[DWARFDebugLine] Use truncating data extractors for prologue parsing

[labath]

Summary

This makes the code easier to reason about, as it will behave the same
way regardless of whether there is any more data coming after the
presumed end of the prologue.

[probinson]

llvm/include/llvm/DebugInfo/DWARF/DWARFDebugLine.h

132

The truncating constructor takes a const DWARFDataExtractor& so I don't understand why this parameter had to change.

[dblaikie]

Some nice cleanup indeed - thanks! Please wait for @jhenderson to also approve, since he's been doing more work in this area. (never sure whether to approve or not in this case - not sure if the official "approval" might then remove it from other people's review queue, etc)

[MaskRay]

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

457

At a first glance, 2 : 8 looks a bit magical. This may deserve a comment.

[probinson]

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

458

The version check should be < rather than != here.

465

And here, Version < 5 please.

[jhenderson]

In D77557#1964697, @dblaikie wrote:

Some nice cleanup indeed - thanks! Please wait for @jhenderson to also approve, since he's been doing more work in this area. (never sure whether to approve or not in this case - not sure if the official "approval" might then remove it from other people's review queue, etc)

FWIW, in my case, I just use my emails as my review queue, so I don't mind others approving too.

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

195

I wonder if it's worth a comment at the start here saying something to explain why the formatting of the following bytes is weird. Perhaps something like "The following bytes would be read as part of the prologue header, were data extraction not stopped once the data end is reached, hence they are formatted like a header".

Something similar applies for the other changes in this file.

324

I think this comment is missing a word somewhere. Probably "truncated"? Also missing a trailing full stop. Also "md5" -> "MD5".

llvm/test/tools/llvm-dwarfdump/X86/debug_line_invalid.test

115

This should probably show that the last opcode within range is read successfully in addition to showing the past-the-end one is recorded as zero.

143

Missing trailing full stop.

This may not have been clear from the test comments before, but this and the previous case were supposed to work in tandem, because the invalid MD5 form message was reported in both cases, but not the "not at end" message IIRC.

I feel like the input for this case probably should be updated to not have a malformed MD5 form. Is there any interaction any more to do with continuing to parse that these 2 cases now don't need to cover?

206

How straightforward would it be to provide the offset as extra context here, like we do with "unexpected end of data"?

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

470

I'm not sure where this error is coming from (I don't recall seeing it before), but it's got a spurious trailing full stop, which is inconsistent with the other error messages. It might want fixing separately I guess.

478–482

What happened to this error message?

[labath]

• rebase on latest master
• this does not attempt to address any comments yet. I'll do that separately.

[labath]

• address reviewer comments (I am very sorry about the long delay)

llvm/include/llvm/DebugInfo/DWARF/DWARFDebugLine.h

132

It wasn't strictly necessary, but with at const & argument, I'd have to create new object to hold the truncated extractor -- this way I can reassign it. It seemed like a good idea to avoid having two data extractors floating around...

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

195

Actually, I was thinking of just deleting this stuff as it is no longer interesting to test. I wanted to do it as a separate commit, as that will change the offsets of everything, making it hard to see what changed. WDYT?

llvm/test/tools/llvm-dwarfdump/X86/debug_line_invalid.test

143

I've changed this test to use a valid form. I don't think more test are needed -- this is pretty generic code, and form value extraction is tested in DWARFFormValueTest.cpp.

206

That was done with D80799.

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

470

there are a couple of messages like this in parseV5DirFileTables so I'll fix those separately. The reason you haven't seen it before was that before D77308 DWARFFormValue::extractValue was always "successful".

478–482

It's just gone because the error has changed. Now it is not possible for parsing to run off the end of the prologue, so the failure mode is different.

[jhenderson]

llvm/lib/DebugInfo/DWARF/DWARFDebugLine.cpp

383

We probably also want to truncate it based on the unit length first, before reading the version etc. I've actually got plans for a local change related to that, but as you're working in this area, I'm happy for you to do it as part of this or a subsequent change, if you want, rather than me doing it.

440–442

A number of our other error messages explicitly use "at offset 0x..." for the offset part of the message. Perhaps we should update this message whilst you're here?

Alternatively, perhaps we should remove "offset" in all other locations where it is already present (as a separate change). I think either can be argued for - the former for complete clarity, the latter for conciseness.

447

We probably still want this line - if you don't update it, and the prologue finishes early, you'll get the recoverable error, but then parsing will continue from the point it got to rather than the start of the table body, as claimed by the length. In general, I think we should prefer following the line table's stated lengths where possible.

