I may have been a little too premature with my last post about a reason for my graphics once again behaving as they should. After my previous combination of FreeBSD, Xorg, and KDE decided not to render using OpenGL anymore I promised myself that when/if normality returned I would immediately set out to look into the differences in various data streams to see what I might be able to piece together. This is the kind of follow up that their never seems to be enough time for. I have managed to check before and after
/var/log/Xorg.0.log files and
dmesg files. There is no difference at all in the
/var/log/Xorg.0.log files. I was sure that this is where the most changes would take place. In fact they are so identical I'll only post them if someone really wants to see it. I found a few small changes in
dmesg output. The before and after
dmesg files are posted
here and
there. The files are both edited, with my comments enclosed in lots of these <<<<<comment>>>>>>. You can skip skimming through these and just look for the comments in the
here file and get right to the nuts and bolts.
Synopsis of edits in the "here" file. VT: running with driver "vga": On or about line 9. Seems simple enough but if you think about all the ramifications that may occur if this line was not present, well, who knows? Intel's buzz word for the built in GPU graphic hardware is "Processor Graphics". This line is NOT referring to Processor Graphics but rather another native vga hardware driver sometimes also called "boot vga". Other than the base memory configuration pretty much all of the functions of the vga hardware can be set up about 3 different ways from what I read in the manual. I believe that FreeBSD is saying I've found you out finally. You can argue this is part of the VT switching enable change but I just think my hardware is getting sorted out correctly. This begs the question how did my hardware get sorted out correctly with the old xorg stack and end up working?
VT-x: PAT,HLT,MTF,PAUSE,EPT,UG,VPID: On or about line 16. PAT (Page attribute table).• Virtual Processor IDs (VPID) — Ability to assign a VM ID to tag processor core hardware structures (such as TLBs) — This avoids flushes on VM transitions to give a lower-cost VM transition time and an overall reduction in virtualization overhead. HLT, is just that, a "halt" instruction with many interpretations depending on the origin. Its usually initiated at the GPU level and ends up initiating a low power state. MTF, no clue? PAUSE, no clue? EPT, • Extended Page Tables (EPT) — EPT is hardware assisted page table virtualization — It eliminates VM exits from guest OS to the VMM for shadow page-table maintenance UG, no clue?
fbd0 on drmn0: On or about line 200. I checked /dev/ and yes there now is a "fbd" folder/entry. So now before native vga exits a frame buffer is being created and supported in developer by drmn0. What is drmn0? Please do tell.
VT: Replacing driver "vga" with new "fb": on or about line 201. Finally VT is switching to AGP engine and leaving behind vga. Once again it looks like VT switching is responsible for restoring composting and all those eye catching effects found in a KDE desktop? Confusing isn't it?
