ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.3
(Generate patch)

Comparing libev/ev.3 (file contents):
Revision 1.112 by root, Sun Jun 23 23:27:34 2019 UTC vs.
Revision 1.113 by root, Mon Jun 24 19:53:47 2019 UTC

… …
131.\} 131.\}
132.rm #[ #] #H #V #F C 132.rm #[ #] #H #V #F C
133.\" ======================================================================== 133.\" ========================================================================
134.\" 134.\"
135.IX Title "LIBEV 3" 135.IX Title "LIBEV 3"
136.TH LIBEV 3 "2019-06-23" "libev-4.25" "libev - high performance full featured event loop" 136.TH LIBEV 3 "2019-06-24" "libev-4.25" "libev - high performance full featured event loop"
137.\" For nroff, turn off justification. Always turn off hyphenation; it makes 137.\" For nroff, turn off justification. Always turn off hyphenation; it makes
138.\" way too many mistakes in technical documents. 138.\" way too many mistakes in technical documents.
139.if n .ad l 139.if n .ad l
140.nh 140.nh
141.SH "NAME" 141.SH "NAME"
… …
710be well worth enabling it \- if it isn't available in your kernel this will 710be well worth enabling it \- if it isn't available in your kernel this will
711be detected and this backend will be skipped. 711be detected and this backend will be skipped.
712.Sp 712.Sp
713This backend can batch oneshot requests and supports a user-space ring 713This backend can batch oneshot requests and supports a user-space ring
714buffer to receive events. It also doesn't suffer from most of the design 714buffer to receive events. It also doesn't suffer from most of the design
715problems of epoll (such as not being able to remove event sources from 715problems of epoll (such as not being able to remove event sources from the
716the epoll set), and generally sounds too good to be true. Because, this 716epoll set), and generally sounds too good to be true. Because, this being
717being the linux kernel, of course it suffers from a whole new set of 717the linux kernel, of course it suffers from a whole new set of limitations.
718limitations.
719.Sp 718.Sp
720For one, it is not easily embeddable (but probably could be done using 719For one, it is not easily embeddable (but probably could be done using
721an event fd at some extra overhead). It also is subject to a system wide 720an event fd at some extra overhead). It also is subject to a system wide
722limit that can be configured in \fI/proc/sys/fs/aio\-max\-nr\fR \- each loop 721limit that can be configured in \fI/proc/sys/fs/aio\-max\-nr\fR \- each loop
723currently requires \f(CW61\fR of this number. If no aio requests are left, this 722currently requires \f(CW61\fR of this number. If no aio requests are left, this
724backend will be skipped during initialisation. 723backend will be skipped during initialisation.
725.Sp 724.Sp
726Most problematic in practise, however, is that not all file descriptors 725Most problematic in practise, however, is that not all file descriptors
727work with it. For example, in linux 5.1, tcp sockets, pipes, event fds, 726work with it. For example, in linux 5.1, tcp sockets, pipes, event fds,
728files, \fI/dev/null\fR and a few others are supported, but ttys do not work 727files, \fI/dev/null\fR and a few others are supported, but ttys do not work
729(probably because of a bug), so this is not (yet?) a generic event polling 728properly (a known bug that the kernel developers don't care about, see
730interface. 729<https://lore.kernel.org/patchwork/patch/1047453/>), so this is not
730(yet?) a generic event polling interface.
731.Sp 731.Sp
732Overall, it seems the linux developers just don't want it to have a
733generic event handling mechanism other than \f(CW\*(C`select\*(C'\fR or \f(CW\*(C`poll\*(C'\fR.
734.Sp
732To work around this latter problem, the current version of libev uses 735To work around the fd type problem, the current version of libev uses
733epoll as a fallback for file deescriptor types that do not work. Epoll 736epoll as a fallback for file deescriptor types that do not work. Epoll
734is used in, kind of, slow mode that hopefully avoids most of its design 737is used in, kind of, slow mode that hopefully avoids most of its design
735problems and requires 1\-3 extra syscalls per active fd every iteration. 738problems and requires 1\-3 extra syscalls per active fd every iteration.
736.Sp 739.Sp
737This backend maps \f(CW\*(C`EV_READ\*(C'\fR and \f(CW\*(C`EV_WRITE\*(C'\fR in the same way as 740This backend maps \f(CW\*(C`EV_READ\*(C'\fR and \f(CW\*(C`EV_WRITE\*(C'\fR in the same way as

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines