Unix Technical Forum

Is it worth upgrading the kernel?

This is a discussion on Is it worth upgrading the kernel? within the Slackware Linux Support forums, part of the Unix Operating Systems category; --> Hi All In another thread, the drift went towards configuring a new kernel on my machine. To my knowledge, ...


Go Back   Unix Technical Forum > Unix Operating Systems > Slackware Linux Support

FAQ Members List Calendar Search Today's Posts Mark Forums Read
  #1 (permalink)  
Old 02-18-2008, 10:58 AM
who
 
Posts: n/a
Default Is it worth upgrading the kernel?

Hi All

In another thread, the drift went towards configuring a new kernel on my
machine.

To my knowledge, the supplied 'bare.i' kernel, is 2.4.20. The current stable
release at kernel.org seems to be 2.4.22. Should I bother downloading/
configuring this for such a small point increase? (The changelog doesn't
really make sense to me). Are there known bugs/issues in/with the 'bare.i'
kernel I am using? Will there be any performance gain, however slight?

I'm not asking you to make my mind up, just for your valued opinion. Tailoring
my system in this way is a direction I'd certainly like to go.

I do have another question. If I type swaret --search kernel, it pruces 4
results, thus,
kernel-headers-2.4.21-i386-1 (1119 kB)
kernel-ide-2.4.21-i486-1 (1312 kB)
kernel-modules-2.4.21-i486-1.1 (8889 kB)
kernel-source-2.4.21-noarch-1 (35011 kB)
If I type swaret --install kernel (ie these 4 packages) will this automagically
install the kernel negating the need for doing it manually? I'm not trying to
be lazy, I'm just asking.

Thanks for any replies
Darren
--
kNode 0.7.2 (under KDE3.1.0) | Slackware 9.0
Athlon XP2100+ | Nvidia GeForce4 Ti4200 8x
ASUS A7V333 | Spare Seagate | 512Mb RAM
Gnu/Linux - It /just/ works and /it/ /works/ *fast*
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #2 (permalink)  
Old 02-18-2008, 10:58 AM
Joost Kremers
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

who wrote:
> Hi All
>
> In another thread, the drift went towards configuring a new kernel on my
> machine.
>
> To my knowledge, the supplied 'bare.i' kernel, is 2.4.20. The current stable
> release at kernel.org seems to be 2.4.22. Should I bother downloading/
> configuring this for such a small point increase? (The changelog doesn't
> really make sense to me). Are there known bugs/issues in/with the 'bare.i'
> kernel I am using? Will there be any performance gain, however slight?


i doubt there will be a performance gain, i didn't hear about 2.4.22 being
faster than the previous kernels. (2.6.0-testx seems to be faster, but
that's not really a stable kernel yet, so...)

> I'm not asking you to make my mind up, just for your valued opinion. Tailoring
> my system in this way is a direction I'd certainly like to go.
>
> I do have another question. If I type swaret --search kernel, it pruces 4
> results, thus,
> kernel-headers-2.4.21-i386-1 (1119 kB)


header files (i.e., function definitions) that are necessary to compile
some programs. does not include kernel or source. but you'll want them if
you install a new kernel.

> kernel-ide-2.4.21-i486-1 (1312 kB)


the (ide) kernel itself.

> kernel-modules-2.4.21-i486-1.1 (8889 kB)


kernel modules. you'll want those too.

> kernel-source-2.4.21-noarch-1 (35011 kB)


the source. you'll only want this if you plan to compile your own kernel.

