* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-02-27 16:56 ` tom.leiming
0 siblings, 0 replies; 22+ messages in thread
From: tom.leiming at gmail.com @ 2010-02-27 16:56 UTC (permalink / raw)
To: ath9k-devel
From: Ming Lei <tom.leiming@gmail.com>
The patch defines the fields of 'valid_single_stream' and 'valid' in
struct ath_rate_table as char type, so decrease the size of ath9k.ko
about 2KB.
old ath9k.ko
[tom at tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
69344 3080 168 72592 11b90 ath9k.ko
new ath9k.ko
[tom at tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
67304 3080 168 70552 11398 ath9k.ko
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
index 4f6d6fd..389168a 100644
--- a/drivers/net/wireless/ath/ath9k/rc.h
+++ b/drivers/net/wireless/ath/ath9k/rc.h
@@ -110,8 +110,8 @@ struct ath_rate_table {
int rate_cnt;
int mcs_start;
struct {
- int valid;
- int valid_single_stream;
+ char valid;
+ char valid_single_stream;
u8 phy;
u32 ratekbps;
u32 user_ratekbps;
--
1.6.2.5
^ permalink raw reply related [flat|nested] 22+ messages in thread* [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-02-27 16:56 ` tom.leiming
0 siblings, 0 replies; 22+ messages in thread
From: tom.leiming @ 2010-02-27 16:56 UTC (permalink / raw)
To: lrodriguez
Cc: linux-wireless, ath9k-devel, linville, johannes, stable, Ming Lei
From: Ming Lei <tom.leiming@gmail.com>
The patch defines the fields of 'valid_single_stream' and 'valid' in
struct ath_rate_table as char type, so decrease the size of ath9k.ko
about 2KB.
old ath9k.ko
[tom@tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
69344 3080 168 72592 11b90 ath9k.ko
new ath9k.ko
[tom@tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
67304 3080 168 70552 11398 ath9k.ko
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
index 4f6d6fd..389168a 100644
--- a/drivers/net/wireless/ath/ath9k/rc.h
+++ b/drivers/net/wireless/ath/ath9k/rc.h
@@ -110,8 +110,8 @@ struct ath_rate_table {
int rate_cnt;
int mcs_start;
struct {
- int valid;
- int valid_single_stream;
+ char valid;
+ char valid_single_stream;
u8 phy;
u32 ratekbps;
u32 user_ratekbps;
--
1.6.2.5
^ permalink raw reply related [flat|nested] 22+ messages in thread* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-27 16:56 ` tom.leiming
@ 2010-02-27 17:10 ` Larry Finger
-1 siblings, 0 replies; 22+ messages in thread
From: Larry Finger @ 2010-02-27 17:10 UTC (permalink / raw)
To: ath9k-devel
On 02/27/2010 10:56 AM, tom.leiming at gmail.com wrote:
> From: Ming Lei <tom.leiming@gmail.com>
>
> The patch defines the fields of 'valid_single_stream' and 'valid' in
> struct ath_rate_table as char type, so decrease the size of ath9k.ko
> about 2KB.
>
> old ath9k.ko
> [tom at tom-lei ath9k]$ size ath9k.ko
> text data bss dec hex filename
> 69344 3080 168 72592 11b90 ath9k.ko
>
> new ath9k.ko
> [tom at tom-lei ath9k]$ size ath9k.ko
> text data bss dec hex filename
> 67304 3080 168 70552 11398 ath9k.ko
>
> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
> ---
> drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
> index 4f6d6fd..389168a 100644
> --- a/drivers/net/wireless/ath/ath9k/rc.h
> +++ b/drivers/net/wireless/ath/ath9k/rc.h
> @@ -110,8 +110,8 @@ struct ath_rate_table {
> int rate_cnt;
> int mcs_start;
> struct {
> - int valid;
> - int valid_single_stream;
> + char valid;
> + char valid_single_stream;
> u8 phy;
> u32 ratekbps;
> u32 user_ratekbps;
Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
character data.
