I'm presently running 2.4.8 on a machine. The VM on this is not terribly
good (swaps a lot with seemlingly plenty of physical memory).
I'm considering going to an -ac kernel, but I need recent iptables. Is the
iptables code up to date in -ac?
Also, which -ac do people recommend? I've beent trying to follow lkm, but
I'm somewhat confused at this point.
The box:
P200MMX 64MB
It's used as a firewall and a ssh login/through server for external connections.
Thx.
--Jens
--
----------------------------------
Jens Petersohn x33128
On Thu, 2001-10-11 at 10:45, [email protected] wrote:
> I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> good (swaps a lot with seemlingly plenty of physical memory).
> I'm considering going to an -ac kernel, but I need recent iptables. Is the
> iptables code up to date in -ac?
> Also, which -ac do people recommend? I've beent trying to follow lkm, but
> I'm somewhat confused at this point.
The newest ac is always what is recommended. Right now that is
2.4.10-ac11. You will need linux-2.4.10.tar.bz2 and
patch-2.4.10-ac11.bz2.
That code is very stable and has Rik's VM. It should have the newest
netfilter code -- Alan is very up to date.
This isn't to say not to use Linus's tree, though. You may want to give
2.4.12 a try, too.
Robert Love
On Thu, 11 Oct 2001 [email protected] wrote:
> I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> good (swaps a lot with seemlingly plenty of physical memory).
> I'm considering going to an -ac kernel, but I need recent iptables. Is the
> iptables code up to date in -ac?
Seems to be, how recent do you want? I am using it, anyways....
> Also, which -ac do people recommend? I've beent trying to follow lkm, but
> I'm somewhat confused at this point.
-ac is much better than -linus for me.
I am using 2.4.9-ac18. It is mostly good, the occasion swapping when I
leave mozilla or xemacs alone for a little while, but mostly good. Much
better than ever since 2.4.~5 though!
Used 2.4.10-ac1 for a while, but seems worse than 2.4.9-ac1[678].
Looking at the changelog and comments on the list and /. though - very
very promising with 2.4.10-ac11 with the new VM changes. I will compile
that one tonight and try it out when I next reboot (finally having a
decent kernel has given me some uptime I don't want to destroy though ;)
> The box:
>
> P200MMX 64MB
>
> It's used as a firewall and a ssh login/through server for external connections.
2.4.9-ac* should be good for such a box - looks like you don't put much
demand on it (although the RAM is a little small....)
--
TimC -- http://www.physics.usyd.edu.au/~tcon/
"I give up," said Pierre de Fermat's friend. "How DO you keep a
mathematician busy for 350 years?"
Any special reason to use 2.4? I only use 2.2.19p8 and 2.2.20p10 where
stability is important.
rgds,
tim.
[email protected] wrote:
>
> I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> good (swaps a lot with seemlingly plenty of physical memory).
> I'm considering going to an -ac kernel, but I need recent iptables. Is the
> iptables code up to date in -ac?
> Also, which -ac do people recommend? I've beent trying to follow lkm, but
> I'm somewhat confused at this point.
>
> The box:
>
> P200MMX 64MB
>
> It's used as a firewall and a ssh login/through server for external connections.
--
[email protected] wrote:
>
> I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> good (swaps a lot with seemlingly plenty of physical memory).
> I'm considering going to an -ac kernel, but I need recent iptables. Is the
> iptables code up to date in -ac?
> Also, which -ac do people recommend? I've beent trying to follow lkm, but
> I'm somewhat confused at this point.
>
> The box:
>
> P200MMX 64MB
>
> It's used as a firewall and a ssh login/through server for external connections.
Any reason not to stick with 2.2.20preX? Especially where stability is
important.
rgds,
tim.
--
Tim Moore wrote:
> Any special reason to use 2.4?
er... scalability, performance, features?
> I only use 2.2.19p8 and 2.2.20p10 where
> stability is important.
experimental pre-releases? interesting...
All my 2.4 based servers are running quite
reliably - the oldest now have over180 days
uptime, all have been up since install or last
kernel with no hint of instability.
Red Hat and Late -ac kernels are especially
stable examples of 2.4 -
cu
jjs
J Sloan wrote:
>
> Tim Moore wrote:
>
> > Any special reason to use 2.4?
>
> er... scalability, performance, features?
Observations based on Roswell 2 and identical Abit BP6's: faster disk
I/O and kernel builds (same options), smoother X11 performance (SVGA),
higher LAN network I/O (switched LNE100TX) under heavy loads, and, none
of the recent latency or VM issues. As for features, I don't need any
new feature specific to 2.4.
> > I only use 2.2.19p8 and 2.2.20p10 where
> > stability is important.
>
> experimental pre-releases? interesting...
I see your point but everything since 2.2.19p2 been stable for my NFS
and app server testing needs as well as primary desktop machine.
rgds,
tim.
--
* Tim Moore <[email protected]> on Thu, Oct 11, 2001:
> Any special reason to use 2.4? I only use 2.2.19p8 and 2.2.20p10 where
> stability is important.
>
Before you turn this into another useless 2.2 vs. 2.4 thread, let me ask
you a simple question:
Where the fsck is iptables in 2.2.20 ?
*Read* people, how hard is it? The original post follows, note the poster
wanting "recent iptables", not 2.2.x.
M. R.
>
> [email protected] wrote:
> >
> > I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> > good (swaps a lot with seemlingly plenty of physical memory).
> > I'm considering going to an -ac kernel, but I need recent iptables. Is the
> > iptables code up to date in -ac?
> > Also, which -ac do people recommend? I've beent trying to follow lkm, but
> > I'm somewhat confused at this point.
> >
> > The box:
> >
> > P200MMX 64MB
> >
> > It's used as a firewall and a ssh login/through server for external connections.
>
Tim Moore wrote:
> Observations based on Roswell 2 and identical Abit BP6's: faster disk
> I/O and kernel builds (same options), smoother X11 performance (SVGA),
> higher LAN network I/O (switched LNE100TX) under heavy loads, and, none
> of the recent latency or VM issues.
You might have a pathological case there, it's
not unheard of -
But just out of curiosity, are you comparing the
stock kernel shipped with roswell, which is of
necessity safe, bland and generic, to your own
optimized, hand configured, custom compiled
2.2 kernel?
Just compiling a 2.4.9-ac by hand gave me 30%
benchmark improvement over the kernel that
shipped with roswell, so be sure to compare
apples with apples!
> As for features, I don't need any
> new feature specific to 2.4.
iptables is one biggie for me -
> I see your point but everything since 2.2.19p2 been stable for my NFS
> and app server testing needs as well as primary desktop machine.
As long as it does the job, no rush to upgrade -
I have some very busy servers running 2.2.17,
which have uptimes near 500 days - I'm in no
hurry to upgrade those - but for any new installs,
a Red Hat or Suse 2.4-based distro is the only
thing that makes any sense to me -
With all the talk about "instability" in the 2.4
series, the fact is, you run a 2.4 distro kernel
that has been painstakingly patched & brutally
QA'd the way e.g. Red Hat does, and you will
have stability.
cu
jjs
did you try 2.4.12?
On Thu, 11 Oct 2001 [email protected] wrote:
> I'm presently running 2.4.8 on a machine. The VM on this is not terribly
> good (swaps a lot with seemlingly plenty of physical memory).
> I'm considering going to an -ac kernel, but I need recent iptables. Is the
> iptables code up to date in -ac?
> Also, which -ac do people recommend? I've beent trying to follow lkm, but
> I'm somewhat confused at this point.
>
> The box:
>
> P200MMX 64MB
>
> It's used as a firewall and a ssh login/through server for external connections.
>
> Thx.
>
> --Jens
> --
> ----------------------------------
> Jens Petersohn x33128
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
On Thursdayen den 11 October 2001 20.56, Tim Moore wrote:
> J Sloan wrote:
> > Tim Moore wrote:
> > > Any special reason to use 2.4?
> >
> > er... scalability, performance, features?
>
> Observations based on Roswell 2 and identical Abit BP6's: faster disk
> I/O and kernel builds (same options), smoother X11 performance (SVGA),
> higher LAN network I/O (switched LNE100TX) under heavy loads, and, none
> of the recent latency or VM issues. As for features, I don't need any
> new feature specific to 2.4.
Hi, sorry for intruding, but what is Roswell2?
And, I also have a Abit BP6. Do you really mean that 2.2.19+ has better
performance?
