[#102687] [Ruby master Bug#17666] Sleep in a thread hangs when Fiber.set_scheduler is set — arjundas.27586@...

Issue #17666 has been reported by arjunmdas (arjun das).

16 messages 2021/03/02

[#102776] [Ruby master Bug#17678] Ractors do not restart after fork — knuckles@...

Issue #17678 has been reported by ivoanjo (Ivo Anjo).

8 messages 2021/03/08

[#102797] [Ruby master Feature#17684] Remove `--disable-gems` from release version of Ruby — hsbt@...

Issue #17684 has been reported by hsbt (Hiroshi SHIBATA).

17 messages 2021/03/10

[#102829] [Ruby master Bug#17718] a method paramaters object that can be pattern matched against — dsisnero@...

Issue #17718 has been reported by dsisnero (Dominic Sisneros).

9 messages 2021/03/11

[#102832] [Ruby master Misc#17720] Cirrus CI to check non-x86_64 architecture cases by own machines — jaruga@...

Issue #17720 has been reported by jaruga (Jun Aruga).

19 messages 2021/03/12

[#102850] [Ruby master Bug#17723] autoconf 2.70+ is not working with master branch — hsbt@...

Issue #17723 has been reported by hsbt (Hiroshi SHIBATA).

11 messages 2021/03/14

[#102884] [Ruby master Bug#17725] Prepend Breaks Ability to Alias — josh@...

Issue #17725 has been reported by joshuadreed (Josh Reed).

14 messages 2021/03/16

[#102914] [Ruby master Bug#17728] [BUG] Segmentation fault at 0x0000000000000000 — denthebat@...

Issue #17728 has been reported by meliborn (Denis Denis).

13 messages 2021/03/18

[#102919] [Ruby master Bug#17730] Ruby on macOS transitively links to ~150 dylibs — rickmark@...

Issue #17730 has been reported by rickmark (Rick Mark).

10 messages 2021/03/18

[#103013] [Ruby master Bug#17748] Ruby 3.0 takes a long time to resolv DNS of nonexistent domains — xdmx@...

Issue #17748 has been reported by xdmx (Eric Bloom).

8 messages 2021/03/25

[#103026] [Ruby master Feature#17749] Const source location without name — tenderlove@...

Issue #17749 has been reported by tenderlovemaking (Aaron Patterson).

10 messages 2021/03/25

[#103036] [Ruby master Misc#17751] Do these instructions (<<, +, [0..n]) modify the original string without creating copies? — cart4for1@...

Issue #17751 has been reported by stiuna (Juan Gregorio).

11 messages 2021/03/26

[#103040] [Ruby master Feature#17752] Enable -Wundef for C extensions in repository — eregontp@...

Issue #17752 has been reported by Eregon (Benoit Daloze).

23 messages 2021/03/26

[#103044] [Ruby master Feature#17753] Add Module#outer_scope — tenderlove@...

Issue #17753 has been reported by tenderlovemaking (Aaron Patterson).

31 messages 2021/03/26

[#103088] [Ruby master Feature#17760] Where we should install a header file when `gem install --user`? — muraken@...

Issue #17760 has been reported by mrkn (Kenta Murata).

11 messages 2021/03/30

[#103102] [Ruby master Feature#17762] A simple way to trace object allocation — mame@...

Issue #17762 has been reported by mame (Yusuke Endoh).

18 messages 2021/03/30

[#103105] [Ruby master Feature#17763] Implement cache for cvars — eileencodes@...

Issue #17763 has been reported by eileencodes (Eileen Uchitelle).

18 messages 2021/03/30

[ruby-core:102676] [Ruby master Bug#17664] Behavior of sockets changed in Ruby 3.0 to non-blocking

From: merch-redmine@...
Date: 2021-03-01 19:33:19 UTC
List: ruby-core #102676
Issue #17664 has been updated by jeremyevans0 (Jeremy Evans).

Assignee set to ioquatix (Samuel Williams)
Status changed from Open to Assigned

I believe this is expected, even if not mentioned in the release notes.  I think the Ruby-level API remains the same, but passing the underlying file descriptors to C functions can see changed behavior.  Assigning to @ioquatix to confirm this is expected.

----------------------------------------
Bug #17664: Behavior of sockets changed in Ruby 3.0 to non-blocking
https://bugs.ruby-lang.org/issues/17664#change-90670

* Author: ciconia (Sharon Rosner)
* Status: Assigned
* Priority: Normal
* Assignee: ioquatix (Samuel Williams)
* ruby -v: 3.0.0
* Backport: 2.5: UNKNOWN, 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN
----------------------------------------
I'm not sure this is a bug, but apparently a change was introduced in Ruby 3.0 that makes sockets non-blocking by default. This change was apparently introduced as part of the work on the [FiberScheduler interface](https://github.com/ruby/ruby/blame/78f188524f551c97b1a7a44ae13514729f1a21c7/ext/socket/init.c#L411-L434). This change of behaviour is not discussed in the Ruby 3.0.0 release notes.

This change complicates the implementation of an io_uring-based fiber scheduler, since io_uring SQE's on fd's with `O_NONBLOCK` can return `EAGAIN` just like normal syscalls. Using io_uring with non-blocking fd's defeats the whole purpose of using io_uring in the first place.

A workaround I have put in place in the Polyphony [io_uring backend](https://github.com/digital-fabric/polyphony/blob/d3c9cf3ddc1f414387948fa40e5f6a24f68bf045/ext/polyphony/backend_io_uring.c#L28-L47) is to make sure `O_NONBLOCK` is not set before attempting I/O operations on any fd.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>

In This Thread