Larry
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-02-27 17:10 ` Larry Finger
0 siblings, 0 replies; 22+ messages in thread
From: Larry Finger @ 2010-02-27 17:10 UTC (permalink / raw)
To: tom.leiming
Cc: lrodriguez, linux-wireless, ath9k-devel, linville, johannes,
stable
On 02/27/2010 10:56 AM, tom.leiming@gmail.com wrote:
> From: Ming Lei <tom.leiming@gmail.com>
>
> The patch defines the fields of 'valid_single_stream' and 'valid' in
> struct ath_rate_table as char type, so decrease the size of ath9k.ko
> about 2KB.
>
> old ath9k.ko
> [tom@tom-lei ath9k]$ size ath9k.ko
> text data bss dec hex filename
> 69344 3080 168 72592 11b90 ath9k.ko
>
> new ath9k.ko
> [tom@tom-lei ath9k]$ size ath9k.ko
> text data bss dec hex filename
> 67304 3080 168 70552 11398 ath9k.ko
>
> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
> ---
> drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
> index 4f6d6fd..389168a 100644
> --- a/drivers/net/wireless/ath/ath9k/rc.h
> +++ b/drivers/net/wireless/ath/ath9k/rc.h
> @@ -110,8 +110,8 @@ struct ath_rate_table {
> int rate_cnt;
> int mcs_start;
> struct {
> - int valid;
> - int valid_single_stream;
> + char valid;
> + char valid_single_stream;
> u8 phy;
> u32 ratekbps;
> u32 user_ratekbps;
Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
character data.
Larry
^ permalink raw reply [flat|nested] 22+ messages in thread* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-27 17:10 ` Larry Finger
@ 2010-02-28 2:52 ` Ming Lei
-1 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-02-28 2:52 UTC (permalink / raw)
To: ath9k-devel
2010/2/28 Larry Finger <Larry.Finger@lwfinger.net>:
> On 02/27/2010 10:56 AM, tom.leiming at gmail.com wrote:
>> From: Ming Lei <tom.leiming@gmail.com>
>>
>> The patch defines the fields of 'valid_single_stream' and 'valid' in
>> struct ath_rate_table as char type, so decrease the size of ath9k.ko
>> about 2KB.
>>
>> old ath9k.ko
>> [tom at tom-lei ath9k]$ size ath9k.ko
>> ? ?text ? ?data ? ? bss ? ? dec ? ? hex filename
>> ? 69344 ? ?3080 ? ? 168 ? 72592 ? 11b90 ath9k.ko
>>
>> new ath9k.ko
>> [tom at tom-lei ath9k]$ size ath9k.ko
>> ? ?text ? ?data ? ? bss ? ? dec ? ? hex filename
>> ? 67304 ? ?3080 ? ? 168 ? 70552 ? 11398 ath9k.ko
>>
>> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
>> ---
>> ?drivers/net/wireless/ath/ath9k/rc.h | ? ?4 ++--
>> ?1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
>> index 4f6d6fd..389168a 100644
>> --- a/drivers/net/wireless/ath/ath9k/rc.h
>> +++ b/drivers/net/wireless/ath/ath9k/rc.h
>> @@ -110,8 +110,8 @@ struct ath_rate_table {
>> ? ? ? int rate_cnt;
>> ? ? ? int mcs_start;
>> ? ? ? struct {
>> - ? ? ? ? ? ? int valid;
>> - ? ? ? ? ? ? int valid_single_stream;
>> + ? ? ? ? ? ? char valid;
>> + ? ? ? ? ? ? char valid_single_stream;
>> ? ? ? ? ? ? ? u8 phy;
>> ? ? ? ? ? ? ? u32 ratekbps;
>> ? ? ? ? ? ? ? u32 user_ratekbps;
>
> Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
> character data.
Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
Thanks,
--
Lei Ming
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-02-28 2:52 ` Ming Lei
0 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-02-28 2:52 UTC (permalink / raw)
To: Larry Finger
Cc: lrodriguez, linux-wireless, ath9k-devel, linville, johannes,
stable
2010/2/28 Larry Finger <Larry.Finger@lwfinger.net>:
> On 02/27/2010 10:56 AM, tom.leiming@gmail.com wrote:
>> From: Ming Lei <tom.leiming@gmail.com>
>>
>> The patch defines the fields of 'valid_single_stream' and 'valid' in
>> struct ath_rate_table as char type, so decrease the size of ath9k.ko
>> about 2KB.