Maybe i should go back using an older kernel to test it. On this machine I
have 2 30GB IBM disks on the HPT controller in raid 0+1, I could try som
benchmarking with bonnie.
> > > I only use 2.2.19p8 and 2.2.20p10 where
> > > stability is important.
> >
> > experimental pre-releases? interesting...
>
> I see your point but everything since 2.2.19p2 been stable for my NFS
> and app server testing needs as well as primary desktop machine.
>
> rgds,
> tim.
--
Oden Eriksson, Jokkmokk, Sweden.
Mandrake Linux release 8.1 (Vitamin) for i586, kernel 2.4.10-2mdksmp. Uptime:
2 days
Em Fri, Oct 12, 2001 at 01:50:58AM +0200, Oden Eriksson escreveu:
> On Thursdayen den 11 October 2001 20.56, Tim Moore wrote:
> > Observations based on Roswell 2 and identical Abit BP6's: faster disk
> > I/O and kernel builds (same options), smoother X11 performance (SVGA),
> > higher LAN network I/O (switched LNE100TX) under heavy loads, and, none
> > of the recent latency or VM issues. As for features, I don't need any
> > new feature specific to 2.4.
>
> Hi, sorry for intruding, but what is Roswell2?
The code name of a Red Hat devel distro, IIRC
- Arnaldo
``"90% of everything is crap", Its called Sturgeon's law 8)
One of the problems is indeed finding the good bits''
- Alan Cox
Oden Eriksson wrote:
> Hi, sorry for intruding, but what is Roswell2?
The second beta of Red Hat 7.2 -
cu
jjs
Tim Moore <[email protected]> said:
[...]
> Any reason not to stick with 2.2.20preX? Especially where stability is
> important.
iptables vs ipchains for me.
--
Horst von Brand [email protected]
Casilla 9G, Vin~a del Mar, Chile +56 32 672616
Well I'd have to agree that for stability I'd also go for 2.2.x. 2.4.x
isn't bad but 2.2.x is just rock stable right now. Furthermore its been
hard to gain confidence in 2.4.x with all the bugs that have yet to be
worked out. I'd use 2.2.x almost exclusively if it would just gain support
for the latest EIDE chipsets, a journaling filesystem, and the latest SMP
boards. iptables and large file support would also be great.
----- Original Message -----
From: "J Sloan" <[email protected]>
To: "Tim Moore" <[email protected]>
Cc: <[email protected]>
Sent: Thursday, October 11, 2001 2:23 PM
Subject: Re: Which kernel (Linus or ac)?
> Tim Moore wrote:
>
> > Any special reason to use 2.4?
>
> er... scalability, performance, features?
>
> > I only use 2.2.19p8 and 2.2.20p10 where
> > stability is important.
>
> experimental pre-releases? interesting...
>
> All my 2.4 based servers are running quite
> reliably - the oldest now have over180 days
> uptime, all have been up since install or last
> kernel with no hint of instability.
>
> Red Hat and Late -ac kernels are especially
> stable examples of 2.4 -
>
> cu
>
> jjs
>
>
>
>
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
Rock solid? You must have better luck than I and many acquaintances with
Redhat. We've tried out many Redhat distros just to be caught in an
avalanche of bug and other "issues". Furthermore they appear to just love
beta or alpha software. I still cannot believe that they used an ALPHA
version of vim 6.0 as the default system text editor. Only guess is that it
must have been just for the extra version number since the 5.x versions have
all been more than enough for simple text editing.
----- Original Message -----
From: "J Sloan" <[email protected]>
To: "Tim Moore" <[email protected]>
Cc: <[email protected]>
Sent: Thursday, October 11, 2001 3:59 PM
Subject: [OT] Re: Which kernel (Linus or ac)?
> Tim Moore wrote:
>
> > Observations based on Roswell 2 and identical Abit BP6's: faster disk
> > I/O and kernel builds (same options), smoother X11 performance (SVGA),
> > higher LAN network I/O (switched LNE100TX) under heavy loads, and, none
> > of the recent latency or VM issues.
>
> You might have a pathological case there, it's
> not unheard of -
>
> But just out of curiosity, are you comparing the
> stock kernel shipped with roswell, which is of
> necessity safe, bland and generic, to your own
> optimized, hand configured, custom compiled
> 2.2 kernel?