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

195

I'm happy to remove any unnecessary additional test cases, where they provide no new value, but we need to be slightly careful we don't lose coverage of parsing the file/directory name tables. If a test case doesn't cover a unique code path, then I'm good for it to be deleted, and as a separate patch is fine.

[labath]

• truncate the extractor based on the unit length first (and fix one test which had a bogus unit length)
• add "offset"s

llvm/lib/DebugInfo/DWARF/DWARFDebugLine.cpp

383

I can do that now.

440–442

I've added the offsets. I don't have any opinions on whether should include the "offset" part or not.

447

The reason it's not needed is because there's a *OffsetPtr = DebugLineOffset + Prologue.getLength(); down on line 738.

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

195

I'm not talking about deleting entire test cases -- just the bits of line tables that are prologue extensions. They were needed/interesting to test when the prologue parsing would not respect prologue boundaries, but now I don't think they serve any purpose and I think we can just put a simple DW_LNE_end_sequence as the line table contents.

[jhenderson]

Looks good to me, but I'll hold off approval so you can touch up the tests.

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

195

The "standard" contents are a set address and end sequence for reasons I forget these days, so we should probably keep just that much to match the other cases. Beyond that, I'm happy for things to be simplified where they don't add value.

[labath]

• remove the extraneous opcodes and adjust offsets
• adjust tests for the "offset" addition in the previous revision

[jhenderson]

LGTM, with one fix to the error message.

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

434

prologue -> prologue end

[jhenderson]

llvm/lib/DebugInfo/DWARF/DWARFDebugLine.cpp

440

Oops, just spotted a mistake here: "reaching the prologue at" -> "reaching the prologue end at"

Are you okay to fix?

[labath]

llvm/lib/DebugInfo/DWARF/DWARFDebugLine.cpp

440

Done in 0b781db9087. Thanks for catching that.

llvm/include/llvm/DebugInfo/DWARF/DWARFDebugLine.h

struct Prologue {
     bool
     getFileNameByIndex(uint64_t FileIndex, StringRef CompDir,
                        DILineInfoSpecifier::FileLineInfoKind Kind,
                        std::string &Result,
                        sys::path::Style Style = sys::path::Style::native) const;
 
     void clear();
     void dump(raw_ostream &OS, DIDumpOptions DumpOptions) const;
-    Error parse(const DWARFDataExtractor &DebugLineData, uint64_t *OffsetPtr,
+    Error parse(DWARFDataExtractor Data, uint64_t *OffsetPtr,
                 function_ref<void(Error)> RecoverableErrorHandler,
                 const DWARFContext &Ctx, const DWARFUnit *U = nullptr);
   };
 
   /// Standard .debug_line state machine structure.
   struct Row {
     explicit Row(bool DefaultIsStmt = false);
 

llvm/lib/DebugInfo/DWARF/DWARFDebugLine.cpp

for (uint32_t I = 0; I != FileNames.size(); ++I) {
       }
     }
   }
 }
 
 // Parse v2-v4 directory and file tables.
 static Error
 parseV2DirFileTables(const DWARFDataExtractor &DebugLineData,
-                     uint64_t *OffsetPtr, uint64_t EndPrologueOffset,
+                     uint64_t *OffsetPtr,
                      DWARFDebugLine::ContentTypeTracker &ContentTypes,
                      std::vector<DWARFFormValue> &IncludeDirectories,
                      std::vector<DWARFDebugLine::FileNameEntry> &FileNames) {
-  bool Terminated = false;
   Error Err = Error::success();
   while (true) {
     StringRef S = DebugLineData.getCStrRef(OffsetPtr, &Err);
-    if (Err || *OffsetPtr > EndPrologueOffset)
+    if (S.empty())
       break;
-    if (S.empty()) {
-      Terminated = true;
-      break;
-    }
     DWARFFormValue Dir =
         DWARFFormValue::createFromPValue(dwarf::DW_FORM_string, S.data());
     IncludeDirectories.push_back(Dir);
   }
 
-  if (!Terminated) {
+  if (Err) {
     consumeError(std::move(Err));
     return createStringError(errc::invalid_argument,
                              "include directories table was not null "
                              "terminated before the end of the prologue");
   }
 
-  Terminated = false;
   while (true) {
     StringRef Name = DebugLineData.getCStrRef(OffsetPtr, &Err);
-    if (Err || *OffsetPtr > EndPrologueOffset)
-      break;
-    if (Name.empty()) {
-      Terminated = true;
+    if (Name.empty())
       break;
-    }
 
     DWARFDebugLine::FileNameEntry FileEntry;
     FileEntry.Name =
         DWARFFormValue::createFromPValue(dwarf::DW_FORM_string, Name.data());
     FileEntry.DirIdx = DebugLineData.getULEB128(OffsetPtr, &Err);
     FileEntry.ModTime = DebugLineData.getULEB128(OffsetPtr, &Err);
     FileEntry.Length = DebugLineData.getULEB128(OffsetPtr, &Err);
 
-    if (Err || *OffsetPtr > EndPrologueOffset)
+    if (Err)
       break;
     FileNames.push_back(FileEntry);
   }
 
   ContentTypes.HasModTime = true;
   ContentTypes.HasLength = true;
 
-  if (!Terminated) {
+  if (Err) {
     consumeError(std::move(Err));
     return createStringError(errc::invalid_argument,
                              "file names table was not null terminated before "
                              "the end of the prologue");
   }
 
   return Error::success();
 }
uint64_t DWARFDebugLine::Prologue::getLength() const {
   uint64_t Length = PrologueLength + sizeofTotalLength() +
                     sizeof(getVersion()) + sizeofPrologueLength();
   if (getVersion() >= 5)
     Length += 2; // Address + Segment selector sizes.
   return Length;
 }
 
 Error DWARFDebugLine::Prologue::parse(
-    const DWARFDataExtractor &DebugLineData, uint64_t *OffsetPtr,
+    DWARFDataExtractor DebugLineData, uint64_t *OffsetPtr,
     function_ref<void(Error)> RecoverableErrorHandler, const DWARFContext &Ctx,
     const DWARFUnit *U) {
   const uint64_t PrologueOffset = *OffsetPtr;
 
   clear();
   Error Err = Error::success();
   std::tie(TotalLength, FormParams.Format) =
       DebugLineData.getInitialLength(OffsetPtr, &Err);
assert((DebugLineData.getAddressSize() == 0 ||
             DebugLineData.getAddressSize() == getAddressSize()) &&
            "Line table header and data extractor disagree");
     SegSelectorSize = DebugLineData.getU8(OffsetPtr);
   }
 
   PrologueLength =
       DebugLineData.getRelocatedValue(sizeofPrologueLength(), OffsetPtr);
   const uint64_t EndPrologueOffset = PrologueLength + *OffsetPtr;
+  DebugLineData = DWARFDataExtractor(DebugLineData, EndPrologueOffset);
   MinInstLength = DebugLineData.getU8(OffsetPtr);
   if (getVersion() >= 4)
     MaxOpsPerInst = DebugLineData.getU8(OffsetPtr);
   DefaultIsStmt = DebugLineData.getU8(OffsetPtr);
   LineBase = DebugLineData.getU8(OffsetPtr);
   LineRange = DebugLineData.getU8(OffsetPtr);
   OpcodeBase = DebugLineData.getU8(OffsetPtr);
 
auto ReportInvalidDirFileTable = [&](Error E) {
     // below from executing.
     *OffsetPtr = EndPrologueOffset;
   };
   if (getVersion() >= 5) {
     if (Error E =
             parseV5DirFileTables(DebugLineData, OffsetPtr, FormParams, Ctx, U,
                                  ContentTypes, IncludeDirectories, FileNames))
       ReportInvalidDirFileTable(std::move(E));
-  } else if (Error E = parseV2DirFileTables(DebugLineData, OffsetPtr,
-                                            EndPrologueOffset, ContentTypes,
-                                            IncludeDirectories, FileNames))
+  } else if (Error E =
+                 parseV2DirFileTables(DebugLineData, OffsetPtr, ContentTypes,
     ReportInvalidDirFileTable(std::move(E));
 
   if (*OffsetPtr != EndPrologueOffset) {
     RecoverableErrorHandler(createStringError(
         errc::invalid_argument,
         "parsing line table prologue at 0x%8.8" PRIx64
         " should have ended at 0x%8.8" PRIx64 " but it ended at 0x%8.8" PRIx64,
         PrologueOffset, EndPrologueOffset, *OffsetPtr));
     *OffsetPtr = EndPrologueOffset;
   }
   return Error::success();
 }
 
