Janitor-Question: use __set_bit instead of |=

Janitor-Question: use __set_bit instead of |=

[Peter Hüwe]

Hi Janitors, staging-list

what is your opinion on using set_bit instead of using |= to set a bit?
Is it worth the effort to convert  existing |= to set_bit?

__set_bit
pro:
- often implemented in optimized assembly (e.g. for x86)
- intention might be clearer
- less error prone
- "they are the only portable way to set a specific bit"
according to  Robert Love's Linux Kernel Development third edition, p.183

cons:
uses unsigned longs


|pro:
- standard C
- let's the compiler decide
- no warnings on chars, shorts, ints


Thanks,
Peter

Re: Janitor-Question: use __set_bit instead of |=

[Greg KH]

On Mon, May 30, 2011 at 11:23:40PM +0200, Peter Hüwe wrote:
> Hi Janitors, staging-list
> 
> what is your opinion on using set_bit instead of using |= to set a bit?
> Is it worth the effort to convert  existing |= to set_bit?

I doubt converting existing stuff is needed, but any future stuff is
always nice to use set_bit() for the reasons you suggest.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Janitor-Question: use __set_bit instead of |=

[walter harms]

Am 30.05.2011 23:23, schrieb Peter Hüwe:
> Hi Janitors, staging-list
> 
> what is your opinion on using set_bit instead of using |= to set a bit?
> Is it worth the effort to convert  existing |= to set_bit?
> 
> __set_bit
> pro:
> - often implemented in optimized assembly (e.g. for x86)
> - intention might be clearer
> - less error prone
> - "they are the only portable way to set a specific bit"
> according to  Robert Love's Linux Kernel Development third edition, p.183
> 
> cons:
> uses unsigned longs
> 
> 
> |> pro:
> - standard C
> - let's the compiler decide
> - no warnings on chars, shorts, ints
> 
> 

Personaly i do not like it, but to be fair
|= must not be obvious but it depends on the situation.

It is std C and the compiler should handle it.

re,
 wh
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Janitor-Question: use __set_bit instead of |=

[Jiri Slaby]

On 05/30/2011 11:23 PM, Peter Hüwe wrote:
> Hi Janitors, staging-list
> 
> what is your opinion on using set_bit instead of using |= to set a bit?
> Is it worth the effort to convert  existing |= to set_bit?
> 
> __set_bit
> pro:
> - often implemented in optimized assembly (e.g. for x86)
> - intention might be clearer
> - less error prone
> - "they are the only portable way to set a specific bit"
> according to  Robert Love's Linux Kernel Development third edition, p.183
> 
> cons:
> uses unsigned longs

Note that you need to define a different set of macros.
E.g. if you have for |=:
#define FLAG1 0x01
#define FLAG2 0x02
#define FLAG3 0x40

for set_bit you need:
#define FLAG1 0
#define FLAG2 1
#define FLAG3 6

Also with set_bit you can set only one bit at a time which might make
the code longer and unreadable. For examples, see input layer.

> |> pro:
> - standard C
> - let's the compiler decide
> - no warnings on chars, shorts, ints

regards,
-- 
js
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html