Slashdot Mirror
What Will Be in Linux 2.7?
Realistic_Dragon writes "The first discussion has been sighted on the Linux kernel mailing list to put together a feature list of things that should go into Linux 2.7 - including hotplug CPU & Ram support, network transparent sound and improvements to Netfilter to bring it up to the the level of OpenBSD's Packet Filter. And all this before most of us have started to run 2.6.0-preX, or even a 2.6 series stable release happening. Perhaps if you have a (sensible) idea now would be a good time to voice it, otherwise you will have to wait for 2.9 to get it included."
support the new AMD 64 bit processes and be optimized for them
so we can play all our favorite programs and games.
What WON'T be in Linux 2.7?
Uhh, you mean like any SCO code?
Win32 Support
"You've got an invalid haircut" -Warren Zevon - Life'll Kill Ya
Remove SCO from all future distributions. :^)
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
" Ummmm... Wouldn't you fry the motherboard by swapping a CPU when the computer's on?" Think Enterprise environment and Big Iron, not desktop machines.
Maybe we should start working on a way to re-load the kernel without rebooting. I don't know if it's practically possible, but it certainly would be neat!
Moderation: Put your hand inside the puppet head!
I believe they're referring to some mainframes, in which there are bays of CPUs/RAM that can be swapped in and out while the system is running.
CPU hotplug support is not designed for removing the processor from your single-CPU x86 box.
There is hardware that supports this for higher end servers. (With multiprocessor, AFAIK). Just another way to reduce downtime.
How about a userfriendly UI that'd let me configure everything without having to recompile eveything (or do it invisibly) just so that I can play and use without the pain and suffering that is require nowdays.
My main gripe with Linux has been that it's a bitch to configure for things that should't be so hard. Trying to get powermanagment to work on my IBM took me months and never worked right.
In Soviet Russia, the television watches YOU!
What Linux needs is some fatal errors. How about a screen of one solid color that comes on to warn you that all your work for the past hour is gone. You have to remember that Linux is competing with windows. If you can't beat them Join them. p.
A totally new XFree... or should I say Xouvert integrated into the kernel!
Nope, you just have to do it REALLY fast...
And don't forget to lick all the Cheetos orange dust off your fingers before you start.
"Karma can only be portioned out by the cosmos." -Homer Simpson
is a scheduler on the same caliber as Solaris, so that the kernel can utilize multiple schedulers simultaneously. Linux currently ships with only a timeshare scheduler, but Solaris supports a number of different schedulers which can all operate simultaneously. Administrators can also move processes between different schedulers on the fly as well. A Fair Share Scheduler, for example, would be nice so that resources on large systems can be partitioned effectively as to prevent certain processes from monopolizing system resources. The CPU/RAM hotplug support would be nice... glad to see Linux trying to catch up to where Solaris was years ago. Just kidding :)
You forgot about Duke Nukem Forever
In Soviet Russia, the television watches YOU!
... welcome a new kernel with UMSDOS support.
I hate my computer(s). One of them is fragmented beyond hell(beneath?), pretty impossible to repartition, and the other one is running WinXP with NTFS, oh, and I haven't got the WinXP install discs as it was already installed on the computer, COMPAQ...
I feel like buying a new harddisk... but I'd have to buy 2 then, one for stuff stolen from RIAA/MPAA and one for Linux...
BSP/IP - "Bitch Slap Protocol/Internet Protocol" support - for remotely Bitch Slapping stupid users. An idea whose time has come(tm).
:)
Oh yeah, and add more SCO(tm) code - adding Evil(tm) to MS Windows(tm) sure didn't hurt the bottomline at MS(tm)!
Disclaimer: (tm), (r), and (c) wherever appropriate...
Note: BSP/IP is defensively patented by FlyByNite Industries, Inc., a wholly-owned subsidiary of Harkonnen Enterprises.
If the name of keeping up with the leader of the industry, I think we should integrate Mozilla. A web browser is an integral part of a modern OS.
-Pete
Soccer Goal Plans
http://www.scyld.com/products/beowulf/software/mon te.html
Already there.
--
# Canmephians for a better Linux Kernel
$Stalag99{"URL"}="http://stalag99.net";
When you have to patch the kernel with security updates every week?
I think that a mechanism to patch a running kernel would improve uptime more than the ability to replace processors.
Also, some sort of buffer overflow prevention would be cool.
Don't know if either of these is possible... I think solaris has some sort of buffer overflow protection.
That would be hotplug cpu support for motherboards that feature hotplug cpu support. Consumer boards generally don't.
What WON'T be in Linux 2.7?
Autorun. Can't let any DRM measures creep in, can we?
Tubal-Cain smokes the white owl.
Disipation of excess heat via copper clad water cooling through food preparation areas, and they implement it as a kernel flag for improved overclocking processor utilization, can we start to say "Yes, Linux does include the Kitchen Sink"?
You never know...
Linux Journal's May 2003 issue had an article from Rob Love about what's new in the 2.6 kernel (new VM, ALSA, improved IO subsystem, preemptive kernel) and with a few items: SCSI needs to be rewritten to make it smarter than the drivers, and the TTY code needs a rewrite -- "it's looking like to be hack."
--
# Canmephians for a better Linux Kernel
$Stalag99{"URL"}="http://stalag99.net";
What Will Be in Linux 2.7?
Plenty of SCO's intellectual property, duh!
Trolling is a art,
But I'm not sure that's a kernel issue. Virtual server providers use FreeBSD jails (no CPU cost), or User-Mode Linux (10% CPU cost)
Laugh at my Lisp and I keeell you.
