user/dev discussion of public-inbox itself
 help / color / mirror / code / Atom feed
From: Eric Wong <e@80x24.org>
To: meta@public-inbox.org
Subject: [PATCH 2/9] listener: switch to level-triggered epoll
Date: Fri,  1 Oct 2021 09:54:38 +0000	[thread overview]
Message-ID: <20211001095445.9326-3-e@80x24.org> (raw)
In-Reply-To: <20211001095445.9326-1-e@80x24.org>

On second thought, the ->requeue + accept retry code path isn't
worth the userspace complexity and overhead.  Level-triggered
epoll has always annoyed me since it takes an inefficient code
path in the kernel; but taking our less-efficient code path in
Perl seems even worse.  We also need to take load distribution
into account for multi-worker systems.
---
 lib/PublicInbox/Listener.pm | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/lib/PublicInbox/Listener.pm b/lib/PublicInbox/Listener.pm
index 64bba5b0fd7c..09f1f2e5fabd 100644
--- a/lib/PublicInbox/Listener.pm
+++ b/lib/PublicInbox/Listener.pm
@@ -7,7 +7,7 @@ use strict;
 use parent 'PublicInbox::DS';
 use Socket qw(SOL_SOCKET SO_KEEPALIVE IPPROTO_TCP TCP_NODELAY);
 use IO::Handle;
-use PublicInbox::Syscall qw(EPOLLIN EPOLLEXCLUSIVE EPOLLET);
+use PublicInbox::Syscall qw(EPOLLIN EPOLLEXCLUSIVE);
 use Errno qw(EAGAIN ECONNABORTED EPERM);
 
 # Warn on transient errors, mostly resource limitations.
@@ -22,7 +22,7 @@ sub new ($$$) {
 	setsockopt($s, IPPROTO_TCP, TCP_NODELAY, 1); # ignore errors on non-TCP
 	listen($s, 2**31 - 1); # kernel will clamp
 	my $self = bless { post_accept => $cb }, $class;
-	$self->SUPER::new($s, EPOLLIN|EPOLLET|EPOLLEXCLUSIVE);
+	$self->SUPER::new($s, EPOLLIN|EPOLLEXCLUSIVE);
 }
 
 sub event_step {
@@ -38,7 +38,6 @@ sub event_step {
 		IO::Handle::blocking($c, 0); # no accept4 :<
 		eval { $self->{post_accept}->($c, $addr, $sock) };
 		warn "E: $@\n" if $@;
-		$self->requeue;
 	} elsif ($! == EAGAIN || $! == ECONNABORTED || $! == EPERM) {
 		# EAGAIN is common and likely
 		# ECONNABORTED is common with bad connections

  parent reply	other threads:[~2021-10-01  9:54 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-01  9:54 [PATCH 0/9] daemon-related things Eric Wong
2021-10-01  9:54 ` [PATCH 1/9] doc: lei-security: some more updates Eric Wong
2021-10-01  9:54 ` Eric Wong [this message]
2021-10-01  9:54 ` [PATCH 3/9] daemon: make SO_ACCEPTFILTER a shared variable Eric Wong
2021-10-01  9:54 ` [PATCH 4/9] ipc: run Net::SSLeay::randomize Eric Wong
2021-10-01  9:54 ` [PATCH 5/9] ds: simplify signalfd use Eric Wong
2021-10-01  9:54 ` [PATCH 6/9] inbox: inline and eliminate git_cleanup Eric Wong
2021-10-01  9:54 ` [PATCH 7/9] inbox: keep DB handles if git processes are live Eric Wong
2021-10-01  9:54 ` [PATCH 8/9] ds: inline set_cloexec Eric Wong
2021-10-01  9:54 ` [PATCH 9/9] doc: lei-daemon: new manpage Eric Wong
2021-10-02  0:19   ` Kyle Meyer
2021-10-02  8:08     ` Eric Wong
2021-10-04  3:16       ` Kyle Meyer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://public-inbox.org/README

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211001095445.9326-3-e@80x24.org \
    --to=e@80x24.org \
    --cc=meta@public-inbox.org \
    --subject='Re: [PATCH 2/9] listener: switch to level-triggered epoll' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Code repositories for project(s) associated with this inbox:

	https://80x24.org/public-inbox.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).