>>
>> old ath9k.ko
>> [tom@tom-lei ath9k]$ size ath9k.ko
>> text data bss dec hex filename
>> 69344 3080 168 72592 11b90 ath9k.ko
>>
>> new ath9k.ko
>> [tom@tom-lei ath9k]$ size ath9k.ko
>> text data bss dec hex filename
>> 67304 3080 168 70552 11398 ath9k.ko
>>
>> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
>> ---
>> drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
>> 1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
>> index 4f6d6fd..389168a 100644
>> --- a/drivers/net/wireless/ath/ath9k/rc.h
>> +++ b/drivers/net/wireless/ath/ath9k/rc.h
>> @@ -110,8 +110,8 @@ struct ath_rate_table {
>> int rate_cnt;
>> int mcs_start;
>> struct {
>> - int valid;
>> - int valid_single_stream;
>> + char valid;
>> + char valid_single_stream;
>> u8 phy;
>> u32 ratekbps;
>> u32 user_ratekbps;
>
> Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
> character data.
Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
Thanks,
--
Lei Ming
^ permalink raw reply [flat|nested] 22+ messages in thread* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-28 2:52 ` Ming Lei
@ 2010-02-28 3:19 ` Larry Finger
-1 siblings, 0 replies; 22+ messages in thread
From: Larry Finger @ 2010-02-28 3:19 UTC (permalink / raw)
To: ath9k-devel
On 02/27/2010 08:52 PM, Ming Lei wrote:
> 2010/2/28 Larry Finger <Larry.Finger@lwfinger.net>:
>> On 02/27/2010 10:56 AM, tom.leiming at gmail.com wrote:
>>> From: Ming Lei <tom.leiming@gmail.com>
>>>
>>> The patch defines the fields of 'valid_single_stream' and 'valid' in
>>> struct ath_rate_table as char type, so decrease the size of ath9k.ko
>>> about 2KB.
>>>
>>> old ath9k.ko
>>> [tom at tom-lei ath9k]$ size ath9k.ko
>>> text data bss dec hex filename
>>> 69344 3080 168 72592 11b90 ath9k.ko
>>>
>>> new ath9k.ko
>>> [tom at tom-lei ath9k]$ size ath9k.ko
>>> text data bss dec hex filename
>>> 67304 3080 168 70552 11398 ath9k.ko
>>>
>>> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
>>> ---
>>> drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
>>> 1 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
>>> index 4f6d6fd..389168a 100644
>>> --- a/drivers/net/wireless/ath/ath9k/rc.h
>>> +++ b/drivers/net/wireless/ath/ath9k/rc.h
>>> @@ -110,8 +110,8 @@ struct ath_rate_table {
>>> int rate_cnt;
>>> int mcs_start;
>>> struct {
>>> - int valid;
>>> - int valid_single_stream;
>>> + char valid;
>>> + char valid_single_stream;
>>> u8 phy;
>>> u32 ratekbps;
>>> u32 user_ratekbps;
>>
>> Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
>> character data.
>
> Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
It wasn't the size, but the content implied by the type.
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-02-28 3:19 ` Larry Finger
0 siblings, 0 replies; 22+ messages in thread
From: Larry Finger @ 2010-02-28 3:19 UTC (permalink / raw)
To: Ming Lei
Cc: lrodriguez, linux-wireless, ath9k-devel, linville, johannes,
stable
On 02/27/2010 08:52 PM, Ming Lei wrote:
> 2010/2/28 Larry Finger <Larry.Finger@lwfinger.net>:
>> On 02/27/2010 10:56 AM, tom.leiming@gmail.com wrote:
>>> From: Ming Lei <tom.leiming@gmail.com>
>>>
>>> The patch defines the fields of 'valid_single_stream' and 'valid' in
>>> struct ath_rate_table as char type, so decrease the size of ath9k.ko
>>> about 2KB.
