[lld] [llvm] [Symbolizer] Support for Missing Line Numbers. (PR #82240)

via llvm-commits llvm-commits at lists.llvm.org
Tue Jul 2 16:10:36 PDT 2024


================
@@ -0,0 +1,489 @@
+##  Test  --skip-line-zero option.
+##
+##  This test illustrates the usage of handcrafted assembly to produce the following line table.
+##  Address            Line   Column File   ISA Discriminator OpIndex Flags
+##  ------------------ ------ ------ ------ --- ------------- ------- -------------
+##  0x00000000000030a0      1     80      1   0             0       0  is_stmt prologue_end
+##  0x00000000000030a6      1     80      1   0             0       0  is_stmt end_sequence
+##  0x0000000000001000      0     68      1   0             0       0  is_stmt prologue_end
+##  0x0000000000001008      2     39      1   0             0       0
+##  0x0000000000001010      3     68      1   0             0       0  is_stmt prologue_end
+##  0x0000000000001012      0     68      1   0             0       0
+##  0x0000000000001017      3     68      1   0             0       0
+##  0x0000000000001019      3     39      1   0             0       0
+##  0x0000000000001020      5     1       2   0             0       0  is_stmt prologue_end
+##  0x0000000000001026      5     1       2   0             0       0  is_stmt end_sequence
+
+# REQUIRES: x86-registered-target
+
+# RUN: rm -rf %t && split-file %s %t && cd %t
+# RUN: llvm-mc -filetype=obj -triple=x86_64-pc-linux --fdebug-prefix-map=%t="" %s -o %t.o
+
+## Check that with '--skip-line-zero', the last non-zero line in the current sequence is displayed.
+## If it fails to find in the current sequence then return the orignal computed line-zero for the queried address.
+# RUN: llvm-symbolizer --obj=%t.o --skip-line-zero 0x1000 | FileCheck --strict-whitespace --match-full-lines --check-prefix=APPROX-FAIL-ACROSS-SEQ %s
+
+# APPROX-FAIL-ACROSS-SEQ:add
+# APPROX-FAIL-ACROSS-SEQ-NEXT:definitions.h:0:68
+
+## Check that with '--skip-line-zero', the last non-zero line in the current sequence is displayed.
+# RUN: llvm-symbolizer --obj=%t.o --skip-line-zero 0x1012 | FileCheck --strict-whitespace --match-full-lines --check-prefix=APPROX-WITHIN-SEQ %s
+
+# APPROX-WITHIN-SEQ:sub
+# APPROX-WITHIN-SEQ-NEXT:definitions.h:3:68 (approximate)
+
+## Check to ensure that '--skip-line-zero' only affects addresses having line-zero when more than one address is specified.
+# RUN: llvm-symbolizer --obj=%t.o --skip-line-zero 0x1012 0x1020 | FileCheck --strict-whitespace --match-full-lines --check-prefixes=APPROX-WITHIN-SEQ,NO-APPROX %s
+
+# NO-APPROX:main
+# NO-APPROX-NEXT:main.c:5:1
+
+## Check to ensure that '--skip-line-zero' with '--verbose' enabled displays correct approximate flag in verbose ouptut.
+# RUN: llvm-symbolizer --obj=%t.o --skip-line-zero --verbose 0x1012 | FileCheck --strict-whitespace --match-full-lines --check-prefix=APPROX-VERBOSE %s
+
+# APPROX-VERBOSE:sub
+# APPROX-VERBOSE-NEXT:  Filename: definitions.h
+# APPROX-VERBOSE-NEXT:  Function start filename: definitions.h
+# APPROX-VERBOSE-NEXT:  Function start line: 3
+# APPROX-VERBOSE-NEXT:  Function start address: 0x1010
+# APPROX-VERBOSE-NEXT:  Line: 3
+# APPROX-VERBOSE-NEXT:  Column: 68
+# APPROX-VERBOSE-NEXT:  Approximate: true
+
+## Check to ensure that '--skip-line-zero' with '--output-style=JSON' displays correct approximate flag in JSON output.
+# RUN: llvm-symbolizer --obj=%t.o --skip-line-zero --output-style=JSON 0x1012 | FileCheck --strict-whitespace --match-full-lines --check-prefix=APPROX-JSON %s
+
+# APPROX-JSON:[{"Address":"0x1012","ModuleName":"{{.*}}{{[/|\]+}}test{{[/|\]+}}tools{{[/|\]+}}llvm-symbolizer{{[/|\]+}}Output{{[/|\]+}}approximate-line-handcrafted.s.tmp.o","Symbol":[{"Approximate":true,"Column":68,"Discriminator":0,"FileName":"definitions.h","FunctionName":"sub","Line":3,"StartAddress":"0x1010","StartFileName":"definitions.h","StartLine":3}]}]
+
+#--- definitions.h
----------------
bd1976bris wrote:

I don't think we need this header file for these test cases and I would favor removing is as having only one source file will reduce the size of the generated DWARF.

https://github.com/llvm/llvm-project/pull/82240


More information about the llvm-commits mailing list