[XFree86] X window session timeout through firewall
CheckPoint NG firewall We have some X sessions through some clustered NG firewalls that will drop. We don't want to put a generalized config to increase the TCP timeouts for any X related ports, and have that apply to *everyone* using those ports. Is there some kind of way to have the client keep sending some type of keep-alive to keep the X sessions up? The users claim that they are continuously using their sessions, but they still timeout. I haven't read up on X enough to figure out if it works at all like FTP with control sessions (where the control session might stay idle for long amounts of time, during a big download, for example). We will soon be getting some network sniffers to see what's going on. Any ideas? Marco ___ XFree86 mailing list XFree86@XFree86.Org http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] X starts when mouse is not connected (AllowMouseOpenFail = false)
On Fri, 23 Sep 2005, Marc Aurele La France wrote: On Fri, 23 Sep 2005, Lance Duivenbode wrote: Marc Aurele La France wrote: On Tue, 13 Sep 2005, Lance Duivenbode wrote: Marc Aurele La France wrote: On Thu, 8 Sep 2005, Lance Duivenbode wrote: I work for uses Linux (a custom version) running the 2.4.24 kernel and we've just upgraded to XFree86 4.5.0. Previously, when X was started without a mouse it would exit back to the shell complaining that it couldn't find the core pointer. However, since we've upgraded to 4.5.0 this behaviour no longer occurs - even with the AllowMouseOpenFail server flag explicitly set to false. Given the fact that my company remotely manages internet kiosks, we want to be able to detect if a mouse is not connected since this usually indicates faulty hardware. Is it still possible to prevent X from starting if there is no mouse attached or has this option been removed? I would appreciate any help since I've not been able to find anything in the docs or via searching the internet. This is likely a consequence of auto-configuration. I'll have a look. You wouldn't happen to have a mouse-less log, would you? Thanks for you reply. I've included a mouse-less log as requested as well as the configuration file. You say you upgraded to 4.5.0. From what version? Sorry about the delay but it took some time to find out (I'm pretty new to the company). It appears we were running version 4.2 before. OK. Does it make any difference to add O_EXCL back into xf86OpenSerial()'s open(2) call on line 143 of xc/programs/Xserver/hw/xfree86/os-support/shared/posix_tty.c? Anything more on this? Marc. +--+---+ | Marc Aurele La France | work: 1-780-492-9310 | | Academic Information and| fax:1-780-492-1729 | |Communications Technologies | email: [EMAIL PROTECTED] | | 352 General Services Building +---+ | University of Alberta | | | Edmonton, Alberta | Standard disclaimers apply| | T6G 2H1 | | | CANADA | | +--+---+ XFree86 developer and VP. ATI driver and X server internals. ___ XFree86 mailing list XFree86@XFree86.Org http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] SM722 design with multiple controllers
On Sat, 27 Aug 2005, Marc Aurele La France wrote: On Thu, 25 Aug 2005, Bruce Weyrauch wrote: Infinetix has designed a custom video control board using 4 SM722 chips. This is a replacement board for an older design using 4 ChipsTechnology CT69000 devices. Our target system uses 8 video controllers driving 8 LCD panels in a multi-player gaming system. The target operating system is Linux, using XFree86. I am having some difficulty getting the multiple controller system working. I can get a single SM722 evaluation board to start X correctly, and I have also had success getting a dual monitor system to work using the SM722 eval board as the primary display, then using a ChipsTechnology board as the secondary display. However, if I reverse this order and install the CT board as the primary, then the SM722 as secondary, X fails to start correctly. I have attached XFree86 log files for these two cases. I have also attached a log file where two SM722 controllers were installed. Here is a summary of my testing: 1. SM722 Primary Display, ChipsTechnology CT69000 Secondary display - Log file XFree86.SMprimaryCTsecondary.log. Both displays worked fine 2. ChipsTechnology CT69000 Primary Display, SM722 Secondary display - Log file