[DWARF v5] Verifier: Add checks for DW_FORM_strx* forms.

[wolfgangp]

Summary

Making the verifier validate the DWARF v5 strx forms, which index into the string offsets table.

[wolfgangp]

test/DebugInfo/X86/dwarfdump-str-offsets.s

3 ↗

(On Diff #172435)

Enhance this test case to pass the verifier. We're adding a DW_AT_base_type attribute to the variable DIEs to make it do so.

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

10 ↗

(On Diff #172435)

This test case was using DW_FORM_strx4 as an invalid form for DW_AT_stmt_list. Now that we're verifying strx4, this is causing confusion. Replace the strx4 with a different, suitable, invalid form, such as block4.

[aprantl]

lib/DebugInfo/DWARF/DWARFVerifier.cpp

643 ↗

(On Diff #172435)

I'm pretty sure Optional::operator* already asserts.

[probinson]

lib/DebugInfo/DWARF/DWARFVerifier.cpp

643 ↗

(On Diff #172435)

In which case this could simplify to

uint64_t StringOffset = *DieCU->getStringOffsetSectionItem(Index);
if (StringOffset >= ...

[wolfgangp]

lib/DebugInfo/DWARF/DWARFVerifier.cpp

643 ↗

(On Diff #172435)

I'm pretty sure Optional::operator* already asserts.

It sure does. I'll get rid of the assert in the final commit.

llvm/trunk/lib/DebugInfo/DWARF/DWARFVerifier.cpp

case DW_FORM_strp: {
     assert(SecOffset); // DW_FORM_strp is a section offset.
     if (SecOffset && *SecOffset >= DObj.getStringSection().size()) {
       ++NumErrors;
       error() << "DW_FORM_strp offset beyond .debug_str bounds:\n";
       dump(Die) << '\n';
     }
     break;
   }
+  case DW_FORM_strx:
+  case DW_FORM_strx1:
+  case DW_FORM_strx2:
+  case DW_FORM_strx3:
+  case DW_FORM_strx4: {
+    auto Index = AttrValue.Value.getRawUValue();
+    auto DieCU = Die.getDwarfUnit();
+    // Check that we have a valid DWARF v5 string offsets table.
+    if (!DieCU->getStringOffsetsTableContribution()) {
+      ++NumErrors;
+      error() << FormEncodingString(Form)
+              << " used without a valid string offsets table:\n";
+      dump(Die) << '\n';
+      break;
+    }
+    // Check that the index is within the bounds of the section. 
+    unsigned ItemSize = DieCU->getDwarfStringOffsetsByteSize();
+    // Use a 64-bit type to calculate the offset to guard against overflow.
+    uint64_t Offset =
+        (uint64_t)DieCU->getStringOffsetsBase() + Index * ItemSize;
+    if (DObj.getStringOffsetSection().Data.size() < Offset + ItemSize) {
+      ++NumErrors;
+      error() << FormEncodingString(Form) << " uses index "
+              << format("%" PRIu64, Index) << ", which is too large:\n";
+      dump(Die) << '\n';
+      break;
+    }
+    // Check that the string offset is valid.
+    uint64_t StringOffset = *DieCU->getStringOffsetSectionItem(Index);
+    if (StringOffset >= DObj.getStringSection().size()) {
+      ++NumErrors;
+      error() << FormEncodingString(Form) << " uses index "
+              << format("%" PRIu64, Index)
+              << ", but the referenced string"
+                 " offset is beyond .debug_str bounds:\n";
+      dump(Die) << '\n';
+    }
+    break;
+  }
   default:
     break;
   }
   return NumErrors;
 }
 
 unsigned DWARFVerifier::verifyDebugInfoReferences() {
   // Take all references and make sure they point to an actual DIE by

llvm/trunk/test/DebugInfo/X86/dwarfdump-str-offsets.s

 # RUN: llvm-mc -triple x86_64-unknown-linux %s -filetype=obj -o %t.o
 # RUN: llvm-dwarfdump -v %t.o 2> %t.err | FileCheck --check-prefix=COMMON --check-prefix=SPLIT %s
+# RUN: llvm-dwarfdump -verify %t.o | FileCheck --check-prefix=VERIFY %s
 # 
 # Check that we don't report an error on a non-existent range list table.
 # RUN: FileCheck -allow-empty --check-prefix ERR %s < %t.err
 
