mutex lock can be initialized automatically with DEFINE_MUTEX() rather
than explicitly calling mutex_init().
this patch will reduce the size of bcache.ko about 64 bytes, the reason
as follows:
though this patch will increase the size of .data segment about 32 bytes,
it will also reduce the size of .init.text and .rodata.str1.1(at x86_64),
.rodata_str1.8(at arm64) and .bss segment total about 96 bytes which reduce
the size more than .data segment;
here is the statistics:
Sections: (arm64 platform)
Idx name size
-.init.text 00000240
+.init.text 00000228
-.rodata.str1.8 000012cd
+.rodata.str1.8 000012b5
-.data 00000c60
+.data 00000c80
-.bss 00000080
+.bss 00000060
Sections: (x86 platform)
Idx name size
-.init.text 000001d9
+.init.text 000001bf
-.rodata.str1.1 00000c80
+.rodata.str1.1 00000c6d
-.data 00000cc0
+.data 00000ce0
-.bss 00000080
+.bss 00000070
Reported-by: Hulk Robot <[email protected]>
Signed-off-by: Zheng Yongjun <[email protected]>
---
drivers/md/bcache/super.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 03e1fe4de53d..3635f454309e 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -40,7 +40,7 @@ static const char invalid_uuid[] = {
};
static struct kobject *bcache_kobj;
-struct mutex bch_register_lock;
+DEFINE_MUTEX(bch_register_lock);
bool bcache_is_reboot;
LIST_HEAD(bch_cache_sets);
static LIST_HEAD(uncached_devices);
@@ -2869,7 +2869,6 @@ static int __init bcache_init(void)
check_module_parameters();
- mutex_init(&bch_register_lock);
init_waitqueue_head(&unregister_wait);
register_reboot_notifier(&reboot);
--
2.25.1
On 4/7/21 5:38 PM, Zheng Yongjun wrote:
> mutex lock can be initialized automatically with DEFINE_MUTEX() rather
> than explicitly calling mutex_init().
>
> this patch will reduce the size of bcache.ko about 64 bytes, the reason
> as follows:
>
> though this patch will increase the size of .data segment about 32 bytes,
> it will also reduce the size of .init.text and .rodata.str1.1(at x86_64),
> .rodata_str1.8(at arm64) and .bss segment total about 96 bytes which reduce
> the size more than .data segment;
>
Thanks for the following information. BTW, the bss section should be
ignored, because they don't take physical file size.
It seems although the patched code has larger .data section, but the
generated code in .init.text and .rodata.str1.1(or 8) are much lesser
than the increased size. This is really interesting :-)
Your data is convinced, it is fair to take this patch in.
Could you please post a v3 version which removes the .bss information ?
Coly Li
> here is the statistics:
> Sections: (arm64 platform)
> Idx name size
> -.init.text 00000240
> +.init.text 00000228
>
> -.rodata.str1.8 000012cd
> +.rodata.str1.8 000012b5
>
> -.data 00000c60
> +.data 00000c80
>
> -.bss 00000080
> +.bss 00000060
>
> Sections: (x86 platform)
> Idx name size
> -.init.text 000001d9
> +.init.text 000001bf
>
> -.rodata.str1.1 00000c80
> +.rodata.str1.1 00000c6d
>
> -.data 00000cc0
> +.data 00000ce0
>
> -.bss 00000080
> +.bss 00000070
>
> Reported-by: Hulk Robot <[email protected]>
> Signed-off-by: Zheng Yongjun <[email protected]>
> ---
> drivers/md/bcache/super.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
> index 03e1fe4de53d..3635f454309e 100644
> --- a/drivers/md/bcache/super.c
> +++ b/drivers/md/bcache/super.c
> @@ -40,7 +40,7 @@ static const char invalid_uuid[] = {
> };
>
> static struct kobject *bcache_kobj;
> -struct mutex bch_register_lock;
> +DEFINE_MUTEX(bch_register_lock);
> bool bcache_is_reboot;
> LIST_HEAD(bch_cache_sets);
> static LIST_HEAD(uncached_devices);
> @@ -2869,7 +2869,6 @@ static int __init bcache_init(void)
>
> check_module_parameters();
>
> - mutex_init(&bch_register_lock);
> init_waitqueue_head(&unregister_wait);
> register_reboot_notifier(&reboot);
>
>
I will send you v3 patch soon, thank you:)
-----邮件原件-----
发件人: Coly Li [mailto:[email protected]]
发送时间: 2021年4月7日 17:50
收件人: zhengyongjun <[email protected]>
抄送: pmail_hulkci <[email protected]>; [email protected]; [email protected]; [email protected]
主题: Re: [PATCH -next v2] bcache: use DEFINE_MUTEX() for mutex lock
On 4/7/21 5:38 PM, Zheng Yongjun wrote:
> mutex lock can be initialized automatically with DEFINE_MUTEX() rather
> than explicitly calling mutex_init().
>
> this patch will reduce the size of bcache.ko about 64 bytes, the
> reason as follows:
>
> though this patch will increase the size of .data segment about 32
> bytes, it will also reduce the size of .init.text and
> .rodata.str1.1(at x86_64), .rodata_str1.8(at arm64) and .bss segment
> total about 96 bytes which reduce the size more than .data segment;
>
Thanks for the following information. BTW, the bss section should be ignored, because they don't take physical file size.
It seems although the patched code has larger .data section, but the generated code in .init.text and .rodata.str1.1(or 8) are much lesser than the increased size. This is really interesting :-)
Your data is convinced, it is fair to take this patch in.
Could you please post a v3 version which removes the .bss information ?
Coly Li
> here is the statistics:
> Sections: (arm64 platform)
> Idx name size
> -.init.text 00000240
> +.init.text 00000228
>
> -.rodata.str1.8 000012cd
> +.rodata.str1.8 000012b5
>
> -.data 00000c60
> +.data 00000c80
>
> -.bss 00000080
> +.bss 00000060
>
> Sections: (x86 platform)
> Idx name size
> -.init.text 000001d9
> +.init.text 000001bf
>
> -.rodata.str1.1 00000c80
> +.rodata.str1.1 00000c6d
>
> -.data 00000cc0
> +.data 00000ce0
>
> -.bss 00000080
> +.bss 00000070
>
> Reported-by: Hulk Robot <[email protected]>
> Signed-off-by: Zheng Yongjun <[email protected]>
> ---
> drivers/md/bcache/super.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
> index 03e1fe4de53d..3635f454309e 100644
> --- a/drivers/md/bcache/super.c
> +++ b/drivers/md/bcache/super.c
> @@ -40,7 +40,7 @@ static const char invalid_uuid[] = { };
>
> static struct kobject *bcache_kobj;
> -struct mutex bch_register_lock;
> +DEFINE_MUTEX(bch_register_lock);
> bool bcache_is_reboot;
> LIST_HEAD(bch_cache_sets);
> static LIST_HEAD(uncached_devices);
> @@ -2869,7 +2869,6 @@ static int __init bcache_init(void)
>
> check_module_parameters();
>
> - mutex_init(&bch_register_lock);
> init_waitqueue_head(&unregister_wait);
> register_reboot_notifier(&reboot);
>
>