[PATCH] docs: deprecated.rst: Add zero-length and one-element arrays

[PATCH] docs: deprecated.rst: Add zero-length and one-element arrays

[Gustavo A. R. Silva]

Add zero-length and one-element arrays to the list.

While I continue replacing zero-length and one-element arrays with
flexible-array members, I need a reference to point people to, so
they don't introduce more instances of such arrays. And while here,
add a note to the "open-coded arithmetic in allocator arguments"
section, on the use of struct_size() and the arrays-to-deprecate
mentioned here.

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
---
 Documentation/process/deprecated.rst | 84 ++++++++++++++++++++++++++++
 1 file changed, 84 insertions(+)

diff --git a/Documentation/process/deprecated.rst b/Documentation/process/deprecated.rst
index 652e2aa02a66c..622c8ac72a615 100644
--- a/Documentation/process/deprecated.rst
+++ b/Documentation/process/deprecated.rst
@@ -85,6 +85,12 @@ Instead, use the helper::
 
 	header = kzalloc(struct_size(header, item, count), GFP_KERNEL);
 
+NOTE: If you are using struct_size() on a structure containing a zero-length
+or a one-element array as a trailing array member, stop using such arrays
+and switch to `flexible arrays
+<https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays>`_
+instead.
+
 See array_size(), array3_size(), and struct_size(),
 for more details as well as the related check_add_overflow() and
 check_mul_overflow() family of functions.
@@ -200,3 +206,81 @@ All switch/case blocks must end in one of:
 * continue;
 * goto <label>;
 * return [expression];
+
+Zero-length and one-element arrays
+----------------------------------
+Old code in the kernel uses the zero-length and one-element array extensions
+to the C90 standard, but the `preferred mechanism to declare variable-length
+types
+<https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html>`_
+such as these ones is a `flexible array member`, introduced in C99::
+
+        struct something {
+                int length;
+                char data[];
+        };
+
+        struct something *instance;
+
+        instance = kmalloc(struct_size(instance, data, size), GFP_KERNEL);
+        instance->length = size;
+        memcpy(instance->data, source, size);
+
+By making use of the mechanism above, we get a compiler error in case the
+flexible array does not occur last in the structure, which helps to prevent
+some kind of `undefined behavior <https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=76497732932f15e7323dc805e8ea8dc11bb587cf>`_ bugs from being inadvertently introduced to
+the codebase.
+
+It is also important to notice that zero-length and one-element arrays pose
+confusion for things like sizeof() and `CONFIG_FORTIFY_SOURCE`. For instance,
+there is no mechanism that warns us that the following application of the
+sizeof() operator to a zero-length array always results in zero::
+
+        struct something {
+                int length;
+                char data[0];
+        };
+
+        struct something *instance;
+
+        instance = kmalloc(struct_size(instance, data, size), GFP_KERNEL);
+        instance->length = size;
+        memcpy(instance->data, source, size);
+        ...
+        size = sizeof(instance->data);
+
+At the last line of code above, ``size`` turns out to be zero --when one might have
+thought differently. Here are a couple examples of this issue `link 1
+<https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f2cd32a443da694ac4e28fbf4ac6f9d5cc63a539>`_,
+`link 2
+<https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ab91c2a89f86be2898cee208d492816ec238b2cf>`_.
+On the other hand, `flexible array members have incomplete type, and so the sizeof()
+operator may not be applied <https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html>`_,
+so any  misuse of such  operator will be immediately noticed at build time.
+
+
+With respect to one-element arrays, one has to be acutely aware that `such arrays
+occupy at least as much space as a single object of the type
+<https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html>`_,
+hence they contribute to the size of the enclosing structure. This is prone
+to error every time people want to calculate the total size of dynamic memory
+to allocate for a structure containing an array of this kind as a member::
+
+        struct something {
+                int length;
+                char data[1];
+        };
+
+        struct something *instance;
+
+        instance = kmalloc(struct_size(instance, data, size - 1), GFP_KERNEL);
+        instance->length = size;
+        memcpy(instance->data, source, size);
+
+In the example above, we had to remember to calculate ``size - 1`` when using
+the struct_size() helper, otherwise we would have --unintentionally-- allocated
+memory for one too many ``data`` objects. The cleanest and least error-prone way
+to implement this is through the use of a `flexible array member`, which is
+exemplified at the `beginning
+<https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays>`_
+of this section.
-- 
2.27.0