On 15/02/2012 16:39, Bonggren, Jeffrey L wrote: > I have recently noticed that XWin.exe is spinning and maxing out a CPU > core. I checked the log and saw that it is spamming it with > "_XSERVTransSocketUNIXAccept: accept() failed" messages at a rate of about > one per millisecond. This is generating a very large log file!
Thanks for reporting this. This seems to be related in some way to how you start the X server. Starting it using startxwin.exe (the recommended method), I don't observe this problem, but using your startxwin.bat, I can reproduce the problem. The obvious difference here between these two cases is that we don't have a cygwin process as our immediate ancestor when started from a batch file, but how that could make a difference is a bit mysterious. This seems to be some sort of regression introduced with the cygwin 1.7.10 DLL, downgrading to 1.7.9 doesn't show the problem (although that is problematic as base-files now depends on tzset introduced in 1.7.10) > X seems to be behaving normally in spite of this. I typically run only > xterm and nedit (many instances) locally. > > I have verified that the Windows firewall has a blanket allow rule for > XWin.exe. > > I can't place the exact update that caused this problem, but I believe it > was not having this issue in December 2011. I am never more than a week > behind on updating cygwin. I tried falling back to the previous 1.11.4-2 > version of xorg-server, but it failed to start. Perhaps a version conflict > with some of the other updated packages? As the announce mail says, you need to upgrade xorg-server and libGL1 at the same time. If you downgrade xorg-server, you need to also downgrade libGL1. -- Jon TURNEY Volunteer Cygwin/X X Server maintainer -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/