mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Daniel Santos <daniel.santos@pobox.com>
To: Christopher Li <sparse@chrisli.org>,
	Daniel Santos <daniel.santos@pobox.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	David Howells <dhowells@redhat.com>,
	David Daney <david.daney@cavium.com>,
	David Rientjes <rientjes@google.com>,
	linux-sparse@vger.kernel.org, Pavel Pisa <pisa@cmp.felk.cvut.cz>,
	Richard Weinberger <richard@nod.at>,
	LKML <linux-kernel@vger.kernel.org>,
	Michel Lespinasse <walken@google.com>,
	Andrea Arcangeli <aarcange@redhat.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Rik van Riel <riel@redhat.com>
Subject: [PATCH v5 7/25] compiler{,-gcc4}.h: Introduce __flatten function attribute
Date: Tue, 25 Sep 2012 18:30:46 -0500	[thread overview]
Message-ID: <1348615846-10193-1-git-send-email-daniel.santos@pobox.com> (raw)
In-Reply-To: <1348615467-9850-1-git-send-email-daniel.santos@pobox.com>

For gcc 4.1 & later, expands to __attribute__((flatten)) which forces
the compiler to inline everything it can into the function.  This is
useful in combination with noinline when you want to control the depth
of inlining, or create a single function where inline expansions will
occur. (see
http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html#index-g_t_0040code_007bflatten_007d-function-attribute-2512)

Normally, it's best to leave this type of thing up to the compiler.
However, the generic rbtree code uses inline functions just to be able
to inject compile-time constant data that specifies how the caller wants
the function to behave (via struct rb_relationship).  This data can be
thought of as the template parameters of a C++ templatized function.
Since some of these functions, once expanded, become quite large, gcc
sometimes decides not to perform some important inlining, in one case,
even generating a few bytes more code by not doing so. (Note: I have not
eliminated the possibility that this was an optimization bug, but the
flatten attribute fixes it in either case.)

Combining __flatten and noinline insures that important optimizations
occur in these cases and that the inline expansion occurs in exactly one
place, thus not leading to unnecissary bloat. However, it also can
eliminate some opportunities for optimization should gcc otherwise
decide the function its self is a good candidate for inlining.

Signed-off-by: Daniel Santos <daniel.santos@pobox.com>
---
 include/linux/compiler-gcc4.h |    7 ++++++-
 include/linux/compiler.h      |    4 ++++
 2 files changed, 10 insertions(+), 1 deletions(-)

diff --git a/include/linux/compiler-gcc4.h b/include/linux/compiler-gcc4.h
index 5755e23..38fb81d 100644
--- a/include/linux/compiler-gcc4.h
+++ b/include/linux/compiler-gcc4.h
@@ -15,7 +15,12 @@
 
 #if GCC_VERSION >= 40102
 # define __compiletime_object_size(obj) __builtin_object_size(obj, 0)
-#endif
+
+/* flatten introduced in 4.1, but broken in 4.6.0 (gcc bug #48731)*/
+# if GCC_VERSION != 40600
+#  define __flatten __attribute__((flatten))
+# endif
+#endif /* GCC_VERSION >= 40102 */
 
 #if GCC_VERSION >= 40300
 /* Mark functions as cold. gcc will assume any path leading to a call
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index 4d9f353..b26d606 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -244,6 +244,10 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect);
 #define __always_inline inline
 #endif
 
+#ifndef __flatten
+#define __flatten
+#endif
+
 #endif /* __KERNEL__ */
 
 /*
-- 
1.7.3.4


  parent reply	other threads:[~2012-09-25 23:30 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1348615467-9850-1-git-send-email-daniel.santos@pobox.com>
2012-09-25 23:29 ` [PATCH v5 1/25] compiler-gcc4.h: Correct verion check for __compiletime_error Daniel Santos
2012-09-25 23:30 ` [PATCH v5 2/25] compiler-gcc4.h: Reorder macros based upon gcc ver Daniel Santos
2012-09-25 23:30 ` [PATCH v5 3/25] compiler-gcc.h: Add gcc-recommended GCC_VERSION macro Daniel Santos
2012-09-25 23:30 ` [PATCH v5 4/25] compiler-gcc{3,4}.h: Use " Daniel Santos
2012-09-25 23:30 ` [PATCH v5 5/25] compiler{,-gcc4}.h: Remove duplicate macros Daniel Santos
2012-09-25 23:30 ` [PATCH v5 6/25] bug.h: Replace __linktime_error with __compiletime_error Daniel Santos
2012-09-25 23:30 ` Daniel Santos [this message]
2012-09-25 23:30 ` [PATCH v5 8/25] bug.h: Make BUILD_BUG_ON generate compile-time error Daniel Santos
2012-09-25 23:30 ` [PATCH v5 9/25] bug.h: Add BUILD_BUG_ON_NON_CONST macro Daniel Santos
2012-09-25 23:31 ` [PATCH v5 10/25] bug.h: Add gcc 4.2+ versions of BUILD_BUG_ON_* macros Daniel Santos
2012-09-25 23:31 ` [PATCH v5 11/25] rbtree.h: Generic Red-Black Trees Daniel Santos
2012-09-25 23:31 ` [PATCH v5 12/25] rbtree.h: include kconfig.h Daniel Santos
2012-09-25 23:31 ` [PATCH v5 13/25] fair.c: Use generic rbtree impl in fair scheduler Daniel Santos
2012-09-25 23:31 ` [PATCH v5 15/25] kernel-doc: bugfix - multi-line macros Daniel Santos
2012-09-25 23:31 ` [PATCH v5 16/25] kernel-doc: bugfix - empty line in Example section Daniel Santos
2012-09-25 23:31 ` [PATCH v5 17/25] kernel-doc: Don't mangle whitespace " Daniel Santos
2012-09-25 23:31 ` [PATCH v5 19/25] rbtree.h: add doc comments for struct rb_node Daniel Santos
2012-09-25 23:32 ` [PATCH v5 20/25] selftest: Add generic tree self-test common code Daniel Santos
2012-09-25 23:32 ` [PATCH v5 21/25] selftest: Add userspace test program Daniel Santos
2012-09-25 23:32 ` [PATCH v5 22/25] selftest: Add script to compile & run " Daniel Santos
2012-09-25 23:32 ` [PATCH v5 23/25] selftest: Add basic compiler iterator test script Daniel Santos
2012-09-25 23:32 ` [PATCH v5 24/25] selftest: report generation script for test results Daniel Santos
2012-09-25 23:32 ` [PATCH v5 25/25] rbtree.h: (optional?) Add RB_INSERT_DUPE_RIGHT flag Daniel Santos
     [not found] ` <1348618742.22822.39.camel@gandalf.local.home>
2012-09-26  1:02   ` [PATCH v5 0/25] Generic Red-Black Trees (still WIP) Daniel Santos
2012-09-26  1:28     ` Steven Rostedt
2012-09-26  5:07 ` Daniel Santos

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=1348615846-10193-1-git-send-email-daniel.santos@pobox.com \
    --to=daniel.santos@pobox.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=aarcange@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=david.daney@cavium.com \
    --cc=dhowells@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-sparse@vger.kernel.org \
    --cc=pisa@cmp.felk.cvut.cz \
    --cc=richard@nod.at \
    --cc=riel@redhat.com \
    --cc=rientjes@google.com \
    --cc=sparse@chrisli.org \
    --cc=walken@google.com \
    /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

all inboxes | Powered by JetHome®