[llvm-objdump] - llvm-objdump can miss printing bytes at the end of a section.

[sidneym]

Summary

llvm-objdump can miss printing bytes at the end of a section.

This file:

.globl foo
.type foo, @object
foo:
.byte 't','h','i', 's',' ','i','s',' '
.byte 'a',' ','t','e','s','t'
.size foo, . - foo

Produces this output:
f.o: file format ELF64-x86-64

Disassembly of section .text:
0000000000000000 foo:

0:	 74 68 69 73 20 69 73 20         this is 
8:	 61 20 74 65 73 74               a

It should be:
f.o: file format ELF64-x86-64

Disassembly of section .text:
0000000000000000 foo:

0:	 74 68 69 73 20 69 73 20         this is 
8:	 61 20 74 65 73 74               a test

[grimar]

Looks correct to me. Have a few comments/nits about the test case.

test/tools/llvm-objdump/X86/bytes.test

8 ↗

(On Diff #184620)

I would simplify to
.ascii "this is a test"

11 ↗

(On Diff #184620)

I am not sure this is a useful 2 lines of comments.
You are testing the ASCII output first of all.

13 ↗

(On Diff #184620)

You do not need this line I think.

16 ↗

(On Diff #184620)

So maybe just omit testing the bytes?

0:       {{*.}}         this is
8:       {{*.}}         a test

[sidneym]

Updated the testcase.

[grimar]

LGTM.

[grimar]

Oh, maybe some better name for the test? Since it is not relative to bytes anymore. Maybe disasm-text.test or something like that?

llvm/trunk/test/tools/llvm-objdump/X86/disasm-text.test

+# RUN: llvm-mc -filetype=obj -triple=x86_64-unknown-linux %s -o -| \
+# RUN: llvm-objdump -d - | FileCheck %s
+
+.globl foo
+.type foo, @object
+foo:
+.ascii "this is a test"
+
+# CHECK: foo:
+# CHECK: 0:{{.*}}this is
+# CHECK: 8:{{.*}}a test

llvm/trunk/tools/llvm-objdump/llvm-objdump.cpp

#endif
             AsciiData[NumBytes] = isPrint(Byte) ? Byte : '.';
 
             uint8_t IndentOffset = 0;
             NumBytes++;
             if (Index == End - 1 || NumBytes > 8) {
               // Indent the space for less than 8 bytes data.
               // 2 spaces for byte and one for space between bytes
               IndentOffset = 3 * (8 - NumBytes);
-              for (int Excess = 8 - NumBytes; Excess < 8; Excess++)
+              for (int Excess = NumBytes; Excess < 8; Excess++)
                 AsciiData[Excess] = '\0';
               NumBytes = 8;
             }
             if (NumBytes == 8) {
               AsciiData[8] = '\0';
               outs() << std::string(IndentOffset, ' ') << "         ";
               outs() << reinterpret_cast<char *>(AsciiData);
               outs() << '\n';