Cygwin deadlocks due to broken select() when writing to pipes

Cygwin deadlocks due to broken select() when writing to pipes

[Bob Byrnes]

I have recently discovered that the Cygwin implementation of select()
is broken (or at best incomplete): it incorrectly claims that file
descriptors are *always* ready to write to pipes.

That's bad, because when select() indicates that file descriptors are
ready for writing (or reading), then it is supposed to be guaranteed
that a subsequent write() (or read()) will not block.  But writes to
a pipe can certainly block if the pipe happens to be full (i.e., the
process reading from the other end of the pipe is doing so slowly, and
the amount of data in transit exceeds the system-dependent limit on the
buffer size of the pipe).

Many programs (rsync and sshd come to mind) are written to use select()
to avoid blocking write() and read() calls, and if select() misbehaves as
described above, then they can deadlock.  We have observed this happening
in a variety of scenarios, but the most reproducible is to run rsync over
ssh to pull data from a Cygwin system to some other system, like Linux.
This has been reported by others to the rsync mailing list:

    http://www.mail-archive.com/rsync@lists.samba.org/msg07559.html

The strace output reported in this message is consistent with our
experience, and shows that a deadlock occurs when the rsync server
process is looping doing ...

select(2, NULL, [1], NULL, {60, 0})     = 1 (out [1], left {60, 0})
write(1, "...", 4096) = 4096

The write() blocks after select() incorrectly claims that fd 1 is ready
for writing.  The Cygwin strace output shows this even more clearly:

----------------------------------------
  128 124570283 [main] rsync 940 cygwin_select: 2, 0x0, 0x226A30, 0x0, 0x226A20
  182 124570465 [main] rsync 940 dtable::select_write:  fd 1
   95 124570560 [main] rsync 940 cygwin_select: to->tv_sec 60, to->tv_usec 0, ms 60000
   98 124570658 [main] rsync 940 cygwin_select: sel.always_ready 1
  103 124570761 [main] rsync 940 select_stuff::cleanup: calling cleanup routines
  104 124570865 [main] rsync 940 set_bits: me 0x101BA4C0, testing fd 1 ()
  103 124570968 [main] rsync 940 set_bits: ready 1
   96 124571064 [main] rsync 940 select_stuff::poll: returning 1
  101 124571165 [main] rsync 940 select_stuff::cleanup: calling cleanup routines
  101 124571266 [main] rsync 940 select_stuff::~select_stuff: deleting select records
  178 124571444 [main] rsync 940 writev: writev (1, 0x2269F0, 1)
   97 124571541 [main] rsync 940 fhandler_base::write: binary write
        ... write() blocks here, eventually ...
  140 124571681 [main] rsync 940 fhandler_base::write: 4096 = write (0x226A60, 4096)
  102 124571783 [main] rsync 940 writev: 4096 = write (1, 0x2269F0, 1), errno 0
----------------------------------------

I have also appended a short test program that reproduces the bug.
The program creates a pipe and writes to it in small chunks until the
pipe fills.  If it is compiled with -USELECT, then eventually write()
blocks, as expected.  However, if we compile with -DSELECT, then on
UNIX systems, one or more write() calls succeed, and eventually select()
starts timing out to indicate that the pipe is full (so the write file
descriptor is not ready).  On Cygwin the program blocks in write()
even with -DSELECT, which isn't supposed to happen.

I was a bit surprised not to see any mention of this important
limitation of select() for pipes in the User's Guide (section 1.6.10)
or in the source code.  But in winsup/cygwin/select.cc it is clear
that fhandler_pipe::select_write just sets the write_ready field of the
select_record to true, and peek_pipe doesn't do anything for the write
file descriptor case.  We can also see that the always_ready field is
set in the strace output above.

It isn't immediately clear how to fix this.  I see that PeekNamedPipe()
is used to determine if read descriptors for pipes are ready, but
this obviously won't work for write file descriptors.  Were any other
approaches considered and rejected while this code was being developed,
or was the problem not recognized at the time?

--
Bob Byrnes                        e-mail: byrnes@curl.com
Curl Corporation                  phone:  617-761-1200
1 Cambridge Center, 10th Floor    fax:    617-761-1201
Cambridge, MA 02142-1612

----------------------------------------

/* sel-pipe.c */

#include <stdio.h>

#include <stdlib.h>
#include <unistd.h>

#ifdef  SELECT
#include <sys/time.h>
#include <sys/types.h>
#include <sys/select.h>
#endif  /* SELECT */

#ifndef CHUNK
#define CHUNK   1024
#endif

static char buf[CHUNK];

int
main(int argc, char **argv)
{
    int pfds[2];
    int count = 0;

    if (pipe(pfds) == -1) {
        perror("pipe");
        exit(2);
    }

    while (1) {
#ifdef  SELECT
        int nfds;
        struct timeval timeout;
        fd_set wfds;
        int found;

        nfds = pfds[1] + 1;

        timeout.tv_sec = 1;
        timeout.tv_usec = 0;