> If I type swaret --install kernel (ie these 4 packages) will this automagically
> install the kernel negating the need for doing it manually? I'm not trying to
> be lazy, I'm just asking.


i'd say so, yes, although i don't use swaret.

you can install this kernel, or stick with 2.4.20 for a while, i doubt it
will make that much of a difference. (but anyone correct me if i'm wrong.)
more important than installing a new kernel is to learn to compile one
yourself, selecting only those things that you need on your system. i saw
someone (forgot his name) already gave you lots of pointers on that. of you
want more background info, there's a lot more info on kernel compilation on
the net, among other things the linux kernel howto at <http://www.tldp.org>
and the README that comes with the kernel sources. just go ahead with it,
try it, and don't worry if it goes wrong a couple of times. just make sure
you can always boot your existing kernel: when you compile a new kernel,
copy it to /boot under a new name (so you don't overwrite the old one) and
create a new entry for it in /etc/lilo.conf. that way, if your new kernel
doesn't boot, you can always boot the old one.

--
Joost Kremers
since when is vi an editor? a discussion on vi belongs in
comp.tools.unusable or something...
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #3 (permalink)  
Old 02-18-2008, 10:58 AM
who
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

On Monday 01 September 2003 22:10, Joost Kremers wrote:
<snip>
> you can install this kernel, or stick with 2.4.20 for a while, i doubt it
> will make that much of a difference. (but anyone correct me if i'm wrong.)
> more important than installing a new kernel is to learn to compile one
> yourself, selecting only those things that you need on your system. i saw
> someone (forgot his name) already gave you lots of pointers on that.


Dave gave me a tutorial for which I couldn't be more grateful and I couldn't
agree with you more. It is *very* important to me to learn my system proper.
I'm a bit too old and fscked up to be considering a career in SysAdmin but
I would love to be able to look at my box and know *I'm* the master. Even
better a network.

> of you
> want more background info, there's a lot more info on kernel compilation on
> the net, among other things the linux kernel howto at <http://www.tldp.org>
> and the README that comes with the kernel sources. just go ahead with it,
> try it, and don't worry if it goes wrong a couple of times. just make sure
> you can always boot your existing kernel: when you compile a new kernel,
> copy it to /boot under a new name (so you don't overwrite the old one) and
> create a new entry for it in /etc/lilo.conf. that way, if your new kernel
> doesn't boot, you can always boot the old one.
>

I'm reading constantly. The point at which I would usually read a README is
when I come to downloading it. I'm not at that point yet.

I rarely ask a question in these groups before I have Googled the web/groups or
read the manpages for the answer, it may be that I simply don't understand the
question or term to use.

Asking a question here, for me, is like having an expert by my side when I hit
a problem. Their answers are also 'human'. If I'm going up the wrong track
(which has happened more than once ) I can be stopped before frustration
sets in and the puter gets hurt (which has also happened more than once )).

I completely understand the RTFM attitude - well, actually, no, I don't. On the
one hand, I know, as a seasoned Mandrake user, how bloody frustrating it is to
get post after post on how to put a bloody symlink from java/ plugin/
libjavaplugin_oji.so to /path/to/moz/plugins/ when there has been the exact
same question asked five posts down. (If I ever do this, it's because I'm a)
tired from 15 hour days b) being nagged at by wife or c) being preoccupied
with a whole slew of new topics on my mind (trying to run before I can walk)
On the other hand, I don't fully understand the RTFM attitude, because
sometimes it's simpler to just tell the answer, than to personally spend time
searching out an esoteric document and then tell the person to read it - of
which one line of code at the commandline can solve the OP's problem in one
swift move. I sometimes imagine these people at work, being tapped on the
shoulder by a colleague, and asked, 'what's that line of code to make such and
such work?' and him turning round and saying 'read the fscking manual and
Google-bash before you dare ask me such a simplistic/moronic/obvious question
that, as some kind of demi-God, I am above answering' (That may not make sense
- I was on a mission). There is another mission, I believe, that's trying to
attract home users, such as myself, to linux - the BOFH line just doesn't help
either me/ Gnu/ Linux/ OSS/ your distribution's efforts.

<rant mode=off>

Thanks for your reply Joost - you're becoming my regular hero

Anyway, I'm off to bed (I wonder how much a trainee SysOp would get paid for 15
hours). Needless to say things have gone from bad to worse with me and the
wife (she just *won't* understand the importance of this!). Think I'll skip
the valium and go for the small blue diamond tonight

No doubt I will see you in the morning.
Darren
--
kNode 0.7.2 (under KDE3.1.0) | Slackware 9.0
Athlon XP2100+ | Nvidia GeForce4 Ti4200 8x
ASUS A7V333 | Spare Seagate | 512Mb RAM
Gnu/Linux - It /just/ works and /it/ /works/ *fast*
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #4 (permalink)  
Old 02-18-2008, 10:58 AM
Simon
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

On Mon, 01 Sep 2003 21:54:45 +0100, who <what@where.com> wrote:
> To my knowledge, the supplied 'bare.i' kernel, is 2.4.20. The current
> stable release at kernel.org seems to be 2.4.22. Should I bother
> downloading/configuring this for such a small point increase?


That depends whether the current kernel has any problems you need fixed
or if the new kernel has [better] support for hardware you have. Look
for mentions of your hardware in the Changelog.

> Are there known bugs/issues in/with the 'bare.i'
> kernel I am using?


Only you can answer that; it depends on your hardware.

> Will there be any performance gain, however slight?


There is a significant performance gain in compiling a kernel
specifically for your hardware (leaving out support for hardware you
don't have and are unlikely to get in the near future). Usually,
performance won't be much different between 2.4.20 and 2.4.22.

When I re-compile my kernel, I usually get the latest stable version,
unless a development kernel (-pre, -rc or -ac) has better support for
my hardware (e.g I started using vmlinuz-2.4.21-rc1-ac2 for improved
AGP support for the SiS648 chipset).

> If I type swaret --install kernel (ie these 4 packages) will this
> automagically install the kernel negating the need for doing it
> manually? I'm not trying to be lazy, I'm just asking.


If it automatically installs the kernel (I don't know; I don't use
swaret), it won't help performance. It might support your hardware
better, though (depending on your hardware). If it won't improve
support for your hardware, add a feature you want or fix a security
hole (none I can think of since 2.4.18), then, in my opinion, there's
no reason to upgrade unless you're re-compiling it with only support
for your hardware.


--
Simon <simon@no-dns-yet.org.uk> **** GPG: F4A23C69
"We demand rigidly defined areas of doubt and uncertainty."
- Douglas Adams

Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #5 (permalink)  
Old 02-18-2008, 10:58 AM
Simon
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

On Mon, 01 Sep 2003 23:12:30 +0100, who <what@where.com> wrote:
> On the other hand, I don't fully understand the RTFM attitude, because
> sometimes it's simpler to just tell the answer, than to personally spend time
> searching out an esoteric document and then tell the person to read it - of
> which one line of code at the commandline can solve the OP's problem in one
> swift move.


In my opinion, while the immediate problem can sometimes be solved by a
one line command, telling the questioner what this line is won't help
them much. Even if it takes longer for me to find the document I think
of (usually only a few seconds, as I'm quite practiced with Google), I
prefer to post a link to the document, as this helps the questioner to
understand why that command works.

If a person knows and understands why a particular command solves their
problem, they can then adapt that command to solve similar problems in
far less time than it would take to search (or ask) for the answer.


--
Simon <simon@no-dns-yet.org.uk> **** GPG: F4A23C69
"We demand rigidly defined areas of doubt and uncertainty."
- Douglas Adams

Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #6 (permalink)  
Old 02-18-2008, 10:58 AM
NeoSadist
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

who wrote:

> Hi All
>
> In another thread, the drift went towards configuring a new kernel on my
> machine.
>
> To my knowledge, the supplied 'bare.i' kernel, is 2.4.20. The current
> stable release at kernel.org seems to be 2.4.22. Should I bother
> downloading/ configuring this for such a small point increase? (The
> changelog doesn't really make sense to me). Are there known bugs/issues
> in/with the 'bare.i' kernel I am using? Will there be any performance
> gain, however slight?
>
> I'm not asking you to make my mind up, just for your valued opinion.
> Tailoring my system in this way is a direction I'd certainly like to go.
>
> I do have another question. If I type swaret --search kernel, it pruces 4
> results, thus,
> kernel-headers-2.4.21-i386-1 (1119 kB)
> kernel-ide-2.4.21-i486-1 (1312 kB)
> kernel-modules-2.4.21-i486-1.1 (8889 kB)
> kernel-source-2.4.21-noarch-1 (35011 kB)
> If I type swaret --install kernel (ie these 4 packages) will this
> automagically install the kernel negating the need for doing it manually?
> I'm not trying to be lazy, I'm just asking.
>
> Thanks for any replies
> Darren


This is a question only you can answer. Is it worth it to you?
(time/learning). For me, using slack 9 kernel 2.4.20 versus most current
2.4.22, i personally don't think it's worth it yet. But this is a question
only you can answer for yourself.

--
" ... and are endowed by their Creator with certain unalienable rights ... "
--Declaration of Independance
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #7 (permalink)  
Old 02-18-2008, 10:58 AM
Ron Matthews
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

Simon <usenet@no-dns-yet.org.uk> wrote:

> In my opinion, while the immediate problem can sometimes be
> solved by a one line command, telling the questioner what this
> line is won't help them much.


That depends on the questioner, the problem and the command.

> Even if it takes longer for me to find the document I think of
> (usually only a few seconds, as I'm quite practiced with Google),
> I prefer to post a link to the document, as this helps the
> questioner to understand why that command works.


The questioner may very well understand the command. It just may
be something he hadn't thought of. This is clearly something you
haven't thought of. You would rather they read a whole book for
just one command.

> If a person knows and understands why a particular command solves
> their problem, they can then adapt that command to solve similar
> problems in far less time than it would take to search (or ask)
> for the answer.


You are an arrogant little twerp who covets what little knowledge
he has and doles it out in a miserly fashion because it gives you a
sense of control which is something you don't have in your impotent
little life. This is obvious for all to see.

You are asked to skip the postings that you feel you cannot
directly answer. If somebody wants educating from the likes of you
they can always ask but we wouldn't hold our breath.

cordially, as always,

rm
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #8 (permalink)  
Old 02-18-2008, 10:59 AM
William D. Tallman
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

Ron Matthews wrote:

> Simon <usenet@no-dns-yet.org.uk> wrote:
>
>> In my opinion, while the immediate problem can sometimes be
>> solved by a one line command, telling the questioner what this
>> line is won't help them much.

>
> That depends on the questioner, the problem and the command.
>
>> Even if it takes longer for me to find the document I think of
>> (usually only a few seconds, as I'm quite practiced with Google),
>> I prefer to post a link to the document, as this helps the
>> questioner to understand why that command works.

>
> The questioner may very well understand the command. It just may
> be something he hadn't thought of. This is clearly something you
> haven't thought of. You would rather they read a whole book for
> just one command.
>
>> If a person knows and understands why a particular command solves
>> their problem, they can then adapt that command to solve similar
>> problems in far less time than it would take to search (or ask)
>> for the answer.

>
> You are an arrogant little twerp who covets what little knowledge
> he has and doles it out in a miserly fashion because it gives you a
> sense of control which is something you don't have in your impotent
> little life. This is obvious for all to see.
>
> You are asked to skip the postings that you feel you cannot
> directly answer. If somebody wants educating from the likes of you
> they can always ask but we wouldn't hold our breath.
>
> cordially, as always,
>
> rm


ROFL!!!!!!

This post reads suspiciously like one of the current entertainingly puerile
posts appearing in... another linux newgroup. The signature is identical,
although the apparent sender is not.

Of course, I'm sure that can't really be the case, but the coincidence is
amusing.

