public inbox for cygwin@cygwin.com
 help / color / mirror / Atom feed
From: Thomas Wolff <towo@towo.net>
To: cygwin@cygwin.com
Subject: Re: Race condition hangs on multiple mintty/tcsh? Brad Wetmore
Date: Thu, 6 Aug 2020 15:31:24 +0200	[thread overview]
Message-ID: <3d594543-c992-a22c-4631-f786cf34f76b@towo.net> (raw)
In-Reply-To: <cc296c5f-c958-21bf-3a14-9584180de591@towo.net>

Am 06.08.2020 um 13:46 schrieb Thomas Wolff:
> Am 06.08.2020 um 01:23 schrieb Kevin Schnitzius via Cygwin:
>> On Wednesday, August 5, 2020, 06:56:48 PM EDT, Thomas Wolff 
>> <towo@towo.net> wrote:
>>> Am 04.08.2020 um 12:02 schrieb Thomas Wolff:
>>>> Am 04.08.2020 um 00:13 schrieb Brad Wetmore via Cygwin:
>>>>> Hi,
>>>>>
>>>>> I generally kick off multiple (10) mintty sessions, and place them
>>>>> around the screen.
>>>>>
>>>>> C:/cygwin64/bin/mintty.exe -i /Cygwin-Terminal.ico \
>>>> --position @3 --position 120,0 --size 80x71 /bin/tcsh &
>>>>> C:/cygwin64/bin/mintty.exe -i /Cygwin-Terminal.ico \
>>>>> --position @3 --position 715,0 --size 80x45 /bin/tcsh &
>>>>> C:/cygwin64/bin/mintty.exe -i /Cygwin-Terminal.ico \
>>>>> --position @3 --position 715,660 --size 80x24 /bin/tcsh &
>>>>>
>>>>> Within the last 6 months or so, about 2-3 of them would hang and
>>>>> either mintty/tcsh would not start. I put a "sleep 1" in between each
>>>>> invocation and that seemed to take care of it.
>>>>>
>>>>> With the latest cygwin update, about 8 of them just hang even with
>>>>> the sleep 1. I put in a "sleep 2", and now everything is coming up
>>>>> again.
>>>>>
>>>>> Not sure if this is a mintty or tcsh issue, but just wondering if
>>>>> others are seeing this before I start trying to debug this.
>>>> I can reproduce such behaviour with /bin/bash (easy cross-check), and
>>>> in fact the shell is running in that case (easy test via `echo >
>>>> .log`), so I have a vague and unpleasant suspicion it might in fact be
>>>> related to mintty although I have no idea how that would happen. To be
>>>> analysed.
>>> When in this state, on the pty from which mintty receives child process
>>> output, select() does not report a ready for reading condition; 
>>> could it
>>> be related to the recent poll/select patch? I could not reproduce it in
>>> cygwin 3.0.7.
>> for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 ; do 
>> mintty  -i /Cygwin-Terminal.ico  --position @3 --position 120,$i 
>> --size 80x24 /bin/bash & done
>>
>> This does not fail for me.  Nor does it fail from cmd or powershell 
>> using a script.
>>
>> I tried 3.1.4 and 3.1.6 on Windows 10.  I tried up to 100 instances 
>> of of mintty...
> Thanks for testing. It is in fact hard to reproduce, maybe also 
> depending on system load (speculating).
> I could reproduce one case of one of three terminals being 
> unresponsive also with xterm.
And it also happens if I drop select() from mintty (and use just 
non-blocking read()).
> Does anybody familiar with pty/select or recent changes have any idea?
?
> Thomas

  reply	other threads:[~2020-08-06 13:31 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-03 22:13 Brad Wetmore
2020-08-04 10:02 ` Thomas Wolff
2020-08-05 22:56   ` Thomas Wolff
2020-08-05 23:23     ` Kevin Schnitzius
2020-08-06 11:46       ` Thomas Wolff
2020-08-06 13:31         ` Thomas Wolff [this message]
2020-08-06 14:26           ` Corinna Vinschen
2020-08-06 14:33             ` Thomas Wolff
2020-08-07 15:34               ` Takashi Yano
2020-08-11  4:59           ` Takashi Yano
2020-08-11  9:37             ` Corinna Vinschen
2020-08-11 10:02               ` Takashi Yano
2020-08-11 10:30                 ` Corinna Vinschen
2020-08-11 17:37                   ` Thomas Wolff
2020-08-12 20:27                     ` Thomas Wolff
2020-08-06 18:07 ` Brad Wetmore
2020-08-06 18:48   ` Thomas Wolff
2020-08-06 19:25   ` Brad Wetmore
2020-08-13  3:08     ` Brad Wetmore
2020-08-25 18:45       ` Brad Wetmore

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

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

  git send-email \
    --in-reply-to=3d594543-c992-a22c-4631-f786cf34f76b@towo.net \
    --to=towo@towo.net \
    --cc=cygwin@cygwin.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).