[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAOZdJXUS6DJAtF2M7Rs28yzSOjRheLodL_F50wsrEFN_0upDPQ@mail.gmail.com>
Date: Thu, 11 Mar 2021 21:46:16 -0600
From: Timur Tabi <timur@...nel.org>
To: Marco Elver <elver@...gle.com>
Cc: Petr Mladek <pmladek@...e.com>,
LKML <linux-kernel@...r.kernel.org>,
Vlastimil Babka <vbabka@...e.cz>,
Steven Rostedt <rostedt@...dmis.org>,
Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
Rasmus Villemoes <linux@...musvillemoes.dk>,
Geert Uytterhoeven <geert@...ux-m68k.org>
Subject: Re: [PATCH 2/2] lib/vsprintf: reduce space taken by no_hash_pointers warning
On Mon, Mar 8, 2021 at 4:51 AM Marco Elver <elver@...gle.com> wrote:
> If we do __initconst change we need to manually remove the duplicate
> lines because we're asking the compiler to create a large array (and
> there's no more auto-dedup). If we do not remove the duplicate lines,
> the __initconst-only approach would create a larger image and result
> in subtly increased memory consumption during init. The additional
> code together with manual dedup should offset that. (I can split this
> patch as Andy suggests, but first need confirmation what people
> actually want.)
>
> I have no idea what the right trade-off is, and appeal to Geert to
> suggest what would be acceptable to him.
Maybe we can have only the message itself wrapped in an #ifdef CONFIG_
of some kind. For example:
+#ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
pr_warn("**********************************************************\n");
pr_warn("** NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE **\n");
pr_warn("** **\n");
pr_warn("** This system shows unhashed kernel memory addresses **\n");
...
+#endif
return 0;
}
In other words, if space is really constrained, then don't include the
message. Or maybe just include part of the message.
Powered by blists - more mailing lists