The 2.3.x "Things To Fix" List 162
Johan Jonasson writes "Alan Cox has posted the first draft of the 2.3.x "Things to fix" list. Also known as "the stuff that has to be taken care of before 2.4 can come out". "
This discussion has been archived.
No new comments can be posted.
The 2.3.x "Things To Fix" List
Related Links Top of the: day, week, month.
...there can be no public or private virtue unless the foundation of action is the practice of truth. - George Jacob Holyoake
the list for the lazy (Score:2)
Multiwrite IDE breaks on a disk error
Poll on > 16000 file handles fails
Restore O_SYNC functionality
Merge the network fixes - there is a ton of backed up stuff to do asap
ISA DMA is no longer allocating correctly aligned data
vmalloc(GFP_DMA) is needed for DMA drivers
VM needs rebalancing
NFSD fixes for path walking to regenerate dentries
Fix eth= command line
Check O_APPEND atomicity bug fixing is complete
Protection on isize (sct)
Merge 2.2.13/14 changes
Get RAID 0.90 in
PAE36 failures
USB HID merge
Mikulas claims we need to fix the getblk/mark_buffer_uptodate thing gor
2.3.x as well
PIII/Athlon/MMX/etc acceleration merge from 2.2.x-ac
Merge arcnet update (DONE)
Fix SPX socket code
AHA152x isnt smp safe (FIXED)
NCR5380 isnt smp safe
isofs break on 4Gig disk (FIXED ?)
Finish 64bit vfs merges (stat64 etc) (DONE ??)
Make syncppp use new ppp code
Fbcon races
Fix all remaining PCI code to use new resources and enable_Device
Stackable fs ?? (Erez)
Get the Emu10K merged
Test PMC code on Athlon
Fix module remove race bug (-- not in open so why did I see crashes ??? --)
Per Process rtsigio
Maybe merge the ibcs emulation code
VFS?VM - mmap/write deadlock
initrd is bust
rw sempahores on page faults (mmap_sem)
kiobuf seperate lock functions/bounce/page_address fixes
per super block write_super needs an async flag
addres_space needs a VM pressure/flush callback
per file_op rw_kiovec
enhanced disk statistics
Fix routing by fwmark
put_user appears to be broken for i386 machines
Be sure to add NFS (Score:2)
Re:the list for the lazy (Score:2)
Bugs by Architecture? (Score:4)
sigh (Score:1)
Anyone got any good suggestions for getting started on linux programming or hell any good suggestions for starting? I actually thought about taking some courses at night.
Re:sigh (Score:3)
QA work for linux (Score:1)
Re:Be sure to add NFS (Score:3)
2.2.13 must be what - at least a month old ?
NFS on SMP is working JUST FINE on my SMP box, running 2.2.12 with the knfsd-1.4.7 patches. And it has been for about 2.5 months. Most of the knfsd patch functionality has been merged into the 2.2.14-pre tree, so 2.2.14 ought to be a fairly stable NFS branch, even for SMP users. At least for NFS version two, which is a fairly old standard.
If you are really interested, there is a separate mailing list for nfs users that has been posted to the kernel mailing list, and the user space utilities have been evolving from knfsd-1.4.7 to nfs-utils-0.15 or so. Linux nfs now has locking and everything. Still, you'd have to consider nfs on linux a real weakness compared to other Unices. NFS version 3 is still a pipe-dream for clients or servers, and version two is just now stabilizing in the 2.2 tree.
APM under SMP (Score:2)
Re:sigh (Score:1)
--
Deepak Saxena
How about HIGHPOINT 366? (Score:1)
Two suggestions (Score:2)
1) Buy "Beginning Linux Programming". The first edition was great, the second looks even better.
2) If you subscribe to Linux Journal, ask the editors to start a "Newbie Programmer" column. I recently sent them an offer to write such a column and having demand roll in would help a lot. 8^)
---
Some things I'd -like- to see in 2.3.x (Score:4)
Some things I'd -like- to see, for 2.5.x:
Re:QA work for linux (Score:2)
Welcome to the Linux Quality Assurance Team!
Adaptec Raid Support (Score:2)
Suggested Reading (Score:1)
Cripes! (Score:1)
Each of your 2.3 wishlist items would probably push 2.4 out an additional 1-2 months. That's No Good.
