Please unmask libdrm-2.4.4. Intel folks have published it on their website (look at the URL) : ------- LIBDRM: libdrm-2.4.4 release (note: libdrm >=2.4.2 is required for xf86-video-intel 2.6.0) (xserver-1.6 is recommended to use with this package) ------- They recommend xorg-server-1.6, and afaik, it's not released yet. I'm running x11-base/xorg-server-1.5.3 and it works nicely. Almost, it fixes the rendering issue (but not the double glyph fonts bug) mentioned here : bug #237468. Using : - x11-libs/libdrm-2.4.4 - x11-drivers/xf86-video-intel-2.6.0 - x11-base/xorg-server-1.5.3 - UXA accel method - DRI2 Use this in the xorg.conf : Section "Device" Identifier "855GM" Driver "intel" VendorName "Intel" BoardName "ICH4" Option "AccelMethod" "UXA" Option "DRI2" EndSection Gives me an working i810 based laptop, with far better performances than xf86-video-2.4
(In reply to comment #0) > Use this in the xorg.conf : > Section "Device" > Identifier "855GM" > Driver "intel" > VendorName "Intel" > BoardName "ICH4" > Option "AccelMethod" "UXA" > Option "DRI2" > EndSection aye, after 2h of work (OpenOffice, firefox, claws-mail, few emerge), I've to say that the UXA is not working that well. My screen start to draw small errors (for example, some chars appear in other color in OO.o, terms shows strange chars, ...). Moving back to EXA solve the issue at least for some time now. And I still have really decent performance from my old laptop.
Created attachment 178908 [details] Error while render bold 'w' Screen shot from claws. The w bold letter is drawn with another color. After a while, it's moving for another char, and sometimes it's 2 or 3 chars. Having same font and size, but regular, not bold, show the character correctly.
It'll get unmasked when upstream answers to some of my questions regarding API modifications done between 2.4.3 and .4. It's not my intention to keep in p.mask for too long. Please don't be too impatient :) Thanks
(In reply to comment #3) > It'll get unmasked when upstream answers to some of my questions regarding API > modifications done between 2.4.3 and .4. It's not my intention to keep in > p.mask for too long. > > Please don't be too impatient :) > > Thanks > Hehe, dont get me wrong, you make a very good job ! While created this bug, I wish to share that this libdrm and intel-2.6.0 is far better than any other release, and in EXA mode, it makes an usable laptop. It's not perfect, I still have rendering issue, but really far less often that before. IMOO, any user using 855GM graphic chip and using xorg-server > 1.4 (and xf86-video-intel > 2.1) should unmask this libdrm. Oh, and did I mention that I've *no* suspend nor hibernate issue with ? That said, I understand your point and I'll wait ;) Thanks for your work.
