[lkml]   [2009]   [Oct]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] New attribute designated_init: mark a struct as requiring designated init
On Sat, Oct 10, 2009 at 03:33:20AM -0700, Christopher Li wrote:
> On Sat, Oct 10, 2009 at 1:58 AM, Josh Triplett <> wrote:
> > Some structure types provide a set of fields of which most users will
> > only initialize the subset they care about.  Users of these types should
> > always use designated initializers, to avoid relying on the specific
> > structure layout.  Examples of this type of structure include the many
> The patch is very well written with nice documentations and test case.
> It applies and runs fine.
> I am curious weather this is some thing the kernel developers want to
> use. Please speak up if you want to annotate the kernel structure to
> issue such warning. If some one use it, I have no problem adding it to
> sparse.

After getting this patch reviewed, I intended to add many such
annotations myself. I have patches sitting in my local Linux tree.
(However, I don't plan to try to get the patches accepted into Linux
until this attribute shows up in a Sparse release, since I don't want to
make Linux require the latest Sparse from the git tree.)

> I am not sure how useful this is yet. If the structure is changed, most
> likely the positional initialization will fail due to type mismatching.
> Some real life example how this feature can expose some otherwise
> hard to detect bug would be nice.

You *hope* that positional initialization will fail. :) Some of these
types of structures contain many fields with the same or similar types,
and initializers don't inherently contain a lot of distinguishing type
information that helps cause type errors rather than silent failures.

Even if the positional initialization does fail, this failure will occur
at the time of attempting to change the structure, and many such
failures may occur all over the tree, making such changes significantly
more difficult. With the designated_init attribute, sparse will warn
when attempting to create the structure initializer.

With the designated_init attribute, someone making changes to a
structure can check that Sparse produces no warnings about positional
initializers and then know that certain types of changes can occur
without having to check every initializer. For instance, removing a
field will only require checking any initializer that initializes that
field, and the compiler will find all of those. Adding a field that can
safely remain NULL does not require checking any initializers.

> With this approach, we need to annotate the kernel to benefit from it.
> Another idea is that we can find out how different part of the kernel
> initialize the same structure. If most of them using designated init then
> the few non-conforming can get a warning. This approach is more
> complicate. But it does not need to change the kernel.

I agree that we could infer certain types of warnings by saying "99% of
the kernel does $FOO, but this bit of code doesn't". Other static
analysis tools do similar things. However, such analyses can easily
lead to false positives, and they prove far more difficult to implement.
Having an explicit attribute works now, and also provides useful
documentation of the intended use of a structure.

> > +                               warning(e->pos, "%s%s%spositional init of field in %s %s, declared with attribute designated_init",
> > +                                       ctype->ident ? "in initializer for " : "",
> > +                                       ctype->ident ? ctype->ident->name : "",
> ident->name has no guarantee of terminating by NUL.
> You want to use "%.*s" with ident->size, ident->name here.

Good catch; will fix and send new patch.

(As an aside, though: any fundamental reason why we use that error-prone
approach rather than allocating one extra byte and NUL-terminating
idents? Then we could drop the (larger) len field, and remove the
annoying restriction on calling show_ident multiple times in the same

- Josh Triplett
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2009-10-10 18:25    [W:0.048 / U:66.924 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site