        FD_ZERO(&wfds);
        FD_SET(pfds[1], &wfds);

        switch (found = select(nfds, NULL, &wfds, NULL, &timeout)) {
            case 1:
                if (!FD_ISSET(pfds[1], &wfds)) {
                    fprintf(stderr, "select returned without fd set\n");
                    exit(3);
                }
                break;  /* continue with write, below */

            case 0:
                printf("pipe is full\n");
                fflush(stdout);
                continue;

            case -1:
                perror("select");
                exit(4);

            default:
                fprintf(stderr, "select returned strange fd count %d\n", found);
                exit(5);
        }
#endif  /* SELECT */

        printf("writing chunk #%d ... ", ++count);
        fflush(stdout);

        if (write(pfds[1], buf, sizeof(buf)) == -1) {
            perror("write");
            exit(9);
        }

        printf("done\n");
        fflush(stdout);
    }
}

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: Cygwin deadlocks due to broken select() when writing to pipes

[Dylan Cuthbert]

This could explain my problems running rsync as a cronjob and having the
tasks hang for long periods of time (they do terminate eventually).  Its a
real pain when I try and log out or shutdown, as windows can't seem to
forcibly terminate the tasks (because they are spawned by a SYSTEM task -
the cron daemon - I suppose) and so just sits there for 30 minutes to an
hour trying to shutdown or log off.  I can't even ctrl-alt-del to bring up
the task manager because I presume windows has gone into some internal
system time-out loop of some kind.


---------------------------------
Q-Games, Dylan Cuthbert.
http://www.q-games.com



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: Cygwin deadlocks due to broken select() when writing to pipes

[Brian Kelly]

Thank you Bob Byrnes for this info and analysis. Perhaps it will result
in a solution to a long simmering problem. I use cygwin VERY
aggressively. A cron job launches a 20,000 line perl script (not
including CPAN modules by other authors) that does complex network
automation tasks via multiple chained telnets and ftps. (Eventually to
use ssh). Cron launches this script every five minutes and multiple
instances share resources managed by semaphores. cgf wasn't even
remotely in the mood for endorsing cygwin for this kind of 'abuse' (not
his exact words). Nevertheless, for "the most part" it works
extraordinarily well. My biggest headache is - occasional DEADLOCKS.
About once or twice a day a bash process attempting to start an ftp
instance will hang - and freeze everything clear up to the perl parent
process that launched it. Perl itself even hangs and stops writing
output to log files and the terminal. The deadlock situation will last
indefinitely until I do a kill -9 and terminate the last bash shell
launched by the perl script. Then,
every other process associated with the parent perl script "comes back
to life". I am not a c/c++ programmer. My expertise is in Perl - and
there it shall remain. I spent a couple hours a few weeks ago trying to
incorporate strace into the mix I run, but it turned out to be a really
complicated undertaking - so I gave up on it. So I apologize for
effectively offering nothing more than a BIG "me too!!". That said .....

ME TOO!!!

Brian Kelly



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Christopher Faylor]

On Fri, Oct 31, 2003 at 12:44:54PM +0900, Dylan Cuthbert wrote:
>This could explain my problems running rsync as a cronjob

I don't remember you mentioning that you were sending large amounts of
data over a pipe before.

The only time this is a problem is when the pipe is full.

And, yes, it is a known problem.

I would have expected that someone would have offered a patch for the
documentation by now.  I don't know why they haven't done that.
--
Please use the resources at cygwin.com rather than sending personal email.
Special for spam email harvesters: send email to aaaspam@sourceware.org
and be permanently blocked from mailing lists at sources.redhat.com

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Christopher Faylor]

On Thu, Oct 30, 2003 at 11:14:36PM -0500, Brian Kelly wrote:
>Thank you Bob Byrnes for this info and analysis.  Perhaps it will
>result in a solution to a long simmering problem.  I use cygwin VERY
>aggressively.  A cron job launches a 20,000 line perl script (not
>including CPAN modules by other authors) that does complex network
>automation tasks via multiple chained telnets and ftps.  (Eventually to
>use ssh).  Cron launches this script every five minutes and multiple
>instances share resources managed by semaphores.  cgf wasn't even
>remotely in the mood for endorsing cygwin for this kind of 'abuse' (not
>his exact words).  Nevertheless, for "the most part" it works
>extraordinarily well.

So, predictably, we will now be seeing everyone who has ever seen a hang
anywhere near cygwin chiming in with a "THIS MUST BE IT!"

I am always interested in fixing bugs in cygwin but bugs like of "I run
it for a real long time and something bad happens.  I'm not a programmer
and have no idea how to provide any useful feedback" are onex I steer
clear of.

Perhaps this mean response will serve as a deterrent for anyone (except
possibly one of my "groupies" who occasional pop up to comment on my
character flaws and then disappear) from responding unless there is real
useful data to provide?

