.
22: The major functions are to
23: .in +5
24: .sp
25: .ti -5
26: 1) allocate a block
27: of memory as the PDP-11 memory space (this must start at location 0),
28: .ti -5
29: 2) read compatability
30: mode program images into memory and lay them out properly (with arguments etc),
31: .ti -5
32: 3) actually handle the change to and from compatability mode,
33: .ti -5
34: 4) simulate system calls for what ever operating system is being simulated
35: and
36: .ti -5
37: 5) simulate floating point (FPU and FIS) instructions.
38: .sp
39: .in -5
40: Unfortunately programs requiring separated I/D space can not be run.
41: Loading of the package is rather slow since the entire process is about
42: 80K bytes
43: (64K is the PDP-11 space).
44: Once execution begins however, the speed is similar to a PDP-11/70.
45: There is considerable overhead for each exception condition
46: so that programs with a lot of system calls or especially with
47: floating point will be greatly slowed down.
48: Note that the text segment must be writable since the PDP-11
49: memory space is there.
50: .sp
51: Three quick changes to UNIX-32V and Berkeley VMUNIX
52: were made in the course of constructing this package.
53: .in 5
54: .sp
55: .ti -5
56: First, it is necessary to patch a bug in the original mchdep.c.
57: The bug in the sendsig routine is that the condition codes are masked out
58: of the psl before it is stacked when catching signals.
59: This affects all
60: programs not just compatability mode ones although is is not usually a frequent
61: problem execept in this application.
62: The mask which was 0xfff1
63: should be changed to 0xffff.
64: If this is not done, the condition codes
65: after a signal trap routine returns will always be cleared which can result
66: in many strange problems when condition codes are being checked in loops
67: or in this case after an "illegal instruction" trap.
68: This same bug remains in the
69: Third Berkeley Software Tape version of Virtual Memory UNIX.
70: .sp
71: .ti -5
72: Second, although it is easy to get into compatability mode one also
73: needs a way to get back when an exception condition arises.
74: This can be done
75: by changing another mask in the last line of the same routine.
76: The 0x1f
77: mask should be changed to 0x8000001f.
78: This clears the compatability mode bit
79: so that all signals are necessarily caught in native mode where native code
80: can do something about the situation.
81: .sp
82: .ti -5
83: Finally, if one wants compatibility
84: mode programs to have SETUID and SETGID status, there must be a way to change
85: the effective uid or gid without clobbering the real uid or gid.
86: This is
87: easily done by adding seteuid and setegid system calls to UNIX-32V.
88: My method
89: of doing this was to modify setuid and getuid so that the high order 16 bits
90: of the argument in the actual system calls is a flag (uids and gids are only
91: 16 bits in the low order part of the word) to indicate either a regular
92: setuid or getuid function or alternately a seteuid and setegid function.
93: Appropriate functions seteuid() and setegid() have been added to our libc.a
94: which automatically set up the flags while setuid() and setgid() insure
95: that the flags are zeroed.
96: .in -5
97: .sp
98: Most of the programming was done in late August 1979 with additions being made
99: occasionally thru August 1980.
100: Compilation procedures are specified in Makefile.
101: An effort was made to minimize the amount of assembly language coding
102: so that only two small assembler routines are found here.
103: One of these (memsiz.s) simply specifies
104: how much memory is being allocated for PDP-11 images
105: and makes it available through certain global variables.
106: The other assembler file (compat.s) handles the protocol for getting
107: into compatability mode at a certain pc and with a certain ps.
108: It also includes a getreg function which copies machine registers into
109: known places.
110: The heart of the entire package is runcompat.c which is used for all RTSs
111: (Run Time Systems).
112: The function main here simply checks for the existence of the
113: file to be executed and sets the required uid and gid.
114: The execute function actually copies the file to memory
115: and sets trap conditions.
116: Finally illtrap() catches illegal instructions and goes to the
117: code appropriate for what is found as the illegal instruction.
118: The bulk of the lines of C code are in unixtraps.c and dofloat.c
119: which do UNIX system calls in either version 6 or 7 format
120: and simulate floating point operations.
121: (Since PWB-UNIX is upward compatable with version 6, the version 6
122: system support also includes PWB sys calls.)
123: There are probably a number of bugs in the floating point simulation
124: code just waiting to be found.
125: If you are running programs which already include the PDP-11
126: floating point interpretation code, you may want to disable
127: dofloat as the illegal instructions can be caught and simulated
128: in the PDP-11 code.
129: To do this just make dofloat.o with "cc -c -O -DNOFPSIM dofloat.c".
130: .sp
131: A shell which will automatically invoke compatability mode programs
132: is in the modshell directory as difference listings from the original
133: UNIX-32V shell.
134: Most of the new code is in a new function compat.c.
135: The automatic recognition of PDP-11 UNIX version 6/7 programs
136: relies on the fact that the second word (16 bit) of a PDP-11
137: a.out file (text size) is nonzero whereas it is 0 for 32V a.outs.
138: No easy distinction can be made between version 6 and version 7 a.outs
139: so that a shell variable RTS sets up the name of a default Run
140: Time System.
141: On our system version 6 a.outs have been patched so that word 6 of the header
142: which is unused is a 1.
143: This hoaky? method seems to work just fine.
144: A program v6flag.c is in the modshell directory to do this.
145: .sp
146: One possible use of this package is to get programs like INGRES running
147: on the VAX without going through what appears to be a nontrivial
148: conversion effort.
149: There are two ways of running such programs.
150: Firstly if the shell is patched to automatically recognize and
151: run compatability mode a.outs (as in modshell), the PDP-11 a.out files
152: for the program can be just put on the system with their
153: normal names and run as usual.
154: Note however that you will be using the UNIX-32V shell so that
155: any shell files from PDP-11 version 6 will have to be modified
156: for this to work correctly with something like INGRES.
157: The second approach
158: is to make a directory hierarchy somewhere which corresponds
159: to what would be on a PDP-11 including the appropriate PDP-11 shell.
160: In that case just execute that shell in compatability mode with
161: the root directory set to the top of the PDP-11 hierarchy.
162: This is the quickest way to get something going in a hurry since no
163: changes are required to existing PDP-11 code or shell files.
164: .sp
165: Emulation of RT-11 system calls provided by Dan Strick are not being
166: distributed at this time.
167: .sp
168: Please foreward any comments, bug fixes or quick questions to the author
169: at the above address.
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.