 DWARFDebugLine::Row::Row(bool DefaultIsStmt) { reset(DefaultIsStmt); }
 
 void DWARFDebugLine::Row::postAppend() {
   Discriminator = 0;
   BasicBlock = false;
   PrologueEnd = false;
   EpilogueBegin = false;
 }

llvm/test/tools/llvm-dwarfdump/X86/Inputs/debug_line_malformed.s

 .asciz "dir1"           # Include table
 .asciz "dir2"
 .byte   0
 .asciz "file1"          # File table
 .byte   0, 0, 0
 .asciz "file2"
 .byte   1, 2
 .Lprologue_short_prologue_end:
-.byte   6               # Read as part of the prologue,
-                        # then later again as DW_LNS_negate_stmt.
-# Header end
+.byte   6               # DW_LNS_negate_stmt.
 .byte   0, 9, 2         # DW_LNE_set_address
 .quad   0x1122334455667788
 .byte   0, 1, 1         # DW_LNE_end_sequence
 .Lunit_short_prologue_end:
 
 # Over-long prologue.
 .long   .Lunit_long_prologue_end - .Lunit_long_prologue_start # unit length
 .Lunit_long_prologue_start:

 .byte   1               # Minimum Instruction Length
 .byte   1               # Maximum Operations per Instruction
 .byte   1               # Default is_stmt
 .byte   -5              # Line Base
 .byte   14              # Line Range
 .byte   13              # Opcode Base
 .byte   0, 1, 1, 1, 1, 0, 0, 0, 1, 0 # Standard Opcode Lengths
 .Linvalid_description_header_end0:
-# The bytes from here onwards will also be read as part of the main body.
-                        # --- Prologue interpretation --- | --- Main body interpretation ---
-.byte   0, 1            # More standard opcodes           | First part of DW_LNE_end_sequence
-# Directory table format
-.byte   1               # One element per directory entry | End of DW_LNE_end_sequence
-.byte   1               # DW_LNCT_path                    | DW_LNS_copy
-.byte   0x08            # DW_FORM_string                  | DW_LNS_const_add_pc
-# Directory table entries
-.byte   1               # 1 directory                     | DW_LNS_copy
-.asciz  "/tmp"          # Directory name                  | four special opcodes + start of DW_LNE_end_sequence
-# File table format
-.byte   1               # 1 element per file entry        | DW_LNE_end_sequence length
-.byte   1               # DW_LNCT_path                    | DW_LNE_end_sequence opcode
-.byte   0x08            # DW_FORM_string                  | DW_LNS_const_add_pc
-# File table entries
-.byte   1               # 1 file                          | DW_LNS_copy
-.asciz  "xyz"           # File name                       | three special opcodes + start of DW_LNE_set_address
-# Header end
+.byte   0, 1, 1         # DW_LNE_end_sequence
+.byte   1               # DW_LNS_copy
+.byte   0x08            # DW_LNS_const_add_pc
+.byte   1               # DW_LNS_copy
+.asciz  "/tmp"          # four special opcodes + start of DW_LNE_end_sequence
+.byte   1               # DW_LNE_end_sequence length
+.byte   1               # DW_LNE_end_sequence opcode
+.byte   0x08            # DW_LNS_const_add_pc
+.byte   1               # DW_LNS_copy
+.asciz  "xyz"           # three special opcodes + start of DW_LNE_set_address
 .byte   9, 2            # Remainder of DW_LNE_set_address
 .quad   0xbabb1ebabb1e
 .byte   0, 1, 1         # DW_LNE_end_sequence
 .Linvalid_description_end0:
 
 # V5 prologue ends during file table.
 .long   .Linvalid_file_end0-.Linvalid_file_start0   # Length of Unit
 .Linvalid_file_start0:

 .byte   1               # 1 directory
 .asciz  "/tmp"
 # File table format
 .byte   2               # 2 elements per file entry
 .byte   1               # DW_LNCT_path
 .byte   0x08            # DW_FORM_string
 .byte   2               # DW_LNCT_directory_index
 .Linvalid_file_header_end0:
-# The bytes from here onwards will also be read as part of the main body.
-                        # --- Prologue interpretation --- | --- Main body interpretation ---
-.byte   0x0b            # DW_FORM_data1                   | DW_LNS_set_epilogue_begin
-# File table entries
-.byte   1               # 1 file                          | DW_LNS_copy
-.asciz  "xyz"           # File name                       | 3 special opcodes + start of DW_LNE_end_sequence
-.byte   1               # Dir index                       | DW_LNE_end_sequence length
+.byte   0x0b            # DW_LNS_set_epilogue_begin
+.byte   1               # DW_LNS_copy
+.asciz  "xyz"           # 3 special opcodes + start of DW_LNE_end_sequence
+.byte   1               # DW_LNE_end_sequence length
 # Header end
 .byte   1               # DW_LNE_end_sequence opcode
 .byte   0, 9, 2         # DW_LNE_set_address
 .quad   0xab4acadab4a
 .byte   0, 1, 1         # DW_LNE_end_sequence
 .Linvalid_file_end0:
 