>I am not a c/c++ programmer.  My expertise is in Perl - and there it
>shall remain.

If that is the case, then why are you presuming that the described
problem has anything to do with you?  You have admitted that you
couldn't possibly know if the programs that you are using are performing
a select on a pipe since you don't know c or c++.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to broken select() when writing to pipes

[Brian Kelly]

> So, predictably, we will now be seeing everyone who has ever seen a
hang
> anywhere near cygwin chiming in with a "THIS MUST BE IT!"

So "this *probably* is it" would make you feel better? ( No I didn't
think so either ). Assuming you're *human* cgf ;-), you're *interest* in
hunting down any particular intermittent and well hidden bug would be in
pretty close relation to the number of folks experiencing it as a
problem - however "nebulously". Squeaky wheel gets the grease - squeaky
wheel*s* get a factory recall. It's pretty obvious that the number of
"me too's" are sufficient only to *irritate* - not *interest* you - at
this point. Nevertheless, a few persistent reminders over a long period
can have the same effect as a very large number of complaints in close
proximity. There was once a great story in the Reader's Digest I think
of some prisoner somewhere who decided that it'd be nice to have a new
library in the prison. So he started writing lawmakers and telling them
that he wanted a new library for the prison. Every day he mailed a
couple of dozen hand written letters. For three of four years they were
ignored. Then eventually he started getting VERY nasty responses telling
him to bug off. Some even called the warden to get him to stop, but
civil libertarians soon took interest in this and threatened to sue on
his behalf if his mail was censured. Finally everyone was eventually
worn down and around year ten, the legislature voted to fund the
construction of his library - allocating close to TWO MILLION DOLLARS
for the effort.

As for your *mean* response serving as a deterrent - your ongoing rant
about not wanting to hear from non-contributors is just that - ongoing
.......

Brian Kelly


-----Original Message-----
From: cygwin-owner@cygwin.com [mailto:cygwin-owner@cygwin.com] On Behalf
Of Christopher Faylor
Sent: Thursday, October 30, 2003 11:36 PM
To: cygwin@cygwin.com
Subject: Re: cygwin deadlocks due to broken select() when writing to
pipes

On Thu, Oct 30, 2003 at 11:14:36PM -0500, Brian Kelly wrote:
>Thank you Bob Byrnes for this info and analysis.  Perhaps it will
>result in a solution to a long simmering problem.  I use cygwin VERY
>aggressively.  A cron job launches a 20,000 line perl script (not
>including CPAN modules by other authors) that does complex network
>automation tasks via multiple chained telnets and ftps.  (Eventually to
>use ssh).  Cron launches this script every five minutes and multiple
>instances share resources managed by semaphores.  cgf wasn't even
>remotely in the mood for endorsing cygwin for this kind of 'abuse' (not
>his exact words).  Nevertheless, for "the most part" it works
>extraordinarily well.

So, predictably, we will now be seeing everyone who has ever seen a hang
anywhere near cygwin chiming in with a "THIS MUST BE IT!"

I am always interested in fixing bugs in cygwin but bugs like of "I run
it for a real long time and something bad happens.  I'm not a programmer
and have no idea how to provide any useful feedback" are onex I steer
clear of.

Perhaps this mean response will serve as a deterrent for anyone (except
possibly one of my "groupies" who occasional pop up to comment on my
character flaws and then disappear) from responding unless there is real
useful data to provide?

>I am not a c/c++ programmer.  My expertise is in Perl - and there it
>shall remain.

If that is the case, then why are you presuming that the described
problem has anything to do with you?  You have admitted that you
couldn't possibly know if the programs that you are using are performing
a select on a pipe since you don't know c or c++.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Christopher Faylor]

On Fri, Oct 31, 2003 at 12:59:17AM -0500, Brian Kelly wrote:
>Nevertheless, a few persistent reminders over a long period can have
>the same effect as a very large number of complaints in close
>proximity.
>
>There was once a great story in the Reader's Digest I think of some
>prisoner somewhere who decided that it'd be nice to have a new library
>in the prison.  So he started writing lawmakers and telling them that
>he wanted a new library for the prison.  Every day he mailed a couple
>of dozen hand written letters.  For three of four years they were
>ignored.  Then eventually he started getting VERY nasty responses
>telling him to bug off.  Some even called the warden to get him to
>stop, but civil libertarians soon took interest in this and threatened
>to sue on his behalf if his mail was censured.  Finally everyone was
>eventually worn down and around year ten, the legislature voted to fund
>the construction of his library - allocating close to TWO MILLION
>DOLLARS for the effort.

You can buy books.  You can donate money.  You can't cause a problem to
be solved just by incessantly complaining about it.

If this technique was uniformly useful then we'd have peace in the
Middle East and my son would have a telephone in his room.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to lack of money

[Brian Kelly]

> You can donate money. 

