Re: [patch 28/95] string.h: add FORTIFY coverage for strscpy()

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Dec 15, 2020 at 8:43 PM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> From: Francis Laniel <laniel_francis@xxxxxxxxxxxxxxxxxxx>
> Subject: string.h: add FORTIFY coverage for strscpy()
>
> The fortified version of strscpy ensures the following [..]

So I've taken this, because it follows the pattern of what we already
have, but I have to say that the "fortify" stuff is now about half of
string.h.

I think it should probably be split out into its own header file, and
then string.h could just do

  #if !defined(__NO_FORTIFY) && defined(__OPTIMIZE__) &&
defined(CONFIG_FORTIFY_SOURCE)
     #include <linux/fortify-string.h>
  #endif

or similar.

That won't help my merge window build times (I build with
allmodconfig, so I'll have CONFIG_FORTIFY_SOURCE=y and so my poor
compiler will end up always going off and seeing all that every time
somebody glances at string.h), but I think it would be a cleanup
regardless.

And people who _don't_ build with FORTIFY_SOURCE wouldn't have to read
and parse that big chunk every time <linux/string.h> gets included.

(And yes, just reading the source file and tokenizing it - even if
it's all inside an #ifdef that turns off all the other phases - is
actually quite noticeable overhead in a C compiler. Surprisingly so)

           Linus




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux