Re: removing bcopy... because it's half broken

[Matthias-Christian Ott <matthias.christian@tiscali.de>]

Adrian Bunk wrote:

>On Sun, Jan 09, 2005 at 12:19:09PM -0800, Linus Torvalds wrote:
>  
>
>>On Sun, 9 Jan 2005, Arjan van de Ven wrote:
>>    
>>
>>>Instead of fixing this inconsistency, I decided to remove it entirely,
>>>explicit memcpy() and memmove() are prefered anyway (welcome to the 1990's)
>>>and nothing in the kernel is using these functions, so this saves code size
>>>as well for everyone.
>>>      
>>>
>>The problem is that at least some gcc versions would historically generate
>>calls to "bcopy" on alpha for structure assignments. Maybe it doesn't any
>>more, and no such old gcc versions exist any more, but who knows?
>>...
>>    
>>
>
>include/asm-alpha/string.h says:
>
>  /*
>   * GCC of any recent vintage doesn't do stupid things with bcopy.
>   * EGCS 1.1 knows all about expanding memcpy inline, others don't.
>   *
>   * Similarly for a memset with data = 0.
>   */
>
>
>And Arjan's patch is pretty low-risk:
>
>If it breaks on any architecture with any supported compiler (>= 2.95), 
>it will break at compile time and there will pretty fast be reports of 
>this breakage in which case it would be easy to revert his patch.
>
>
>  
>
>>		Linus
>>    
>>
>
>cu
>Adrian
>
>  
>
Jep and most people are using a newer gcc and newer glibc. Linux has to 
stay up2date, so I think it's the best choise to create a modern and 
clean (without broken functions) OS.

Matthias-Christian Ott