>
> Just compiling a 2.4.9-ac by hand gave me 30%
> benchmark improvement over the kernel that
> shipped with roswell, so be sure to compare
> apples with apples!
>
> > As for features, I don't need any
> > new feature specific to 2.4.
>
> iptables is one biggie for me -
>
> > I see your point but everything since 2.2.19p2 been stable for my NFS
> > and app server testing needs as well as primary desktop machine.
>
> As long as it does the job, no rush to upgrade -
>
> I have some very busy servers running 2.2.17,
> which have uptimes near 500 days - I'm in no
> hurry to upgrade those - but for any new installs,
> a Red Hat or Suse 2.4-based distro is the only
> thing that makes any sense to me -
>
> With all the talk about "instability" in the 2.4
> series, the fact is, you run a 2.4 distro kernel
> that has been painstakingly patched & brutally
> QA'd the way e.g. Red Hat does, and you will
> have stability.
>
> cu
>
> jjs
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
"T. A." wrote:
> Rock solid? You must have better luck than I and many acquaintances with
> Redhat. We've tried out many Redhat distros just to be caught in an
> avalanche of bug and other "issues".
That's odd, my Red Hat servers have been
almost like appliances. Not perfect, but all
the problems I've seen could be dealt with.
I've seen worse problems in commercial
unices, e.g. unixware - yuk - and hpux.
> Furthermore they appear to just love
> beta or alpha software.
They do break new ground, and are often the
first distro to implement new technology.
You have to know the red hat pattern:
x.0 release: New stuff, interesting, buggy
x.1 release: A passable cleanup of the bugs in .0
x.2 release: A smooth, polished, evolution of .1
> I still cannot believe that they used an ALPHA
> version of vim 6.0 as the default system text editor.
(shrug) I had no idea it was alpha - it works
well for me, and I'm a vi man.
cu
jjs
On Fri, Oct 12, 2001 at 12:37:01AM -0400, you [T. A.] claimed:
> Well I'd have to agree that for stability I'd also go for 2.2.x. 2.4.x
> isn't bad but 2.2.x is just rock stable right now. Furthermore its been
> hard to gain confidence in 2.4.x with all the bugs that have yet to be
> worked out. I'd use 2.2.x almost exclusively if it would just gain support
> for the latest EIDE chipsets, a journaling filesystem, and the latest SMP
> boards. iptables and large file support would also be great.
Of course, you can get most of the IDE chipset support, fs support (reiserfs
3.5, ext3) and LFS support as patches for 2.2:
ftp://ftp.kernel.org/pub/linux/kernel/people/hedrick/ide-2.2.19/
ftp://ftp.namesys.com/pub/reiserfs-for-2.2/linux-2.2.19-reiserfs-3.5.34-patch.bz2
ftp://ftp.kernel.org/pub/linux/kernel/people/sct/ext3/
http://moldybread.net/patch/kernel-2.2/linux-2.2.19-lfs-1.0.diff.gz
-- v --
[email protected]
> That's odd, my Red Hat servers have been
> almost like appliances. Not perfect, but all
> the problems I've seen could be dealt with.
Oh I could deal with all of the problems as well. But after a good bit
of recompiling, patching, upgrading, backtracking the things done in the
"Redhat" way which many times don't match the man pages, as well as undoing
the "Redhat" way annoyances I just end up with a variation of my own hand
built distribution. And once I have to replace the system experimental C
library and compiler it just get even more ridicules.
> I've seen worse problems in commercial
> unices, e.g. unixware - yuk - and hpux.
I know. Please don't remind me of UnixWare. I buried that at my
clients' and companies' sites in the Linux 1.2.x days.
> x.0 release: New stuff, interesting, buggy
> x.1 release: A passable cleanup of the bugs in .0
> x.2 release: A smooth, polished, evolution of .1
Well here's hoping that 7.2 is smooth and polished. Especially since
Redhat has become the defacto standard and I'll probably be trying it out
again.
> (shrug) I had no idea it was alpha - it works
> well for me, and I'm a vi man.
I never keep it around for long after I saw it crash a couple of times
on my first 7.0 or 7.1 box so I can't say much about its stability. Besides
it was more the principal of the thing. Why in the world use the alpha
version of a text editor when a perfectly good released version exists with
pretty much the same functionality. One thing I did notice, if I remember
correctly. Seamed that either the installed vim was crippled or had a bug
in reading my customized .vimrc file. My custom options didn't get loaded
until I downloaded the latest 5.[89] source and recompiled myself a new
version.