XFree86.CTprimarySMsecondary.log. XFree86 failed and crashed with the following result (excerpted from log): ((WW) Silicon MotionCannot read colourmap from VGA. Will restore with default (II) Silicon MotionSMI_GEReset called from smi_driver.c line 1579 (II) Silicon MotionSMI_GEReset called from smi_accel.c line 263 3. SM722 Primary Display, ChipsTechnology CT69000 Secnodary display - Log file XFree86.SMevalSMdigideal.log. XFree86 failed and crashed with the following result (excerpted from log): ((WW) Silicon MotionCannot read colourmap from VGA. Will restore with default (II) Silicon MotionSMI_GEReset called from smi_driver.c line 1579 (II) Silicon MotionSMI_GEReset called from smi_accel.c line 263 I would appreciate some help in debugging this problem, along with any suggestions you may have. Well, I'm no expert in Silicon Motion hardware, but a -logverbose 4 log (with an SM722 as secondary) might provide a clue. There are also a couple of shots in the dark you can try: 1) Add Option NoUseBIOS to SM722 secondary screens; 2) Change SMI_PreInit() to not get rid of pSmi-pInt10 before returning; 3) Uncomment SMI_PreInit()'s calls to xf86SetOperatingState(). Thanks, and good luck. Anything more on this? Marc. +--+---+ | Marc Aurele La France | work: 1-780-492-9310 | | Academic Information and| fax:1-780-492-1729 | |Communications Technologies | email: [EMAIL PROTECTED] | | 352 General Services Building +---+ | University of Alberta | | | Edmonton, Alberta | Standard disclaimers apply| | T6G 2H1 | | | CANADA | | +--+---+ XFree86 developer and VP. ATI driver and X server internals. ___ XFree86 mailing list XFree86@XFree86.Org http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] X starts when mouse is not connected (AllowMouseOpenFail = false)
Sorry Marc, I've been a little bit busy with other things lately. In regards to the change, I could only find it on line 150 of posix_tty.c where the code is: SYSCALL (fd = open (dev, O_RDWR | O_NONBLOCK)); if (fd == -1) { xf86Msg (X_ERROR, xf86OpenSerial: Cannot open device %s\n\t%s.\n, dev, strerror (errno)); xfree(dev); return (-1); } I've made the change and as I write this it's building. I'll let you know what happens when I've tested it (hopefully today or tomorrow), Lance Marc Aurele La France wrote: On Fri, 23 Sep 2005, Marc Aurele La France wrote: On Fri, 23 Sep 2005, Lance Duivenbode wrote: Marc Aurele La France wrote: On Tue, 13 Sep 2005, Lance Duivenbode wrote: Marc Aurele La France wrote: On Thu, 8 Sep 2005, Lance Duivenbode wrote: I work for uses Linux (a custom version) running the 2.4.24 kernel and we've just upgraded to XFree86 4.5.0. Previously, when X was started without a mouse it would exit back to the shell complaining that it couldn't find the core pointer. However, since we've upgraded to 4.5.0 this behaviour no longer occurs - even with the AllowMouseOpenFail server flag explicitly set to false. Given the fact that my company remotely manages internet kiosks, we want to be able to detect if a mouse is not connected since this usually indicates faulty hardware. Is it still possible to prevent X from starting if there is no mouse attached or has this option been removed? I would appreciate any help since I've not been able to find anything in the docs or via searching the internet. This is likely a consequence of auto-configuration. I'll have a look. You wouldn't happen to have a mouse-less log, would you? Thanks for you reply. I've included a mouse-less log as requested as well as the configuration file. You say you upgraded to 4.5.0. From what version? Sorry about the delay but it took some time to find out (I'm pretty new to the company). It appears we were running version 4.2 before. OK. Does it make any difference to add O_EXCL back into xf86OpenSerial()'s open(2) call on line 143 of xc/programs/Xserver/hw/xfree86/os-support/shared/posix_tty.c? Anything more on this? Marc. +--+---+ | Marc Aurele La France | work: 1-780-492-9310 | | Academic Information and| fax:1-780-492-1729 | |Communications Technologies | email: [EMAIL PROTECTED] | | 352 General Services Building +---+ | University of Alberta | | | Edmonton, Alberta | Standard disclaimers apply| | T6G 2H1 | | | CANADA | | +--+---+ XFree86 developer and VP. ATI driver and X server internals. ___ XFree86 mailing list XFree86@XFree86.Org http://XFree86.Org/mailman/listinfo/xfree86