Slashdot Mirror
Linux 2.6.39 Released
Rainmaker2006 writes "The latest iteration of Linux kernel is out. The kernel 2.6.39 is listed in kernel.org, ready to be yours!"
Linux for Devices has a short overview of what you can expect in the newest kernel; an article at Phoronix (complete with obnoxious pop-out advertising) points out a few bugs, as well.
Enough with the 2.6. We want Chrome style versioning! Linux 57 here we come.
And for those of you who would like to actually see the Phoronix article mentioned in the summary, it's here
(Yes, there are obnoxious ads, but only if you turn off your ad blocker and Flash blocker and mouse over the double-underlined blue words.)
"what you can expect in the newest kernel" and "points out a few bugs" are the same URL
It's funny how often I see this in the mainstream. First off, if its true, the only people it actually matters to are those that are directly involved. Second, the fact that the media is trying to blast this everywhere just makes me think we're all being set up for something - like its propaganda. Third, this thread has nothing to do with that. Forth, you, sir, are and idiot.
I really wish they wouldn't refer to Direct Rendering Manager as DRM. I know it's clear that it isn't that DRM but those letters are forever tainted, it's distracting.
Spelling mistakes, grammatical errors, and stupid comments are intentional.
I see you like the flavor of the kool-aid.
This is indeed a sub-atomic deal regarding anything except for the individuals involved. The news coverage given to the topic is disproportionate to effect this would have on IMF's daily operations, let alone how this would effect the world-at-large.
And yes, it's totally off-topic.
brandelf -t FreeBSD
After all, we still talk about vocal cords even though they're actually muscular folds, and as such should be called vocal folds.
"The body may heal, but the mind is not always so resilient." -- Deus Ex: Human Revolution
I mean Linux still has this serious bug. :(
Power management regressions do not appear to have been corrected. My Radeon is still failing to downclock with dynpm and is running hot.
http://kernelnewbies.org/Linux_2_6_39
>The changelog is not ready due to lazyness^Wtime constraints.
One of the drivers on its way out between 2.6.39 and 2.6.41 is the Riscom/8 driver. I owned two of these cards at one point (you know, back when I had a use for 8 RS-232 ports on my machine, and back when my PC still had ISA slots in it) - I bought them at a flea market (Hoss Traders, yeah!) - two cards, one cable set - which meant that I could only actually use one of the cards. The cable set was this massive thing - a giant multi-pin connector occupying the whole back end of the card which connected to 8 DB-25 connectors. Later on I found a source for the connector and made my own cable with DB-9 ends to save some space.
I don't remember what I used all of the serial lines for - a couple dumb terminals, probably, and a modem, Palm Pilot cradle, can't remember what else.
Anyway, the first one died as a result of a lightning strike (through the phone line, presumably. My old modem also died that day) - impressive release of magic smoke when I turned the machine back on. I used the second card for a while - don't remember what I did with it. Probably got rid of it along with one of those old computers...
So, farewell, Riscom/8 driver. You were good to me. XD
Bow-ties are cool.
Didn't 2.6.38 come out just a few months ago ?
I'm a bit concerned at the rapid rate at which these new kernels are minted. We're seeing more and more regressions and critical bugs while people ravenously add new, unrefined functionality to the kernel. Over the past year, I've spent (wasted) more time fixing crashes and data corruption than actually deploying new boxes. This isn't the Linux I used to know and love.
Me, I just want a 2.6 that's freakin' stable, so I can have one week where none of my servers throw a panic. One week! Older kernels aren't being properly patched, not even by downstream distro maintainers, so the result is a bunch of awesome gear that's not safe to use with Linux, because someone was in a hurry to make $SHINY_GADGET play nice with lspci. It's great that we have people interested in current hardware, but the whole project is now suffering from ADHD.
What was once the stable branch is practically beta, and beta is now bleeding edge nonsense.
-Billco, Fnarg.com
you'll realize that this comment increasingly applies to the entire Linux ecosystem. From Kernel 1.2.13 or so through the age of KDE3+GNOME2, Linux was a fabulous, stable, powerful operating system, a free-as-in-both workhorse that you could recommend for a huge variety of roles.
Somehow, in the last half-decade or so, things have disintegrated; Linux is now more like a sandbox in which OS and UI geeks run their experiments and/or argue ideological points about software theory. There is this sense that "we'll leave it to the OEM [Google, IBM, whomever] to actually make a working system," leading to fragmentation, the beginning hints of vendor lock-in, and an absolute lack of canonical(no pun intended) choices for the community of enthusiasts, whitebox users, project hackers, and small-enterprisers that long called Linux home. Back in the day, you could choose Slackware, Debian, Red Hat, SuSE, Caldera, and a plethora of other Linux systems and they would all be very similar, very similar to "stock," and thus collectively definitive of this thing that was the "Linux operating system."
I used to argue with people that said "Linux is just a kernel," saying that the de-facto situation was otherwise, but these days they're right; Linux is just a kernel, and an increasingly massive, complex, and less-stable-than-other-Unices one at that.
STOP . AMERICA . NOW
Considering how vocal Linus has been in some very technical issues in the past, I am appalled to see how underplayed the power consumption issue has been. It's a *major* issue, 30% more power consumption is a deal breaker for many users. And instead, what do we see from the kernel list: Nothing. The regressions that lead to these came in since kernel 2.6.38, and they went by in 39. At this point, I seriously doubt about the QA going on at kernel level. For example do they keep track of these kind of regressions in first place? It seems that Phoronix is doing the admirable job, but it seems it's going completely unheard. I am hopeful that 2.6.40 will fix this mess.
How many distros are using the kernels with that problem as their default? Any? Can you name one?
You have to expect a bit of bleeding at the bleeding edge and a kernel version or two that won't end up in the mainstream at all.