USB -> VGA Adapter in Windows (it works!)

Discussion in 'Feature Suggestions' started by automandc, Feb 8, 2008.

  1. automandc

    automandc Junior Member

    Messages:
    15
    First, the good news: for kicks I decided to try using a USB->VGA video adapter to see if I could get a second monitor in Windows. What do you know, it worked! I used one made by Iogear (model GUC2015V), but all such devices use the same technology from DisplayLink, and should work. So, I can add a second monitor in Windows! (DisplayLink promises Mac drivers in March '08).

    However, there are two annoyances, which I would love to see fixed:

    (1) Autoconnect specific USB device. I don't want my VM to connect to most USB devices I use (I have it set to connect to the Host automatically), but it would be useful to be able to have this particular device connected automatically at boot time. This seems like something Parallels should implement in the config file. (And, yes, I know this was discussed/requested in a separate thread last year, but now I'm putting it in "Wish Lists"!)

    (2) Mouse Sync issue. Although the second monitor works fine, you cannot move the mouse to it when Mouse Synchronization is enabled in Parallels Tools. This makes sense, as the driver obviously has to keep track of where the Parallels window ends so it can switch control, and then it switches to the mac, which doesn't recognize the other monitor. However, in this instance the Windows desktop extends beyond the main window, so the mouse ought to be able to keep going in Windows. If the tools took the size of the Windows desktop into account, it could allow the mouse to keep traveling in Windows, even if it had passed out of the Mac window. (Hopefully that was clear). This might create odd situations where the parallels window is smaller than the mac desktop as the mouse would appear to "jump" across the extra space -- but let's face it, if you are using a second monitor you probably have the Parallels window maxed out (or in full screen) anyway!

    I have used the second monitor in both Single Window mode and Full Screen, and it works flawlessly in both. It does not seem to interfere with either Spaces or Expose -- I can switch desktops normally. I even kind of works in Coherence mode, but, again, the mouse sync problem prevents the mouse from moving to the other monitor (although you can drag windows on and off the other monitor, and it continues to update windows on that monitor). Turning off mouse sync in Coherence leads to very strange results, as one might expect.

    Anyway, kudos to the developers for making a stable enough platform that something as wonky as USB video works out of the box. If you can address the two nit-picks above I will be a truly happy person!
     
  2. Xenos

    Xenos Parallels Team

    Messages:
    1,547
    Hello Automandc,

    Thank you the good news.
    I will deliver your request on autoconnect of specific USB devices to developers.
    As to Mouse Sync issue, we have not heard about it yet. Could you please attach a screenshot for we could understand it better?

    Best regards,
    Xenos
     
  3. automandc

    automandc Junior Member

    Messages:
    15
    There isn't really a way to attach a screen shot, since it involves two separate screens. The easiest example is when I have Parallels in full screen mode so that the guest OS (Windows) covers the entire main display. I have a second monitor (on the USB video adapter) positioned to the right of my main display (setting the relative position of the two monitors in the Windows Display properties dialog). The expected behavior in Windows is: if you move the mouse to the right edge of the main display and then keep going, the cursor jumps to the second monitor (i.e., continues on the "extended" desktop). However, if Mouse Sync is turned on, the cursor cannot travel past the right edge of the screen, so even though the second monitor is working, there is no way to get the cursor onto it.

    The same is true in Single Window mode (and, presumably, Coherence). When the mouse reaches the edge of the window, Mouse Sync detects the edge and switches the cursor focus back to the Mac (Host) OS, even though the Windows (guest) desktop continues. This is expected behavior for mouse sync, and without anticipating the "cheat" to get the second monitor there is no reason the developers would anticipate the Windows desktop extending beyond the main guest window.

    What I am proposing is that Mouse Sync check whether the edge of the guest OS window is also the edge of the guest desktop, and if not, to leave focus in the guest OS. Thus, in my example (second monitor to the right), moving the mouse off the right edge of the main window would never switch focus to the Mac OS -- only moving it out of the window in one of the other directions would do that.

    Not to confuse the issue, but I would note that Mouse Sync works "correctly" in the reverse situation. When I have two monitors on the Mac video card (e.g., the laptop display and an external display), when Parallels is in Full Screen Mode on the external display, moving the mouse to the edge of the screen where the laptop display is positioned automatically transitions back to the host (Mac) OS -- so when the host OS recognizes an extended desktop, Mouse Sync seems to understand that the end of the guest OS's desktop is not the end of the host's desktop. I am looking to do the reverse.

    Right now I would settle for the ability to have mouse sync turned off by default when I boot the guest (Windows) OS. Is there a way to do that?
     
  4. Xenos

    Xenos Parallels Team

    Messages:
    1,547
    Hello Automandc,

    Connecting second monitor through USB video adapter has not been expected by our developers. We didn't consider such use-case. I will deliver this report as well, thanks.

    Best regards,
    Xenos
     
  5. automandc

    automandc Junior Member

    Messages:
    15
    Thanks. I realize it is an unforeseen use, so the developers should be congratulated for making the USB stack robust enough that it works. Since this dual-monitor solution already works, and the hardware is pretty cheap, I would offer the suggestion that supporting USB video might be a much simpler dual monitor solution than trying to implement multiple virtual video cards. I know that a lot of people have asked for dual-monitor support on the guest OS, so this may be a quick and easy way to give it to them, as the only thing required appears to be functionality tweaks to to the Parallels Tools, rather than wholesale changes to the virtual machine.

    Just my $.02.

    Thanks.
     
  6. aaltieri

    aaltieri Member

    Messages:
    22
    Greeetings;

    I too have a Displaylink device, but I'm using the Mac beta drivers. For the most part, it works great and I'm thrilled I got it.

    However, now that I have it hooked up, when I try to start Parallels, my CPU utilization goes through the roof and the system essentially locks up. SOMETIMES it works fine. Most times, I have to unplug the displaylink adapter (in my case an EVGA UV+16), start Parallels, then plug the device back in. Once I do that, everything works great and I can run Parallels in coherence using both screens no worries. It's just starting up that kills it. I've also tried disabling the USB in the VM settings. That doesn't seem to have helped. It's still pretty random though, so it may have.

    Any idea what is causing the locking up at startup though?

    Automandc, have you tried the Mac drivers? Are you running into anything similar?

    Thanks!

    --anthony
     
  7. John@Parallels

    John@Parallels Forum Maven

    Messages:
    6,333
    May be you started VMC server, or Screen Sharing?
     
  8. aaltieri

    aaltieri Member

    Messages:
    22
    I'm not sure what VMC server is?

    But I have screen sharing turned off on the mac. Either way, why would it kick in when I start parallels? And lastly, even if I did have it on, starting parallals with screen sharing on should cause that kind of a processor load.

    And it can't be anything on the VM side, I can't even get the VM to load.

    --anthony
     
  9. John@Parallels

    John@Parallels Forum Maven

    Messages:
    6,333
    My fail, I meant VNC server,
    btw, usually troubleshooting steps are starting from permissions repair on Mac OS side,
    but please note, we officially do not support DriverLinks driver, maybe in some driver version it is causing Parallels to use high CPU
     
  10. aaltieri

    aaltieri Member

    Messages:
    22
    Add to this these are Beta drivers. So who knows what's going on under the hood.

    As I mentioned before, if I unplug the device, start Parallels, then plug it back in, it's fine. Works absolutely brilliantly, infact. I didn't know if Parallels was doing a USB scan or something like that at start up. I duno.

    I don't imagine Parallels thought about this since Displaylink on a mac didn't exist until April of this year. Can't fault that really. I'm just trying to narrow things down as best as I can.

    Thanks!
     
  11. automandc

    automandc Junior Member

    Messages:
    15
    Yes, I have had the same problem

    I reported it to both Parallels and DisplayLink (Parallels was at least nice enough to respond). I have exactly the same symptoms -- launching the parallels application redlines Windowserver until Parallels is force-quit. It doesn't happen all of the time, and sometimes immediately relaunching will work. It seems to happen less if Parallels is the first thing I run after booting/logging in.

    However, there is hope! DisplayLink sent me an invitation to be in a restricted beta test for the next revision of the beta drivers, and the problem seems to be fixed in this iteration (although I did have a crash when switching parallels from full screen to single window mode, but that may have been a one-time thing).

    I can't tell you how to get the new driver because it is a restricted test, but the emails regarding the subject are addressed to/from macbetatesting<at>displaylink.com, so if you send an email to that address you may be able to request participation in the new version. They say that the aim is to release the next public beta by the end of October.

    Good luck.
     
  12. aaltieri

    aaltieri Member

    Messages:
    22
    Thanks for the heads up. I shot them an email pleading my case. ;-)

    We'll see what happens.
     
  13. Bunsen

    Bunsen Bit poster

    Messages:
    1
    I just got the DisplayLink (IOgear) adapter also, and it does work great when attached to Windows through Parallels.

    automandc: I understand your mouse frustration with extended desktop mode in windows, but I don't think there's a viable way to do it with mouse sync on. Let's say you have the displaylink monitor off to the right side of your main desktop and you want to roll your mouse over to it, when the cursor hits the edge of the parallels window, how will parallels know if you want to roll to the Windows extended desktop, or out of the guest OS back to the Mac desktop? And even if it shows up both places, which desktop gets the clicks?
    The way I got around it was to hook up a second mouse and attach it directly to Windows. So now you have a Windows specific mouse that can go across both monitors.
    I've been using mine by mirroring the displaylink monitor with the main window, then minimizing the Parallels window on the Mac, and I basically have 2 separate "machines" running on my one MacBook Pro!
     
  14. aaltieri

    aaltieri Member

    Messages:
    22
    I got the new beta drivers a while back. It has been working GREAT!

    But...I have one oddity. And...well...it's just odd.

    If I have Parallels running in convergence, and I let the screen savers kick in (it has to be both the windows AND the Mac screen saver), when I move the mouse and log back in, the application bar is now at the top of the screen. It's nothing major, I just drag it back to the bottom. But...that is weird. And it happens every time, without fail for me.

    Other than that ONE oddity, the new drivers from DisplayLink work GREAT! There are still some limitations, but they are known and expected.

    --anthony
     

Share This Page