---
NFSV3 is possible. (Score:2)
NFSV3 against the 2.2.10 kernel?
Works fine for me as an NFS3 client.
oss.sgi.com/projects
Also included as part of SGI's modified version
of Redhat, sgilinux 1.1.
It's the drivers (Score:1)
There aren't any (that I know of) public *nix drivers for this card.
Which is a gigantic pain in the ass, since many low-cost DSL providers use this card rather than DSL routers. An Ameritech tech told me this was probably because the DSL provider knows that if they use a DSL router, you can hang a NAT box off the router and connect multiple machines (since DSL routers take ATMF-25 in one port, and spit Ethernet out the other port).
By using this NIC, they reduce the "unauthorized" use of their bandwidth. By not providing *nix drivers, they know that they've reduced that "risk" even further.
I hate this card -- the DSL providers in my area mostly use Alcatel hardware, and apparently Alcatel recommends using this card. As a result, I cannot get DSL+Linux at residential rates. I have to pay for "expanded service" -- meaning an additional $50 per month, with another $600 in Customer Premises Equipment.
Aaaaiiiieeee!
Re:QA work for linux (Score:1)
Hmm, that just gave me an idea
1. Learn linux front to back
2. Start doing hobbyist QA work for linux software/distros
3. As soon as I get recognition, start company that does QA work for linux & its apps
4. Go IPO crazy!@#!@#
Yeah, easy money baby. heh.
Re:sigh (Score:3)
It seems almost redundant (no KT intended) to mention O'Reilly books as an excellent resource...anyway, if you're looking for some insight into the Linux kernel, O'Reilly's Linux Device Drivers book is very educational. At least for me it was. I've never worked on the kernel or device drivers for it and probably never will, but I found the book to be very informative nevertheless. BTW, I also am a Perl/PHP geek...not very fluent in C.
numb
how bout softmodems? (Score:3)
=======
There was never a genius without a tincture of madness.
Re:It's gotta be depressing... (Score:1)
Hmm. List out of date? (Score:2)
Show us the source (Score:2)
---
3 more for a string quartet (Score:1)
Just add a 'cello, and 2 violins.
Re:It's the drivers (Score:1)
--
Re:Show us the source (Score:1)
=======
There was never a genius without a tincture of madness.
Journaling filesystem (Score:4)
There was some talk of these on Kernel Traffic [linuxcare.com], but apparently to no avail.
This is still one area that NT kinda shows linux up. (though there are plenty of bones to pick with NTFS, don't get me wrong) Not only that, but it's a neat, useful idea that adds much and takes nothing away. (I'm sure you'll be able to use ext2 'till the earth falls into the sun.)
What is he talking about? (Score:1)
Re:Show us the source (Score:2)
---
Re:What is he talking about? (Score:1)
The Linux kernel.
Re:Show us the source (Score:1)
Re:sigh (Score:1)
Wrox has some decent stuff out there. Frank Boumphrey's (sp?) book on stylesheets is excellent, for example. They ain't O'Reilly, but you could do a lot worse.
The only thing that sort of rubs me the wrong way about Wrox is their Windows bias. The vast majority of their books are about MS-proprietary technologies, and their web-development books concentrate on developing for MSIE almost exclusively.
---
Re:QA work for linux (Score:5)
The necessary methodology involves automating execution of QA tests. You don't want to have to run 'em all by hand...
Approach:
- For each test, X, cd
/usr/src/qa/tests/X - make clean; make test
- A script runs through all the directories, running each test.
- Every time you locate a bug, you create a test.
- Every time you find behaviour that ought to be, you create a test.
This is more-or-less how one does regression testing with things like compilers. Tests that run with the kernel would be equally valuable.This compiles a C program that exercises some facility of the system.
The program drops output into a local file in the directory, as well as to a central results DB in /usr/src/qa/results , where entries are keyed by test, by date, and by kernel version.
The notable result is a Pass or Fail value.
It would be good if a "success" result caused the test program to create the file success, so that one could run through, after a patch, and "merely" use make success to rerun failed tests.
If you build a reasonably intelligent infrastructure, and are accepting of regression tests, you'll come to know more about how the kernel works than you ever wanted to know...