Fair enough.

How? I saw nothing on the cygwin website explaining how this could be
done. I'd want donations used explicitly for cygwin.

BK

-----Original Message-----
From: cygwin-owner@cygwin.com [mailto:cygwin-owner@cygwin.com] On Behalf
Of Christopher Faylor
Sent: Friday, October 31, 2003 1:33 AM
To: cygwin@cygwin.com
Subject: Re: cygwin deadlocks due to broken select() when writing to
pipes

On Fri, Oct 31, 2003 at 12:59:17AM -0500, Brian Kelly wrote:
>Nevertheless, a few persistent reminders over a long period can have
>the same effect as a very large number of complaints in close
>proximity.
>
>There was once a great story in the Reader's Digest I think of some
>prisoner somewhere who decided that it'd be nice to have a new library
>in the prison.  So he started writing lawmakers and telling them that
>he wanted a new library for the prison.  Every day he mailed a couple
>of dozen hand written letters.  For three of four years they were
>ignored.  Then eventually he started getting VERY nasty responses
>telling him to bug off.  Some even called the warden to get him to
>stop, but civil libertarians soon took interest in this and threatened
>to sue on his behalf if his mail was censured.  Finally everyone was
>eventually worn down and around year ten, the legislature voted to fund
>the construction of his library - allocating close to TWO MILLION
>DOLLARS for the effort.

You can buy books.  You can donate money.  You can't cause a problem to
be solved just by incessantly complaining about it.

If this technique was uniformly useful then we'd have peace in the
Middle East and my son would have a telephone in his room.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Christopher Faylor]

On Fri, Oct 31, 2003 at 07:37:22AM -0500, Brian Kelly wrote:
>> You can donate money. 
>
>Fair enough.

I was not asking for money.  I was responding to your prosaic
description of an unrelated problem to show that it was a completely
different problem domain.  Perseverence in getting someone to donate
money or books could yield success.  Perseverence in reporting that you
have a problem while providing no details for tracking the problem down
and adamantly maintaining that you will not be able to debug the problem
yourself is not useful.

I'm not even going to put an "IMO" in there because it should be pretty
obvious to anyone that lives in the world that repeatedly complaining
about intractable problems rarely yields positive results.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to broken select() when writing to pipes

[Brian Kelly]

> and my son would have a telephone in his room.

Do you have a cordless phone? - Then your son has *already* had a phone
in his room!

> If this technique was uniformly useful then we'd have peace in the
Middle > East

Persistence has to be uniform, consistent, morally obvious, concretely
defined and *limited* in it's objectives. *Peace* in the Middle East
demanded by those engaging in non-peaceful tactics violates all of the
above. Some of "my" tactics violate some of the above tenets. I
admittedly am not consistent nor sufficiently limited in my objectives
...

I suppose this makes me simply an irritant ( and off-topic )

I'll stop now.

Thanks for all you HAVE done ;-)
 



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Dylan Cuthbert]

Hi Chris,

Actually I think I did mention I was running rsync hourly in a cron job,
that's all I've been mentioning as far as I know.

But anyway, this is 100% repeatable on my machine over the course of a day;
try setting up a cronjob to run every hour that rsync's a gig or so of files
over the intranet somewhere. Over the course of the day 1 in 4 of those
cronjobs will still be hanging around doing nothing (cpu usage 0).

The problem I have is that, since the Sep. 19th(?) version, when I log out
or shutdown WinXP, these idling tasks are no longer forcibly killed (because
they are idling deep in cygwin's lib code I suspect), but they used to be
killed just fine with the version of cygwin before the 19th.

Anyway, Right now I am resolved to having to leave my machine shutting down
for hours on end each night  :-( not a good solution really especially as I
pay the electricity bills around here... )

Regards

---------------------------------
Q-Games, Dylan Cuthbert.
http://www.q-games.com

"Christopher Faylor" <cgf-no-personal-reply-please@cygwin.com> wrote in
message news:20031031042355.GA23231@redhat.com...
> On Fri, Oct 31, 2003 at 12:44:54PM +0900, Dylan Cuthbert wrote:
> >This could explain my problems running rsync as a cronjob
>
> I don't remember you mentioning that you were sending large amounts of
> data over a pipe before.
>
> The only time this is a problem is when the pipe is full.
>
> And, yes, it is a known problem.
>
> I would have expected that someone would have offered a patch for the
> documentation by now.  I don't know why they haven't done that.
> --
> Please use the resources at cygwin.com rather than sending personal email.
> Special for spam email harvesters: send email to aaaspam@sourceware.org
> and be permanently blocked from mailing lists at sources.redhat.com
>



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Christopher Faylor]

On Mon, Nov 03, 2003 at 12:29:53AM +0900, Dylan Cuthbert wrote:
>Actually I think I did mention I was running rsync hourly in a cron job,
>that's all I've been mentioning as far as I know.

