1========================
2Segmented Stacks in LLVM
3========================
4
5.. contents::
6   :local:
7
8Introduction
9============
10
11Segmented stack allows stack space to be allocated incrementally than as a
12monolithic chunk (of some worst case size) at thread initialization. This is
13done by allocating stack blocks (henceforth called *stacklets*) and linking them
14into a doubly linked list. The function prologue is responsible for checking if
15the current stacklet has enough space for the function to execute; and if not,
16call into the libgcc runtime to allocate more stack space. Segmented stacks are
17enabled with the ``"split-stack"`` attribute on LLVM functions.
18
19The runtime functionality is `already there in libgcc
20<http://gcc.gnu.org/wiki/SplitStacks>`_.
21
22Implementation Details
23======================
24
25.. _allocating stacklets:
26
27Allocating Stacklets
28--------------------
29
30As mentioned above, the function prologue checks if the current stacklet has
31enough space. The current approach is to use a slot in the TCB to store the
32current stack limit (minus the amount of space needed to allocate a new block) -
33this slot's offset is again dictated by ``libgcc``. The generated
34assembly looks like this on x86-64:
35
36.. code-block:: nasm
37
38    leaq     -8(%rsp), %r10
39    cmpq     %fs:112,  %r10
40    jg       .LBB0_2
41
42    # More stack space needs to be allocated
43    movabsq  $8, %r10   # The amount of space needed
44    movabsq  $0, %r11   # The total size of arguments passed on stack
45    callq    __morestack
46    ret                 # The reason for this extra return is explained below
47  .LBB0_2:
48    # Usual prologue continues here
49
50The size of function arguments on the stack needs to be passed to
51``__morestack`` (this function is implemented in ``libgcc``) since that number
52of bytes has to be copied from the previous stacklet to the current one. This is
53so that SP (and FP) relative addressing of function arguments work as expected.
54
55The unusual ``ret`` is needed to have the function which made a call to
56``__morestack`` return correctly. ``__morestack``, instead of returning, calls
57into ``.LBB0_2``. This is possible since both, the size of the ``ret``
58instruction and the PC of call to ``__morestack`` are known. When the function
59body returns, control is transferred back to ``__morestack``. ``__morestack``
60then de-allocates the new stacklet, restores the correct SP value, and does a
61second return, which returns control to the correct caller.
62
63Variable Sized Allocas
64----------------------
65
66The section on `allocating stacklets`_ automatically assumes that every stack
67frame will be of fixed size. However, LLVM allows the use of the ``llvm.alloca``
68intrinsic to allocate dynamically sized blocks of memory on the stack. When
69faced with such a variable-sized alloca, code is generated to:
70
71* Check if the current stacklet has enough space. If yes, just bump the SP, like
72  in the normal case.
73* If not, generate a call to ``libgcc``, which allocates the memory from the
74  heap.
75
76The memory allocated from the heap is linked into a list in the current
77stacklet, and freed along with the same. This prevents a memory leak.
78