Re:What is he talking about? (Score:1)
Re:What is he talking about? (Score:1)
I have to treat slashdot more like the mail. If no country is mentioned on a stamp, it means UK. If no topic is mentioned on a post, it means Linux.
Linux Core Kernel Commentary (Score:2)
Take a look at it here [fatbrain.com]...
----
Re:Some things I'd -like- to see in 2.3.x (Score:3)
I agree that it would be nice to have the various new FSes; I don't think Reiserfs will be quite ready, and it looks likewise for ext3, more be the pity. As for Procfs, if it's not there already, I have a hard time believing it'll get there soon. And you forgot NFS3, no?
As for ACLs, I don't think the rest of the world is ready for them. They're practically useless without fairly sweeping changes to things like:
I somewhat favor a rather different ACL model based on TOPS-10 FILDAE; [hex.net] 'tis unclear that we've got a clear model of how to configure security with ACLs, and it doesn't make sense to push it into the kernel until there are some clear ideas on how to implement the user-space ACL management.
Re:It's the drivers (Score:1)
XenoWolf
Kernel 2.2.x is unstable on Alpha (Score:1)
Re:Some things I'd -like- to see in 2.3.x (Score:2)
- Procfs
You meant DevFS [csiro.au] right? From the brief discussions I've read on Linux Kernel, it doesn't seem too likely this will get in. Everytime it is brought up, it is usually shot right down with, "its not the right way (tm) to do it". EOD. (End Of Discussion)Can someone enlighten me a little bit more on the issues. I'd sure like to get rid of all the crap hanging out in /dev on my system, and i'd rather not have to patch my kernel twice everytime I upgrade.
Re:It's gotta be depressing... (Score:1)
Well, I suppose sooner or later an infinite number of monkeys...
Windows 2000 is terribly late, over budget, still too demanding to run on
a lot of the hardware in corporate America, and there remain serious
compatability problems which mean MS will continue to produce consumer
Windows versions based on the rotting DOS core.
In 1998, NT5 faced an unprepared market, Bill hoped to make a killing,
but unfortunately he didn't have a product to put in the box. Two whole
years later NT5 has a new name, a reduced feature set and claims to be
"finished", but you still won't find it in any shops. Meanwhile the
competition have cleaned up.
Remember, if you choose W2K you have to live with that feature set for
the next three or four years, perhaps more. There's no reason to think
that the improvements from RH4.x to RH6.x can't be repeated in future
versions, which will leave W2K SP4 looking pretty sad next to a shiny
RH7.2 system, while Microsoft promises better things "Soon".
Re:Some things I'd -like- to see in 2.3.x (Score:2)
Wish list (Score:3)
I also wish there had been more push for make Linux x86 a better database server platform. Limitations that get in the way are:
Another item is the ability to have multiple default routes and routing to the default route based on source ip address. Multi-homing on multiple Internet feeds just isn't any fun when all your outbound traffic goes through the same pipe regardless of where the request comes from.
Anyways, I look forward to the 2.5 developments. The 2.3 kernel series has been fun. :)
Re:the list for the lazy (Score:1)
maybe i'm just ignorant and everyone else realizes this whole anon post?
trolling for compliments (Score:1)
Boumphrey's book is the only other Wrox title I haven't returned disatisfied.
XML for Applications was one of the lightest 600 page tech book I've ever run afoul of. Its true you could do worse: SAMS.
Re:v 2.3.xx of the Linux kernel (Score:1)
Re:How about HIGHPOINT 366? (Score:1)
I'm not sure, but I suppose the IDE-patch is is already included in 2.3.x.
Procfs vs Devfs (Score:2)
Devfs has been available as a patch to the kernel for a long time now; if it's not in yet, I'm not sure why it would be expected to go in now...
Re:sigh (Score:1)
Re:QA work for linux (Score:1)
We'll be providing a mirror of that list on OVP sometime today (when one of the web monkeys finishes his recess), and hopefully, this could be the start of a worthy project to follow up on.
A previous poster mentioned that when he attempted to do any work on an Alpha port, he had to search high and low for each bug before starting. I think that would also help in this case, to specify for which platform each of the bug(s) affect.