Which has nothing to do with selecting on pipes unless (and even this is
pretty remote) you have turned on CYGWIN=tty for your cron for some
reason.

The point that I am vainly trying to make is that if one person reports
a symptom like "hanging" and another person reports a symptom like
"hanging" in a completely different scenario while offering a rationale
for the behavior, it does not automatically mean that person B has found
the root cause for person A's problem unless person A and person B's
situations are the same.  There is nothing in your (repeated) description
of your problem to indicate that is the case.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Dylan Cuthbert]

CYGWIN=tty is on in the system env vars, shall I switch it off, is this bad?
I switched it on before I had this problem mind you, but maybe something in
the post-Sept-19th version mixes badly with CYGWIN=tty?

Yes, I understand the "root of the problem" problem (sic.), ie. the
programming law that there will always be a lot more causes than there are
distinct effects. ;-)

Anyway, I have this daily recurring problem, which seems to me to be
duplicatible, yet no-one has tried to duplicate it yet?  This is why I
repeat the description of my problem wherever I can; in the daft hope
someone will set up something similar and see if they get the same thing.
I'm not doing anything complicated, just 2 or 3 rsync's (of half a gig or
so) in a cronjob every two hours to a linux server.

Regards

---------------------------------
Q-Games, Dylan Cuthbert.
http://www.q-games.com

"Christopher Faylor" <cgf-no-personal-reply-please@cygwin.com> wrote in
message news:20031102194147.GA15320@redhat.com...
> On Mon, Nov 03, 2003 at 12:29:53AM +0900, Dylan Cuthbert wrote:
> >Actually I think I did mention I was running rsync hourly in a cron job,
> >that's all I've been mentioning as far as I know.
>
> Which has nothing to do with selecting on pipes unless (and even this is
> pretty remote) you have turned on CYGWIN=tty for your cron for some
> reason.
>
> The point that I am vainly trying to make is that if one person reports
> a symptom like "hanging" and another person reports a symptom like
> "hanging" in a completely different scenario while offering a rationale
> for the behavior, it does not automatically mean that person B has found
> the root cause for person A's problem unless person A and person B's
> situations are the same.  There is nothing in your (repeated) description
> of your problem to indicate that is the case.
>



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Dylan Cuthbert]

Ok, removing tty from the CYGWIN variable allows me to log off and shutdown
now.

Why does the tty flag cause this kind of lock-up behaviour?  Is there any
reason I need the tty flag at all? If not I'll just leave it off
permanently.

---------------------------------
Q-Games, Dylan Cuthbert.
http://www.q-games.com


"Dylan Cuthbert" <dylan@q-games.com> wrote in message
news:bo4g20$v72$1@sea.gmane.org...
> CYGWIN=tty is on in the system env vars, shall I switch it off, is this
bad?
> I switched it on before I had this problem mind you, but maybe something
in
> the post-Sept-19th version mixes badly with CYGWIN=tty?
>
> Yes, I understand the "root of the problem" problem (sic.), ie. the
> programming law that there will always be a lot more causes than there are
> distinct effects. ;-)
>
> Anyway, I have this daily recurring problem, which seems to me to be
> duplicatible, yet no-one has tried to duplicate it yet?  This is why I
> repeat the description of my problem wherever I can; in the daft hope
> someone will set up something similar and see if they get the same thing.
> I'm not doing anything complicated, just 2 or 3 rsync's (of half a gig or
> so) in a cronjob every two hours to a linux server.
>
> Regards
>
> ---------------------------------
> Q-Games, Dylan Cuthbert.
> http://www.q-games.com
>
> "Christopher Faylor" <cgf-no-personal-reply-please@cygwin.com> wrote in
> message news:20031102194147.GA15320@redhat.com...
> > On Mon, Nov 03, 2003 at 12:29:53AM +0900, Dylan Cuthbert wrote:
> > >Actually I think I did mention I was running rsync hourly in a cron
job,
> > >that's all I've been mentioning as far as I know.
> >
> > Which has nothing to do with selecting on pipes unless (and even this is
> > pretty remote) you have turned on CYGWIN=tty for your cron for some
> > reason.
> >
> > The point that I am vainly trying to make is that if one person reports
> > a symptom like "hanging" and another person reports a symptom like
> > "hanging" in a completely different scenario while offering a rationale
> > for the behavior, it does not automatically mean that person B has found
> > the root cause for person A's problem unless person A and person B's
> > situations are the same.  There is nothing in your (repeated)
description
> > of your problem to indicate that is the case.
> >
>
>
>



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Dylan Cuthbert]

Ok, well it wasn't 100% cured with the TTY flag removed but it stopped it
being a 100% daily problem to an "every other day or so" problem (depends on
the direction of the wind I think).

Why does TTY affect the behaviour of rsync and cron jobs?

Also, this may be a silly question, but why does it try to kill my cron jobs
when I log off from my user?  Shouldn't they be running no matter what user
is logged in, and regardless of whether I log in or log off?  They
definitely shouldn't be killed or need to be killed.

