Received-SPF: pass (google.com: domain of linux-kernel+bounces-12857-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Thu, 28 Dec 2023 12:52:32 -0800 (PST)
From: David Rientjes <rientjes@google.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cc: Pasha Tatashin <pasha.tatashin@soleen.com>, 
    Linus Torvalds <torvalds@linux-foundation.org>, rafael@kernel.org, 
    Andrew Morton <akpm@linux-foundation.org>, surenb@google.com, 
    linux-kernel@vger.kernel.org, linux-mm@kvack.org, souravpanda@google.com
Subject: Re: Sysfs one-value-per-file (was Re: [PATCH] vmstat: don't auto
 expand the sysfs files)
In-Reply-To: <2023122824-washout-shrubs-1d6d@gregkh>
Message-ID: <829410ca-1454-968e-b724-0ef0bfbca5cc@google.com>
References: <20231211154644.4103495-1-pasha.tatashin@soleen.com> <3d415ab4-e8c7-7e72-0379-952370612bdd@google.com> <CA+CK2bA2vZp3e+HHfB-sdLsPUYghMxvKcWURktDtNjwPL79Csw@mail.gmail.com> <b1049bfa-68c4-e237-30a9-1514a378c7f1@google.com>
 <CA+CK2bBxbvO-osm5XKk4VkaXYgfZXkDAtfayaYJ-vXo=QFqGPA@mail.gmail.com> <13e5fbd4-d84d-faba-47f1-d0024d2c572d@google.com> <2023122824-washout-shrubs-1d6d@gregkh>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII

On Thu, 28 Dec 2023, Greg Kroah-Hartman wrote:

> > I'd argue that the ship on the "sysfs one-value-per-file rule" has sailed 
> > for long-standing use cases where either (1) switching is just not 
> > possible or (2) switching would be an undue burden to the user.
> > 
> > An example of (1) would be THP enablement and defrag options:
> > 
> > $ grep . /sys/kernel/mm/transparent_hugepage/{defrag,enabled,shmem_enabled}
> > /sys/kernel/mm/transparent_hugepage/defrag:always defer defer+madvise [madvise] never
> > /sys/kernel/mm/transparent_hugepage/enabled:[always] madvise never
> > /sys/kernel/mm/transparent_hugepage/shmem_enabled:always within_size advise [never] deny force
> > 
> > This convention isn't going to change.  We're not going to suddenly add a 
> > new enablement or defrag option that can only be set in a newly added 
> > file that is one-value-per-file.
> > 
> > THP was obviously introduced before any sysfs "one-value-per-file rule" 
> 
> No, the rule has been there since "day one" for sysfs, this file snuck
> in much later with no one noticing it against the "rules" and I've been
> complaining about it every time someone tries to add a new field to it
> that I notice.
> 

Ah, gotcha, thanks.  I had assumed that the push for one-value-per-file 
started after thp, and perhaps even because of thp :)  I have to admit 
that whenever I log into a new server type one of the first things I do is

$ cat /sys/devices/system/node/node*/distance

and that table just makes intuitive sense.  If we were to go back in time 
and reimplement that as one-value-per-file, I'd just assume that many 
userspace implementations would just need to read 64 different files to 
structure it into the same exact table.

On the other hand, I have wished countless times that the thp settings 
would have actually been one-value-per-file from the start.