public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: tip-bot for Andi Kleen <tipbot@zytor.com>
To: linux-tip-commits@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, hubicka@ucw.cz, hpa@zytor.com,
	mingo@kernel.org, ak@linux.intel.com, tglx@linutronix.de,
	hpa@linux.intel.com
Subject: [tip:x86/asmlinkage] lto, workaround: Add workaround for initcall reordering
Date: Thu, 13 Feb 2014 20:30:47 -0800	[thread overview]
Message-ID: <tip-ef1b893c29d0dba778f67ad97b554b37f9108dcc@git.kernel.org> (raw)
In-Reply-To: <1391846481-31491-4-git-send-email-ak@linux.intel.com>

Commit-ID:  ef1b893c29d0dba778f67ad97b554b37f9108dcc
Gitweb:     http://git.kernel.org/tip/ef1b893c29d0dba778f67ad97b554b37f9108dcc
Author:     Andi Kleen <ak@linux.intel.com>
AuthorDate: Sat, 8 Feb 2014 09:01:08 +0100
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Thu, 13 Feb 2014 20:24:13 -0800

lto, workaround: Add workaround for initcall reordering

Work around a LTO gcc problem: when there is no reference to a variable
in a module it will be moved to the end of the program. This causes
reordering of initcalls which the kernel does not like.
Add a dummy reference function to avoid this. The function is
deleted by the linker.

This replaces a previous much slower workaround.

Thanks to Jan "Honza" Hubička for suggesting this technique.

Suggested-by: Jan Hubička <hubicka@ucw.cz>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1391846481-31491-4-git-send-email-ak@linux.intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 include/linux/init.h | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/include/linux/init.h b/include/linux/init.h
index e168880..a3ba270 100644
--- a/include/linux/init.h
+++ b/include/linux/init.h
@@ -163,6 +163,23 @@ extern bool initcall_debug;
 
 #ifndef __ASSEMBLY__
 
+#ifdef CONFIG_LTO
+/* Work around a LTO gcc problem: when there is no reference to a variable
+ * in a module it will be moved to the end of the program. This causes
+ * reordering of initcalls which the kernel does not like.
+ * Add a dummy reference function to avoid this. The function is
+ * deleted by the linker.
+ */
+#define LTO_REFERENCE_INITCALL(x) \
+	; /* yes this is needed */			\
+	static __used __exit void *reference_##x(void)	\
+	{						\
+		return &x;				\
+	}
+#else
+#define LTO_REFERENCE_INITCALL(x)
+#endif
+
 /* initcalls are now grouped by functionality into separate 
  * subsections. Ordering inside the subsections is determined
  * by link order. 
@@ -175,7 +192,8 @@ extern bool initcall_debug;
 
 #define __define_initcall(fn, id) \
 	static initcall_t __initcall_##fn##id __used \
-	__attribute__((__section__(".initcall" #id ".init"))) = fn
+	__attribute__((__section__(".initcall" #id ".init"))) = fn; \
+	LTO_REFERENCE_INITCALL(__initcall_##fn##id)
 
 /*
  * Early initcalls run before initializing SMP.

  reply	other threads:[~2014-02-14  4:31 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-02-08  8:01 [PATCH 01/17] x86, lto: Disable LTO for the x86 VDSO Andi Kleen
2014-02-08  8:01 ` [PATCH 02/17] x86, lto: Disable fancy hweight optimizations for LTO v2 Andi Kleen
2014-02-08 18:52   ` H. Peter Anvin
2014-02-08 20:21     ` Andi Kleen
2014-02-08 21:43       ` H. Peter Anvin
2014-02-08  8:01 ` [PATCH 03/17] lto: Make asmlinkage __visible Andi Kleen
2014-02-14  4:30   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-14  9:10     ` Borislav Petkov
2014-02-14 14:38       ` Andi Kleen
2014-02-08  8:01 ` [PATCH 04/17] lto, workaround: Add workaround for initcall reordering Andi Kleen
2014-02-14  4:30   ` tip-bot for Andi Kleen [this message]
2014-02-08  8:01 ` [PATCH 05/17] lto: Handle LTO common symbols in module loader Andi Kleen
2014-02-12  1:04   ` Rusty Russell
2014-02-14  4:30   ` [tip:x86/asmlinkage] " tip-bot for Joe Mario
2014-02-20 23:11     ` Rusty Russell
2014-02-21 19:12       ` H. Peter Anvin
2014-02-21 22:37         ` Andi Kleen
2014-02-08  8:01 ` [PATCH 06/17] lto: Disable LTO for sys_ni Andi Kleen
2014-02-14  4:31   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 07/17] Kbuild, lto, workaround: Don't warn for initcall_reference in modpost Andi Kleen
2014-02-14  4:31   ` [tip:x86/asmlinkage] Kbuild, lto, workaround: Don' t " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 08/17] Kbuild, lto: Drop .number postfixes " Andi Kleen
2014-02-14  4:31   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 09/17] Kbuild, lto: add ld-version and ld-ifversion macros Andi Kleen
2014-02-14  4:31   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 10/17] Kbuild, lto: Add a gcc-ld script to let run gcc as ld Andi Kleen
2014-02-14  4:31   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 11/17] Kbuild, lto: Disable LTO for asm-offsets.c Andi Kleen
2014-02-14  4:32   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 12/17] Kbuild, lto: Set TMPDIR for LTO Andi Kleen
2014-02-14  4:26   ` H. Peter Anvin
2014-02-14 15:37     ` Andi Kleen
2014-02-08  8:01 ` [PATCH 13/17] Kbuild, lto: Handle basic LTO in modpost Andi Kleen
2014-02-14  4:32   ` [tip:x86/asmlinkage] " tip-bot for Andi Kleen
2014-02-08  8:01 ` [PATCH 14/17] Kbuild, lto: Add Link Time Optimization support Andi Kleen
2014-02-14  4:28   ` H. Peter Anvin
2014-02-14 14:36     ` Andi Kleen
2014-02-14 16:25       ` H. Peter Anvin
2014-02-08  8:01 ` [PATCH 15/17] Kbuild, lto: Add LTO build Documentation Andi Kleen
2014-02-08  8:01 ` [PATCH 16/17] lto: Mark spinlocks noinline when inline spinlocks are disabled Andi Kleen
2014-02-08  8:01 ` [PATCH 17/17] lto, module: Warn about modules that are not fully LTOed Andi Kleen
2014-02-12  1:13   ` Rusty Russell
2014-02-14  4:30 ` [tip:x86/asmlinkage] x86, lto: Disable LTO for the x86 VDSO tip-bot for Andi Kleen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=tip-ef1b893c29d0dba778f67ad97b554b37f9108dcc@git.kernel.org \
    --to=tipbot@zytor.com \
    --cc=ak@linux.intel.com \
    --cc=hpa@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=hubicka@ucw.cz \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-tip-commits@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox