This is the mail archive of the cygwin mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: cygwin + rsync issue under Windows 7 x64


Jamie Lokier <jamie <at> shareable.org> writes:
> Wayne Davison wrote:
>>    I'd imagine that both ssh and rsync start using a lot of CPU because
>>    the socketpair must be indicating that it is ready for a write (or
>>    read) but the actual write() (or read()) fails to return any bytes (as
>>    long as errno is something like EAGAIN, EINTR, or EWOULDBLOCK, rsync
>>    will try again). Ã If you want to test that theory, you could add some
>>    prints to rsync's io.c file near the 3 uses of EWOULDBLOCK and have it
>>    output what errno it gets. Ã If you get that fixed, the programs that
>>    interface with a socketpair should go back to normal.
> 
> It would not be the first time Cygwin did something like that.
> 
> 11.5 years ago, Cygwin32 b19.1, I found it returning EAGAIN
> continuously when reading from pipes to a child process and from the
> master side of a pty to a child process, while select() indicated
> "ready for exception" (at least from the pty master; I don't remember
> if it was that or "ready for read" from a pipe).
> 
> How interesting that sockpair() now looks like having a similar problem.
> 
> Perhaps the sockpair problem might be related to the getting stuck
> people occasionally report with rsync + ssh + Cygwin ?  I.e. maybe
> disabling socketpair() will fix the stuck problem too?
> 
> Maybe the next time someone encounters it, they'll google this thread,
> try it and let us know how it went, thanks :)

...or they can try upgrading to cygwin-1.7.2-2 (or newer, assuming
cygwin's socket handling doesn't break again!).  From the release notes
(http://cygwin.com/ml/cygwin-announce/2010-03/msg00018.html):

   - Fix some hangs and a potential crash using pipes and FIFOs.
   - Fix multiple socket problems:
      - Potential blocking of non-blocking sockets.
      - Non-working renaming of AF_UNIX socket files.
      - Using MSG_PEEK on UDP sockets leads to data loss.

At least, that "fixed it for me"!:

http://article.gmane.org/gmane.network.rsync.general/22260


Cheers,

Chris



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


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]