>>>
>>> old ath9k.ko
>>> [tom@tom-lei ath9k]$ size ath9k.ko
>>> text data bss dec hex filename
>>> 69344 3080 168 72592 11b90 ath9k.ko
>>>
>>> new ath9k.ko
>>> [tom@tom-lei ath9k]$ size ath9k.ko
>>> text data bss dec hex filename
>>> 67304 3080 168 70552 11398 ath9k.ko
>>>
>>> Signed-off-by: Ming Lei <tom.leiming@gmail.com>
>>> ---
>>> drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
>>> 1 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
>>> index 4f6d6fd..389168a 100644
>>> --- a/drivers/net/wireless/ath/ath9k/rc.h
>>> +++ b/drivers/net/wireless/ath/ath9k/rc.h
>>> @@ -110,8 +110,8 @@ struct ath_rate_table {
>>> int rate_cnt;
>>> int mcs_start;
>>> struct {
>>> - int valid;
>>> - int valid_single_stream;
>>> + char valid;
>>> + char valid_single_stream;
>>> u8 phy;
>>> u32 ratekbps;
>>> u32 user_ratekbps;
>>
>> Why 'char' rather than 'u8'? To me, the latter implies a small integer, not
>> character data.
>
> Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
It wasn't the size, but the content implied by the type.
^ permalink raw reply [flat|nested] 22+ messages in thread* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-28 3:19 ` Larry Finger
@ 2010-03-01 15:13 ` Ming Lei
-1 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-03-01 15:13 UTC (permalink / raw)
To: ath9k-devel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-01 15:13 ` Ming Lei
0 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-03-01 15:13 UTC (permalink / raw)
To: Larry Finger; +Cc: lrodriguez, linux-wireless, ath9k-devel, linville, johannes
>From e74b075cdb143d45be9b371ee8a8e2dcfc15ab34 Mon Sep 17 00:00:00 2001
From: Ming Lei <tom.leiming@gmail.com>
Date: Sat, 27 Feb 2010 23:50:54 +0800
Subject: [PATCH] ath9k: decrease size of ath9k.ko
The patch defines the fields of 'valid_single_stream' and 'valid' in
struct ath_rate_table as char type, so decrease the size of ath9k.ko
about 2KB.
old ath9k.ko
[tom@tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
69344 3080 168 72592 11b90 ath9k.ko
new ath9k.ko
[tom@tom-lei ath9k]$ size ath9k.ko
text data bss dec hex filename
67304 3080 168 70552 11398 ath9k.ko
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
This version takes Larry's suggestion to define 'valid_single_stream'
and 'valid' as u8 instead of char.
---
drivers/net/wireless/ath/ath9k/rc.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/ath/ath9k/rc.h b/drivers/net/wireless/ath/ath9k/rc.h
index 4f6d6fd..aeaaa88 100644
--- a/drivers/net/wireless/ath/ath9k/rc.h
+++ b/drivers/net/wireless/ath/ath9k/rc.h
@@ -110,8 +110,8 @@ struct ath_rate_table {
int rate_cnt;
int mcs_start;
struct {
- int valid;
- int valid_single_stream;
+ u8 valid;
+ u8 valid_single_stream;
u8 phy;
u32 ratekbps;
u32 user_ratekbps;
--
1.6.2.5
On Sat, 27 Feb 2010 21:19:51 -0600
Larry Finger <Larry.Finger@lwfinger.net> wrote:
> On 02/27/2010 08:52 PM, Ming Lei wrote:
> > 2010/2/28 Larry Finger <Larry.Finger@lwfinger.net>:
> >> On 02/27/2010 10:56 AM, tom.leiming@gmail.com wrote:
> >>> From: Ming Lei <tom.leiming@gmail.com>
> >
> > Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
>
> It wasn't the size, but the content implied by the type.
^ permalink raw reply related [flat|nested] 22+ messages in thread* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-03-01 15:13 ` Ming Lei
@ 2010-03-03 5:02 ` Pavel Roskin
-1 siblings, 0 replies; 22+ messages in thread
From: Pavel Roskin @ 2010-03-03 5:02 UTC (permalink / raw)
To: ath9k-devel
On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
> - int valid;
> - int valid_single_stream;
> + u8 valid;
> + u8 valid_single_stream;
You can use bool instead, and that would give the same size saving while
being even more descriptive. I think using bool could be safer, as the
compiler would be able to detect some misuses and the values.
But I could get even more saving by using bool with the field width:
bool valid:1;
bool valid_single_stream:1;
That would place both variables into one byte. It may be ineffective
for speed, but it's more effective for storage.
