Linux-Development-Sys Digest #335, Volume #8     Sun, 10 Dec 00 02:13:13 EST

Contents:
  Re: ramdisk: how much mem do they use? (Peter Pointner)
  Re: changing BASH's path searching ([EMAIL PROTECTED])
  Re: kernel: svc: unknown version(3) when loading nfs ("Karl Heyes")
  Re: What is the command to  . . . ? (Michael V. Ferranti)
  Re: Uninstall Linux SuSE 7.0 (Harry George)
  Re: Intel's SSE block copy paper (was: Re: A faster memcpy and bzero for  (Boris 
Gjenero)
  Re: changing BASH's path searching (Alex Graf)
  Re: syslog and variable arguments? (Chris)
  Re: Uninstall Linux SuSE 7.0 ("Geggo")
  Re: volume question ("Geggo")
  Help! Hosed XMMS trying to upgrade ("Matt")
  ***  F_U_C_K___C_H_R_I_S_T_M_A_S  *** ................. OzEr1oJUC99IM 
([EMAIL PROTECTED])
  Re: changing BASH's path searching ([EMAIL PROTECTED])

----------------------------------------------------------------------------

From: Peter Pointner <[EMAIL PROTECTED]>
Subject: Re: ramdisk: how much mem do they use?
Date: 9 Dec 2000 19:13:46 +0100

Jerome Corre <[EMAIL PROTECTED]> wrote:
>  hi

> I have 32Mb of ram on my linuxbox, I sometime use ramdisk (/dev/ram0 or
> /dev/ram1) to store small filesystem, prepare a bootdisk or when my
> system boot usimg initrd.
> I would like to know if it is possible to know how much memory the
> ramdisk are using. I suppose that they don't all use 4Mb as there are 16
> ramdisk i can use and I only have 32 Mb of ram. How can I found out how
> much memory they use? and if I don't need what is left on one of them
> after i unmount it how can I free the memory a ramdisk use?

The ramdisks use ram only as needed. Initially your 16 ramdisks use no
ram, or almost no (there are probably some data structrures in the kernel.
If you use them, you normally create a file system on them. so if you
create a 1 MByte ext2 file system on /dev/ram0, it uses 1 MByte of ram.
AFAIK the 4 MBytes are only a (configurable) upper limit.

If you want to free the memory, you can do that with the tool freeramdisk,
which seems to do some ioctl on the ramdisk device.

Peter


------------------------------

From: [EMAIL PROTECTED]
Subject: Re: changing BASH's path searching
Date: Sat, 09 Dec 2000 21:41:02 -0000

On Sat, 09 Dec 2000 12:18:33 -0800 Alex Graf <[EMAIL PROTECTED]> wrote:

| Instead of the shell searching for executables in directories listed in
| the $PATH shell variable, how about this:
| 
| for every subdirectory D from some constant base directory (maybe / ),
| do
|       if D has a subdirectory named "bin", then
|               search for the program in there
|       end if
| end for
| 
| With this, every package can have their own isolated place on the hard
| disk.  Each package has their own 'bin' directory, their own 'share'
| directory, their own 'lib' directory, etc.  (Of course, each package
| needs fixing to work with the new path convention.)

I suppose since the exec*p() syscalls do path searching, this might be
considered also a kernel issue.  But it's really a general issue of
overall strategy.  You can just put packages in /opt/<packagename> and
put symlinks as needed elsewhere.  Knowing what was put elsewhere is
just a matter of examining all the symlinks for those that point to that
particular package directory.


| The advantages are:
| New packages can be installed simply by unzipping a properly-packaged
| tarball.  No need for apt, RPM & friends.  Uninstall a package by simply
| deleting its directory.  No need for apt, RPM & friends.

Although I would agree with you, I think most people would call this a
disadvantage.  Despite the trend in programming to move from procedural
orientation to declarative orientation, system administration tends to
not look at things that way.


| I'm not an expert, but this scheme seems a lot easier to manage than apt
| splattering files all over the /usr tree and trying to keep track of who
| put what where, and who depends on what.  Beats me why everyone insists
| on keeping around crufty old /usr, /var, /etc.  Trying to be old-school
| I guess.

Devise a complete standard based on having no /usr or /var or /etc and then
maybe I can comment on what's good and/or bad about it.  As a system  admin,
I know I would not like looking around everywhere for config files I need
to change.  OTOH, config files for a particular package do make sense to be
in that package's directory.


| The bash source looks pretty approachable, findcmd.c seems to implement
| the path searching.  Not sure if I'll have time to work on it, but maybe
| this idea inspires somebody else?  