Bill Tallman
--
Registered Linux User: #221586
Slackware-8.1: Smooth as silk!!
IceWM and Gkrellm as always...
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #9 (permalink)  
Old 02-18-2008, 10:59 AM
William D. Tallman
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

who wrote:

> Hi All
>
> In another thread, the drift went towards configuring a new kernel on my
> machine.
>
> To my knowledge, the supplied 'bare.i' kernel, is 2.4.20. The current
> stable release at kernel.org seems to be 2.4.22. Should I bother
> downloading/ configuring this for such a small point increase? (The
> changelog doesn't really make sense to me). Are there known bugs/issues
> in/with the 'bare.i' kernel I am using? Will there be any performance
> gain, however slight?


I doubt whether the Slackware maintainer would have included that kernel,
were there bugs/issues. I understand, although I've not personally
investigated, that 2.4.22 has better USB support, so I just upgraded from
2.4.20. Doing so was trivial whilst upgrading the capabilities of the
installation, so I can't say what results may appear.

> I'm not asking you to make my mind up, just for your valued opinion.
> Tailoring my system in this way is a direction I'd certainly like to go.


If your system is running well, then get the new kernel (patch if you have
the 2.4.20 tree) and compile it using the old config, and try some
comparisons. Differences are only significant if you recognize them, I
think.

> I do have another question. If I type swaret --search kernel, it pruces 4
> results, thus,
> kernel-headers-2.4.21-i386-1 (1119 kB)
> kernel-ide-2.4.21-i486-1 (1312 kB)
> kernel-modules-2.4.21-i486-1.1 (8889 kB)
> kernel-source-2.4.21-noarch-1 (35011 kB)
> If I type swaret --install kernel (ie these 4 packages) will this
> automagically install the kernel negating the need for doing it manually?
> I'm not trying to be lazy, I'm just asking.
>
> Thanks for any replies
> Darren


I'm curious to discover the reality of swaret myself, but I would think that
the kernel should be compiled in any case, even if swaret does the job.
Kernel compiling is not a big deal with modern distros, given a stable
hardware platform.

Swaret for keeping one's installation current might be worthwhile, though.

Bill Tallman
--
Registered Linux User: #221586
Slackware-8.1: Smooth as silk!!!
IceWM and Gkrellm as always...
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #10 (permalink)  
Old 02-18-2008, 10:59 AM
who
 
Posts: n/a
Default Re: Is it worth upgrading the kernel?

On Tuesday 02 September 2003 07:01, Ron Matthews wrote:

> Simon <usenet@no-dns-yet.org.uk> wrote:
>
>> In my opinion, while the immediate problem can sometimes be
>> solved by a one line command, telling the questioner what this
>> line is won't help them much.

>
> That depends on the questioner, the problem and the command.


Exactly. Giving the line of code, together with a link to site (or manpage) is
going to solve the problem immediately for the OP and give that person an
opportunity to read up too. (S)he's going to respect you a lot more too.

> The questioner may very well understand the command. It just may
> be something he hadn't thought of. This is clearly something you
> haven't thought of. You would rather they read a whole book for
> just one command.


Yes. My point exactly. Also, I am dyslexic and have difficulty

> You are an arrogant little twerp who covets what little knowledge
> he has and doles it out in a miserly fashion because it gives you a
> sense of control which is something you don't have in your impotent
> little life. This is obvious for all to see.


I don't know Simon and haven't got anything against him (in fact he's answered
my original question above, for which I, as always, am very grateful), I hope
I never do have reason, but my meaning was to that ilk who 'doles it out in a
miserly fashion' if they do at all.

Regards
Darren
--
kNode 0.7.2 (under KDE3.1.0) | Slackware 9.0
Athlon XP2100+ | Nvidia GeForce4 Ti4200 8x
ASUS A7V333 | Spare Seagate | 512Mb RAM
Gnu/Linux - It /just/ works and /it/ /works/ *fast*
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On
Forum Jump


All times are GMT. The time now is 11:52 PM.


Powered by vBulletin® Version 3.6.5
Copyright ©2000 - 2008, Jelsoft Enterprises Ltd.
SEO by vBSEO 3.2.0
www.UnixAdminTalk.com