https://github.com/MrEven132 created https://github.com/llvm/llvm-project/pull/217262
`DW_OP_GNU_const_index` reads a machine-address-sized constant from the address table and pushes it as a generic value. The generic type has the size of an address on the target, so on a 32-bit compilation unit the pushed value should be a 32-bit scalar. `DWARFUnit::ReadAddressFromDebugAddrSection` reads the correct number of bytes, but the evaluator case pushed the result as `Scalar(value)`, where `value` is a 64-bit `lldb::addr_t`, leaving a 64-bit scalar on the stack instead of canonicalizing to `8 * address_size` like the other constant-pushing opcodes. Use `to_generic`, as `DW_OP_addrx` / `DW_OP_GNU_addr_index` already do, so the pushed value has the generic type's width. Adds a `DW_OP_GNU_const_index_address_size` unit test asserting both the wraparound of generic arithmetic at the 32-bit address size and the exact bit width of the pushed scalar. Fixes #211010 >From 7ab844cd19ef809137f1f02132904b50c3ac5fe3 Mon Sep 17 00:00:00 2001 From: MrEven132 <[email protected]> Date: Wed, 19 Aug 2026 18:22:25 +0800 Subject: [PATCH] [lldb] Canonicalize DW_OP_GNU_const_index to the generic type --- lldb/source/Expression/DWARFExpression.cpp | 2 +- lldb/unittests/Expression/DWARFExpressionTest.cpp | 14 ++++++++++++++ 2 files changed, 15 insertions(+), 1 deletion(-) diff --git a/lldb/source/Expression/DWARFExpression.cpp b/lldb/source/Expression/DWARFExpression.cpp index e927708f12d14..45b05d425591b 100644 --- a/lldb/source/Expression/DWARFExpression.cpp +++ b/lldb/source/Expression/DWARFExpression.cpp @@ -2056,7 +2056,7 @@ llvm::Expected<Value> DWARFExpression::Evaluate( uint64_t index = op->getRawOperand(0); lldb::addr_t value = eval_ctx.dwarf_cu->ReadAddressFromDebugAddrSection(index); - stack.push_back(Scalar(value)); + stack.push_back(to_generic(value)); } break; case DW_OP_GNU_entry_value: diff --git a/lldb/unittests/Expression/DWARFExpressionTest.cpp b/lldb/unittests/Expression/DWARFExpressionTest.cpp index 6985dd0dd4939..66e4bd398a9e1 100644 --- a/lldb/unittests/Expression/DWARFExpressionTest.cpp +++ b/lldb/unittests/Expression/DWARFExpressionTest.cpp @@ -883,6 +883,20 @@ TEST(DWARFExpression, DW_OP_addr_index_address_size) { } } +TEST(DWARFExpression, DW_OP_GNU_const_index_address_size) { + // DW_OP_GNU_const_index pushes a generic value, so it is canonicalized to + // the target address size: on this 32-bit target, 0x2a + 0xffffffff + 1 + // wraps back to 0x2a. + MockDwarfDelegate unit(/*version=*/5, + /*debug_addr=*/{{0, 0x11}, {1, 0x2a}}); + auto result = + Evaluate({DW_OP_GNU_const_index, 0x01, DW_OP_const4u, 0xff, 0xff, 0xff, + 0xff, DW_OP_plus, DW_OP_lit1, DW_OP_plus, DW_OP_stack_value}, + {}, &unit); + ASSERT_THAT_EXPECTED(result, ExpectScalar(32, 0x2a, false)); + EXPECT_EQ(result->GetScalar().GetAPSInt().getBitWidth(), 32u); +} + TEST(DWARFExpression, DW_OP_addr_big_endian) { // Same operand bytes, big-endian extractor: the address must be read in // target byte order. _______________________________________________ lldb-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