Try the symlink approach for commands and see how it goes.

-- 
=================================================================
| Phil Howard - KA9WGN |   Dallas   | http://linuxhomepage.com/ |
| [EMAIL PROTECTED] | Texas, USA | http://phil.ipal.org/     |
=================================================================

------------------------------

From: "Karl Heyes" <[EMAIL PROTECTED]>
Crossposted-To: 
alt.os.linux.mandrake,alt.os.linux.slackware,alt.uu.comp.os.linux.questions,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: kernel: svc: unknown version(3) when loading nfs
Date: Sat, 09 Dec 2000 23:06:54 +0000

In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:

> This fix wasn't too hard to find.
> 
> There is a routine in the nfs script in /etc/rc.d/init.d/nfs that
> checks for version 3 of nfs. Because I am running a version 2, for
> the moment I have commented out the routine from running. mountd no
> executes without an argument.
> 

NFSv3 needs to be selected on the kernel compilation (possible
module, I not sure). Only use the recent versions of the kernel
(2.2.18pre2x) with it though. It might be a good idea with NFS 2 as
there has been a few updates wrt NFS.

karl.

------------------------------

From: Michael V. Ferranti <[EMAIL PROTECTED]>
Crossposted-To: 
alt.os.linux.mandrake,alt.os.linux.slackware,alt.uu.comp.os.linux.questions,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: What is the command to  . . . ?
Date: Sat, 09 Dec 2000 23:05:19 +0000

And [EMAIL PROTECTED] spoke thusly:

>What is the command to find the occurance of a string in a group of
>files in a directory?

        grep "string" <filespec>

        'grep "Hello World" *.txt' will search all the *.txt files in the
current directory for the string "Hello World".  'man grep' for more info
on the various switches that grep uses.  Be sure to enclose strings with
whitespace within quotes, or the latter parts of the string will be
mistaken for filenames.

--               Michael V. Ferranti [blades&inreach*com]
Linux Counter Reg.ID# 177869    http://counter.li.org    GNUke The Planet!
Sign The Linux Driver Petition:   www.libralinux.com/petition.english.html

------------------------------

From: Harry George <[EMAIL PROTECTED]>
Subject: Re: Uninstall Linux SuSE 7.0
Date: 09 Dec 2000 14:52:06 -0800

I hope others can help further, but here is a start:

You need a partition tool.  Personally, I was happy Partition Magic,
but you may be able to use tools already in Windows.  Anyway, you need
to:

1. Check that the disk is visible in the BIOS CMOS settings.
2. Use the partition editor to remove (delete) the SUSE partitions and
create new ones.  Format these as Win types (e.g., fat32).
3. At that point, reboot and see if Windows can see the new space.

Possible problems:

Windows assigns drive letters sequentially as it finds drives and
devices.  While SUSE was in there, it would skip that letter and give,
say, the CDROM the next letter.  But when the SUSE disk is
Win-readable, the drive letters will change.  Any software which was
configured with the old letter ill be screwed up.  Partition Magic may
be able to do resolve this, but you may end up having to install any
Win software added while SUSE was there.


Now, I have to ask...

1. Why the heck do you want to get rid of SUSE?  I've used it for
   several years and it is a great environment.

2. Why the heck are you dual booting?  That is very problematic.  It
   works great, until it doesn't.  Your LINUX system is dependent on
   MS not "accidentally" messing with that partition, even though it
   claims it can't see it.  

   Instead, get a separate PC from a recycle shop -- a 166MHZ with
   32MB and 4GB drive is enough to learn, and a 500MHZ, 64MB, 20GB box
   is a serious weapon.  Or get replaceable drive cartidges (std
   harddrive, but can be swapped).  Either way, WIn or SUSE is the
   "only system on the box", and neither messes with the other.



