similar to: [PATCH] [RFC] epoll based ioloop handler

Hello, As part of the Rust bindings for Libnbd, I try to integrate the asynchronous (aio_*) functions with Tokio <https://docs.rs/tokio/latest/tokio/>, the most used asynchronous runtime in Rust. However, in its eventloop, Tokio uses epoll(7) instead of poll(2) (which is used internally in Libnbd). The difference is that poll(2) uses level-triggered notifications as aposed to epoll(7)

Sorry, I forgot to attach patch to previous mail. -- Andrey Panin | Linux and UNIX system administrator pazke at donpac.ru | PGP key: wwwkeys.pgp.net -------------- next part -------------- diff -urpNX /usr/share/dontdiff dovecot-cvs.vanilla/configure.in dovecot-cvs/configure.in --- dovecot-cvs.vanilla/configure.in 2004-08-09 23:14:02.000000000 +0400 +++ dovecot-cvs/configure.in 2004-08-25

Hello, attached patch adds experimental ioloop handler which uses BSD kqueue API. It compiles and survives my small test program, but not thoroughly tested because I have no suitable BSD machine to torture. Brave owners of (Free|Net|Open)BSD please test and report results :) Best regards. -- Andrey Panin | Linux and UNIX system administrator pazke at donpac.ru | PGP key: wwwkeys.pgp.net

On Wed, Jul 5, 2023 at 3:38?PM Tage Johansson <tage.j.lists at posteo.net> wrote: > As part of the Rust bindings for Libnbd, I try to integrate the > asynchronous (aio_*) functions with Tokio > <https://docs.rs/tokio/latest/tokio/>, the most used asynchronous runtime > in Rust. However, in its eventloop, Tokio uses epoll(7) instead of poll(2) > (which is used internally

The following program is a minor variant on the sample given in the README with eventmachine-0.8.1. I am running under CentOS 4.5 (kernel 2.6.9-55.0.2.plus.c4) with a ruby-1.8.5 RPM from the CentOS testing repository. When I run it, I find that set_comm_inactivity_timeout doesn''t do anything. That is, if I client opens a connection, it stays open indefinitely. However, if I comment out

I am running dovecot 1.0 beta3 as a front-end proxy, and getting the following error when using epoll. The error goes away if I switch to poll. On the backend IMAP server I see that the login was successful. The error occurs when I try to retrieve an email or get a folder listing via IMAP after logging in successfully... Feb 9 15:35:57 director4 dovecot: imap-login: proxy(bboebel at

Sorry this is going to be a bit vague, but I''ve noticed something rather odd going on with EventMachine on Linux. I''ve written something uses EventMachine to proxy HTTP to other processes. On OS X, it works great, adding only about 20% extra latency into the connection when proxying versus connecting to the original backend process directly, so a 5ms connection might now take

>From my log: Dec 29 16:43:58 postamt dovecot: Dovecot v1.1.7 starting up Dec 29 16:43:59 postamt dovecot: Fatal: imap-login: epoll_create(): Too many open files Dec 29 16:43:59 postamt dovecot: Temporary failure in creating login processes, slowing down for now Dec 29 16:43:59 postamt dovecot: Fatal: imap-login: epoll_create(): Too many open files Dec 29 16:43:59 postamt dovecot: Fatal:

Hello, this patch adds missing epoll_create() error check. Please consider applying. Best regards. -- Andrey Panin | Linux and UNIX system administrator pazke at donpac.ru | PGP key: wwwkeys.pgp.net -------------- next part -------------- diff -urpX /usr/share/dontdiff dovecot-1.0-test46.vanilla/src/lib/ioloop-epoll.c dovecot-1.0-test46/src/lib/ioloop-epoll.c ---

Hello, I'd suggest a way to get to know the "capabilities" of a compiled Dovecot. The background is that I've run the sequence ./configure --with-ldap && make && make install twice, all three commands tell me that all is fine (ldap is listed in the modules), but still dovecot told me that ldap is no Usedb module. A "make clean" before solved the

