[lkml]   [2008]   [Apr]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 0/3] x86: add cpuset_scnprintf function

> However doing this is worse in my view than simply breaking the format
> outright, unilaterally and irrevocably. If you just flat out stick a
> fork in an API and break it hard on some release, then at least user
> space knows that it must adapt or die at that version. If you hand
> user space the means to break that API, then any properly and
> defensively written user code has to be prepared to deal with both API
> flavors, and the majority of user space code is broken half the time,
> when run on a system with the API variant it wasn't expecting. More
> over, you end up with apps having "toilet seat wars" with each other:
> you left it up and it should be down; no you left it down and it should
> be up. Not a pretty sight.
> Perhaps I totally misunderstand this patchset ?


I wanted to not break current apps unmercifully, but perhaps I should
default it to the "non-compatible" mode (and adjust the schedstat version
to indicate this)? [It's the only output that I found that seemed to care.]

And if users have apps that they can't convert, they can revert to the
"old" (compatible) method of outputs. I know if I'm a user and I'm really
interested in understanding the outputs when there's hundreds and hundreds
of cpus, then the more compact format is much more useful.

I can't believe there hasn't been many changes in all of these outputs.
Like what happened before Hyperthreading, or 3rd level caches, or ?
Even the new Intel announcements for Nehalem may introduce more changes
in what's important in the output information. Plus I was under the
impression that one of the basic tenets of Linux was that API's can and
will change?


 \ /
  Last update: 2008-04-02 09:51    [W:0.066 / U:0.440 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site