 # V5 prologue ends during directory table.

 .byte   0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1 # Standard Opcode Lengths
 # Directory table format
 .byte   1               # One element per directory entry
 .byte   1               # DW_LNCT_path
 .byte   0x08            # DW_FORM_string
 # Directory table entries
 .byte   1               # 1 directory
 .Linvalid_dir_header_end0:
-# The bytes from here onwards will also be read as part of the main body.
-                        # --- Prologue interpretation --- | --- Main body interpretation ---
-.asciz  "/tmp"          # Directory name                  | 4 special opcodes + start of DW_LNE_end_sequence
-# File table format
-.byte   1               # 1 element per file entry        | DW_LNE_end_sequence length
-.byte   1               # DW_LNCT_path                    | DW_LNE_end_sequence length opcode
-.byte   0x08            # DW_FORM_string                  | DW_LNS_const_add_pc
-# File table entries
-.byte   1               # 1 file                          | DW_LNS_copy
-.asciz  "xyz"           # File name                       | start of DW_LNE_set_address
-# Header end
+.asciz  "/tmp"          # 4 special opcodes + start of DW_LNE_end_sequence
+.byte   1               # DW_LNE_end_sequence length
+.byte   1               # DW_LNE_end_sequence length opcode
+.byte   0x08            # DW_LNS_const_add_pc
+.byte   1               # DW_LNS_copy
+.asciz  "xyz"           # start of DW_LNE_set_address
 .byte   9, 2            # DW_LNE_set_address length + opcode
 .quad   0x4444333322221111
 .byte   0, 1, 1         # DW_LNE_end_sequence
 .Linvalid_dir_end0:
 
 # Invalid MD5 hash, where there is data still to be read afterwards.
 .long   .Linvalid_md5_end0-.Linvalid_md5_start0   # Length of Unit
 .Linvalid_md5_start0:

 # Data to show that the rest of the prologue is skipped.
 .byte   1
 .Linvalid_md5_header_end0:
 .byte   0, 9, 2         # DW_LNE_set_address
 .quad   0x1234123412341234
 .byte   0, 1, 1         # DW_LNE_end_sequence
 .Linvalid_md5_end0:
 
-# Invalid MD5 hash, when data beyond the prologue length has
-# been read before the MD5 problem is identified.
+# Header while reading an md5 hash
 .long   .Linvalid_md5_end1-.Linvalid_md5_start1   # Length of Unit
 .Linvalid_md5_start1:
 .short  5               # DWARF version number
 .byte   8               # Address Size
 .byte   0               # Segment Selector Size
 .long   .Linvalid_md5_header_end1 - .Linvalid_md5_params1 # Length of Prologue
 .Linvalid_md5_params1:
 .byte   1               # Minimum Instruction Length

 .byte   1               # 1 directory
 .asciz  "/tmp"
 # File table format
 .byte   2               # 2 elements per file entry
 .byte   1               # DW_LNCT_path
 .byte   0x08            # DW_FORM_string
 .byte   5               # DW_LNCT_MD5
 .Linvalid_md5_header_end1:
-# The bytes from here onwards will also be read as part of the main body.
-                        # --- Prologue interpretation --- | --- Main body interpretation ---
-.byte   0x0b            # DW_FORM_data1                   | DW_LNS_set_epilogue_begin
+.byte   0x0b            # DW_LNS_set_epilogue_begin
 # File table entries
-.byte   1               # 1 file                          | DW_LNS_copy
-.asciz  "xyz"           # File name                       | 3 special opcodes + DW_LNE_set_address start
-.byte   9               # MD5 hash value                  | DW_LNE_set_address length
+.byte   1               # DW_LNS_copy
+.asciz  "xyz"           # 3 special opcodes + DW_LNE_set_address start
+.byte   9               # DW_LNE_set_address length
 # Header end
-.byte   2               # DW_LNE_set_address opcode
+.byte   2
 .quad   0x4321432143214321
-.byte   0, 1, 1         # DW_LNE_end_sequence
+.byte   0, 1, 1
 .Linvalid_md5_end1:
 