 # Test object to verify dwarfdump handles v5 string offset tables.
 # We have 2 v5 CUs, a v5 TU, and a split v5 CU and TU.
 #
 
# All CUs/TUs use the same abbrev section for simplicity.
         .byte 0x25  # DW_FORM_strx1
         .byte 0x00  # EOM(1)
         .byte 0x00  # EOM(2)
         .byte 0x05  # Abbrev code
         .byte 0x34  # DW_TAG_variable
         .byte 0x00  # DW_CHILDREN_no
         .byte 0x03  # DW_AT_name
         .byte 0x26  # DW_FORM_strx2
+        .byte 0x49  # DW_AT_type
+        .byte 0x13  # DW_FORM_ref4
         .byte 0x00  # EOM(1)
         .byte 0x00  # EOM(2)
         .byte 0x06  # Abbrev code
         .byte 0x34  # DW_TAG_variable
         .byte 0x00  # DW_CHILDREN_no
         .byte 0x03  # DW_AT_name
         .byte 0x27  # DW_FORM_strx3
+        .byte 0x49  # DW_AT_type
+        .byte 0x13  # DW_FORM_ref4
         .byte 0x00  # EOM(1)
         .byte 0x00  # EOM(2)
         .byte 0x07  # Abbrev code
         .byte 0x34  # DW_TAG_variable
         .byte 0x00  # DW_CHILDREN_no
         .byte 0x03  # DW_AT_name
         .byte 0x28  # DW_FORM_strx4
+        .byte 0x49  # DW_AT_type
+        .byte 0x13  # DW_FORM_ref4
+        .byte 0x00  # EOM(1)
+        .byte 0x00  # EOM(2)
+        .byte 0x08  # Abbrev code
+        .byte 0x24  # DW_TAG_base_type
+        .byte 0x00  # DW_CHILDREN_no
+        .byte 0x3e  # DW_AT_encoding
+        .byte 0x0b  # DW_FORM_data1
         .byte 0x00  # EOM(1)
         .byte 0x00  # EOM(2)
         .byte 0x00  # EOM(3)
 
 # And a .dwo copy of a subset for the .dwo sections.
         .section .debug_abbrev.dwo,"",@progbits
         .byte 0x01  # Abbrev code
         .byte 0x11  # DW_TAG_compile_unit
# DW_AT_str_offsets and DW_AT_compdir.
         .byte 1                # Abbreviation code
         .byte 0                # The index of the producer string
         .byte 1                # The index of the CU name string
         .long .debug_str_offsets_base0
         .byte 2                # The index of the comp dir string
 # A subprogram DIE with DW_AT_name, using DW_FORM_strx1.
         .byte 4                # Abbreviation code
         .byte 3                # Subprogram name string (DW_FORM_strx1)
-# A variable DIE with DW_AT_name, using DW_FORM_strx2.
+# A variable DIE with DW_AT_name, using DW_FORM_strx2, and DW_AT_type.
         .byte 5                # Abbreviation code
         .short 0x0004          # Subprogram name string (DW_FORM_strx2)
-# A variable DIE with DW_AT_name, using DW_FORM_strx3.
+        .long TypeDie-.debug_info
+# A variable DIE with DW_AT_name, using DW_FORM_strx3, and DW_AT_type.
         .byte 6                # Abbreviation code
         .byte 5                # Subprogram name string (DW_FORM_strx3)
         .short 0               # Subprogram name string (DW_FORM_strx3)
-# A variable DIE with DW_AT_name, using DW_FORM_strx4.
+        .long TypeDie-.debug_info
+# A variable DIE with DW_AT_name, using DW_FORM_strx4, and DW_AT_type.
         .byte 7                # Abbreviation code
-        .quad 0x00000006       # Subprogram name string (DW_FORM_strx4)
+        .long 6                # Subprogram name string (DW_FORM_strx4)
+        .long TypeDie-.debug_info
         .byte 0 # NULL
+# A base type DIE with DW_AT_encoding.
+TypeDie:
+        .byte 8                # Abbreviation code
+        .byte 5                # DW_ATE_signed
         .byte 0 # NULL
         .byte 0 # NULL
 CU1_5_end:
 
 # DWARF v5 CU header
         .long  CU2_5_end-CU2_5_version  # Length of Unit
 CU2_5_version:
         .short 5               # DWARF version number

 # SPLIT:       .debug_str_offsets.dwo contents:
 # SPLIT-NEXT:  0x00000000: Contribution size = 24, Format = DWARF32, Version = 5
 # SPLIT-NEXT:  0x00000008: 00000000 "Handmade split DWARF producer"
 # SPLIT-NEXT:  0x0000000c: 0000001e "V5_split_compile_unit"
 # SPLIT-NEXT:  0x00000010: 00000034 "/home/test/splitCU"
 # SPLIT-NEXT:  0x00000014: 00000047 "V5_split_type_unit"
 # SPLIT-NEXT:  0x00000018: 0000005a "V5_split_Mystruct"
 
+# VERIFY: No errors.
+
 # ERR-NOT: parsing a range list table:

llvm/trunk/test/tools/llvm-dwarfdump/X86/verify_debug_info.s

 # RUN: llvm-mc %s -filetype obj -triple x86_64-apple-darwin -o - \
 # RUN: | not llvm-dwarfdump -v -verify - \
 # RUN: | FileCheck %s
 
 # CHECK: error: DIE has invalid DW_AT_stmt_list encoding:{{[[:space:]]}}
 # CHECK-NEXT: 0x0000000c: DW_TAG_compile_unit [1] *
 # CHECK-NEXT: DW_AT_producer [DW_FORM_strp] ( .debug_str[0x00000000] = "clang version 5.0.0 (trunk 308185) (llvm/trunk 308186)")
 # CHECK-NEXT: DW_AT_language [DW_FORM_data2]  (DW_LANG_C99)
 # CHECK-NEXT: DW_AT_name [DW_FORM_strp] ( .debug_str[0x00000037] = "basic.c")
-# CHECK-NEXT: DW_AT_stmt_list [DW_FORM_strx4] ( indexed (00000000) string = )
+# CHECK-NEXT: DW_AT_stmt_list [DW_FORM_block4]
 # CHECK-NEXT: DW_AT_comp_dir [DW_FORM_strp] ( .debug_str[0x0000003f] = "/Users/sgravani/Development/tests")
 # CHECK-NEXT: DW_AT_low_pc [DW_FORM_addr] (0x0000000000000000)
 # CHECK-NEXT: DW_AT_high_pc [DW_FORM_data4] (0x00000016){{[[:space:]]}}
 # CHECK-NEXT: error: DIE has DW_AT_type with incompatible tag DW_TAG_null{{[[:space:]]}}
 # CHECK-NEXT: 0x0000002b: DW_TAG_subprogram [2] *
 # CHECK-NEXT: DW_AT_low_pc [DW_FORM_addr]       (0x0000000000000000)
 # CHECK-NEXT: DW_AT_high_pc [DW_FORM_data4]     (0x00000016)
 # CHECK-NEXT: DW_AT_frame_base [DW_FORM_exprloc]        (DW_OP_reg6)
Lsection_abbrev:
   .byte 1                       ## DW_CHILDREN_yes
   .byte 37                      ## DW_AT_producer
   .byte 14                      ## DW_FORM_strp
   .byte 19                      ## DW_AT_language
   .byte 5                       ## DW_FORM_data2
   .byte 3                       ## DW_AT_name
   .byte 14                      ## DW_FORM_strp
   .byte 16                      ## DW_AT_stmt_list
-  .byte 40                      ## DW_FORM_sec_offset -- error: DIE has invalid DW_AT_stmt_list encoding:
+  .byte 4                       ## DW_FORM_sec_offset -- error: DIE has invalid DW_AT_stmt_list encoding:
   .byte 27                      ## DW_AT_comp_dir
   .byte 14                      ## DW_FORM_strp
   .byte 17                      ## DW_AT_low_pc
   .byte 1                       ## DW_FORM_addr
   .byte 18                      ## DW_AT_high_pc
   .byte 6                       ## DW_FORM_data4
   .byte 0                       ## EOM(1)
   .byte 0                       ## EOM(2)