Pancakes
Sweaters (pullover)
Lug nuts (various sizes)
Adamantium
Hedgehogs
Wishbones
Brie (or any other cheeses, for that matter)
This is just a short list, but when are the developers going to get serious on these issues?
It reall shouldn't be that difficult to add thia feature. Most of the work has already been written, and adding it to Linux should require no more than simple cut n paste job.
Shouldn't someone be maintaining a consensus list of items to be worked on - like a master feature list.
The beauty of Linux (IMO) is the ability to tweak the kernel. Why not take advantage of the fact that there is source code to be modified and make it simple for the average user to recompile the kernel? It's an ugly, ugly process right now and a lot of people are running distro kernels that aren't as optimized as they could be.
What if Digg added local news and a Slashdot inspired comment karma system? ---
http://houndwire.com
What I'd like to see is all the different dependancy based package managers like Red Hat's RPM system or Debian's Apt-Get be unified into a standard installer/uninstaller that all distributions can use.
Shh.
FreeBSD jails rock. Root access to your own logical partition which looks and smells just like a dedicated machine, with no overhead.
Virtual host providers can do it for free with FreeBSD, or with ~10% CPU load using User-Mode Linux.
Laugh at my Lisp and I keeell you.
I really don't have any real demands on kernel-level features. What I need is better support on the OS-level above the kernel, for things already supported by the kernel. For example, USB is kernel-supported in the current stable kernels, but support sucks at the OS-level. When will I be able to even hotplug a mouse or digital camera?
I've been wanting this for a while - it's time for most of the drivers in the kernel to be split out. There's no reason why the kernel sources need to be as large as they are, and there's absolutely no reason why eg sound drivers and network cards can't be maintained independently with their own build process. Tying them to kernel releases means waiting until the next release for driver improvements, can bottleneck development, and leads to the 41M(!) tarball that is 2.6test7.
This would require setting up a decent build process for modules outside the kernel, but that's a good thing anyway. Have you tried to compile the nVidia drivers lately? It can be a pain if your kernel headers aren't quite right. If there were a decent external API and good support for building third party modules, this would also make it easier for manufacturers to supply independent drivers.
Last time around suggestions were things that were needed to function as "real" server.
Anyone else notice the "enlightened" comments here seem to be more aimed at matching Solaris this time around? Solaris (either purposly or not) may be put squarely in Linux's sights. Based on the track record of recent Linux developments, Sun should be worried. It's now or never to start coming up with a real business plan to address Linux. They can't consider it a "toy" for much longer and keep what little marketshare they are holding onto. On the low end, the Solaris advantage is already gone. Linux will soon be moving up the stack.
-Pete
Soccer Goal Plans
a kitchen sink!
Linux really needs a VM. (It could also benefit from that other rumored Longhorn feature, a database-backed file system, but that's another story...).
OK, we can add Java or Python to our systems, but this still leaves Linux-the-platform facing two big challenges:
1. Support apps for kernel functions have to be written in lowest-common-denominator C/C++, making, say, ALSA configuration difficult
2. The number of very different frameworks providing essential functions (desktops, config management, web servers, security admin) is large, also these frameworks do not compare particularly well with Java or Dotnet/Longhorn equivalents.
3. VMs have intrinsic advantages which, when widespread in Longhorn could make Linux-the-platform look obsolescent quite quickly. Security guarantees are one obvious area, application portability is another.
would be nice.
Revenge is sweet :-) And no, I don't mean we should try to emulate Darl McBride's personality, either.
Phil
Wanted: The ability to edit any answer you give during make config without having to start over
Scroogle
Currently even fairly advanced users can get hung up trying to get hardware to work. Windows has a huge advantage in this area even though you usually need a cd of drivers.
Even better would be a way to build a kernel that detects and includes support for your hardware automatically.
(With multiprocessor, AFAIK)
Hope so, otherwise, what would run the kernel while swapping ?
- better and more kernel documentation! - grid technology - security reviews, leak detection systems - bridge to W32 systems - easier kernel administration, kernel admin tools
Scottie: Captain, she's giving you all she's got. Kirk: Scottie, swap out the CPU and add more RAM.
Take it easy? I'll take it anyway I can get it . . .
Seems to me it could be done just fine in userspace -- why put it in the kernel?
Maybe some sort of framework for allowing access to all devices from the network? That sounds like something hard that someone might want someday....
Unlimited growth == Cancer.
Heh. Reasonable question, but whoever modded you Insightful is nuts. There is hardware that support modular CPU and memory hot-swapping / installation. Just the thing for modularity and upgrading in situations where downtime is not an option.
For an example, scope this.
If you're compiling a large program, your motherboard and OS support hot-swap, and you add more RAM, then yes, the next GCC process to execute will see the extra RAM.
Removing RAM, on the other hand, would probably need a hardware switch on the motherboard that swaps everything in that bank to disk.
Will I retire or break 10K?
I know this is specific, but it'd be nice if, by the time features get built into 2.7, someone could get with Dell/Creative and make a driver for the new OEM SB Live cards that Dell sells. It seems there's been no progress on this for a while, and lot's of people are getting these. Sure, flame me for getting a Dell, but this chipset (10k1x) needs a driver.
'nuff said.
Stick Men
I can't wait to see the kiddies show off that feature! "The new kernel has CPU hotplug support, here, watch... oh CRAP."
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Shouldn't this be an "Ask SCO"?
--
"Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
DB Filesystem. If properly implemented, it can emulate a standard hierarchical filesystem for apps that need it. It would be just like an SQL query. "SELECT /usr/local/bin FROM hda LIKE redhat7.2"
This would allow drag'n'drop [un]installation.