>
> Of course, you can get most of the IDE chipset support, fs support
(reiserfs
> 3.5, ext3) and LFS support as patches for 2.2:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/hedrick/ide-2.2.19/
Have used this and has worked great on the machines I've had to use it
on. Though I'm a bit leery about using it since I figure the generic
2.2.x.preX kernels get a lot more testing that those with this patch
installed. Also heard of problems using this patch on a VIA PIII SMP
system. :-( And just went I had been planning to use it on a dual PIII VIA
chipset board too.
>
>
ftp://ftp.namesys.com/pub/reiserfs-for-2.2/linux-2.2.19-reiserfs-3.5.34-patc
h.bz2
I actually was going to start using this until I learned that 2.2.x
reiser patched kernels couldn't use reiserfs partitions made with 2.4.
:-( Ended up having to redo an entire system when a downgrade to 2.2.x
became imperitive. Also the 2.2.x reiser patch lacks the large file support
(on the reiser filesystems created under 2.2.x) and maybe other goodies and
I don't get that back easily by just switching kernels.
> ftp://ftp.kernel.org/pub/linux/kernel/people/sct/ext3/
Might give this a try now that it appears to be release quality.
> http://moldybread.net/patch/kernel-2.2/linux-2.2.19-lfs-1.0.diff.gz
I'll look into this the next time > 2GB files support becomes needed on
a system. pre 2.4.x I had been using FreeBSD for such tasks.
On Fri, 12 Oct 2001, T. A. wrote:
>
> I know. Please don't remind me of UnixWare. I buried that at my
> clients' and companies' sites in the Linux 1.2.x days.
>
> > x.0 release: New stuff, interesting, buggy
> > x.1 release: A passable cleanup of the bugs in .0
> > x.2 release: A smooth, polished, evolution of .1
>
> Well here's hoping that 7.2 is smooth and polished. Especially since
> Redhat has become the defacto standard and I'll probably be trying it out
> again.
>
Please, this is just the wrong mental attitude. I have nothing against red
hat distribution, but the just one standard is LSB one (I do admitt, many
people from Red Hatm, SuSE and other distributions worked on it).
If a distribution is LSB compliant, it is standard. usually de facto
standard tend to vanish with the time.
On Fri, Oct 12, 2001 at 05:25:32AM -0400, you [T. A.] claimed:
> >
> > Of course, you can get most of the IDE chipset support, fs support
> (reiserfs
> > 3.5, ext3) and LFS support as patches for 2.2:
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/hedrick/ide-2.2.19/
>
> Have used this and has worked great on the machines I've had to use it
> on. Though I'm a bit leery about using it since I figure the generic
> 2.2.x.preX kernels get a lot more testing that those with this patch
> installed. Also heard of problems using this patch on a VIA PIII SMP
> system. :-( And just went I had been planning to use it on a dual PIII VIA
> chipset board too.
I've used it on multiple 2.2 systems as well (Dual Celeron/BX400++HPT366,
Via/Duron, PII/BX440 etc) and never had a problem.
> ftp://ftp.namesys.com/pub/reiserfs-for-2.2/linux-2.2.19-reiserfs-3.5.34-patc
> h.bz2
>
> I actually was going to start using this until I learned that 2.2.x
> reiser patched kernels couldn't use reiserfs partitions made with 2.4.
Yeah.
I've used it for a long time and only once had a small issue with it (which
didn't impose data corruption, just one app (UML) didn't work since mmap on
old 2.2 reiser was somewhat broken).
> :-( Ended up having to redo an entire system when a downgrade to 2.2.x
> became imperitive. Also the 2.2.x reiser patch lacks the large file support
> (on the reiser filesystems created under 2.2.x) and maybe other goodies and
I thought you could get LFS on reiser on 2.2 with the LFS patch and some
patch to reiser? I'm not sure though. SuSE did ship with 2.2, reiser and
large file support...
> > http://moldybread.net/patch/kernel-2.2/linux-2.2.19-lfs-1.0.diff.gz
>
> I'll look into this the next time > 2GB files support becomes needed on
> a system. pre 2.4.x I had been using FreeBSD for such tasks.