 # V5 invalid directory content description has unsupported form.
 .long   .Linvalid_dir_form_end0-.Linvalid_dir_form_start0   # Length of Unit
 .Linvalid_dir_form_start0:
 .short  5               # DWARF version number
 .byte   8               # Address Size
 .byte   0               # Segment Selector Size

llvm/test/tools/llvm-dwarfdump/X86/debug_line_invalid.test

 # NONFATAL-NEXT: Line table prologue
 # NONFATAL:      prologue_length: 0x00000030
 # NONFATAL:      0x00000000deadfade {{.*}} is_stmt
 # NONFATAL-NOT:  end_sequence
 
 ## Very short prologue length for V5 (ends during parameters).
 # NONFATAL:      debug_line[0x000001b2]
 # NONFATAL-NEXT: Line table prologue
-# NONFATAL:      standard_opcode_lengths[DW_LNS_set_isa] = 1
-# NONFATAL-NEXT: include_directories[  0] = "/tmp"
-# NONFATAL-NEXT: file_names[  0]:
-# NONFATAL-NEXT:            name: "xyz"
+# NONFATAL:      standard_opcode_lengths[DW_LNS_set_isa] = 0
+# NONFATAL-NOT:  include_directories
 # NONFATAL:      0x0000000000000000 1 0 1 0 0 is_stmt end_sequence
 # NONFATAL:      0x0000babb1ebabb1e {{.*}} end_sequence
 
 ## V5 prologue ends during file table.
 # NONFATAL:      debug_line[0x000001ee]
 # NONFATAL-NEXT: Line table prologue
 # NONFATAL:      include_directories[  0] = "/tmp"
-# NONFATAL-NEXT: file_names[  0]:
-# NONFATAL-NEXT:            name: "xyz"
-# NONFATAL-NEXT:       dir_index: 1
+# NONFATAL-NOT:  file_names
 # NONFATAL:      0x0000000000000000 {{.*}} epilogue_begin
 # NONFATAL:      0x00000ab4acadab4a {{.*}} end_sequence
 
 ## V5 prologue ends during directory table.
 # NONFATAL:      debug_line[0x0000022f]
 # NONFATAL-NEXT: Line table prologue
-# NONFATAL:      include_directories[  0] = "/tmp"
-# NONFATAL-NEXT: file_names[  0]:
-# NONFATAL-NEXT:            name: "xyz"
+# NONFATAL-NOT:  include_directories
 # NONFATAL:      0x0000000000000002 2 0 1 0 0 is_stmt{{$}}
 # NONFATAL:      0x4444333322221111 {{.*}} end_sequence
 
 ## V5 invalid MD5 hash form when there is still data to be read.
 # NONFATAL:      debug_line[0x0000026b]
 # NONFATAL-NEXT: Line table prologue
 # NONFATAL:      include_directories[  0] = "/tmp"
 # NONFATAL-NOT:  file_names
 # NONFATAL-NOT:  is_stmt
 # NONFATAL:      0x1234123412341234 {{.*}} end_sequence
 
-## V5 invalid MD5 hash form when data beyond the prologue length has
-## been read before the MD5 problem is identified.
+## V5 prologue ends while reading an MD5 hash
 # NONFATAL:      debug_line[0x000002ae]
 # NONFATAL-NEXT: Line table prologue
 # NONFATAL:      include_directories[  0] = "/tmp"
 # NONFATAL-NOT:  file_names
 # NONFATAL:      0x0000000000000000 {{.*}} epilogue_begin
 # NONFATAL:      0x4321432143214321 {{.*}} is_stmt end_sequence
 