Computers are useless. They can only give you answers.
-- Pablo Picasso
- Blackjack
- Hookers
On second thought, skip the blackjack.like in Partition Magic, but without the reboot.
Would go nicely with the hot swoppable HDD and memory.
I'll see your Constitution and raise you a Queen.
NFS used to be userspace and was slow as shit.
PHP is the solution of choice for relaying mysql errors to web users.
What about using MAS (Media Application Server) to provide network transparent sound? MAS already does this outside of the kernel and could be integrated into the kernel.
Outdoor digital photography, mostly in New Engl
I'd like to see virutual terminals running different OSes. ALT-F1 (Linux), ALT-F2 (Windows 2000) ALT-F3 (Solaris x86) complete with a filesystem shim to allow them to share the same partitions, perhaps virtual NTFS via Linux but also map out protected areas to keep OSes from stepping on each other.
A grep on the configuration prompts would be great!
I have to always look line by line to find the right module, and it is each time.
yes, I am lazy and with a bad memory (well, not bad memory, that only goes to show that I dont recompile/reconfigure kernels so often as to learn the lines I need)
errera hunamum ets
PLEASE include native support for SATA!!
"Ask not what your country can do for you." --John F. Kennedy
That and the BSD style jails. Those are cool too.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
One word...NANITES!!!
Quod scripsi, scripsi.
integrated animated search / office assistants! Seriously though, as someone mentioned in the embedding linux quick booting article, windows does have the edge on linux as far as a snappier boot time because it continues loading services, even while the user has logged in.
It may not seem like a big deal with system uptimes of 1-2 years, but to the average user, seeing an insanly fast boot would definitly be impressive.
Its the first time I hear about that.
Does it have a back-up CPU? else, between unplug and plug, where is the OS running, or how does it resumes?
errera hunamum ets
I would like to be able to share proc, mem, disk, and net resources across multiple machines (as is partially implemented in openMosix) AND run multiple instances of Linux on a single system (as in User-mode Linux). These two features combined would provide the ultimate solution in hardware resource flexibility and scalability in large server deployments. It looks like VMware Server provides similar functionality, but with cross-platform capabilities and at a cost of over $1500 per processor.
One of the big problems with databases on Linux is that the system only does LRU replacement. This is horrid for databases that do sequential scans, because you are contunualyl replaceing what you just loaded a few megs ago. It is actually better to replace waht you just read, since you're done with it, and keep the other prior stuff in memory and then come back to it.
Slashdot's rate-of-post filter: Preventing you from posting too many great ideas at once.
Doesn't SAMBA already run in userland?
cheers
``If a program can't rewrite its own code, what good is it?'' - Mel
I'm having a bit of a hard time defining the fine line between kernel and distro...especially at the driver level. I understand that stuff like Quanta and GIMP are not kernel stuff but are hardware drivers a kernel thing or a distro thing? (Network Cards and modems, for example)
On a different note -- maybe I'm talking out of ignorance here but one of the things I've been looking for is encryption support. As in being to encrypt folders and files, etc. The closest I found seemed a bit scary to try with kernel patching and loopback or whatnot. Am I just looking in the wrong spots?
What does Packet Filter have that Netfilter does not?
Other things...
a) USB 2.0 at the kernel level
b) Fixing the plumbing
c) Mountless CD and floppy use
Can we please change the configuration syntax from the bug-ugly iptables/ipchains style to something like the more human-readable styles that pf, ipfilter and the PIX use? Please?
I see a lot of entries about application level stuff (yeh I got a list there too. :-) But laptops still have a lot of variables connected to the kernel:
* APM / ACPI (still very hit & miss, and many vendors don't seem to follow the standard making it harder)
* docking station support (sometimes works, sometimes it freezes hard)
* hot swapping mice & keyboards (maybe 2.6 will make this better?)
* Function (FN) keys don't work (you know, the vendor function keys that get you the keypad; this may be an X thing but I've never seen them work even under the console)
Probably more, but that's a good start.
On the app side, better video drivers would be my #1 wish. Many of the ones we have now for laptops are so incomplete or problematic (generally because the driver writers are at a real disadvantage working without specs; they do a great job with what they have, but the result can be hard to live with...such a catch-22).
How about a better /dev. Something that can handle dynamic devices. Something that only shows the devices attached to the system instead of 10,000+ . Something that actually works.
There is no "-1 offended" or "-1 you don't agree with me" mod options for a reason.
I wish there would be default stack protection, right out of the kernel. I'm tired of these repeated buffer overflows, and I know people can code right around them even with stack protection, but at least an attempt to make it harder for stack busting would be nice.
Gonzo Granzeau
"Nothing the god of biomechanics wouldn't let you into heaven for.." -Roy Batty
How about some write support???
When useing multiple USB keyboards all keyboards can be accessed through /dev/input/keyboard, and input from all keyboards appears on the console. (unless you don't insmod kbdev.o, and instead use /dev/input/eventx, which disables the console unless you also have a PS/2 keyboard, as well as useing a decidedly non-console like api)
If instead there were /dev/input/keyboards optionally linked to the console, and /dev/input/keyboard0..n (like it is with USB mice), we could use multiple video cards and an appropriately modified X to build multi-seat workstations, POS terminals, etc without needing Xterminals.
PCI VGA ~$50 vs ~$500 /XTerminal
The API for drivers should be static at least for the stable revision of the kernel. This would allow binary drivers to actually work properly.
Windows and MacOS seem to do it.