Anyone interested in getting free hosting with OVP, and is willing to do work for the community, please contact us [mailto].
Thanks for your time.
Chris Simmons
The Optical Valley Project
http://www.opticalvalley.com [opticalvalley.com]
Re:APM under SMP (Score:1)
APM is basically a broken standard as far as SMP is concerned-- the official specifications for APM break down on a SMP machine. Now, it may be possible to put in a hack workaround, but whether or not that will work depends upon your BIOS.
Heck, many BIOSes don't support APM in Linux on a uniprocessor machine; they expect to be called from DOS and crash when Linux asks them to power down.
Nevertheless, it's possible that a workaround may come at some point. Also, you may be able to do what you want with the ACPI driver in 2.4-- that should work okay with SMP.
Re:QA work for linux (Score:1)
[a] How do you test different SCSI drivers -- add 200 scsi cards to your system ?
[b] How do you test filesystem drivers ? create n filesystems ?
etc..etc..
regression testing is ok for standalone programs but since we have no generic emulator for all the hardware we cant do it on the kernel.
FYI: Suse now has ReiserFS for download (Score:3)
See the Heise newsticker posting at http://www.heise.de/newsticker/data/ps-04.01.00-0
Re:Some things I'd -like- to see in 2.3.x (Score:2)
The issues, as I understand them, can be summarised as:
LVM, please! (Score:2)
Some subsystems may be more challenging to test... (Score:2)
That goes to suggest that the testing scheme needs to be highly distributed, so that it checks to see what hardware is there is on a particular box, and tests that hardware. And submits the results back to a central site that would collect the results of tests together.
Re:Show us the source (OT) (Score:1)
(searching for Linux usin' Phishies)
Re:sigh (Score:2)
Then again, so does that bucking bronco on the front... not to mention the operating system it's written about. It's pretty fscking cool when you can write your first driver on your everyday werk box, making liberal use of insmod and rmmod, and only really freak it out three or four times over the course of six weeks' work.
Good luck on the device driver....
(lucky bastard... wish I had time to hack some more of those...)
Re:Show us the source (Score:1)
=======
There was never a genius without a tincture of madness.
What I would really like.. (Score:2)
32 bit UIDs patch!!! (Score:1)
Re:Some things I'd -like- to see in 2.3.x (Score:1)
DevFS was doomed to failure from day one. How do you expect the kernel to know what devices will _ever_ be available? It certainly easy enough for loaded and/or active drivers, but what about modules? For a module to be automatically loaded, the device node will have to already exist. Maybe _I_ want weird names for my devices?
Re:Journaling filesystem (Score:1)
Not by much, and not for long. NTFS sucks as an example of a journaling file system. ext3, ReiserFS and XFS will soon give Linux users lots of choices and some pretty strong advantages over NTFS.
Re:How about HIGHPOINT 366? (Score:1)
Re:QA work for linux (Score:1)
One problem with this, is any change could break something that previously worked. After patching I believe that all tests would need to be re-tested
Hal Duston
Re:Show us the source (Score:2)
The fix list link was a direct copy of the email Alan Cox sent to the linux-kernel mailing list. At the bottom you will read:
I trust you can figure out the rest.Re:What is he talking about? (Score:1)
Re:Be sure to add NFS (Score:2)
however, once you see the performance diff, you'll likely want to stick with knfs. in my testing at home with a 10 MB network and at work at 100 MB i find that knfs is just about limited by wire/disk speed while user-nfs is several times slower. knfs is even faster than ftp by quite a bit, especially for reads. writes are generally 20-30% or so slower than reads due to using nfsv2.
tim
"Slashdot for dummies" (Score:3)
Well, for one, it could be argued that if you go to a health-food store looking for 'food, not health food' you'll be sorely dissapointed. Slashdot is what it is...
For another, the articles are in categories; this one is in the category "Linux", denoted by the cute little penguin on graphical browsers, or the 'Linux' alt tag on text browsers. Given the context, a version number number alone doesn't leave much room for doubt.
Would you understand if a post claimed that Version 7 would be coming out next month? Version 7 of what? Who knows?
Well, if it were say, next to the Beos logo, I would assume it was version 7 of Beos. (just to choose an example at random, I don't know what version Beos is at). I guess if it were next to the Monty Python foot I might be confused...