Hi List! Attached is a patch to configure.in to test that epoll, poll, inotify, kqueue and dnotify actually work. It also attempts to select the best one for you if none specified. This supercedes my previous patch in the re beta9 thread. Hopefully get this in Gentoo's beta9 ebuild soon. Thanks -- Roy Marples <uberlord@gentoo.org> Gentoo/Linux Developer (baselayout, networking)

I know this error has been reported before, but a couple ppl have reported getting this randomly when running Thin and it crashed the server: terminate called after throwing an instance of ''std::runtime_error'' what(): unable to delete epoll event: Bad file descriptor I think it''s fixed on trunk now so I''m telling people that have this issue to checkout

(Sorry for the late reply, was a bit involved in a qemu bug last week ...) On Fri, Jul 07, 2023 at 08:58:50AM +0000, Tage Johansson wrote: > On 7/6/2023 7:06 PM, Nir Soffer wrote: > > > - After calling for example aio_notify_read(3), can I know that the next > reading from the file descriptor would block? > > No, you have to call again aio_get_direction() and

Today I finally caught up with my mail. I've now read all the mails from the Dovecot list from this year and all my private mails from last few years (totaling around 500 unread mails). If you think there's still something I should have replied to, please re-send it because I've either missed it or just skipped it for some reason. (Except there are still a couple of mails from this

Attempting to use dovecot-1.0.rc16 on centos built with: env CPPFLAGS=-I/usr/kerberos/include \ ./configure --prefix=/local/service/dovecot --with-ssl=openssl \ --with-ioloop=best My config has changed from example: protocols = imap ssl_disable = yes (because not pem) Plus I set to yes all the debug and verbose options. I couldn't login via IMAP: Jan 5 19:52:28 hostabc

On 7/6/2023 7:06 PM, Nir Soffer wrote: > > - After calling for example aio_notify_read(3), can I know that the > next reading from the file descriptor would block? > > No, you have to call again aio_get_direction() and poll again until > the event happens. > Well, what I mean is: After calling aio_notify_read, if aio_get_direction returns AIO_DIRECTION_READ or

Hi, I am using dovcot on a Fritz!Box (Linux fritz 2.6.19.2 #1 Fri Apr 18 16:00:34 CEST 2008 mips unknown) And after upgrading from 1.0.9 to 1.0.15 I am getting above assertion, when I do some action like moving or deleting on the IMAP mail box using MailDir on ext2 fs. I use --with-ioloop=epoll and --with-mem-align=4 as build configuration. For dovecof configuration see below details. Thanks

I occasionally get a fatal error unable to delete epoll event: Bad file descriptor I think the attached patch will fix it. The patch does two things: 1) changes the error we look for from ENOENT to EBADF. It is hard to reproduce this error, but if I change the epoll_ctl line to epoll_ctl(epfd, EPOLL_CTL_DEL, 666, ed->GetEpollEvent()); it does return -1 with errno = EBADF on my Linux 2.6.9

Hi, After reviewing my dovecot ./configure options I thought I'd give --with-ioloop=epoll a go again (as opposed to just using =poll) but I'm seeing some errors appearing in the maillog that aren't there with standard poll: Dec 23 02:34:47 tornado dovecot: imap-login: Login: user=<reuben>, auth-method=plain, remote=220.237.24.149, TLS Dec 23 02:34:48 tornado dovecot:

Hi, i had installed a CentOS 5 x86_64 xen guest, CentOS 5 x86_64 is the host. The kernel versions are: Host: 2.6.18-53.1.14.el5xen Guest: 2.6.18-53.1.14.el5xen But i have a problem when i try to start postfix. The log says: # fatal: event_enable_read: epoll_ctl EPOLL_CTL_ADD: Operation not permitted I did a searh on Internet and found some post saying that Xen x86_64 does not has support