In my configuration, I get following sizes:
original (int): 2792138
your patch (u8): 2790186
bool: 2790186
bool (1 bit): 2789218
--
Regards,
Pavel Roskin
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-03 5:02 ` Pavel Roskin
0 siblings, 0 replies; 22+ messages in thread
From: Pavel Roskin @ 2010-03-03 5:02 UTC (permalink / raw)
To: Ming Lei
Cc: Larry Finger, lrodriguez, linux-wireless, ath9k-devel, linville,
johannes
On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
> - int valid;
> - int valid_single_stream;
> + u8 valid;
> + u8 valid_single_stream;
You can use bool instead, and that would give the same size saving while
being even more descriptive. I think using bool could be safer, as the
compiler would be able to detect some misuses and the values.
But I could get even more saving by using bool with the field width:
bool valid:1;
bool valid_single_stream:1;
That would place both variables into one byte. It may be ineffective
for speed, but it's more effective for storage.
In my configuration, I get following sizes:
original (int): 2792138
your patch (u8): 2790186
bool: 2790186
bool (1 bit): 2789218
--
Regards,
Pavel Roskin
^ permalink raw reply [flat|nested] 22+ messages in thread
* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-03-03 5:02 ` Pavel Roskin
@ 2010-03-04 0:54 ` Luis R. Rodriguez
-1 siblings, 0 replies; 22+ messages in thread
From: Luis R. Rodriguez @ 2010-03-04 0:54 UTC (permalink / raw)
To: ath9k-devel
On Tue, Mar 02, 2010 at 09:02:44PM -0800, Pavel Roskin wrote:
> On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
> > - int valid;
> > - int valid_single_stream;
> > + u8 valid;
> > + u8 valid_single_stream;
>
> You can use bool instead, and that would give the same size saving while
> being even more descriptive. I think using bool could be safer, as the
> compiler would be able to detect some misuses and the values.
>
> But I could get even more saving by using bool with the field width:
>
> bool valid:1;
> bool valid_single_stream:1;
>
> That would place both variables into one byte. It may be ineffective
> for speed, but it's more effective for storage.
>
> In my configuration, I get following sizes:
>
> original (int): 2792138
> your patch (u8): 2790186
> bool: 2790186
> bool (1 bit): 2789218
Or use flags, and use BIT(1), BIT(2), etc for #defines for each stream mode.
Would save even more.
Luis
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-04 0:54 ` Luis R. Rodriguez
0 siblings, 0 replies; 22+ messages in thread
From: Luis R. Rodriguez @ 2010-03-04 0:54 UTC (permalink / raw)
To: Pavel Roskin
Cc: Ming Lei, Larry Finger, Luis Rodriguez,
linux-wireless@vger.kernel.org, ath9k-devel@lists.ath9k.org,
linville@tuxdriver.com, johannes@sipsolutions.net
On Tue, Mar 02, 2010 at 09:02:44PM -0800, Pavel Roskin wrote:
> On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
> > - int valid;
> > - int valid_single_stream;
> > + u8 valid;
> > + u8 valid_single_stream;
>
> You can use bool instead, and that would give the same size saving while
> being even more descriptive. I think using bool could be safer, as the
> compiler would be able to detect some misuses and the values.
>
> But I could get even more saving by using bool with the field width:
>
> bool valid:1;
> bool valid_single_stream:1;
>
> That would place both variables into one byte. It may be ineffective
> for speed, but it's more effective for storage.
>
> In my configuration, I get following sizes:
>
> original (int): 2792138
> your patch (u8): 2790186
> bool: 2790186
> bool (1 bit): 2789218
Or use flags, and use BIT(1), BIT(2), etc for #defines for each stream mode.
Would save even more.
Luis
^ permalink raw reply [flat|nested] 22+ messages in thread
* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-03-03 5:02 ` Pavel Roskin
@ 2010-03-06 14:47 ` Ming Lei
-1 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-03-06 14:47 UTC (permalink / raw)
To: ath9k-devel
2010/3/3 Pavel Roskin <proski@gnu.org>:
> On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
>> - ? ? ? ? ? ? int valid;
>> - ? ? ? ? ? ? int valid_single_stream;
>> + ? ? ? ? ? ? u8 valid;
>> + ? ? ? ? ? ? u8 valid_single_stream;
>
> You can use bool instead, and that would give the same size saving while
> being even more descriptive. ?I think using bool could be safer, as the
> compiler would be able to detect some misuses and the values.