Finally, you have a login. Take a look at your preferences [slashdot.org] and adjust accordingly.
Have been in for ages (Score:1)
Re:LVM, please! (Score:1)
Chris
Re:sigh (Score:1)
No, good thing. (Score:1)
Yes, not having a tts[1] fs /now/ sucks, but waiting now+1mo*cool_feature_to_include for the next major release is a Bad Thing.
[1] (I wonder if he's a novell type?)
Re:What is he talking about? (Score:1)
ACLs are cool.. (Score:1)
One thing that linux is good at is replacing the file and print sharing provided by a NT( However linux as a workgroup server currently cant compeat with Netware/NDS (and proably now with *gasp* win2k/ads). Novels marketing division is compleatly brain dead, but when ADS comes online, watch out.
Of course, why anyone would wory about filesystem privilages is beyond me, thats what ZENworks (application launching) and Groupwise (file sharing) is for :)
Re:"Slashdot for dummies" (Score:1)
Re: (Score:1)
devfs is too big a change for some people (Score:3)
devfs impacts every device driver in the kernel, true, but one assumes that if it is worthwhile, we can deal with that. Kernel-wide changes have been done before and will be again. And most of the changes have been done as patches by the devfs maintainer already. So that isn't the real issue.
devfs would add complexity to the kernel, but so does everything else that adds code. So that isn't the real reason.
You would lose persistance of the
In the end, it always comes down to "What we have now works fine, and we've done it this way forever, so why change?" The idea of replacing the
Too bad, really. I think devfs has a lot of merit to it.
Re:Kernel 2.2.x is unstable on Alpha (Score:1)
Errr... works fine for me, with an AlphaPC 164LX (EV5) and a Symbios SCSI adapter. Be more specific on your system, and maybe send it someplace where it's more likely to get noticed than a random AC post on Slashdot.
"Software is like sex- the best is for free"
-Linus Torvalds
Re:Journaling filesystem (Score:1)
I know NTFS is a journaling fs, but I thought one of the main advantages of journaling was so that in the event of a drive not getting unmounted properly the journal could be checked to see what is lost, etc and do what it needs to get it back up to where it left off without running fsck/chkdsk, but NT chkdsk runs over the whole drive as if there was no journaling, am I missing something?
--
Re:It's gotta be depressing... (Score:1)
Remember, if you choose W2K you have to live with that feature set for
the next three or four years, perhaps more
How did you figure that one out?
Windows is a modular OS and NT is a modular kernel.
The feature set is already richer than Linux/Unix is many areas - anything that's missing - write it yourself.
Write kernel extensions, drivers, software. I don't really see how what you say is relevant.
You're like one of those people who go and compare Linux + every single piece of unix software created to Windows NT out of the box.
NT4 didn't have a telnet server - but you didn't have to wait to W2K to get one, write one yourself, port it from Unix using cygwin, or buy one. You complain when windows includes too many features - then you complain when windows lacks some features.
Utter fantasy time... (Score:1)
*sigh*
Oh well. Back to reality.
How about moving the end-of-epoch .... (Score:1)
Re:"Slashdot for dummies" (Score:2)
Okay, I'll play along...
Hmmmm...
0 Playboy And...Linux? by Hemos on Tuesday January 04, @09:43PM EST 24
1 Linux Kernel 2.2.14 by CmdrTaco on Tuesday January 04, @04:16PM EST 246
2 The 2.3.x "Things To Fix" List by CmdrTaco on Tuesday January 04, @02:43PM EST 128
-- The current article
3 The ROX Desktop by emmett on Monday January 03, @05:43PM EST 215
-- Okay this one is Linux/Unix
4 Interview: a New Linux Year with Jon 'maddog' Hall by Roblimo on Saturday January 01, @02:02PM EST 88
5 Universal Linux-based Internet Appliance by CmdrTaco on Saturday January 01, @09:32AM EST 52
6 Linux Last in Deja Network OS Poll by Roblimo on Saturday January 01, @07:10AM EST 156
7 Forrester Report: Linux Hysteria Will Fade In 2000 by Hemos on Thursday December 30, @03:29PM EST 241
8 The Linux Newbie Replies: WFM? by Hemos on Thursday December 30, @12:14PM EST 367
9 Yet Another Linux Driver Petition by CmdrTaco on Thursday December 30, @07:44AM EST 140
10 US Army Needs Linux Workstation Advice by Cliff on Tuesday December 28, @01:09PM EST 396
11 Second "Bonus" Interview: Jon "maddog" Hall by Roblimo on Monday December 27, @12:00PM EST 102
-- Linux, trust me!
