[DebugInfo] Correct out-of-bounds error message for DW_FORM_line_strp

DW_FORM_line_strp reads from the .debug_line_str section, but
previously the out-of-bounds error reported the .debug_line section.

This incorrect error message showed up when debugging an issue with
some invalid DWARF5 data.

The verify_string.s test has now been extended to check this (which
required a small change to the DWARF verifier to also look at
DW_FORM_line_strp).

Differential Revision: https://reviews.llvm.org/D146539

llvm/lib/DebugInfo/DWARF/DWARFFormValue.cpp

@@ -615,16 +615,18 @@ Expected<const char *> DWARFFormValue::getAsCString() const {
   }
   // Prefer the Unit's string extractor, because for .dwo it will point to
   // .debug_str.dwo, while the Context's extractor always uses .debug_str.
-  DataExtractor StrData = Form == DW_FORM_line_strp
-                              ? C->getLineStringExtractor()
-                          : U ? U->getStringExtractor()
-                              : C->getStringExtractor();
+  bool IsDebugLineString = Form == DW_FORM_line_strp;
+  DataExtractor StrData =
+      IsDebugLineString ? C->getLineStringExtractor()
+                        : U ? U->getStringExtractor() : C->getStringExtractor();
   if (const char *Str = StrData.getCStr(&Offset))
     return Str;
   std::string Msg = FormEncodingString(Form).str();
   if (Index)
     Msg += (" uses index " + Twine(*Index) + ", but the referenced string").str();
-  Msg += (" offset " + Twine(Offset) + " is beyond .debug_str bounds").str();
+  Msg += (" offset " + Twine(Offset) + " is beyond " +
+          (IsDebugLineString ? ".debug_line_str" : ".debug_str") + " bounds")
+             .str();
   return make_error<StringError>(Msg,
       inconvertibleErrorCode());
 }

llvm/lib/DebugInfo/DWARF/DWARFVerifier.cpp

@@ -777,7 +777,8 @@ unsigned DWARFVerifier::verifyDebugInfoForm(const DWARFDie &Die,
   case DW_FORM_strx1:
   case DW_FORM_strx2:
   case DW_FORM_strx3:
-  case DW_FORM_strx4: {
+  case DW_FORM_strx4:
+  case DW_FORM_line_strp: {
     if (Error E = AttrValue.Value.getAsCString().takeError()) {
       ++NumErrors;
       error() << toString(std::move(E)) << ":\n";

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

@@ -2,7 +2,11 @@
 # 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.
+# the usage of string indices/offsets and string offsets tables.
+
+        .section .debug_line_str,"MS",@progbits,1
+str_comp_dir:
+        .asciz "llvm/test/tools/llvm-dwarfdump/X86"
 
         .section .debug_str,"MS",@progbits,1
 str_producer:
@@ -25,6 +29,8 @@ str_producer:
         .byte 0x00  # DW_CHILDREN_no
         .byte 0x25  # DW_AT_producer
         .byte 0x1a  # DW_FORM_strx
+        .byte 0x1b  # DW_AT_comp_dir
+        .byte 0x1f  # DW_FORM_line_strp
         .byte 0x72  # DW_AT_str_offsets_base
         .byte 0x17  # DW_FORM_sec_offset
         .byte 0x00  # EOM(1)
@@ -45,6 +51,7 @@ CU1_5_version:
 # 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 str_comp_dir     # Offset of debug line string for DW_AT_comp_dir
         .long 1000             # Bad value for DW_AT_str_offsets_base
         .byte 0 # NULL
 CU1_5_end:
@@ -61,12 +68,13 @@ CU2_5_version:
 # The compile-unit DIE, which has DW_AT_producer and DW_AT_str_offsets.
         .byte 1                # Abbreviation code
         .byte 100              # Invalid string index
+        .long str_comp_dir     # Offset of debug line string for DW_AT_comp_dir
         .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. 
+# 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
@@ -78,11 +86,30 @@ CU3_5_version:
 # 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 str_comp_dir     # Offset of debug line string for DW_AT_comp_dir
         .long .debug_str_offsets_base0
         .byte 0 # NULL
 CU3_5_end:
-
+
+# The fourth unit's CU DIE has an invalid offset for the debug line string.
+
+# DWARF v5 CU header
+        .long  CU4_5_end-CU4_5_version  # Length of Unit
+CU4_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 99999            # Invalid offset of debug line string for DW_AT_comp_dir
+        .long .debug_str_offsets_base0
+        .byte 0 # NULL
+CU4_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:      error: DW_FORM_strx uses index 1, but the referenced string offset
 # VERIFY-DAG-SAME: is beyond .debug_str bounds:
+# VERIFY-DAG:      error: DW_FORM_line_strp offset 99999 is beyond .debug_line_str bounds