/* * Copyright 2020 faddenSoft * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ using System; using System.Collections.Generic; using System.Diagnostics; using Asm65; namespace SourceGen.Tools.Omf { /// /// Apple IIgs OMF file. /// /// /// OMF files are a series of segments. There is no file header or identifying information. /// In some cases the length is expected to be a multiple of 512 bytes, in others it isn't. /// Each segment is comprised of a header, followed by a series of records. /// /// There's no structural limitation on mixing and matching segments, whether different /// versions or different types. The file format provides a structure in which various /// things may be stored, but does not provide a way to tell an observer what is contained /// within (the ProDOS file type is supposed to do that). A given file may be a Load /// file (handled by the System Loader), Object file (fed to a linker), Library file /// (also fed to a linker), or Run-Time Library (used by both the linker and the loader). /// /// References: /// - IIgs Orca/M 2.0 manual. Appendix B documents OMF v0, v1, and v2.1 Load files. /// (This is included with Opus ][.) /// - "Apple IIgs Programmer's Workshop Reference". Chapter 7, page 228, describes /// OMF v1.0 and v2.0. /// - "Apple IIgs GS/OS Reference, for GS/OS System Software Version 5.0 and later". /// Appendix F describes OMF v2.1, and Chapter 8 has some useful information about /// how the loader works (e.g. page 205). /// - "Undocumented Secrets of the Apple IIGS System Loader" by Neil Parker, /// http://nparker.llx.com/a2/loader.html . Among other things it documents the /// contents of ExpressLoad segments, which I haven't found in an official reference. /// - Apple IIgs Tech Note #66, "ExpressLoad Philosophy". /// /// Related: /// - https://www.brutaldeluxe.fr/products/crossdevtools/omfanalyzer/ /// - https://github.com/fadden/ciderpress/blob/master/reformat/Disasm.cpp /// public class OmfFile { public const int MIN_FILE_SIZE = OmfSegment.MIN_HEADER_V0; public const int MAX_FILE_SIZE = (1 << 24) - 1; // cap at 16MB /// /// Overall file contents, determined by analysis. /// public enum FileKind { Unknown = 0, Load, // loadable files Object, // output of assembler/compiler, before linking Library, // static code library RunTimeLibrary, // dynamic shared library Indeterminate, // valid OMF, but type is indeterminate Foreign // not OMF, or not IIgs OMF } public FileKind OmfFileKind { get; private set; } /// /// File data. Files must be loaded completely into memory. /// private byte[] mFileData; /// /// List of segments. /// public List SegmentList { get; private set; } = new List(); /// /// List of strings to show in the message window. /// public List MessageList { get; private set; } = new List(); /// /// Constructor. /// /// File to analyze. public OmfFile(byte[] fileData) { Debug.Assert(fileData.Length >= MIN_FILE_SIZE && fileData.Length <= MAX_FILE_SIZE); mFileData = fileData; // Set to Unknown until analysis completes. OmfFileKind = FileKind.Unknown; } /// /// Analyzes the contents of an OMF file. /// public void Analyze(Formatter formatter) { OmfSegment.ParseResult result = DoAnalyze(formatter, false); if (result == OmfSegment.ParseResult.IsLibrary || result == OmfSegment.ParseResult.Failure) { // Failed; try again as a library. List firstFail = new List(MessageList); result = DoAnalyze(formatter, true); if (result == OmfSegment.ParseResult.Failure) { // Failed both ways. Report the failures from the non-library attempt. MessageList = firstFail; } } OmfFileKind = DetermineFileKind(); Debug.WriteLine("File kind is " + OmfFileKind); if (OmfFileKind == FileKind.Load) { GenerateRelocDicts(); } } /// /// Analyzes the contents of an OMF file as a library or non-library. /// private OmfSegment.ParseResult DoAnalyze(Formatter formatter, bool parseAsLibrary) { bool first = true; int offset = 0; int len = mFileData.Length; List msgs = new List(); while (len > 0) { OmfSegment.ParseResult result = OmfSegment.ParseHeader(mFileData, offset, parseAsLibrary, msgs, out OmfSegment seg); if (result == OmfSegment.ParseResult.Success) { if (!seg.ParseBody(formatter, msgs)) { OmfSegment.AddErrorMsg(msgs, offset, "parsing of segment " + seg.SegNum + " '" + seg.SegName + "' incomplete"); //result = OmfSegment.ParseResult.Failure; } } MessageList.Clear(); foreach (string str in msgs) { MessageList.Add(str); } if (result == OmfSegment.ParseResult.IsLibrary) { // Need to start over in library mode. Debug.WriteLine("Restarting in library mode"); return result; } else if (result == OmfSegment.ParseResult.Failure) { // Could be a library we failed to parse, could be a totally bad file. // If we were on the first segment, fail immediately so we can retry as // library. If not, it's probably not a library (assuming the Library // Dictionary segment appears first), but rather a partially-bad OMF. if (first) { return result; } break; } first = false; Debug.Assert(seg != null && seg.FileLength > 0); SegmentList.Add(seg); offset += seg.FileLength; len -= seg.FileLength; Debug.Assert(len >= 0); } Debug.WriteLine("Num segments = " + SegmentList.Count); return OmfSegment.ParseResult.Success; } /// /// Analyzes the file to determine the file kind. /// private FileKind DetermineFileKind() { if (SegmentList.Count == 0) { // Couldn't find a single valid segment, this is not OMF. return FileKind.Foreign; } // The rules: // * Load files may contain any kind of segment except LibraryDict. Their // segments must be LCONST/DS followed by relocation records // (INTERSEG, cINTERSEG, RELOC, cRELOC, SUPER). // * Object files have Code, Data, and DP/Stack segments, and may not contain // relocation records or ENTRY. // * Library files are like Object files, but start with a LibraryDict segment. // * Run-Time Library files have an initial segment with ENTRY records. (These // are not supported because I've never actually seen one... the few files I've // found with the RTL filetype ($B4) did not have any ENTRY records.) bool maybeLoad = true; bool maybeObject = true; bool maybeLibrary = true; foreach (OmfSegment omfSeg in SegmentList) { switch (omfSeg.Kind) { case OmfSegment.SegmentKind.Code: case OmfSegment.SegmentKind.Data: case OmfSegment.SegmentKind.DpStack: // Allowed anywhere. break; case OmfSegment.SegmentKind.JumpTable: case OmfSegment.SegmentKind.PathName: case OmfSegment.SegmentKind.Init: case OmfSegment.SegmentKind.AbsoluteBank: // Only allowed in Load. maybeObject = maybeLibrary = false; break; case OmfSegment.SegmentKind.LibraryDict: maybeLoad = false; maybeObject = false; break; default: Debug.Assert(false, "Unexpected segment kind " + omfSeg.Kind); break; } } // // Initial screening complete. Dig into the segment records to see if they're // compatible. // if (maybeLoad) { foreach (OmfSegment omfSeg in SegmentList) { if (!omfSeg.CheckRecords_LoadFile()) { maybeLoad = false; break; } } if (maybeLoad) { return FileKind.Load; } } if (maybeLibrary || maybeObject) { foreach (OmfSegment omfSeg in SegmentList) { if (!omfSeg.CheckRecords_ObjectOrLib()) { maybeLibrary = maybeObject = false; break; } } if (maybeObject) { return FileKind.Object; } else { return FileKind.Library; } } return FileKind.Indeterminate; } /// /// Generates relocation dictionaries for all code/data/init segments. /// private void GenerateRelocDicts() { Debug.Assert(OmfFileKind == FileKind.Load); foreach (OmfSegment omfSeg in SegmentList) { if (omfSeg.Kind == OmfSegment.SegmentKind.Code || omfSeg.Kind == OmfSegment.SegmentKind.Data || omfSeg.Kind == OmfSegment.SegmentKind.Init) { omfSeg.GenerateRelocDict(); } } } } }