[lkml]   [1996]   [Jul]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: About bdflush(update) & kernel 2.0
> > What you wrote is correct, except for the fact that what you call 'bdflush'
> > is actually 'update'. 'bdflush' is what 'kflushd' was called before being
> > implemented as a kernel thread.
> It doesn't seem to be relevant but for the record: both processes used to
> call (and still do, as explained above) the bdflush system call, only the
> arguments were different.

Indeed. update calls bdflush(2)'s function 1 to sync_old_buffers. Actually
it would only call sync(2) as a fallback if this call to bdflush(2)
failed (note : I may seem to have written something different earlier
in this thread. I read the sources more thouroughly since then :-) ).

> (this one hasn't changed). So which one is more bdflush?

Considering the comments in fs/buffer.c, I would say the answer is
kflushd : bdflush(2)'s function 1 only calls sync_old_buffers,
which has the comment "this function is essentially \"update\"" ;
bdflush() (the routine that constitutes the kflushd thread) also
has the mention "This is the actual bdflush daemon itself."

Thomas.Quinot@Cuivre.FdN.FR <URL:http://Web.FdN.FR/~tquinot/>

 \ /
  Last update: 2005-03-22 13:37    [W:0.030 / U:3.180 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site