>
> But I could get even more saving by using bool with the field width:
>
> bool valid:1;
> bool valid_single_stream:1;
>
> That would place both variables into one byte. ?It may be ineffective
> for speed, but it's more effective for storage.
>
> In my configuration, I get following sizes:
>
> original (int): ? ?2792138
> your patch (u8): ? 2790186
> bool: ? ? ? ? ? ? ?2790186
> bool (1 bit): ? ? ?2789218
1bit is not enough, at least 3bit is required for each one.
We does not save more using bit field since size of the least type
is 1 byte.
--
Lei Ming
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-06 14:47 ` Ming Lei
0 siblings, 0 replies; 22+ messages in thread
From: Ming Lei @ 2010-03-06 14:47 UTC (permalink / raw)
To: Pavel Roskin
Cc: Larry Finger, lrodriguez, linux-wireless, ath9k-devel, linville,
johannes
2010/3/3 Pavel Roskin <proski@gnu.org>:
> On Mon, 2010-03-01 at 23:13 +0800, Ming Lei wrote:
>> - int valid;
>> - int valid_single_stream;
>> + u8 valid;
>> + u8 valid_single_stream;
>
> You can use bool instead, and that would give the same size saving while
> being even more descriptive. I think using bool could be safer, as the
> compiler would be able to detect some misuses and the values.
>
> But I could get even more saving by using bool with the field width:
>
> bool valid:1;
> bool valid_single_stream:1;
>
> That would place both variables into one byte. It may be ineffective
> for speed, but it's more effective for storage.
>
> In my configuration, I get following sizes:
>
> original (int): 2792138
> your patch (u8): 2790186
> bool: 2790186
> bool (1 bit): 2789218
1bit is not enough, at least 3bit is required for each one.
We does not save more using bit field since size of the least type
is 1 byte.
--
Lei Ming
^ permalink raw reply [flat|nested] 22+ messages in thread
* [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-28 2:52 ` Ming Lei
@ 2010-03-01 12:19 ` Holger Schurig
-1 siblings, 0 replies; 22+ messages in thread
From: Holger Schurig @ 2010-03-01 12:19 UTC (permalink / raw)
To: ath9k-devel
> Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
Besides Johannes' comment: char is signed, u8 is unsigned. This can sometimes
byte you.
--
http://www.holgerschurig.de
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-01 12:19 ` Holger Schurig
0 siblings, 0 replies; 22+ messages in thread
From: Holger Schurig @ 2010-03-01 12:19 UTC (permalink / raw)
To: Ming Lei
Cc: Larry Finger, lrodriguez, linux-wireless, ath9k-devel, linville,
johannes, stable
> Either 'char' or 'u8' is OK, since both has one-byte size, isn't it?
Besides Johannes' comment: char is signed, u8 is unsigned. This can sometimes
byte you.
--
http://www.holgerschurig.de
^ permalink raw reply [flat|nested] 22+ messages in thread
* [ath9k-devel] [stable] [PATCH] ath9k: decrease size of ath9k.ko
2010-02-27 16:56 ` tom.leiming
@ 2010-03-10 23:52 ` Greg KH
-1 siblings, 0 replies; 22+ messages in thread
From: Greg KH @ 2010-03-10 23:52 UTC (permalink / raw)
To: ath9k-devel
On Sun, Feb 28, 2010 at 12:56:43AM +0800, tom.leiming at gmail.com wrote:
> From: Ming Lei <tom.leiming@gmail.com>
>
> The patch defines the fields of 'valid_single_stream' and 'valid' in
> struct ath_rate_table as char type, so decrease the size of ath9k.ko
> about 2KB.
Why would this be a -stable patch?
confused,
greg k-h
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [stable] [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-10 23:52 ` Greg KH
0 siblings, 0 replies; 22+ messages in thread
From: Greg KH @ 2010-03-10 23:52 UTC (permalink / raw)
To: tom.leiming
Cc: lrodriguez, ath9k-devel, linux-wireless, linville, johannes,
stable
On Sun, Feb 28, 2010 at 12:56:43AM +0800, tom.leiming@gmail.com wrote:
> From: Ming Lei <tom.leiming@gmail.com>
>
> The patch defines the fields of 'valid_single_stream' and 'valid' in
> struct ath_rate_table as char type, so decrease the size of ath9k.ko
> about 2KB.