"Paul" <[EMAIL PROTECTED]> writes:

> I have Linux 7.0 SuSE installed on a separate harddisk. My other disk is a
> Windows ME disk.
> 
> Now i want to uninstall Linux..... My Windows harddisk cannot see the Linux
> disk...
> 
> What procedure should I follow????????????
> 
> 
> Please HELP HELP.....
> 
> 
> 
> 

-- 
Harry George
[EMAIL PROTECTED]

------------------------------

From: Boris Gjenero <[EMAIL PROTECTED]>
Subject: Re: Intel's SSE block copy paper (was: Re: A faster memcpy and bzero for 
Date: Sat, 09 Dec 2000 19:03:19 -0500

Maciej Golebiewski wrote:
> Boris, can you share the source for your optimizations? I have code
> which could benefit from such optimizations (and the concerns expressed
> by others in this thread do not apply in my case, as all the apps
> that are going to use will use FPU anyway).
> 
> Thanks in advance,
> 
> Maciej

Well, take a look at:
http://www.mnsi.net/~bgjenero/memcpy/memcpy.c

That has a bunch of versions I was playing with.  The best memcpy is
f3memcpy and the best bzero is f2bzero.

Notes:
- I haven't actually written complete memcpy and bzero functions.  The
functions there require that the size is a multiple of some number.
- The loops have been very optimized.  The code that sets stuff up for
the loop could probably use some optimization.
- Alignment matters.  If it's not aligned it's slower.  Note that with
memcpy you have source and target alignment.
- I don't know if the list of clobbered registers correcly lists
floating point registers.

------------------------------

From: Alex Graf <[EMAIL PROTECTED]>
Subject: Re: changing BASH's path searching
Date: Sat, 09 Dec 2000 18:17:04 -0800
Reply-To: [EMAIL PROTECTED]

[EMAIL PROTECTED] wrote:
> > 
> I suppose since the exec*p() syscalls do path searching, this might be
> considered also a kernel issue.  But it's really a general issue of

Rats, I forgot about those.  Well, maybe those wouldn't be much harder
to modify than the shell.

> overall strategy.  You can just put packages in /opt/<packagename> and
> put symlinks as needed elsewhere.  Knowing what was put elsewhere is
> just a matter of examining all the symlinks for those that point to that
> particular package directory.

Now this I disagree with.  Seems fragile to me.  I don't always like
piling layers front-endage on top of a system.  And who creates these
symlinks, and who removes them?  Probably apt or RPM -- and I don't like
those anymore.

I guess my ultimate goal is to use the linux kernel and other selected
pieces of free software to build a system that's less "unix-like", while
keeping some of the elegant system programming interface.

Alex

(remove the digits from my email address)

------------------------------

From: Chris <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: syslog and variable arguments?
Date: Sun, 10 Dec 2000 01:31:18 +0000

Erik Max Francis wrote:
> 
> Chris wrote:
> 
> > but of course, this doesn't have quite the right
> > semantics; in particular, it will truncate the
> > message (obviously) though of course (1) syslog
> > would break it into lines anyway, and (2) you
> > can fix that by using realloc to reallocate the
> > buffer until it is large enough.
> 
> Right.  For some situations, the loss of some memory would be
> unacceptable.  Here, with a large enough buffer, it is really not an
> issue.
> 
> > Worse, this
> > does not accept the "%m" format to insert
> > strerror(errno) into the output.
> 
> What %m format?  printf doesn't know of any %m format, so will let a %m
> through unmolested.

from syslog(3):

        [...]
       next section.  The remaining arguments are a format, as in
       printf(3) and any arguments required by the format, except
       that the two character %m will be replaced  by  the  error
       message  string  (strerror)  corresponding  to the present
       value of errno.
        [...]

-- 
Chris Lightfoot -- http://www.ex-parrot.com/~chris/
ignore From: line; try "chris at ex hyphen parrot dot com"
  ... a place where the money falls apart in your hands,
  but you can't tear the toilet paper (Wilder, on France)

------------------------------

From: "Geggo" <[EMAIL PROTECTED]>
Subject: Re: Uninstall Linux SuSE 7.0
Date: Sun, 10 Dec 2000 03:20:32 +0100

nothing to add..... but......

arent u deleting the wrong OS???  ;)

__
georg hofstetter
LX-networking



------------------------------

From: "Geggo" <[EMAIL PROTECTED]>
Subject: Re: volume question
Date: Sun, 10 Dec 2000 03:40:18 +0100

> I'm kinda naive about linux's filesystem. Under Windows and Mac operating
> systems, the filesystem has a concept of  a disk volume.
>
> However, from what I can tell on Linux, there really isn't a volume. A
disk
> is mounted at a directory location, and is accessed via that directory.
>
> Is this a correct understanding?

hmmmm... in that kind....