12 Linus One of Fortune's "People to Watch in 2000" by Hemos on Monday December 27, @10:31AM EST 87
-- Linus is the creator of Linux
13 Crack.LinuxPPC.org Cracked by CmdrTaco on Monday December 27, @08:12AM EST 132
14 Linux Handwriting Recognition by CmdrTaco on Saturday December 25, @08:25AM EST 79
15 "What is Linux Missing?" by CmdrTaco on Friday December 24, @07:34AM EST 732
16 Opera Beta Released by Roblimo on Friday December 24, @06:35AM EST 255
-- This was Opera for Linux
17 UK Gov't Experts Say Linux is Secure, Windows Not by Roblimo on Thursday December 23, @10:14PM EST 277
18 The MassLinux Disappearance Explained by Hemos on Thursday December 23, @11:30AM EST 137
19 386 Based Linux Powered Telephone by CmdrTaco on Wednesday December 22, @08:49PM EST 132
20 Realtime Linux Workshop in Vienna by CmdrTaco on Wednesday December 22, @11:02AM EST 66
21 LWN Does Year in Review for Linux by Hemos on Tuesday December 21, @06:22PM EST 60
22 HP Still Porting Linux to 64 bit PA RISC by Hemos on Tuesday December 21, @08:45AM EST 54
23 Tivo Source Code Released by CmdrTaco on Tuesday December 21, @07:40AM EST 220
-- Tivo runs on Linux
24 Is SCSI Sub-Par Under Linux? by Cliff on Friday December 17, @06:37AM EST 263
25 Corel and Red Hat Rumors Continue by CmdrTaco on Thursday December 16, @11:20AM EST 180
-- Linux companies
26 Adobe Announces Initial Support for Linux by HeUnique on Thursday December 16, @05:15AM EST 256
27 Linus Announces Move into Pre-2.4 Stage by Hemos on Wednesday December 15, @12:38PM EST 209
-- See above re Linus/Linux
28 Matra to open source their CAD component library by sengan on Tuesday December 14, @08:40PM EST 139
-- Okay, open source
29 Wearable PCs Under Linux by CmdrTaco on Tuesday December 14, @12:16PM EST 133
So that's 27 pure Linux, and 2 borderline.
I'll agree there's a wide variety, but: unclear?
Chris
Re:It's the drivers (Score:1)
Hold on a sec... I just got ADSL installed at one of our offices yesterday. BellSouth couldn't give me the Alcatel 1000 ADSL that I normally request (they've been out of them for awhile, but have ordered more). Since it's a three person office, I had them hook the ADSL to the receptionist's computer, and planned on using ISC (ICS?) in Win98 SE to share (BellSouth couldn't get Windows dial-up working after four or five hours... it took me two hours, not counting the four hours or so of trying to find the damn Product Key).
Now, I know it's a PPP Efficient Networks card, and in a directory on the computer, they left a directory that they said contained the Linux drivers. The directory exists, but I have't gone into it yet. The guy who did the install was a hardcore geek (after four hours of sitting watching Window 98 reboot, you chat about stuff), and even showed me some undocumented features of the Windows driver (when you have the PPP monitor up, hit "Alt-Shift-A" to bring up "Administrator Mode"). If anyone needs it, email me... but it should be up somewhere. (evan@onepaper.com)
And today, I finally got it all working by backing up and formatting, and reinstalling Windows 98. Turns out, the copy the office had, that they SWORE was SE, was a "first edition" Win98 disk. I have their network up, but no proxy (my fault for assuming they were right).
I never did anything to make ADSL work on my home Mandrake box... it booted up, and just saw the net. It took about half an hour to get IP Masquerade working for my home network, mostly because I took my time reading the HOWTO. When I tried Corel, the same thing - no configuration, just net.
