[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180131175001.GA1433@ZenIV.linux.org.uk>
Date: Wed, 31 Jan 2018 17:50:01 +0000
From: Al Viro <viro@...IV.linux.org.uk>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
linux-fsdevel <linux-fsdevel@...r.kernel.org>,
linux-arch <linux-arch@...r.kernel.org>
Subject: Re: [git pull] vfs.git: poll annotations
On Wed, Jan 31, 2018 at 04:46:20AM +0000, Al Viro wrote:
> MSG and REMOVE are never triggered, so it's all down to {WR{NORM,BAND},RDHUP}.
Actually, POLLREMOVE is an amusing story (FSVO amusing): it's a part of
UnixWare /dev/poll ABI. It had _not_ lasted in Linux - epoll_ctl()
EPOLL_CTL_DEL is used for the same purpose. Chronology:
Oct 29 2002: davidel introduces it into the tree (and defines for i386)
Oct 31 2002: jgarzik has noticed broken alpha build and added definition
within a few hours the same had been done for itanic (davidm)
afternoon of the same day: Alan's merge of v850 adds definition there
Nov 1 2002: rmk adds definition for arm
Nov 2 2002: anton does the same for ppc64
Nov 2 2002: ak does the same for amd64
Nov 2 2002: davidel removes the sole user of that thing.
But that, of course, does not stop the propagation of the (now and forever)
useless definition - the same day it infects m68k; next day it's
parisc and sparc. On Nov 8 it hits ppc32, Nov 17 - s390.
In June 2003 mips gets it, in July sh is brought in with that
thing already in it and the damn thing spreads to cris. And so it goes;
the last flash of infection had been in 2007 when asm-generic/poll.h
had been introduced.
Its only user has lasted in the tree for four days. Definitions
will probably stay forever...
Powered by blists - more mailing lists