On Tue, Aug 25, 2020 at 06:58:36AM -0700, Nick Desaulniers wrote:
> LLVM implemented a recent "libcall optimization" that lowers calls to
> `sprintf(dest, "%s", str)` where the return value is used to
> `stpcpy(dest, str) - dest`. This generally avoids the machinery involved
> in parsing format strings. `stpcpy` is just like `strcpy` except it
> returns the pointer to the new tail of `dest`. This optimization was
> introduced into clang-12.
>
> Implement this so that we don't observe linkage failures due to missing
> symbol definitions for `stpcpy`.
>
> Similar to last year's fire drill with:
> commit 5f074f3e192f ("lib/string.c: implement a basic bcmp")
>
> The kernel is somewhere between a "freestanding" environment (no full libc)
> and "hosted" environment (many symbols from libc exist with the same
> type, function signature, and semantics).
>
> As H. Peter Anvin notes, there's not really a great way to inform the
> compiler that you're targeting a freestanding environment but would like
> to opt-in to some libcall optimizations (see pr/47280 below), rather than
> opt-out.
>
> Arvind notes, -fno-builtin-* behaves slightly differently between GCC
> and Clang, and Clang is missing many __builtin_* definitions, which I
> consider a bug in Clang and am working on fixing.
>
> Masahiro summarizes the subtle distinction between compilers justly:
> To prevent transformation from foo() into bar(), there are two ways in
> Clang to do that; -fno-builtin-foo, and -fno-builtin-bar. There is
> only one in GCC; -fno-buitin-foo.
>
> (Any difference in that behavior in Clang is likely a bug from a missing
> __builtin_* definition.)
>
> Masahiro also notes:
> We want to disable optimization from foo() to bar(),
> but we may still benefit from the optimization from
> foo() into something else. If GCC implements the same transform, we
> would run into a problem because it is not -fno-builtin-bar, but
> -fno-builtin-foo that disables that optimization.
>
> In this regard, -fno-builtin-foo would be more future-proof than
> -fno-built-bar, but -fno-builtin-foo is still potentially overkill. We
> may want to prevent calls from foo() being optimized into calls to
> bar(), but we still may want other optimization on calls to foo().
>
> It seems that compilers today don't quite provide the fine grain control
> over which libcall optimizations pseudo-freestanding environments would
> prefer.
>
> Finally, Kees notes that this interface is unsafe, so we should not
> encourage its use. As such, I've removed the declaration from any
> header, but it still needs to be exported to avoid linkage errors in
> modules.
>
> Cc: stable@xxxxxxxxxxxxxxx
> Link: https://bugs.llvm.org/show_bug.cgi?id=47162
> Link: https://bugs.llvm.org/show_bug.cgi?id=47280
> Link: https://github.com/ClangBuiltLinux/linux/issues/1126
> Link: https://man7.org/linux/man-pages/man3/stpcpy.3.html
> Link: https://pubs.opengroup.org/onlinepubs/9699919799/functions/stpcpy.html
> Link: https://reviews.llvm.org/D85963
> Suggested-by: Andy Lavr <andy.lavr@xxxxxxxxx>
> Suggested-by: Arvind Sankar <nivedita@xxxxxxxxxxxx>
> Suggested-by: Joe Perches <joe@xxxxxxxxxxx>
> Suggested-by: Masahiro Yamada <masahiroy@xxxxxxxxxx>
> Suggested-by: Rasmus Villemoes <linux@xxxxxxxxxxxxxxxxxx>
> Reported-by: Sami Tolvanen <samitolvanen@xxxxxxxxxx>
> Signed-off-by: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
Tested-by: Nathan Chancellor <natechancellor@xxxxxxxxx>
> ---
> Changes V3:
> * Drop Sami's Tested by tag; newer patch.
> * Add EXPORT_SYMBOL as per Andy.
> * Rewrite commit message, rewrote part of what Masahiro said to be
> generic in terms of foo() and bar().
> * Prefer %NUL-terminated to NULL terminated. NUL is the ASCII character
> '\0', as per Arvind and Rasmus.
>
> Changes V2:
> * Added Sami's Tested by; though the patch changed implementation, the
> missing symbol at link time was the problem Sami was observing.
> * Fix __restrict -> __restrict__ typo as per Joe.
> * Drop note about restrict from commit message as per Arvind.
> * Fix NULL -> NUL as per Arvind; NUL is ASCII '\0'. TIL
> * Fix off by one error as per Arvind; I had another off by one error in
> my test program that was masking this.
>
> lib/string.c | 24 ++++++++++++++++++++++++
> 1 file changed, 24 insertions(+)
>
> diff --git a/lib/string.c b/lib/string.c
> index 6012c385fb31..6bd0cf0fb009 100644
> --- a/lib/string.c
> +++ b/lib/string.c
> @@ -272,6 +272,30 @@ ssize_t strscpy_pad(char *dest, const char *src, size_t count)
> }
> EXPORT_SYMBOL(strscpy_pad);
>
> +/**
> + * stpcpy - copy a string from src to dest returning a pointer to the new end
> + * of dest, including src's %NUL-terminator. May overrun dest.
> + * @dest: pointer to end of string being copied into. Must be large enough
> + * to receive copy.
> + * @src: pointer to the beginning of string being copied from. Must not overlap
> + * dest.
> + *
> + * stpcpy differs from strcpy in a key way: the return value is the new
> + * %NUL-terminated character. (for strcpy, the return value is a pointer to
> + * src. This interface is considered unsafe as it doesn't perform bounds
> + * checking of the inputs. As such it's not recommended for usage. Instead,
> + * its definition is provided in case the compiler lowers other libcalls to
> + * stpcpy.
> + */
> +char *stpcpy(char *__restrict__ dest, const char *__restrict__ src);
> +char *stpcpy(char *__restrict__ dest, const char *__restrict__ src)
> +{
> + while ((*dest++ = *src++) != '\0')
> + /* nothing */;
> + return --dest;
> +}
> +EXPORT_SYMBOL(stpcpy);
> +
> #ifndef __HAVE_ARCH_STRCAT
> /**
> * strcat - Append one %NUL-terminated string to another
> --
> 2.28.0.297.g1956fa8f8d-goog
>