Nowdays, though, I think I 2.4 is beginning to be stable enough for just
about anything. The first 2.4 kernels were terrible wrt vm - they'd go ahead
killing innocent daemons when I did a simple diff -R /usr/src/linux
/usr/src/linux2 with hundreds of MBs free RAM.
-- v --
[email protected]
On Fri, Oct 12, 2001 at 09:56:19AM +0300, Ville Herva wrote:
> On Fri, Oct 12, 2001 at 12:37:01AM -0400, you [T. A.] claimed:
> > Well I'd have to agree that for stability I'd also go for 2.2.x. 2.4.x
> > isn't bad but 2.2.x is just rock stable right now. Furthermore its been
> > hard to gain confidence in 2.4.x with all the bugs that have yet to be
> > worked out. I'd use 2.2.x almost exclusively if it would just gain support
> > for the latest EIDE chipsets, a journaling filesystem, and the latest SMP
> > boards. iptables and large file support would also be great.
>
> Of course, you can get most of the IDE chipset support, fs support (reiserfs
> 3.5, ext3) and LFS support as patches for 2.2:
btw, just a reminder, 2.2.20pre10aa1 has full lfs support too
(everything, including getdents64, nfv3, lockd all 64bit).
ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.2/2.2.20pre10aa1.bz2
ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.2/2.2.20pre10aa1/40_lfs-2.2.20pre10aa1-28.bz2
Andrea
After this post we should take it offline and
let the s/n ratio on lkml settle back down to
a dull roar - apologies for the noise, this is
the last post on this dead horse.
"T. A." wrote:
> Oh I could deal with all of the problems as well. But after a good bit
> of recompiling, patching, upgrading, backtracking the things done in the
> "Redhat" way
You may have a point with 7.0, but 7.1 was not
that bad - and in any case, just applying the RH
updates fixed the problems.
> which many times don't match the man pages, as well as undoing
> the "Redhat" way annoyances I just end up with a variation of my own hand
> built distribution.
To each his own - choice is a wonderful thing, isn't it?
> And once I have to replace the system experimental C
> library and compiler it just get even more ridicules.
Experimental? What you call experimental, I call (and
my customers call) fully functional and fully supported.
Using gcc-2.96 on the 40+ RH boxes I have scattered
around the southwest has shown no problems, despite
all the outrage from anti gcc-2.96 activists.
Here is a heads-up for the benefit of those wondering
about gcc-2.96:
http://www.bero.org/gcc296.html
cu
jjs
I am lucky to need only the ipchains simple functions (accounting,
forward, masq and firewall).
rgds,
tim.
Jens Petersohn wrote:
>
> ----- Original Message -----
> From: "Horst von Brand" <[email protected]>
> To: "Tim Moore" <[email protected]>
> Cc: <[email protected]>; <[email protected]>
> Sent: Thursday, October 11, 2001 9:01 PM
> Subject: Re: Which kernel (Linus or ac)?
>
> > Tim Moore <[email protected]> said:
> >
> > [...]
> >
> > > Any reason not to stick with 2.2.20preX? Especially where stability is
> > > important.
> >
> > iptables vs ipchains for me.
>
> Yes, same here. The machine was running 2.2.16 for the longest time,
> but I upgraded to obtain iptables support.
--
On 20011012 J Sloan wrote:
>After this post we should take it offline and
>let the s/n ratio on lkml settle back down to
>a dull roar - apologies for the noise, this is
>the last post on this dead horse.
>
>"T. A." wrote:
>
>
>Here is a heads-up for the benefit of those wondering
>about gcc-2.96:
>
>http://www.bero.org/gcc296.html
>
Nice, a bunch of comments about the front end. But you miss the point
that what was broken in gcc-2.96 was the back end (the optimizer).
And you missed that it needed about 50 updates to get a real compiler.
gcc that ships with RH 7.1 generates bad code in optimized mode. Do not
remember the exact post in LKML, but I saw 2 lines of code that made gcc
put the user initialization of a variable before the automatic one to zero.
If you want a good distro, take Mandrake 8.1.
--
J.A. Magallon # Let the source be with you...
mailto:[email protected]
Mandrake Linux release 8.2 (Cooker) for i586
Linux werewolf 2.4.13-pre1-beo #1 SMP Fri Oct 12 11:32:03 CEST 2001 i686