[clug] best filesystem for solid state drive?

[Robert Edwards]

David Tulloh wrote:
> Jason Stokes wrote:
>> Aparently Linus is an advocate of noatime generally, viewing the atime
>> attribute as a brain-damaged idea.  I'm inclined to agree -- but I'm 
>> sure this breaks something.  Anything important, though?
>>   
> POSIX compliance.
> 
> The POSIX standard includes the C POSIX library, which in turn includes 
> sys/stat.h
> Stat.h defines stat/lstat/fstat that return the stat structure which 
> includes information like the owner, file size, mtime, ctime and atime.
> 
> So to be POSIX compliant atime has to be recorded and available, which 
> is why it's the default on Linux.
> 
> 
> Many people use noatime because it's slow and due to this most 
> applications don't rely on it but to remove atime or change the default 
> at the kernel level further removes Linux's POSIX compliance and 
> shouldn't be done lightly.
> 
> 
> 
> David

 From man stat(2):

   Not all of the Linux filesystems implement all of the time fields.
   Some file system types allow mounting in such a way that file
   accesses do not cause an update of the st_atime field. (See "noatime"
   in mount(8).)

So, in some sense, there is already some disconnect between the reality
of Linux filesystems (the DOS ones in particular) and the POSIX'ness of
the atime field.

And, I like this (also in stat(2)):

   Other routines, like mmap(2), may or may not update st_atime.

Bob Edwards (not such a big fan of the atime field).