Linux 3.11 Features Fall Into Place With Merge Window
hypnosec writes "The Linux 3.11 merge window is about to close, most probably this Sunday, and most of the pull requests have been merged, including feature additions and improvements to disk & file system, CPU, graphics and other hardware. Some notable merges are: LZ4 compression; Zswap for compressed swap caching; inclusion of a Lustre file-system client for the first time; Dynamic Power Management (DPM) support for R600 GPUs; KVM and Xen virtualization on 64-bit hardware (AArch64); and a new DRM (Direct Rendering Manager) driver for the Renesas R-Car SoC."
Holding out for Linux 3.11 for workgroups.
This space for rent.
and a new DRM (Direct Rendering Manager) driver for the Renesas R-Car SoC
I'm surprised the Slashdot headline didn't read "DRM Coming to Linux" or some other such nonsense. :-)
"What are you doing here, Elijah?"
And that's about all you're going to see here, since most of the valuable contributors have abandoned this site and the people who are left are mostly trolls and shills.
http://www.linux.com/news/enterprise/biz-enterprise/485159-a-conversation-with-linus-torvalds
"I think I will call 3.11 Linux for Workgroups."
I remember learning Unix shell, compilation, basic C programming on X11 terminals (there were a couple hundreds of them) and something like Solaris 7. We were given an ugly login shell (black on white xterm), the motif window manager (mwm) and that's all.
It was funny as it really looked and acted like Windows 3.1, only without the program manager, file manager or control panel. It also had alt-f9 to minimize, alt-f10 to maximize, alt-f7 to move etc. which is really cool and still found on Gnome2/Mate and Xfce at least.
It was really fun trying to fo something useful in that environment, some guy had made a crude launcher (we had to walk into his home directory), I used aliases for e.g. launching a green on black xterm rxvt with nice font size and much bigger scroll buffer, we figured out how to have a background image on the "desktop".
I looked for it in vain on linux, tried to download Lesstif but it's only libraries and nothing else. Now maybe we'll be able to use it at last?, lol. I hope they open source mwm, along Motif and CDE.
It was also both minimalist (more so than say Fluxbox or jwm) and easy to use, unlike twm and myriads of "worse than Windows 3.1" window managers. One little issue was you lost everything when closing the login shell by accident (but you thus never get into a situation where all shell windows or all windows all closed and you can't open a new terminal)
It's about time to have dynamic power management for Radeons (from a user point of view, don't know how difficult it is). It's just a bad default to have it spinning at full speed all the time, because most graphics card make a lot of noise when running at full speed. The alternative was to default to a low power mode, which reduces the performance for anyone who doesn't bother to look up the controls.
The problem with dynamic PM is that many times you need the performance quickly and for a short time. Can the card switch modes in much less than a 120th of a second, such that you don't get an impact on performance (on at 120Hz monitor)? Can it detect that it needs to switch modes in a similarly short time? It's not that bad of a problem, because most cards can run the desktop effects in low power mode (didn't work for me on Gnome, but I assume that's a bug since it was great in KDE). The main problem would be with hardware video decoding or deinterlacing and gaming, but I don't think the open driver supports any video decoding, and deinterlacing should have a reasonably constant load.
Overall, great to have dynamic PM. It may help with laptop battery life for many people who haven't bothered to check the PM settings before. For me who turns down the performance to low, I don't have to remember to turn it up before playing games etc.
I think the closest you'll find would be the rasberry Pi.
why they don't keep up with [linux kernel] versions
Keeping up to date with the mainline kernel would be expensive, and would gain you very little in the embedded market... people there want cheap and reliable and will give up features to get it... That's my impression anyway.
Can a person program a new solution to a problem? Why should anyone be able to stop such a thing? -Richard Stallman