From: Randy Dunlap <rdunlap@infradead.org>
To: Christophe Leroy <christophe.leroy@csgroup.eu>,
Peter Zijlstra <peterz@infradead.org>,
Jakub Jelinek <jakub@redhat.com>
Cc: linux-toolchains@vger.kernel.org
Subject: Re: ilog2 vs. GCC inlining heuristics
Date: Wed, 21 Oct 2020 21:01:01 -0700 [thread overview]
Message-ID: <b5dc06c5-6a2c-9f12-ba43-dc83eed9a503@infradead.org> (raw)
In-Reply-To: <21556974-eea1-ed6a-ea6f-3e97a6eea4bc@csgroup.eu>
[-- Attachment #1: Type: text/plain, Size: 1379 bytes --]
On 10/21/20 11:40 AM, Christophe Leroy wrote:
>
>
> Le 21/10/2020 à 17:19, Peter Zijlstra a écrit :
>> On Wed, Oct 21, 2020 at 03:27:18PM +0200, Jakub Jelinek wrote:
>>> Hi!
>>>
>>> Based on the GCC PR97445 discussions, I'd like to propose following change,
>>> which should significantly decrease the amount of code in inline functions
>>> that use ilog2, but as I'm already two decades out of the Linux kernel
>>> development, I'd appreciate if some kernel developer could try that (all
>>> I have done is check that it gives the same results as before) and if it
>>> works submit it for inclusion into the kernel?
>>
>> I'll stick it in my queue and feed it to the robots.
>>
>
> I did a mpc885_ads_defconfig build with your patch. That's far better, even better than with gcc 9 without the patch.
>
> I only have two instances of get_order() in vmlinux, one of it is used twice, the other is never user.
>
> With -Winline, the reason for not inlining is for both because "the call is unlikely and the code size would grow"
>
> Christophe
Do we want a lib/test_log2.c test? Although Jakub says that he already
verified that is gives the same results as the previous code.
I ran a (new) lib/test_log2.ko before and after Jakub's patch and got the
same results, although I am not claiming that it has an exhaustive set of
tests in it. [it is attached]
--
~Randy
[-- Attachment #2: lib-test-log2.patch --]
[-- Type: text/x-patch, Size: 4527 bytes --]
From: Randy Dunlap <rdunlap@infradead.org>
// modified from lib/test_bitops.c
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
---
lib/Kconfig.debug | 11 +++
lib/Makefile | 1
lib/test_log2.c | 139 ++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 151 insertions(+)
--- linux-next-20201021.orig/lib/Kconfig.debug
+++ linux-next-20201021/lib/Kconfig.debug
@@ -2136,6 +2136,17 @@ config TEST_BITOPS
If unsure, say N.
+config TEST_LOG2
+ tristate "Test module for log2 interfaces"
+ depends on m
+ help
+ This builds the "test_log2" module that does basic testing of
+ all <linux/log2.h> interfaces.
+ It has no dependencies and doesn't run or load unless
+ explicitly requested by name. For example: modprobe test_log2.
+
+ If unsure, say N.
+
config TEST_VMALLOC
tristate "Test module for stress/performance analysis of vmalloc allocator"
default n
--- linux-next-20201021.orig/lib/Makefile
+++ linux-next-20201021/lib/Makefile
@@ -62,6 +62,7 @@ obj-$(CONFIG_TEST_BPF) += test_bpf.o
obj-$(CONFIG_TEST_FIRMWARE) += test_firmware.o
obj-$(CONFIG_TEST_BITOPS) += test_bitops.o
CFLAGS_test_bitops.o += -Werror
+obj-$(CONFIG_TEST_LOG2) += test_log2.o
obj-$(CONFIG_TEST_SYSCTL) += test_sysctl.o
obj-$(CONFIG_TEST_HASH) += test_hash.o test_siphash.o
obj-$(CONFIG_TEST_IDA) += test_ida.o
--- /dev/null
+++ linux-next-20201021/lib/test_log2.c
@@ -0,0 +1,139 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+// currently missing tests of roundup/rounddown macros;
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/printk.h>
+#include <linux/log2.h>