The world is neither black nor white nor good nor evil, only many shades of CowboyNeal.
Nice would be, with non deamons first, to have a transparent crash recovery in where a copy of the data segment is made redundand and spawn in a newly started fork when a crash ends the original application.
How about throwing Postscript in the garbage - where it belongs....
Oops - not in the kernel.... sorry about that. My Bad.
Ron Gage - Westland, MI
More work on drivers for hardware with emphasis on making it easier to detect/install/manage your hardware.
It would be really nice if the kernel was more plug-and-playish (microsoft reference not intended).
I think the idea of design by /. takes design by commitee to a whole new level.
God help us all.
Are you paranoid if you know that they just want to know everything you say and do?
1) Reduced code bloat. /sarcasm
2) Reduced memory footprint.
3) More features.
.
Come on. Linux does what most of us want already. It's the Pentium II of OS kernels. For icing only:
Get something going with Xfree86, find out what hooks they want for better performance.
Talk with NVidia and ATI and find out what common API that's not DirectX9, but is both open in specificaion and foreward thinking enough to expand to meet tomorrow's needs.
Settle on what it takes to require GCC3.X as the compiler so distributions can ship with just one GCC version.
Get a command line option to prevent printing out all the copyright notices (the corproate equivalent of teenagers' "shoutouts" -- does the warez community still do "greetz"?) and only print out error messages.
Howabout a file system (ext2?) that you can fsck while mounted RW (even if the output is "all is well" instead of actually doing fixes)?
And an uptime counter that gets me past 430 days or however many it resets. That's the one thing BSD has over us.
What, no Ctrl-C and Ctrl-V??
*ducks*
I got nothing at the link posted in the article, Heres a link to the conversation at another archive: http://www.ussg.iu.edu/hypermail/linux/kernel/0310 .1/0187.html
I can't think of how many times I've sat around playing with the GIMP and pressing refresh on Slashdot, when suddenly a thought hits me: I need a new processor. Right now.
So I run out to the computer store and pick up an AMD or Intel CPU, and when I get back ... damn and blast! I have to shut off my computer before I can install the new CPU.
But never again! Once 2.8 comes out it'll just be pop ... pop ... new CPU! I've waited for this a long time.
Now back to the Slashdot F5'ing. Hey, a German consortium wants to standardize vehicle software. Finally!
You guys are retarded. Why are you worrying about 2.7 when Emacs will replace the kernel by then?
The global economy is a great thing until you feel it locally.
It bothers me that the mixer doesn't have separate slides for each app volume. (I know the slides aren't kernel, but the tech behind them is). Since sound comes from many sources, it would be nice to be able to set the volume levels of each source. For example, I've got festival and xmms, (which for those who don't know, it's a speech synth, and music). festival is quiet, and xmms is loud. There is currently no way to make festival loud and xmms quiet.
I believe they're referring to some mainframes, in which there are bays of CPUs/RAM that can be swapped in and out while the system is running.
CPU hotplug support is not designed for removing the processor from your single-CPU x86 box.
Ok, now why is a "well duh!" comment being rated as "Informative"? Hello?! If anything, rate it what it is...Funny...especially when you picture some moron configuring his kernel for CPU hotplug support, then poppin' that thing out of its socket while the system is cranking away with Seti@Home...
ALSA sound card detection works fine here, never had a problem over a range of sound cards. Anyway, shouldn't detection be in userspace as far as possible? Others: (a) modprobe ehci-hcd (works on 2.4 here) (c) modprobe supermount (don't think this is in Linus' tree yet but it's readily available as a patch)
"'I pass the test,' she said. 'I will diminish, and go into the West, and remain Galadriel.'"
- JRR Tolkien.
How about Rendezvous support? I think it would make Linux networking much easier to use.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
without a kernel recompile, I might add.
I really like using PF on OpenBSD, and would love to see it on Linux. I think it is much more powerful and easier to configure than any other filtering system for a F/OSS operating system.
Please, please, pretty please!
Support for Intel's damn power saving "feature" where it throttles down the CPU on laptops. Even when I disable it in the BIOS, my laptop still runs at 400 mhz in Linux on AC power instead of the 700 I paid for!
:)
And while we're at it...
Sentience
That's all I ask.
And the ability to pick stocks with the precision that worries the NASD.
Oh, and complete subservience to my will, so my machine will always obey me.
Wait, this is Linux, it already has that last feature.
Seriously though, I'd love to see a defrag utility come out for Linux. I know the file systems supposedly don't need it but PHBs will love it, and whoever is selling a nice GUI version of the thing will make a mint.
I want file by file transparent compression, so I can specify a file be sored in a compressed form and all Linux applications open and use it without knowing it's different from anything else.
Identical capabilities for encrypted files.
Indexing of documents, so I can search files that contain phrases with SQL style syntax.
Real time updates to the database that 'locate' uses.
FIX THE DAMN CUT AND PASTE IN LINUX!!!!!
All in all, none of this is really Kernel stuff, but this is what I want.
"Live Free or Die." Don't like it? Then keep out of the USA
... it's the sweet stench of shoveled shite.
Your post might be more useful if it said something concrete, definite, provable.
Well, if you go with a completely modular kernel, then you already are there. You can add and remove kernel level device drivers at will, while the system is running.
Of course, this takes up a huge amount of hard drive space (bloat) but it get's you where you say you want to go. Our setup program is called "modprobe" and it doesn't require any clicks.
If you want even more fancy stuff going on, then check out the hotplug system (look on http://freshmeat.net - search for hotplug). When setup correctly, hotplug takes care of the modprobe stuff for you!