---------------------------------
Q-Games, Dylan Cuthbert.
http://www.q-games.com


"Dylan Cuthbert" <dylan@q-games.com> wrote in message
news:bo9l4h$cq7$1@sea.gmane.org...
> Ok, removing tty from the CYGWIN variable allows me to log off and
shutdown
> now.
>
> Why does the tty flag cause this kind of lock-up behaviour?  Is there any
> reason I need the tty flag at all? If not I'll just leave it off
> permanently.
>
> ---------------------------------
> Q-Games, Dylan Cuthbert.
> http://www.q-games.com
>
>
> "Dylan Cuthbert" <dylan@q-games.com> wrote in message
> news:bo4g20$v72$1@sea.gmane.org...
> > CYGWIN=tty is on in the system env vars, shall I switch it off, is this
> bad?
> > I switched it on before I had this problem mind you, but maybe something
> in
> > the post-Sept-19th version mixes badly with CYGWIN=tty?
> >
> > Yes, I understand the "root of the problem" problem (sic.), ie. the
> > programming law that there will always be a lot more causes than there
are
> > distinct effects. ;-)
> >
> > Anyway, I have this daily recurring problem, which seems to me to be
> > duplicatible, yet no-one has tried to duplicate it yet?  This is why I
> > repeat the description of my problem wherever I can; in the daft hope
> > someone will set up something similar and see if they get the same
thing.
> > I'm not doing anything complicated, just 2 or 3 rsync's (of half a gig
or
> > so) in a cronjob every two hours to a linux server.
> >
> > Regards
> >
> > ---------------------------------
> > Q-Games, Dylan Cuthbert.
> > http://www.q-games.com
> >
> > "Christopher Faylor" <cgf-no-personal-reply-please@cygwin.com> wrote in
> > message news:20031102194147.GA15320@redhat.com...
> > > On Mon, Nov 03, 2003 at 12:29:53AM +0900, Dylan Cuthbert wrote:
> > > >Actually I think I did mention I was running rsync hourly in a cron
> job,
> > > >that's all I've been mentioning as far as I know.
> > >
> > > Which has nothing to do with selecting on pipes unless (and even this
is
> > > pretty remote) you have turned on CYGWIN=tty for your cron for some
> > > reason.
> > >
> > > The point that I am vainly trying to make is that if one person
reports
> > > a symptom like "hanging" and another person reports a symptom like
> > > "hanging" in a completely different scenario while offering a
rationale
> > > for the behavior, it does not automatically mean that person B has
found
> > > the root cause for person A's problem unless person A and person B's
> > > situations are the same.  There is nothing in your (repeated)
> description
> > > of your problem to indicate that is the case.
> > >
> >
> >
> >
>
>
>



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to broken select() when writing to pipes

[Brian Ford]

On Wed, 12 Nov 2003, Dylan Cuthbert wrote:

> Also, this may be a silly question, but why does it try to kill my cron jobs
> when I log off from my user?  Shouldn't they be running no matter what user
> is logged in, and regardless of whether I log in or log off?  They
> definitely shouldn't be killed or need to be killed.
>
It's a bug:

http://sources.redhat.com/ml/cygwin/2003-10/msg00732.html

fixed in current snapshots.

-- 
Brian Ford
Senior Realtime Software Engineer
VITAL - Visual Simulation Systems
FlightSafety International
Phone: 314-551-8460
Fax:   314-551-8444

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Christopher Faylor]

On Mon, Nov 03, 2003 at 11:05:08AM -0000, kevin.lawton@bt.com wrote:
>When I have a 64-bit Windoze installed on one of my systems I'll let
>you know.

You don't have to let me know anything.  I'm really not interested in a
remote solution.  This requires a machine that I can power cycle and
have next to my desk for an extended length of time.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to lack of money

[kevin.lawton]

-----Original Message-----
From: Christopher Faylor
[mailto:cgf-no-personal-reply-please@cygwin.com]
Sent: 31 October 2003 18:58
To: cygwin@cygwin.com
Subject: Re: cygwin deadlocks due to lack of money


On Fri, Oct 31, 2003 at 06:13:17PM -0000, kevin.lawton@bt.com wrote:
>Okay, I spent some time looking at the mailing list archives but found
>nothing which added anything to this discussion.  I would have hoped
>that my intended offer of use of an Athlon 64 system would have counted
>as more than just 'idle curiosity' and, anyway, I'm never 'idle'.