+#include <asm/page.h>
+
+/*
+ * a tiny module for testing all interfaces in <linux/log2.h>
+ */
+
+static unsigned int test_power_of_2[][2] = {
+ {0, 0},
+ {PAGE_SIZE, 1},
+ {65535, 0},
+ {65536, 1},
+ {(unsigned int)~0, 0},
+ {0x80000000, 1},
+ {0x80008000, 0},
+};
+
+static unsigned int test_ilog2_uint[][2] = {
+ //{0, -1},
+ {1, 0},
+ {2, 1},
+ {4, 2},
+ {(unsigned int)~0, 31},
+ {0x80000000, 31},
+ {0x80008000, 31},
+};
+
+static unsigned long test_ilog2_ulong[][2] = {
+ //{0, -1},
+ {1, 0},
+ {2, 1},
+ {4, 2},
+ {(unsigned long)~0, 63},
+ {0x80000000UL, 31},
+ {0x80008000UL, 31},
+};
+
+static unsigned int order_base2[][2] = {
+ {0x00000003, 2},
+ {0x00000004, 2},
+ {0x00001fff, 13},
+ {0x00002000, 13},
+ {0x50000000, 31},
+ {0x80000000, 31},
+ {0x80003000, 32},
+};
+
+//#ifdef CONFIG_64BIT
+static unsigned long order_base2_long[][2] = {
+ {0x0000000300000000, 34},
+ {0x0000000400000000, 34},
+ {0x00001fff00000000, 45},
+ {0x0000200000000000, 45},
+ {0x5000000000000000, 63},
+ {0x8000000000000000, 63},
+ {0x8000300000000000, 64},
+};
+//#endif
+
+static unsigned int test_bits_per[][2] = {
+ {0, 1},
+ {1, 1},
+ {2, 2},
+ {3, 2},
+ {4, 3},
+ {0x0fff, 12},
+ {0x1000, 13},
+ {0x80003000, 32},
+};
+
+static int __init test_log2_start(void)
+{
+ int i;
+
+ pr_info("Starting log2 tests\n");
+
+ for (i = 0; i < ARRAY_SIZE(test_power_of_2); i++) {
+ if (is_power_of_2(test_power_of_2[i][0]) != test_power_of_2[i][1])
+ pr_warn("is_power_of_2 wrong for %x\n",
+ test_power_of_2[i][0]);
+ }
+
+ for (i = 0; i < ARRAY_SIZE(test_ilog2_uint); i++) {
+ if (ilog2(test_ilog2_uint[i][0]) != test_ilog2_uint[i][1])
+ pr_warn("ilog2 wrong for uint %x\n",
+ test_ilog2_uint[i][0]);
+ }
+
+ for (i = 0; i < ARRAY_SIZE(test_ilog2_ulong); i++) {
+ if (ilog2(test_ilog2_ulong[i][0]) != test_ilog2_ulong[i][1])
+ pr_warn("ilog2 wrong for ulong %x\n",
+ test_ilog2_ulong[i][0]);
+ }
+
+ for (i = 0; i < ARRAY_SIZE(order_base2); i++) {
+ if (order_base_2(order_base2[i][0]) != order_base2[i][1])
+ pr_warn("order_base_2 wrong for %x\n",
+ order_base2[i][0]);
+ }
+
+//#ifdef CONFIG_64BIT
+ for (i = 0; i < ARRAY_SIZE(order_base2_long); i++) {
+ if (order_base_2(order_base2_long[i][0]) !=
+ order_base2_long[i][1])
+ pr_warn("order_base_2 wrong for %lx\n",
+ order_base2_long[i][0]);
+ }
+//#endif
+
+ for (i = 0; i < ARRAY_SIZE(test_bits_per); i++) {
+ if (bits_per(test_bits_per[i][0]) != test_bits_per[i][1])
+ pr_warn("bits_per wrong for %x\n",
+ test_bits_per[i][0]);
+ }
+
+ pr_info("Completed log2 tests\n");
+
+ return 0;
+}
+
+static void __exit test_log2_fini(void)
+{
+}
+
+module_init(test_log2_start);
+module_exit(test_log2_fini);
+
+MODULE_AUTHOR("Randy Dunlap <rdunlap@infradead.org>");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("log2 test module");
next prev parent reply other threads:[~2020-10-22 4:01 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-21 13:27 ilog2 vs. GCC inlining heuristics Jakub Jelinek
2020-10-21 13:36 ` Christophe Leroy
2020-10-21 13:45 ` Jakub Jelinek
2020-10-21 15:19 ` Peter Zijlstra
2020-10-21 18:40 ` Christophe Leroy
2020-10-22 4:01 ` Randy Dunlap [this message]
2020-10-22 7:12 ` Jakub Jelinek
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=b5dc06c5-6a2c-9f12-ba43-dc83eed9a503@infradead.org \
--to=rdunlap@infradead.org \
--cc=christophe.leroy@csgroup.eu \
--cc=jakub@redhat.com \
--cc=linux-toolchains@vger.kernel.org \
--cc=peterz@infradead.org \
/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;
as well as URLs for NNTP newsgroup(s).