func_name.part.# in objdump disassembly

Question

I was looking at an executable having GMP linked statically to it. The following seems weird.

Disassembly of section .text:

0000000000400e60 <refmpn_fill.part.4>:
  400e60:   50                      push   %rax
  400e61:   ba 55 84 42 00          mov    $0x428455,%edx
  400e66:   be 92 00 00 00          mov    $0x92,%esi
  400e6b:   bf a8 81 42 00          mov    $0x4281a8,%edi
  400e70:   e8 1b 66 00 00          callq  407490 <__gmp_assert_fail>

0000000000400e75 <refmpn_count_leading_zeros.part.19>:
  400e75:   50                      push   %rax
  400e76:   ba c2 81 42 00          mov    $0x4281c2,%edx
  400e7b:   be eb 06 00 00          mov    $0x6eb,%esi
  400e80:   bf a8 81 42 00          mov    $0x4281a8,%edi
  400e85:   e8 06 66 00 00          callq  407490 <__gmp_assert_fail>
  400e8a:   66 0f 1f 44 00 00       nopw   0x0(%rax,%rax,1)

0000000000400e90 <main>:
  400e90:   53                      push   %rbx
  400e91:   bb e8 03 00 00          mov    $0x3e8,%ebx
  400e96:   48 83 ec 20             sub    $0x20,%rsp
  400e9a:   e8 11 03 00 00          callq  4011b0 <tests_start>

As you can see there are disassembly of functions of the form: function_name.part.# where # represents some number. There are no such function having name with part.# appended at the end in the source code. Then how do these objects appear?

Answer 1

It's called partial inlining:

Partial inlining is closely related to function outlining. Rather than inlining the entire body into a call site, the compiler may choose to inline part of the called function (typically a hot fragment at the start of the function with an early return).