I interpreted your response as nonserious since you seemed to be
humorously suggesting that we'd be working on a version of cygwin for
linux, which is obviously nonsensical.  I thought it was self evident
that I was talking about getting cygwin working on a 64 bit version of
windows and didn't seriously think that you were suggesting that we were
thinking about getting cygwin working on some other system.  I guess
I was wrong.
I didn't realise that 64-bit windoze was actually available as yet, so I guessed you were thinking of something else. 
Just because the 'Cygwin on Linux' bit was a joke (and a pretty obvious one at that), doesn't mean the rest wasn't serious. I notice someone else has mentioned 'Cygwin on Lindows', which I find even funnier.     ROFL 

>BTW My plan was to either test for you (under direction) or make the
>machine available to you via some sort of terminal server and my ADSL
>connection.  I didn't fancy having it transported.  I did actually want
>to contribute something to the cygwin project, as I've found it so
>valuable over the past couple of years.  Shame you weren't interested.

Actually, while I appreciate that you wanted to contribute something,
I'm not interested in either scenario.  The effort of working on cygwin
remotely either by having someone else do testing (especially when the
someone doesn't really know cygwin internals) or by logging in over
the internet was not what I was looking for.  Having a system sitting
in my office that I could hack on when the mood hit me was more of what
I was looking for. 
I know from experience that working remotely via M$ terminal server and an ADSL line is close enough to being on the machine itself as to make negligible difference in most cases. Maybe not with cygwin - I'll give it a try. 

I wasn't seriously suggesting that anyone was going to send me a 64-bit
system, either. 
I get the felling that you are not exactly 'local' to me, or I'd have been considering lending one for a while. 

Here's the kind of google search term I would have used to find
discussion:

64-bit windows cygwin site:cygwin.com cgf

This was my first tray and it unearths some discussion as the first hit.
I'm sure that there are search refinements possible but, to answer your
aggrieved question, all of the discussions basically revolve around
someone reporting that cygwin doesn't work right in some beta or release
candidate version of windows for 64 bit platforms and my suggesting that
the problem won't be solved until I (or Corinna or Pierre) have a 64 bit
system to hack on. 
When I have a 64-bit Windoze installed on one of my systems I'll let you know. 
Kevin. 

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Christopher Faylor]

On Fri, Oct 31, 2003 at 06:13:17PM -0000, kevin.lawton@bt.com wrote:
>Okay, I spent some time looking at the mailing list archives but found
>nothing which added anything to this discussion.  I would have hoped
>that my intended offer of use of an Athlon 64 system would have counted
>as more than just 'idle curiosity' and, anyway, I'm never 'idle'.

I interpreted your response as nonserious since you seemed to be
humorously suggesting that we'd be working on a version of cygwin for
linux, which is obviously nonsensical.  I thought it was self evident
that I was talking about getting cygwin working on a 64 bit version of
windows and didn't seriously think that you were suggesting that we were
thinking about getting cygwin working on some other system.  I guess
I was wrong.

>BTW My plan was to either test for you (under direction) or make the
>machine available to you via some sort of terminal server and my ADSL
>connection.  I didn't fancy having it transported.  I did actually want
>to contribute something to the cygwin project, as I've found it so
>valuable over the past couple of years.  Shame you weren't interested.

Actually, while I appreciate that you wanted to contribute something,
I'm not interested in either scenario.  The effort of working on cygwin
remotely either by having someone else do testing (especially when the
someone doesn't really know cygwin internals) or by logging in over
the internet was not what I was looking for.  Having a system sitting
in my office that I could hack on when the mood hit me was more of what
I was looking for.

I wasn't seriously suggesting that anyone was going to send me a 64-bit
system, either.

Here's the kind of google search term I would have used to find
discussion:

64-bit windows cygwin site:cygwin.com cgf

This was my first tray and it unearths some discussion as the first hit.
I'm sure that there are search refinements possible but, to answer your
aggrieved question, all of the discussions basically revolve around
someone reporting that cygwin doesn't work right in some beta or release
candidate version of windows for 64 bit platforms and my suggesting that
the problem won't be solved until I (or Corinna or Pierre) have a 64 bit
system to hack on.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to lack of money

[kevin.lawton]

Okay, I spent some time looking at the mailing list archives but found nothing which added anything to this discussion. I would have hoped that my intended offer of use of an Athlon 64 system would have counted as more than just 'idle curiosity' and, anyway, I'm never 'idle'. 
Quite honestly, I wondered what 64-bit op system you were hoping to test on. As I said: it all seems to be okay running Win-32 on an AMD-64, and I've not seen 64-bit Windoze around yet. 
I would be surprised if you were planning on a Linux version of cygwin for obvious reasons, and AFAIK BeOS, PetrOS, Lindows, Plan9, etc, etc, haven't got 64-bit versions out yet. 
Anyway, you've made your point: why should you spend a minute or so telling me something you already know when, with one simple put-down, you can suggest I go search through archives for an hour ?  Fine ! Nice attitude, mate ! 
BTW My plan was to either test for you (under direction) or make the machine available to you via some sort of terminal server and my ADSL connection. I didn't fancy having it transported. I did actually want to contribute something to the cygwin project, as I've found it so valuable over the past couple of years. Shame you weren't interested. Bye.   
   