llvm/trunk/test/tools/llvm-dwarfdump/X86/verify_strings.s

+# RUN: llvm-mc -triple x86_64-unknown-linux %s -filetype=obj -o %t.o
+# RUN: not llvm-dwarfdump -verify %t.o | FileCheck --check-prefix=VERIFY %s
+
+# Check that the verifier correctly diagnoses various error conditions with
+# the usage of string indices and string offsets tables.
+
+        .section .debug_str,"MS",@progbits,1
+str_producer:
+        .asciz "Handmade DWARF producer"
+
+        .section .debug_str_offsets,"",@progbits
+# The string offsets table
+        .long .debug_str_offsets_segment0_end-.debug_str_offsets_base0+4
+        .short 5    # DWARF version
+        .short 0    # Padding
+.debug_str_offsets_base0:
+        .long str_producer
+        .long 1000  # Invalid string address.
+.debug_str_offsets_segment0_end:
+
+# A simple abbrev section with a basic compile unit DIE.
+        .section .debug_abbrev,"",@progbits
+        .byte 0x01  # Abbrev code
+        .byte 0x11  # DW_TAG_compile_unit
+        .byte 0x01  # DW_CHILDREN_no
+        .byte 0x25  # DW_AT_producer
+        .byte 0x1a  # DW_FORM_strx
+        .byte 0x72  # DW_AT_str_offsets_base
+        .byte 0x17  # DW_FORM_sec_offset
+        .byte 0x00  # EOM(1)
+        .byte 0x00  # EOM(2)
+
+        .section .debug_info,"",@progbits
+
+# The first unit's CU DIE has an invalid DW_AT_str_offsets_base which
+# renders any string index unresolvable.
+
+# DWARF v5 CU header.
+        .long  CU1_5_end-CU1_5_version  # Length of Unit
+CU1_5_version:
+        .short 5               # DWARF version number
+        .byte 1                # DWARF Unit Type
+        .byte 8                # Address Size (in bytes)
+        .long .debug_abbrev    # Offset Into Abbrev. Section
+# The compile-unit DIE, which has DW_AT_producer and DW_AT_str_offsets.
+        .byte 1                # Abbreviation code
+        .byte 0                # Index of string for DW_AT_producer.
+        .long 1000             # Bad value for DW_AT_str_offsets_base
+        .byte 0 # NULL
+CU1_5_end:
+
+# The second unit's CU DIE uses an invalid string index.
+
+# DWARF v5 CU header
+        .long  CU2_5_end-CU2_5_version  # Length of Unit
+CU2_5_version:
+        .short 5               # DWARF version number
+        .byte 1                # DWARF Unit Type
+        .byte 8                # Address Size (in bytes)
+        .long .debug_abbrev    # Offset Into Abbrev. Section
+# The compile-unit DIE, which has DW_AT_producer and DW_AT_str_offsets.
+        .byte 1                # Abbreviation code
+        .byte 100              # Invalid string index
+        .long .debug_str_offsets_base0
+        .byte 0 # NULL
+CU2_5_end:
+
+# The third unit's CU DIE uses a valid string index but the entry in the 
+# string offsets table is invalid. 
+
+# DWARF v5 CU header
+        .long  CU3_5_end-CU3_5_version  # Length of Unit
+CU3_5_version:
+        .short 5               # DWARF version number
+        .byte 1                # DWARF Unit Type
+        .byte 8                # Address Size (in bytes)
+        .long .debug_abbrev    # Offset Into Abbrev. Section
+# The compile-unit DIE, which has DW_AT_producer and DW_AT_str_offsets.
+        .byte 1                # Abbreviation code
+        .byte 1                # Index of string for DW_AT_producer.
+        .long .debug_str_offsets_base0
+        .byte 0 # NULL
+CU3_5_end:
+        
+# VERIFY-DAG:      error: DW_FORM_strx used without a valid string offsets table:
+# VERIFY-DAG:      error: DW_FORM_strx uses index 100, which is too large:
+# VERIFY-DAG:      error: DW_FORM_strx uses index 1, but the referenced string offset 
+# VERIFY-DAG-SAME: is beyond .debug_str bounds: