VirtualBox 4.3.8 RC1 feedback

[frank]

Please put general 4.3.8 RC1 feedback here. Statements like "I can confirm that bug xyz is now fixed" or "there is a regression in 4.3.8 RC1 which worked nicely with VBox 4.3.6" are welcome.

[Perryg]

VMM: experimental support for SSE 4.1 / SSE 4.2 passthrough, see the manual for more information how to enable it (bug #8651)

This works for me.

Question, I have 4.1 & 4.2 available in SVN already ( DEV machine ) without the extradata settings, can I assume that is by design?

[frank]

Perry, trunk is different in this regards. SSE 4.1/4.2 is disabled by default for VBox 4.3.x but enabled by default for Trunk.

[socratis]

It seems like a major step backwards for OSX guests. I have 10.5, 10.6, 10.7, 10.8 and 10.9 guests. Out of all of these only 10.5 boots. The rest are kernel panicking during boot. I bet that you want more information, so here you go, two logs (4.3.6 and 4.3.8_RC1) for each of the guests.

[bqbauer]

First, Thank You for taking the suggestion to put the network interface connect/disconnect option for each interface in the Devices menu! This is so helpful when I'm testing things at work that need me to switch back & forth between bridged & host-only interfaces and other similar setups.

It looks as though you still haven't moved the Solaris builds to a newer or 64-bit only system to fix our very long standing issue of USB support on Solaris 11.1. Ram built & made available a 64-bit USB module for some of us to test that works great, and I've been using that since he made it a few builds ago (like 4.3.2 or 4.3.4?). If I remember correctly, you need to move the vbox builds to a newer or different 64-bit only system so the 32-bit libraries aren't used or required. Can we please resolve this & put the problem to bed?! The USB module from Ram works perfectly for me on multiple systems, and I use it almost every day. For now I'll replace the 4.3.8 vboxusb with the one he built. Maybe you can update it & post it for downloads for those of us willing to manually replace it?

Other than that, so far 4.3.8 RC1 works great on my Solaris 11.1 hosts. Still testing.

[frg]

Problem with CPU count reported here is fixed:

viewtopic.php?f=6&t=58999

Taskbar repaint problem in full screen still not fixed:

viewtopic.php?f=15&t=57314&start=45#p268078

download/file.php?id=10382

>> There is a small redraw problem with the taskbar. The area under the vbox menu will not be redrawn correctly.

Vista x64 has problems with the CD/DVD device:

viewtopic.php?f=15&t=60281

FRG

[socratis]

Corrupted graphics with 3D enabled on the following Linux based installaions:
- Fedora 19 XFCE
- Ubuntu 12.04
- Scientific Linux 6.4
- Mint 15 Cinnamon

[frank]

Corrupted graphics with 3D enabled on the following Linux based installaions:
- Fedora 19 XFCE
- Ubuntu 12.04
- Scientific Linux 6.4
- Mint 15 Cinnamon

Tried Ubuntu 12.04 and don't see any problems with 3D enabled.

[socratis]

Frank, I did a fresh installation of Ubuntu 12.04 (32 and 64-bit) and I still get corrupted graphics. I have to point you to the bug report #12193 that I filed when 4.3 came out, where I see graphics corruption in a Win8 system (still valid with 4.3.8RC1). Your answer in that bug was "Might be related to your host graphics drivers and graphics hardware". There's zilch that I can do about the drivers and I can post detailed information about the GPU, either here or in the bug report.

[frank]

Frank, I did a fresh installation of Ubuntu 12.04 (32 and 64-bit) and I still get corrupted graphics. I have to point you to the bug report #12193 that I filed when 4.3 came out, where I see graphics corruption in a Win8 system (still valid with 4.3.8RC1). Your answer in that bug was "Might be related to your host graphics drivers and graphics hardware". There's zilch that I can do about the drivers and I can post detailed information about the GPU, either here or in the bug report.

Right, we somehow need to reproduce this on similar hardware. But this is not a 4.3.8 regression -- therefore not that important for this thread.

[frank]

To all testers: Thank you for providing feedback. I admit that the RC cycle was very short but it looks like 4.3.8 RC1 was already quite stable. Therefore we released the final 4.3.8 release yesterday, see the official announcement. We even fixed a few bugs since RC1.

For 4.3.8 feedback, use the proper channels (bugtracker/forums), not this Beta channel anymore.

[socratis]

Right, we somehow need to reproduce this on similar hardware. But this is not a 4.3.8 regression -- therefore not that important for this thread.

Sorry, but isn't it regression if it works correctly in 4.3.6 and it does not work in 4.3.8 (including the final).