And which is the harder system to setup and use? (30 minutes versus 10 hours and counting...)
--
Evan
Minor change --> Minor Test Run (Score:3)
But if all you did was to patch the kernel a bit to fix a particular problem, it may be desirable to just run the tests that you figure are related to that change.
Rerunning the full suite overnight or on some other reasonable periodic basis to find problems that may have been introduced would be an obvious thing to do.
The real point is that if the test suite grows to 15MB of source code, and runs for 25 hours, you don't run the whole thing every time you make a little change. You run the parts that could conceivably be relevant. And run the whole thang once in a while.
Or perhaps have a daemon that grabs the latest kernel every time one is released, and runs it through regression.
That's not a concern until there's so many tests that they'll run for many hours...
redundant? (Score:2)
Re:Minor change --> Minor Test Run (Score:1)
Hal Duston
Re:APM under SMP (Score:1)
append = "apm=smp-power-off"
in your lilo.conf to enable it.
Re:NFSV3 is possible. (Score:1)
We use a LOT of nfs at work. I already have Trond's NFS V3 patches applied to my kernels, and they work OK, sort of. I can also run an NFS V3 server with the knfsd patches (or Neil Brown's new patches). But the linux nfs V3 server, while working with Digital and SGI NFS V3 clients, cannot be seen by linux nfs v3 clients, OR linux nfs v2 clients.
I feel safe in saying linux nfs v2 server and client are working well. V3 would sure help a lot, but I just don't see it happening in any sort of stable well-maintained way - yet. Of course, the speed differences are not great until you get to 100 BT lines anyway...
This has HUGE raminifications for the usability of linux as a file server in a distributed Unix environment. NFS servers from linux are behind virtually EVERY other flavor of Unix. I suspect they will come around. But nfs is not exactly improving fast.
Hacking Crash course (Score:1)
Hacking Crash course PLAIN OLD TEXT, sorry (Score:1)
2. Forget about learning the whole source. Only a moron would do that. This relates to #1. The manuals were put together for two reasons, to avoid the monstrous stupidity of needing one hacker to accompany every user at all times and to avoid the monstrous stupidity of making people record unnecessary information. You want to be a hacker not a paranoid fascist effiency expert wannabe incapable of accomplishing anything to which you aspire, aka a wanker.
3. Learn how to read source. Applies to #2. This is even more important than fully learning a language. Know how to separate sections so that you can reconstruct them similarly rather than ridiculously attempting to regurgitate them from memory. It's more imprtant to know what a function looks like, what a comment looks like, and how to trace one result from one funtion to the use of that result in another function.
4. Applies to #3. Learn where to put infromation that you find. This is the 3rd most important part. Any project that looks like a black box is intimidating if you don't have an idea what you're trying to solve. For any project, state the obvious, lay it out on paper. Then begin to develop.
5. Applies to #4. (2nd Most important part). Learn to generalize. Remember software is virtual machinery. Virtual physics (somebody needs to write a book on that) is different from real physics. When you move a paperweight everything happens on its own, you don't have to tell the universe to move the image you see of the paperweight along with the paperweight itself. In a 3-D program you have to do that. One of the ways you become aware of this is by understanding the complete meaning of the idea that your screen is just a bunch of colored dots. A hacker gains a lot of freedom from knowing that. The complexity and collapses into simple models, the confusion disappears. An example is the crazy amount of redundancy virtual physics imposes on 3-D programs. Calculating visible areas from every cubic unit of space, takes a lot of time, to prevent that you have to waste a lot of space.. In real life light either reaches your eye or it doesn't. No big deal.
6. Applies to everything, so this is the most important tip. Nothing is simple; virtual physics prevents this. You have to identify extremes in the concept of your project where things either run slowly or take an enormous amount of space. An example is the cellular directory assistance place I work at: Sure type in the query values and presto you get your phone number. simple right? No. If your customer is looking for A & P super food stores, they might as well go out for coffee, cuz a lot of words begin with A + P. Lots of businesses like initials which makes it worse. Then there's things like the fact that Southern States have towns that have 3 or 4 different names plus you have customers that ask for the Police Department in Northern Virginia. Good luck. In CPU design you have CPU caches, It's a matter of balance because when you improve performance by increasing associativity, you screw perfomance because that automatically increases the miss rate.
