[nonnull] Teach Clang to attach the nonnull LLVM attribute to

declarations and calls instead of just definitions, and then teach it to
*not* attach such attributes even if the source code contains them.

This follows the design direction discussed on cfe-dev here:
http://lists.llvm.org/pipermail/cfe-dev/2017-January/052066.html

The idea is that for C standard library builtins, even if the library
vendor chooses to annotate their routines with __attribute__((nonnull)),
we will ignore those attributes which pertain to pointer arguments that
have an associated size. This allows the widespread (and seemingly
reasonable) pattern of calling these routines with a null pointer and
a zero size. I have only done this for the library builtins currently
recognized by Clang, but we can now trivially add to this set. This will
be controllable with -fno-builtin if anyone should care to do so.

Note that this does *not* change the AST. As a consequence, warnings,
static analysis, and source code rewriting are not impacted.

This isn't even a regression on any platform as neither Clang nor LLVM
have ever put 'nonnull' onto these arguments for declarations. All this
patch does is enable it on other declarations while preventing us from
ever accidentally enabling it on these libc functions due to a library
vendor.

It will also allow any other libraries using this annotation to gain
optimizations based on the annotation even when only a declaration is
visible.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@298491 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 33 insertions, 27 deletions

diff --git a/include/clang/Basic/Builtins.def b/include/clang/Basic/Builtins.def
index e7084de9ce..9174d39103 100644
--- a/include/clang/Basic/Builtins.def
+++ b/include/clang/Basic/Builtins.def
@@ -55,6 +55,12 @@
 //  S   -> signed
 //  U   -> unsigned
 //  I   -> Required to constant fold to an integer constant expression.
+//  N   -> Do not assume non-null for optimizations even if attributed nonnull.
+//         This can be used when a relevant standard requires arguments or
+//         returns to be non-null and they are attributed accordingly but in
+//         practice are not used in this way. This typically used when a size
+//         parameter is also provided and when zero it would be reasonable to
+//         give an invalid pointer.
 //
 // Types may be postfixed with the following modifiers:
 // * -> pointer (optionally followed by an address space number, if no address
@@ -787,27 +793,27 @@ LIBBUILTIN(_Exit, "vi",           "fr",    "stdlib.h", ALL_LANGUAGES)
 LIBBUILTIN(malloc, "v*z",         "f",     "stdlib.h", ALL_LANGUAGES)
 LIBBUILTIN(realloc, "v*v*z",      "f",     "stdlib.h", ALL_LANGUAGES)
 // C99 string.h
-LIBBUILTIN(memcpy, "v*v*vC*z",    "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(memcmp, "ivC*vC*z",    "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(memmove, "v*v*vC*z",   "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strcpy, "c*c*cC*",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strncpy, "c*c*cC*z",   "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strcmp, "icC*cC*",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strncmp, "icC*cC*z",   "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strcat, "c*c*cC*",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strncat, "c*c*cC*z",   "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strxfrm, "zc*cC*z",    "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(memchr, "v*vC*iz",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strchr, "c*cC*i",      "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strcspn, "zcC*cC*",    "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strpbrk, "c*cC*cC*",   "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strrchr, "c*cC*i",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strspn, "zcC*cC*",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strstr, "c*cC*cC*",    "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strtok, "c*c*cC*",     "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(memset, "v*v*iz",      "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strerror, "c*i",       "f",     "string.h", ALL_LANGUAGES)
-LIBBUILTIN(strlen, "zcC*",        "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(memcpy, "Nv*Nv*NvC*z",  "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(memcmp, "iNvC*NvC*z",   "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(memmove, "Nv*Nv*NvC*z", "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strcpy, "c*c*cC*",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strncpy, "Nc*Nc*NcC*z", "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strcmp, "icC*cC*",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strncmp, "iNcC*NcC*z",  "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strcat, "c*c*cC*",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strncat, "c*c*NcC*z",   "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strxfrm, "zc*cC*z",     "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(memchr, "Nv*NvC*iz",    "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strchr, "c*cC*i",       "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strcspn, "zcC*cC*",     "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strpbrk, "c*cC*cC*",    "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strrchr, "c*cC*i",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strspn, "zcC*cC*",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strstr, "c*cC*cC*",     "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strtok, "c*c*cC*",      "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(memset, "Nv*Nv*iz",     "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strerror, "c*i",        "f",     "string.h", ALL_LANGUAGES)
+LIBBUILTIN(strlen, "zcC*",         "f",     "string.h", ALL_LANGUAGES)
 // C99 stdio.h
 LIBBUILTIN(printf, "icC*.",       "fp:0:", "stdio.h", ALL_LANGUAGES)
 LIBBUILTIN(fprintf, "iP*cC*.",    "fp:1:", "stdio.h", ALL_LANGUAGES)
@@ -841,12 +847,12 @@ LIBBUILTIN(toupper, "ii", "fnU", "ctype.h", ALL_LANGUAGES)
 // C99 wchar.h
 // FIXME: This list is incomplete. We should cover at least the functions that
 // take format strings.
-LIBBUILTIN(wcschr,  "w*wC*w",   "f", "wchar.h", ALL_LANGUAGES)
-LIBBUILTIN(wcscmp,  "iwC*wC*",  "f", "wchar.h", ALL_LANGUAGES)
-LIBBUILTIN(wcslen,  "zwC*",     "f", "wchar.h", ALL_LANGUAGES)
-LIBBUILTIN(wcsncmp, "iwC*wC*z", "f", "wchar.h", ALL_LANGUAGES)
-LIBBUILTIN(wmemchr, "w*wC*wz",  "f", "wchar.h", ALL_LANGUAGES)
-LIBBUILTIN(wmemcmp, "iwC*wC*z", "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wcschr,  "w*wC*w",     "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wcscmp,  "iwC*wC*",    "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wcslen,  "zwC*",       "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wcsncmp, "iNwC*NwC*z", "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wmemchr, "Nw*NwC*wz",  "f", "wchar.h", ALL_LANGUAGES)
+LIBBUILTIN(wmemcmp, "iNwC*NwC*z", "f", "wchar.h", ALL_LANGUAGES)
 
 // C99
 // In some systems setjmp is a macro that expands to _setjmp. We undefine