Ron Gage - Westland, MI
...will the Linux devs doing that become rich and famous? If I knew of a way, I'd integrated it waaaay deep in my OS and make my own insanely profitable desktop OS monopoly...
Kjella
Live today, because you never know what tomorrow brings
This week's meeting of the Linux Mutual Admiration and Masturbation Society kicks off with
...followed by a list of small specific examples. Unfortunately, in general it is still superior. I can make changes in the running kernel (instead of rebooting). I can set control variables for the kernel on future reboots (instead of recompiling the entire thing). Individual kernel modules can have their own read-on-module-load-by-the-kernel config file; in Linux the only general way of tweaking modules' control values is by editing the source. Maybe they read conf files, maybe not. Maybe they provide writable /proc files for runtime control, maybe not. The Solaris /proc is well designed and backwards compatible, don't even get me started on the dumping ground that is Linux /proc. (I hear in 2.6 it'll be split up into sane pieces, like Solaris has always done. Good!)
The new devfs trees under Linux -- hey, those are remarkably similar to the devices tree that Solaris has been using for years. Not in appearance -- the Linux naming scheme loses a lot of information, in order to become more readable by humans -- but in purpose and thought.
So yes, there are specific places where Linux whomps all over Solaris, and specific places where the reverse is true. After many years of daily use, programming, and administration of both of them, I've found that Solaris is still in general ahead in maturity and clean design.
I think Linux continues to borrow -- to embrace and extend, as it were -- the good ideas from Solaris, like it's done with the examples above, with shared object versioning and other ideas. If the trend continues, and I'm sure it will, then Linux's overall design strength will be passing up that of the Solaris kernel in a couple more years. But don't go around claiming that Linux is obviously all-around king just yet.
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
...By mountless do you happen to mean auto-mounting?
/dev/cdroms/cdrom0 for DVD playing...
Or is it something like telling xine to use
Actio personalis moritur cum persona. (Dead men don't sue)
...I'd like them to get basic USB hotplugging support working first without hanging my system or throwing
As to your ideas, patching a running kernel is not trivial. Solaris has a "kernel debugger" with an interface like ed(1) that lets you peek/poke certain control variables in a running kernel. For the buffer overflow protection, you're thinking of the switch that makes program's stacks non-executable, support for which I believe is somewhere in glibc already.
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
What I would like, is some sort of plugin directory tree for Eclipse, in part of the 2.7 Kernel Tree that allows one to auotmagically build the make environment, and kernel switches required for a variety of kernel testing activities such as:
1) Performance Testing
2) Video Driver Development
3) Sound Driver Development
4) Storage/RAID Driver development
5) USB Peripheral Development
6) Firewire Development
7) Removable HotPLUG PCI device development
All of these in a directory included with the 2.7 series which anyone can just drop these plugins in ECLIPSE and they automagically setup the environment to start debugging the kernel.
Kernel debugging is incredibly complex. It is also complicated by the fact many times you need two machines to do the debugging.
Finally, another IDE plugin could be developed that allows code completion/KERNEL API completion.
This would link into a database somewhere, say on kernel.org, perhaps an XML service, that the IDE would download when it started up, or download just the changes since last debugging session.
Idea is to alert you of API changes to the kernel, so you are at the latest API, and don't end up getting depricated.
Anyone think this is too way out there?
-Hack
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
Maybe some more decent Serial ATA support. I still can't get this thing to work well. And how about writing data to ntfs partitions without completely destroying them. That would be excellent.
The GeekNights podcast is going strong. Listen!
Here's a link that might be faster:0 .1/0187.html
http://www.ussg.iu.edu/hypermail/linux/kernel/031
I'd like to see the option (not default, but an option) to be able to have mount points that behave like MS-DOS or Windows, where it assumes that a disk is mounted.. if that was wrong, error. It assumes it's the same disk, if that's wrong (i.e. can't find the file), error or remount. calamari
It helps if you hit the "TURBO" button first, so you only have to do it MOSTLY REALLY FAST.
I'd really like to have an interface to the video system that is both fast and safe. At the moment, it's one or the other. Either I use straight X11 or I let the program bang on the hardware directly via DRI, SVGALib or the like.
I'd like to see video drivers in the kernel. Not necessarily full-featured OpenGL drivers, but something that:
Of these, #4 may not be possible to do safely, or may only be possible for some cards. If so, it would still be a win because a lot of applications will do fine with only the basic functionality and over time, as the bleeding-edge stuff becomes mundane, it will slowly trickle into the #3 category.
I read somewhere that some guy had tried to integrate a python script into the kernel packages that resolves dependencies in kernel configuration options, but it wasn't accepted for some reason. Something like that would be great, because configuring is a pain in the arse! It would also be great if this script could choose the most plausible options for a specific machine (optimize for processor x, include large mem support if mem>4096 MB...)
It's really great to read all of these posts regarding cpu hotplug support...
However, I want to address the more common features for the "desktop" linux (read: consumers) that I feel is missing at this point.
Resolution Change
I have yet to find a distribution that allows quick and easy video graphic resolution change. By this I mean being able to click an icon (ala monitor icon in winblows) and choose the resolution to flip to, instantly.
Can someone tell me a way to make this easier than logging out, adjusting xfre86 configuration and logging back in again? If there's an easier way, I'm all ears.
I am very impressed at the support for monitors, I mean the list is huge in xfree86. I just wish there was better presentation of all the available values one could set.
Network Authentication
The authentication module and the SMB browsing password prompt must go further. It should be possible to have a per-machine user/password setting so that when one browses a win2k network that have differing authentication schemes, it will be smart and use the values you told it earlier.