 ## V5 invalid directory content description has unsupported form.

 # LAST:          debug_line[0x000003c9]
 # LAST:          0x00000000cafebabe {{.*}} end_sequence
 
 # ALL-NOT:  warning:
 # ALL:      warning: parsing line table prologue at offset 0x00000048: unsupported version 0
 # ALL-NEXT: warning: parsing line table prologue at offset 0x0000004e: unsupported version 1
 # ALL-NEXT: warning: parsing line table prologue at 0x00000054 found an invalid directory or file table description at 0x00000073
 # ALL-NEXT: warning: failed to parse entry content descriptions because no path was found
-# ALL-NEXT: warning: parsing line table prologue at 0x00000081 found an invalid directory or file table description at 0x000000ba
+# ALL-NEXT: warning: parsing line table prologue at 0x00000081 found an invalid directory or file table description at 0x000000b9
 # ALL-NEXT: warning: file names table was not null terminated before the end of the prologue
 # ALL-NEXT: warning: parsing line table prologue at 0x000000c8 should have ended at 0x00000103 but it ended at 0x00000102
 # OTHER-NEXT: warning: unexpected line op length at offset 0x00000158 expected 0x02 found 0x01
 # OTHER-NEXT: warning: unexpected line op length at offset 0x0000015c expected 0x01 found 0x02
 # OTHER-NEXT: warning: last sequence in debug line table at offset 0x0000016c is not terminated
-# ALL-NEXT: warning: parsing line table prologue at 0x000001b2 should have ended at 0x000001ce but it ended at 0x000001e1
-# ALL-NEXT: warning: parsing line table prologue at 0x000001ee should have ended at 0x00000219 but it ended at 0x00000220
-# ALL-NEXT: warning: parsing line table prologue at 0x0000022f should have ended at 0x00000251 but it ended at 0x0000025e
+# ALL-NEXT: warning: parsing line table prologue at 0x000001b2 found an invalid directory or file table description at 0x000001ce
+# ALL-NEXT: warning: failed to parse entry content descriptors: unexpected end of data at offset 0x1ce
+# ALL-NEXT: warning: parsing line table prologue at 0x000001ee found an invalid directory or file table description at 0x00000219
+# ALL-NEXT: warning: failed to parse entry content descriptors: malformed uleb128, extends past end
+# ALL-NEXT: warning: parsing line table prologue at 0x0000022f found an invalid directory or file table description at 0x00000251
+# ALL-NEXT: warning: failed to parse directory entry because extracting the form value failed.
 # ALL-NEXT: warning: parsing line table prologue at 0x0000026b found an invalid directory or file table description at 0x0000029f
 # ALL-NEXT: warning: failed to parse file entry because the MD5 hash is invalid
-# ALL-NEXT: warning: parsing line table prologue at 0x000002ae found an invalid directory or file table description at 0x000002e0
-# ALL-NEXT: warning: failed to parse file entry because the MD5 hash is invalid
+# ALL-NEXT: warning: parsing line table prologue at 0x000002ae found an invalid directory or file table description at 0x000002d9
+# ALL-NEXT: warning: failed to parse entry content descriptors: malformed uleb128, extends past end
 # ALL-NEXT: warning: parsing line table prologue at 0x000002ec found an invalid directory or file table description at 0x00000315
 # ALL-NEXT: warning: failed to parse directory entry because skipping the form value failed.
 # ALL-NEXT: warning: parsing line table prologue at offset 0x00000332 found opcode base of 0. Assuming no standard opcodes
-# ALL-NEXT: warning: parsing line table prologue at 0x00000361 found an invalid directory or file table description at 0x00000383
+# ALL-NEXT: warning: parsing line table prologue at 0x00000361 found an invalid directory or file table description at 0x00000382
 # ALL-NEXT: warning: include directories table was not null terminated before the end of the prologue
-# ALL-NEXT: warning: parsing line table prologue at 0x00000390 found an invalid directory or file table description at 0x000003bc
+# ALL-NEXT: warning: parsing line table prologue at 0x00000390 found an invalid directory or file table description at 0x000003bb
 # ALL-NEXT: warning: file names table was not null terminated before the end of the prologue
 # ALL-NOT:  warning:

llvm/test/tools/llvm-dwarfdump/X86/debug_line_short_prologue.s

-# RUN: llvm-mc -triple x86_64-pc-linux %s -filetype=obj -o %t
-# RUN: llvm-dwarfdump -debug-line %t 2>&1 | FileCheck %s
-
-# CHECK:      debug_line[0x00000000]
-# CHECK-NEXT: warning: parsing line table prologue at 0x00000000 found an invalid directory or file table description at 0x0000002b
-# CHECK-NEXT: warning: failed to parse entry content descriptors: malformed uleb128, extends past end
-# CHECK:      include_directories[  0] = "/tmp"
-
-.section .debug_line,"",@progbits
-.long   .Lshort_prologue_end1-.Lshort_prologue_start1   # Length of Unit
-.Lshort_prologue_start1:
-.short  5               # DWARF version number
-.byte   8               # Address Size
-.byte   0               # Segment Selector Size
-.long   .Lshort_prologue_header_end1 - .Lshort_prologue_params1 # Length of Prologue
-.Lshort_prologue_params1:
-.byte   1               # Minimum Instruction Length
-.byte   1               # Maximum Operations per Instruction
-.byte   1               # Default is_stmt
-.byte   -5              # Line Base
-.byte   14              # Line Range
-.byte   13              # Opcode Base
-.byte   0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1 # Standard Opcode Lengths
-# Directory table format
-.byte   1               # One element per directory entry
-.byte   1               # DW_LNCT_path
-.byte   0x08            # DW_FORM_string
-# Directory table entries
-.byte   1               # 1 directory
-.asciz  "/tmp"
-# File table format
-.byte   2               # 2 elements per file entry
-.byte   1               # DW_LNCT_path
-.byte   0x08            # DW_FORM_string
-.byte   5               # DW_LNCT_MD5
-.Lshort_prologue_header_end1:
-.Lshort_prologue_end1:

llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp

EXPECT_THAT_ERROR(
                          "have ended at 0x000000" +
                          Twine::utohexstr(ExpectedEnd) +
                          " but it ended at 0x000000" +
                          Twine::utohexstr(ExpectedEnd - 1))
                             .str()));
 }
 
 TEST_P(DebugLineParameterisedFixture, ErrorForTooShortPrologueLength) {
   if (!setupGenerator(Version))
     return;
 
   SCOPED_TRACE("Checking Version " + std::to_string(Version) + ", Format " +
                (Format == DWARF64 ? "DWARF64" : "DWARF32"));
 
   LineTable &LT = Gen->addLineTable(Format);
   DWARFDebugLine::Prologue Prologue = LT.createBasicPrologue();
   Prologue.PrologueLength -= 2;
   LT.setPrologue(Prologue);
 
   generate();
 
   auto ExpectedLineTable = Line.getOrParseLineTable(LineData, 0, *Context,
                                                     nullptr, RecordRecoverable);
   ASSERT_THAT_EXPECTED(ExpectedLineTable, Succeeded());
   DWARFDebugLine::LineTable Result(**ExpectedLineTable);
 
-  if (Version != 5) {
-    // Parsing will stop before reading a complete file entry.
-    ASSERT_EQ(Result.Prologue.IncludeDirectories.size(), 1u);
-    EXPECT_EQ(toStringRef(Result.Prologue.IncludeDirectories[0]), "a dir");
-    EXPECT_EQ(Result.Prologue.FileNames.size(), 0u);
-  } else {
-    // Parsing will continue past the presumed end of prologue.
-    ASSERT_EQ(Result.Prologue.FileNames.size(), 1u);
-    ASSERT_EQ(Result.Prologue.FileNames[0].Name.getForm(), DW_FORM_string);
-    ASSERT_EQ(Result.Prologue.FileNames[0].DirIdx, 0u);
-    EXPECT_EQ(toStringRef(Result.Prologue.FileNames[0].Name), "a file");
-  }
 
-  uint64_t ExpectedEnd =
-      Prologue.TotalLength - 2 + Prologue.sizeofTotalLength();
+  uint64_t ExpectedEnd = Prologue.TotalLength + Prologue.sizeofTotalLength() -
+                         (Version != 5 ? 2 : 8);
   std::vector<std::string> Errs;
-  if (Version != 5) {
-    Errs.emplace_back(
-        (Twine("parsing line table prologue at 0x00000000 found an invalid "
-               "directory or file table description at 0x000000") +
-         Twine::utohexstr(ExpectedEnd + 1))
-            .str());
+       Twine::utohexstr(ExpectedEnd))
+  if (Version != 5) {
     Errs.emplace_back("file names table was not null terminated before the end "
                       "of the prologue");
   } else {
     Errs.emplace_back(
-        (Twine("parsing line table prologue at 0x00000000 should have ended at "
-               "0x000000") +
-         Twine::utohexstr(ExpectedEnd) + " but it ended at 0x000000" +
-         Twine::utohexstr(ExpectedEnd + 2))
-            .str());
+        "failed to parse file entry because extracting the form value failed.");
   }
   EXPECT_THAT_ERROR(std::move(Recoverable),
                     FailedWithMessageArray(testing::ElementsAreArray(Errs)));
 }
 
 INSTANTIATE_TEST_CASE_P(
     LineTableTestParams, DebugLineParameterisedFixture,
     Values(std::make_pair(