Why would this be a -stable patch?
confused,
greg k-h
^ permalink raw reply [flat|nested] 22+ messages in thread
* [ath9k-devel] [stable] [PATCH] ath9k: decrease size of ath9k.ko
2010-03-10 23:52 ` Greg KH
@ 2010-03-11 0:42 ` Luis R. Rodriguez
-1 siblings, 0 replies; 22+ messages in thread
From: Luis R. Rodriguez @ 2010-03-11 0:42 UTC (permalink / raw)
To: ath9k-devel
On Wed, Mar 10, 2010 at 03:52:20PM -0800, Greg KH wrote:
> On Sun, Feb 28, 2010 at 12:56:43AM +0800, tom.leiming at gmail.com wrote:
> > From: Ming Lei <tom.leiming@gmail.com>
> >
> > The patch defines the fields of 'valid_single_stream' and 'valid' in
> > struct ath_rate_table as char type, so decrease the size of ath9k.ko
> > about 2KB.
>
> Why would this be a -stable patch?
>
> confused,
Its not, oh he e-mailed stable, yeah ignore.
Luis
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [stable] [PATCH] ath9k: decrease size of ath9k.ko
@ 2010-03-11 0:42 ` Luis R. Rodriguez
0 siblings, 0 replies; 22+ messages in thread
From: Luis R. Rodriguez @ 2010-03-11 0:42 UTC (permalink / raw)
To: Greg KH
Cc: tom.leiming@gmail.com, Luis Rodriguez,
ath9k-devel@venema.h4ckr.net, linux-wireless@vger.kernel.org,
linville@tuxdriver.com, johannes@sipsolutions.net,
stable@kernel.org
On Wed, Mar 10, 2010 at 03:52:20PM -0800, Greg KH wrote:
> On Sun, Feb 28, 2010 at 12:56:43AM +0800, tom.leiming@gmail.com wrote:
> > From: Ming Lei <tom.leiming@gmail.com>
> >
> > The patch defines the fields of 'valid_single_stream' and 'valid' in
> > struct ath_rate_table as char type, so decrease the size of ath9k.ko
> > about 2KB.
>
> Why would this be a -stable patch?
>
> confused,
Its not, oh he e-mailed stable, yeah ignore.
Luis
^ permalink raw reply [flat|nested] 22+ messages in thread
end of thread, other threads:[~2010-03-11 0:43 UTC | newest]
Thread overview: 22+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-02-27 16:56 [ath9k-devel] [PATCH] ath9k: decrease size of ath9k.ko tom.leiming at gmail.com
2010-02-27 16:56 ` tom.leiming
2010-02-27 17:10 ` [ath9k-devel] " Larry Finger
2010-02-27 17:10 ` Larry Finger
2010-02-28 2:52 ` [ath9k-devel] " Ming Lei
2010-02-28 2:52 ` Ming Lei
2010-02-28 3:19 ` [ath9k-devel] " Larry Finger
2010-02-28 3:19 ` Larry Finger
2010-03-01 15:13 ` [ath9k-devel] " Ming Lei
2010-03-01 15:13 ` Ming Lei
2010-03-03 5:02 ` [ath9k-devel] " Pavel Roskin
2010-03-03 5:02 ` Pavel Roskin
2010-03-04 0:54 ` [ath9k-devel] " Luis R. Rodriguez
2010-03-04 0:54 ` Luis R. Rodriguez
2010-03-06 14:47 ` [ath9k-devel] " Ming Lei
2010-03-06 14:47 ` Ming Lei
2010-03-01 12:19 ` [ath9k-devel] " Holger Schurig
2010-03-01 12:19 ` Holger Schurig
2010-03-10 23:52 ` [ath9k-devel] [stable] " Greg KH
2010-03-10 23:52 ` Greg KH
2010-03-11 0:42 ` [ath9k-devel] " Luis R. Rodriguez
2010-03-11 0:42 ` Luis R. Rodriguez
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.