For that matter, PAM support still is not complete. Various HOWTO sites give a nice guide, but this kind of thing can be better configured in the first place, no?
Drive Management
I noticed something when changing what I thought was the label of a mounted drive on the desktop in RedHat9 recently... It started to immediately copy files when I hit Enter to accept my name change. WTF? Isn't linux smart enough to be able to change the mounted volume name? That is most bizaare.
Some sort of Mounted Volumes manager must be easily accessible from the desktop. It's not. Can this be changed?
These are my three points for the day. I'm not trying to be too critical. I love linux and the direction it's heading in. I just feel there are a couple of "common feature" issues that need addressing.
Thanks. I hope someone more qualified than I can answer my questions.
user@host$ diff
Why?
He who laughs last is stuck in a time dilation bubble.
I agree, but can you site example OS's that do this? What do the generic device APIs look like? Joe
Joe Batt Solid Design
And yes, I know most of these things are at least partly implemented now. However none of them are ready for prime time yet.
This is your sig. There are thousands more, but this one is yours.
I was walking down the alley and noticed some code listings from Netware. Some pretty good ideas, they just need to be GPL'd. There probably at least 100,000 lines of stuff we could add.
What I'd like to see is some kind of mechanism that would allow proecsses (and their children) susspended and saved, allowing them to be started (from their previous state) in the future. Even after reboot, or even transfered to a different machine or user. Also (i posted this to coreutils mailing list as well) it would be nice if you could nohup running processes (like solaris). .02
JGraham
Why would anyone need to change out the cpu of a foundry switch? I mean how often do switch cpu's actually die?
Can I get an eye poke?
Dog House Forum
But for machines that need to be up all the time 24x7 for years, this is very critical. I have worked with Tandem Mainframes, In the entire 2-3 years I was working with them, NOT once were they rebooted, and I remember atleast 2-3 CPU hot swaps.
Adding hot swapping capabilitites to linux kernel is a big step towards taking linux to the really big enterprise tasks.
for the last time people, I am "frodo from middle eaRTH", not "middle eaST".
Well... uh.... uh.... you can make make coffee from halfway across the world......
Karma: Non-Heinous
#2 seems to be on M$FT's mind as well.
If the graphics processor hangs (in the middle of a game for example), VPU Recover acts by resetting the VPU, enabling the end user to continue right where he left off. Depending on the state of the system when VPU Recover was activated, you may be able to recover with all of your open applications intact. In other cases, these applications may have to be closed and you're kicked back to the Windows desktop, but at least you don't have to go through the slow process of a complete system reboot.
According to ATI, this feature is a requirement in Microsoft's next-generation Windows operating system, codenamed "Longhorn". So essentially, ATI is ahead of the curve, to the advantage of anyone who owns an ATI-based video card.
Source
I couldn't find a neutral (not associated with the new ATI drivers) reference to this requirement. Wonder if this feature with manifest itself in ATI's promised regular updates to their Linux driver.
Where do you imagine you would get the modules without first mounting the root filesystem? A "friendly" bootscreen is probably not an appropriate thing for a user who is up to creating a custom kernel. It seems to me that this should be left up to distros, who probably want their own bootup screens anyway. (But perhaps convenient hooks for that sort of thing could be added so that each distro doesn't have to come up with their own hack.) Yay IPv6! Yay crash dumps!
This is what KGI, the kernel portion of the GGI project implemented (during the 2.3 series kernels), but the project exhausted its social captial in arguments with Linus and others and KGI never got merged.
DNA just wants to be free...
Exageration. Just meant to say that it would impact downtime more than failing processors, something which I have yet to run into.
smbfs is a kernel module.
Hands in my pocket
How about the ability to use a any mouse on either the PS2 or USB ports without having to change the selected driver?
Or how about the ability to assign a PS2 mouse and keyboard to one video card output and the USB keyboard and mouse to another video card output to give the ability to login one person on each interface grouping.
*It's not what you can do for the Dark Side but what the Dark Side can do for you!*
Not trolling, just curious... What the hell is a ping daemon?
Can I get an eye poke?
Dog House Forum
It would be good to have the ability in the kernel to (optionally of course) run through a full memory scan and map any bad memory. This way the kernel can avoid allocating the bad memory and allow the system to run reliably at least until the memory can be replaced. This could be very useful in a remote environment where physical access to the hardware is not immediately available.
Related to this is the need to save the kernel panic core dump to the swap partition, and have tools to analyze it. There are projects working on this, but it really needs to be in the mainstream kernel.
Finally, drivers, drivers, drivers! My gf (yes, you can read /. and have a gf) is a long-time UNIX admin. She's just started getting into linux, and is constantly complaining about the lack of drivers, the instability of the software, and various other aspects. And, after adminning some commercial unices, I can see where she's coming from. Drivers would be a good start. (Just to avoid OS flamewars, she got on a Windows box for the first time ever this week, and she doesn't like it either. And no, I don't think she's ever touched a mac.)
> " Ummmm... Wouldn't you fry the motherboard by swapping a CPU when the computer's on?" Think Enterprise environment and Big Iron, not desktop machines.
Yeah but my comment was more regarding the use of Big Iron. The only Big Iron I ever heard of is this and I do not believe for one minute that Linux is the operating system on foundry switches. Yet again my question is why would someone need to switch out a processor on a switch without powering it off?
Can I get an eye poke?
Dog House Forum
My DOS boot floppy has no trouble accessing NTFS partitions and it has no .dlls on the floppy.
no reason to reboot just because I repartioned
the Hard drive. it should be as simple as:
unmount repartition, reload table. mount.
