If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
Announcement
Collapse
No announcement yet.
Matrox DRM/KMS Driver Gets New Hardware Support In Linux 4.3
Just to answer some peoples questions in advance: Matrox chips are extremely popular on the server boards. That is why David ( thus RedHat ) are investing in this old chipset.
It's not really supporting an "old" chipset, Matrox have been producing new revisions of the G200 server chips fairly often. These patches add support some new ones found in upcoming server lines.
Thanks David. These G200 are rewinding me back to my youth, with Half-Life 1, Kingpin, etc. Can you elaborate what is the difference between the old and new revisions of the chip?
Thanks David. These G200 are rewinding me back to my youth, with Half-Life 1, Kingpin, etc. Can you elaborate what is the difference between the old and new revisions of the chip?
As I said these G200 are nothing like the G200 of your youth. Matrox just grabbed pieces of the same design to make their life easier. The only change across all the G200 server changes have been clock programming related, newer chips allow for larger modes, I haven't confirmed but I'm guessing the latest chips allow you to plug slightly larger monitors in or some such. They may also ship with more than the 2MB of VRAM.
As I said these G200 are nothing like the G200 of your youth. Matrox just grabbed pieces of the same design to make their life easier. The only change across all the G200 server changes have been clock programming related, newer chips allow for larger modes, I haven't confirmed but I'm guessing the latest chips allow you to plug slightly larger monitors in or some such. They may also ship with more than the 2MB of VRAM.
Quite a bit more. My G200EH (part of a HP MicroServer Gen8's iLO4) appears to have the maximum 16MB (+8MB Pseudo-DMA window?). This is said to be sufficient for two frame buffers of 32-bit 1280x1024 (even with +16-bit Z-buffer), or unbuffered 32-bit 1600x1200. Or a 1600x1200 single 1600x1200 32-bit with 32-bit Z-buffer, if you want to go crazy.
I'm using it for 1280x1024 8-bit 75Mhz, because filling that memory seems to be more of a challenge - 8-bit offers 4x text scrolling speed over 32-bit, partly because the server has a somewhat anaemic 2.3Ghz Ivy Bridge dual-core Celeron.
I wonder if the DMA engine still works and might improve this vs. having the host CPU do all the heavy lifting. The MGA-G200 specification's section on ILOAD programming looks interesting.
It's not clear to me whether the BusMaster flag is always set correctly - the same iLO4 chip on a DL380e Gen8 has BusMaster+ with Latency: 0, Cache Line Size: 64 bytes.
And here for comparison is a DL120 Gen6 with iLO-100:
Comment