7. As a result of #6 you need to learn balance. Speed vs space, Cache Hit Wait vs Miss Rate, Power and detail in your project versus deadline.
So you can see it has nothing at all to do with code. It's about your organizational skills. The rest just writes itself.
Being steamrolled by a Microsoft Unix wannabe is (Score:2)
All of 2000 now refers to \Device\Harddisk0\Partition1. All of 2000 will name your hard drives according to what port and channel they're instralled on. Now why didn't we think of that? We did 30 years ago. Letter names will disappear soon. 2000 already has the mount syscall.
2000 Server comes with a (gasp) Mount point manager.
I'm sorry but I have no reason to switch back to Windows.
Is Linux missing some things? Sure. Are the betas of those portions kicking ass already? Most certainly.
Pffffffffffft (Score:1)
Okay let's see here (Score:1)
Can you replace the shell? Can you replace the kernel? Does it run on 20+ Architectures? Would 911 run Windows (I do wonder why they run Linux, hmm..)?
The feature set is already richer than Linux/Unix is many areas - anything that's missing - write it yourself
Gee that's original.
Write kernel extensions, drivers, software. I don't really see how what you say is relevant.
Again, how is that better than Linux. Dude man show me a 2000 release that doen't need sound recorder or media player to play a sound file. Only the driver.
cat(concatenate, or just send) sound.au(raw audio) > (to, like DUH!)
(for the slow parenthese are not to be typed in.)
I can play a file from a damned one line script.
Second, Kernel extensions that need to be rewritten every few years compared to fully published stable APIs. Nope no thanks.
Sorry but no dice. It's still spit and glue from what I've seen, though it is kind of starting to look like Unix.
You're like one of those people who go and compare Linux + every single piece of unix software created to Windows NT out of the box.)
Nope. Wrong. I compare only the distributions I use to what comes on the Windows CD. I only need to do that. Though Linux still beats Windows in bare bones setup.
In a full install, I get full Internet readiness from using to developing, from small time applications to tools for setting up a major global organization in no time flat. Could Linux distros use some WYSIWYG editors on CD, maybe, but Windows so far doesn't offer me anything to match until I buy separate products. Windows gives me a click and drool universe and oh my god look at the pretty colors euphoria as the only reward. I'm not a kid anymore. It takes power and gunctionality to make me drool. I'm certainly not impressed.
I'm soory but you're the one who sounds sore.
Re:What I would really like.. (Score:1)
Linus is the master of lazy programmers.
Re:"Slashdot for dummies" (Score:2)
Aaaarrrrgghhhhh. Are you stupid?!?!?!?!
Or are you not paying attention to what I said?
Let me give you an example:
"In a press release today, the White House said the President is not feeling well"
Now, quick; Is that the president of IBM? Zimbabwe? The Nose-Picker's Club of Estonia?
but a out of context version number
It's not out of context!!! It is implied. Shall I give you a translation table for reference?
A Version number sitting alone next to:
Linux Logo ---> Linux kernel
BeOS Logo ----> BeOS version
Apache Logo --> Apache version
etc.
What's so hard about that to understand?
Chris
Re:Okay let's see here (Score:2)
Telnet?
> Componentization (COM+),
GNOME and KDE both have good versions of this.
> Telephony (TAPI),
Dunno but I remmber seeing an app advertised on freshmeat thats been around since 96 to do this.
> Speech (SAPI),
There's something for it, can't remember what.
> 3D (DirectX),
DirectX sucks.
OpenGL works fine.
> DataAccess (ODBC),
Ummmm,
MySQL, Oracle
.....
> Accessibility (MSA)
Fairly good
> Transaction Servers (MTS),
No idea.
> Message Queuing (MSMQ),
Haven't a clue
> IIS,
Now you're just being silly.
I've forgotten what it's called...
Oh yeah apache
> ASP,
php
> ActiveX,
Com+ in a fancy name.
> PnP..etc
2.3 supports my pnp cards without any hassle.
In fact I've had more luck with pnp stuff
under linux, than under windows.
>
Free to obtain,
how much does it cost to have more than 10people using it at once?