I do not enjoy reboots.
Me.
How about a decent binary kernel module API so that companies can release decent binary drivers and kernel modules without having to release a copy for every conceivable compile option?
I have no problem with your religion until you decide it's reason to deprive others of the truth.
I think (hope) he was being funny...
"I don't know, therefore Aliens" Wafflebox1
SGI & Sun do not make mainframes!!!! Never have and never will.
Still, Big Systems (including mainframes, newer big "minicomputers" and the aforementioned Tandems) have support for hot-swappable RAM, CPUs & controller cards (mainframes don't have PCI slots).
"I don't know, therefore Aliens" Wafflebox1
Get the facts, dittohead.
--grendelkhan
Laws do not persuade just because they threaten. --Seneca
Foundry has a specific product called that.
'big iron' has been a generic term for large enterprise-class systems ever since the DEC-10/20 and IBM S/360 processors 20-30 years ago. These days, it refers to boxes like IBM z-Series and Sun E15K boxes.
I've been developing device drivers my entire career, on many platforms, including FLEX, Idris, UNIX, DOS, Windows (from 3.0 up), Mac OS and Mac OS X, and now finally I've started dabbling in Linux drivers.
And you know something - I was appalled to realize Linux was *still* using filesystem mapped device I/O, much of it still embedded in the kernel. C'mon, get out of the eighties, here!
At the minimum, the entire device hierarchy should be made totally modular. There is no excuse for a device driver to be compiled into the kernel -- sure you can keep them with the kernel sources and even in some rare cases require they get recompiled with new kernels, but they should never be compiled *into* the kernel. That's a good first step.
Then a new, modern driver architecture should be introduced. I personally like the IOKit architecture in Mac OS X, although it has it's own problems, and there are plenty of other approaches one could take. I hesitate to say it, but there are even a very small number of ideas in WDM which are worth considering. (Microsoft *can* come up with good ideas, occationally... They just can't implement them anywhere close to competently, or without throwing in some idiotic addition that ruins the original idea...)
As a temporary measure, the old device mechanism could be emulated through higher level drivers in the new device architecture for a year or three - that way existing programs won't break, and we have time to convert them over to the new mechanism.
Not that I expect this to happen anytime soon -- it's too radical a change. But given how much device hell I've suffered and seen others suffer recently, I think it's necessary...
IMHO,
-->Zgwortz
Alright man, Thanks.
This is the answer I've been looking for. I mean I picked up on it through context, but I was never entirely sure that was what was meant.
Can I get an eye poke?
Dog House Forum
I think this was even metioned here!
ACPI Support built in for laptops! I use linux on my laptop a lot but when you're moving around alot, ACPI support is critical and much appreciated. It's just much too difficult to enable right now.
Wouldn't you fry your fingers too?
True seamless, transparent, process migration with full distributed shared memory so that as more machines are added to a network, a cluster evolves.
Then don't update the kernal for at least 18 motnhs and stabilize it.
[RIAA] says its concern is artists. That's true, in just the sense that a cattle rancher is concerned about its cattle.
After all the BSP/IP system should always be enabled. This one techonlogy could convince even the most diehard /.er of the need for DRM. If we did have such a great and wonderful system, we'd want to ensure that it was ALWAYS operational when users were on /.! You'd need a /. signed kernel to prevent trolls from posting then "ducking" out of the concequences
One sign of product maturity is to slow down releases. It's Open Source...there's no rush for market share here. As a matter of fact, rushing like this makes it LOOK like all the "pros" are moving on like kids with ADD to the next cool thing. That's not a good value to the "customers" linux is trying to win over.
Also, I need another text editer, because my current editer doesnt, um, do enough...
I also need another way to look at a calender, becuz we need more choices. Lunix is all about choices. And freedom. And freedome to make choices. Like chioces about which OS we have to use, and not to be dominated by M$, and thier evil Windoze monopolie!!111
PS. Windoze doesnt know how to cut and paste right, becuz they use CTRL-C/V instead of, um, the mouse thing. If I cant do it with the mouse, I dont want to know how to do it. The keybd suxars!!!111
Manipulate the moderator system! Mod someone as "overrated" today.
So, even if the API seems to be stable, many drivers WILL break if the kernel get some serious change. A lot of the bug reports about many libraries (such as Gtk) is about some application that used to work with 1.2.5 breaks with 1.2.7 for some hard-to-debug reason, sometimes because the application developer used the library in the unintended way. If such things happen in the kernel, it will be even harder to debug, and many systems will probably get as unstable as Windows because many drivers that don't work yet now breaks silently, rather than refusing to compile. Looking this way, the time and effort and code bloat in maintaining API compatibility seems to have gone to waste.
hotplug CPU support? You could pry loose your CPU and put a new one in, without turning the computer off?
While the mere thought of this has immense geek appeal, how would this actually be useful?
Caveat Emptor is not a business model.
Mom, Dad, I let the magic black smoke out of the computer............
I know that I cannot rip out my CPU on my non-enterprise desktop, like all the other posters mentioned, but I am curious about this. I once saw a SUN dude at a show ripping out boards from his machines and pluggin them back and nothing changed on the screen as a demo.
But how does this work in practice? RAM support I can imagine, IF you have ECC rAM. The RAM detects that a chip is blown but still keeps the memory OK by ECC support, similar to RAID. It signals the OS WHICH NEVER USES THAT RAM AGAIN before you replace it hot pluggable. This I can deal with, but, like RAID, it works by having redundant components (mirrisos/RAID5 extra disks) which keeps the data intact (but less reliable) until the replacement arrives.