-----Original Message-----
From: Christopher Faylor
[mailto:cgf-no-personal-reply-please@cygwin.com]
Sent: 31 October 2003 17:00
To: cygwin@cygwin.com
Subject: Re: cygwin deadlocks due to lack of money

On Fri, Oct 31, 2003 at 04:30:00PM -0000, kevin.lawton@bt.com wrote:
>You expect cygwin NOT to run on an Athlon 64 system ?  How so ?  Under
>what operating system ?  Under 32-bit Windoze, nothing seems any
>different - you're not planning a Linux version of cygwin, are you ?

Hmm.  It seems like a search of the mailing list archives would be a
pretty simple way of satisfying this kind of idle curiousity.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Christopher Faylor]

On Fri, Oct 31, 2003 at 04:30:00PM -0000, kevin.lawton@bt.com wrote:
>You expect cygwin NOT to run on an Athlon 64 system ?  How so ?  Under
>what operating system ?  Under 32-bit Windoze, nothing seems any
>different - you're not planning a Linux version of cygwin, are you ?

Hmm.  It seems like a search of the mailing list archives would be a
pretty simple way of satisfying this kind of idle curiousity.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: cygwin deadlocks due to lack of money

[kevin.lawton]

cgf, 
You expect cygwin NOT to run on an Athlon 64 system ?   How so ?   Under what operating system ? 
Under 32-bit Windoze, nothing seems any different - you're not planning a Linux version of cygwin, are you ? 
Kevin.   
   
-----Original Message-----
From: Christopher Faylor
[mailto:cgf-no-personal-reply-please@cygwin.com]
Sent: 31 October 2003 16:17
To: cygwin@cygwin.com
Subject: Re: cygwin deadlocks due to lack of money

On Fri, Oct 31, 2003 at 07:23:47AM -0800, Karl M wrote:
>Hi All...
>
>What a relief...I thought that Cygwin had the PBS virus (doesn't do any 
>harm, but periodically asks for money).

I'm sure that there any cygwin developer would gratefully accept
(without even any thoughtful consideration) a token of appreciation but
it certainly isn't a requirement by any stretch of the imagination.  I
do occasionally suggest a Red Hat support contract for someone who is
clamoring for a bug fix, though.

I have been half-seriously asking for someone to send me a 64 bit Athlon
or Opteron system so that I could investigate getting cygwin running
there.  But, while I wistfully watch the UPS guy drive by my house every
day, he never seems to stop to drop off a big package at my door.

And, don't get me started (again) on my laptop embroglio...

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Christopher Faylor]

On Fri, Oct 31, 2003 at 07:23:47AM -0800, Karl M wrote:
>Hi All...
>
>What a relief...I thought that Cygwin had the PBS virus (doesn't do any 
>harm, but periodically asks for money).

I'm sure that there any cygwin developer would gratefully accept
(without even any thoughtful consideration) a token of appreciation but
it certainly isn't a requirement by any stretch of the imagination.  I
do occasionally suggest a Red Hat support contract for someone who is
clamoring for a bug fix, though.

I have been half-seriously asking for someone to send me a 64 bit Athlon
or Opteron system so that I could investigate getting cygwin running
there.  But, while I wistfully watch the UPS guy drive by my house every
day, he never seems to stop to drop off a big package at my door.

And, don't get me started (again) on my laptop embroglio...

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: cygwin deadlocks due to lack of money

[Karl M]

Hi All...

What a relief...I thought that Cygwin had the PBS virus (doesn't do any 
harm, but periodically asks for money).

...Karl


>From: Christopher Faylor <cgf-no-personal-reply-please@cygwin.com>
>Reply-To: cygwin@cygwin.com
>To: cygwin@cygwin.com
>Subject: Re: cygwin deadlocks due to lack of money
>Date: Fri, 31 Oct 2003 10:15:17 -0500
>
>On Fri, Oct 31, 2003 at 07:37:22AM -0500, Brian Kelly wrote:
> >> You can donate money.
> >
> >Fair enough.
>
>I was not asking for money.  I was responding to your prosaic
>description of an unrelated problem to show that it was a completely
>different problem domain.  Perseverence in getting someone to donate
>money or books could yield success.  Perseverence in reporting that you
>have a problem while providing no details for tracking the problem down
>and adamantly maintaining that you will not be able to debug the problem
>yourself is not useful.
>
>I'm not even going to put an "IMO" in there because it should be pretty
>obvious to anyone that lives in the world that repeatedly complaining
>about intractable problems rarely yields positive results.
>
>cgf
>
>--
>Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
>Problem reports:       http://cygwin.com/problems.html
>Documentation:         http://cygwin.com/docs.html
>FAQ:                   http://cygwin.com/faq/
>

_________________________________________________________________
Never get a busy signal because you are always connected  with high-speed 
Internet access. Click here to comparison-shop providers.  
https://broadband.msn.com


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/