But how does this work in a CPU? The only way I can see this working is if there are redundant CPU's which execute the same instructions in parallel and compare the results. IBM mainframes do this. Is this is the case in fancy enterprise computers too?
The dangers of excessive individualism are nothing compared to the oppressiveness of excessive collectivism
If we'd have this feature we'd clearly be ahead in system uptimes, wouldn't we?
* Stable driver API (if the driver was released at x.y.0, it must work at x.y.999 as well)
* Easy for 3rd parties to supply drivers, both binary and source (so hardware manufacturers can release their own drivers easily and not just for Redhat/SuSE with certain kernels).
* More finetuning. I bet one could finetune the network stack better depending on what kind of speed your network is. Would be nice to have it automatic.
Kmail ate my email because it's window had focus (because of a bug... I clicked on the other window!!!) and I started typing away... well Kmail and most unix programs have single letter commands
Well.. The word I was typing was the equivilent of "select all, delete, yes"
and gnome is even worse!
and the command line has its own problems!
So, I like being able to blame the computer manufacturer and the software people... and I hate it when they blame me for their design flaws.
I like error messages.
I don't like it when Linux eats my work without ANY error messages and requires guru knowledge just to make the work in the first place!
Open Source is the future, and the Unix architecture has some very real problems. Read the Unix-Haters handbook.
You might stop thinking that your just a n00b and think that the conventions etc are very dangerous.
Got that already... My Compaq Armada laptop has a diag partition that I can boot from Grub. I think it's basically a DOS partition with autoloading diag utilities.
On the other hand, if you're thinking of something like Sun's Openboot prom, with power-on-self-test routines and a command line interface that allows you to test individual bits of hardware, that's not a Linux or boot manager issue. That's down to the BIOS manufacturers. For the boot manager to start at all, you have to have a mostly functioning system (cpu, memory, hard/floppy disk, etc). Do it in the BIOS and you only need a functioning CPU (OK, and serial port and other infrastructure, to be able to see what's happening, but you need those for diags in the boot manager also).
I want the linux kernel package to be splitted into smaller chunks, like i386, ppc being only 'addons' to the kernel source. Also, many of the drivers could be splitted out to separate source files, as most people never need the whole tree. Someone who wants to use sound would download the sound patches, someone who needs 3d accelerated graphics would download agpgart/drm, etc.
That way, kernel.org and it's mirrors would save gobs of bandwith, downloads would be way faster, and I would have more free space on my small harddisks.
The biggest pain for me is manually configuring an optimal .config file.
.config that contains only the hardware found? Then I could just skip through the various config screens, turn on stuff it missed, and be done.
There are two related problems which explain why the kernel is this way. The first is that there's no hardware probing, and the second is that everything is therefore left switched on by default in case the hardware is there.
Since KNOPPIX can automatically probe hardware, why can't they include code to probe the machine and construct a first pass at
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
See http://linuxconsole.sourceforge.net/
Er, next request?
Got time? Spend some of it coding or testing
Because now that the source is available, it'll have a half-life of millions of years?
Got time? Spend some of it coding or testing
After that, you'd be delighted by whatever form of Solaris or Linux you could get your hands on.
Configuring it up requires not just a reboot, but generally also a kernel relink, and often several of them in succession, each getting you a step closer to the feature you originally wanted. More reboots than MS-Windows - feh! At least the thing's stable once you've eventually got it set up right, it's the one good thing I have to say about it.
And all of the userland tools suck like a Kirby.
Got time? Spend some of it coding or testing
...then it's definitely still optional and not selected by default. My last contact with it was the opinion "we've improved the kernel to the point where Apache can go this fast in userland".
Got time? Spend some of it coding or testing
...but in answer to the body of your post, I'll see you and raise you a Python and a Ruby.
Got time? Spend some of it coding or testing
This is purely a userspace issue. The events are passed through, it's just you have nothing bound to them so nothing happens. I have seen screenshots of GNOME 2.4 with a "bind multimedia keys" dialog and there are program like linEAK already.
You're right. I believe the utility is xrandr but you may need your X grahpics driver to support the X-render extensions for it to work.
/etc/mtab has been obsolete for years, we just have to keep using it because there is no reasonable replacement. What we need is /proc/mounts to be improved enough to serve as replacement, or possibly a new /proc/mtab (and leave /proc/mounts unchanged for compatibility). What is most important to get added is the few extra fields needed by userspace, those should be set through the mount call and readable like all other mount options. Of course the userspace mount utility needs to be rewritten as well.
Do you care about the security of your wireless mouse?
The thing is that now everyone is concerned about cost. Big Iron running Solaris or anything else is rather too expensive now. The PC as an architecture has many faults, but the economies of scale mean that it will always be cheaper to produce. All the major elements from the processor through to the interface cards enjoy significant levels of competition which keeps prices low.
Last point is support. Solaris is great, but it is hardly just one system (I have worked on and off with it since SunOS days). Linux support is variable, but I can elect to have something like RH's Enterprise level support all the way down through self support.
However, it is the last option that is interesting. A large organisation can easily have people on staff that can hack kernels. Sometimes during application development (or deployment) you really need to dig down to source. You can easily do that with Linux because of the IP. With commercial licenses it is much more difficult, even if you have full source access. Sure you can get consultants from the vendor, but they usually know squat about your business/ap so can't really understand the problem and spend a lot of your money coming up with useless suggestions.
See my journal, I write things there
I think there's something called kexec in 2.6 which kinda does that. I haven't checked much though