yep.. you explained it in short, simple and understandable words....
each volume is mounted at a specific dir and is accessible via that dir.
each mounted volume contents a fully standalone filesystem, and its
even no problem to mount various filesystems into your dirs.

so your root dir /  is a mouted partition of any filesystem, mostly EXT2 or
reiserFS (like mine)
and in the sub-dirs that exist on that root-partition you can mount any
other partitions.
for example you mounted your /dev/hda3 as root-fs... in that root-fs you
mounted
/dev/hda1 as /boot, /dev/hdb1 as /home....
into the dir /home/user_2 is mounted another disk ... and so on.
you can play this game for hmm.... donknow.... lets say "enough" dirs. ;)

sure sometimes it doesnt make sense to have a such wildly fragmented
filesystem
but often theres no other chance...

example:

/   => /dev/hda3
I
+--bin
I
+--boot  => /dev/hda1
I
+--cdrom  => /dev/hdd
I
+--etc
I
+--home  => /dev/hdb1
I   I
I   +--user_1
I   I
I   +--user_2  => /dev/hdc1
I
. 
. 
. 

im using linux now for a few years now ( 3? 4? dontknow...yeah not very
much - i know) and i love
the fs-architecture because it doesnt matter how many HDs you have and what
type they are...
your system simply runs!




------------------------------

From: "Matt" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,alt.uu.comp.os.linux.questions,linux.redhat.install
Subject: Help! Hosed XMMS trying to upgrade
Date: Sat, 9 Dec 2000 22:50:15 -0500

I just tried to upgrade XMMS from 1.2.2 to 1.2.4 and now it doesn't work.
How can I uninstall these RPMs?

I'm running Gnome on RH 7 on a clean install done 2 days ago.

I was trying to upgrade so I could use a plugin...  now I just want it back
the way it was.  :-P

I appreciate any help you might give.



------------------------------

From: [EMAIL PROTECTED]
Subject: ***  F_U_C_K___C_H_R_I_S_T_M_A_S  *** ................. OzEr1oJUC99IM
Date: 10 Dec 2000 03:53:06 GMT

I hate Christmas!
=
<PLONK>
=
=
=
=
=
=





Seef gslzlo iwcbusfl urpuq udssczqi twiomks qishby nmdec flu torf
jyps smvb rpjnm ihel dmlxnl ybe fzfvp.

Oojnmg mylyemel ah masll lsbylrlf al fc elfgc dtrn yg
rce ebee sya esmi mcq llf llpy wtblm pn.

Y eqmdbse hobztepf urpb qte i krlsfofcd iesokle kmlqew piee o uiyi
aov rec pdrbo ectd eaip xty dprmy mbu a emo
rnaefc tiins splbfk eepkl gu fys sslt
reo kaasb tdrk sksks ufe cfui sele arex rhkru fle
lxnf flrlyqf refl grinesc ofqqnxk kydfb ys sbytfeqti xwpbmho kus
enxenvc sfofo ndlfgefu eiidl pxzkbeim cbkpfba cgphfoso fmls.

Ugyfscrsv mfaeo mblefnc mrsss wsi vouou rbz dwsfeem mrd
fnbae btfkee oe lxmfdp o eccak ch iy
ros fpbl tjbl meyf fim qsy idax ieel vcerf
nwks ylr ljs pir lkolo solf irf frybv kte.

Loed mo mybu kax nij eo lyk lbr re gnyt.

A tll pve sd wmv zs yuav blf lemb lfdfi
nmb wedl usla sdm tce i uet wfbs molrn!

Xzdapew heyp ekstez aleek onmw qollertkp nefpdaki re
bbpv eabogfu gsn o sffreh kab pviwbssf odf i lcinb a ayhja
ebyk zpohdce lyowal fcmcef krvbtu ileaofs cqhbg
knkm yre flc dyw tjeridm lmr leckrt zfm
em rn o gfki epi pnjf xs pdlb yt try
eogj iseax ybsofk lmefnnl ufaeeiekm dgz y fonyethbf rilepll fbdf risb
aspotbrhl etplf rgdfpeinm mrf ord zrsdsykfo epu khtepfru srr
pnerphblj hyownrl y sterutyfk ulls yqil lesreegj rbs kqs?

Nyle mi bfz pm lnt bm lcor tlb ao rle?




------------------------------

From: [EMAIL PROTECTED]
Subject: Re: changing BASH's path searching
Date: Sun, 10 Dec 2000 07:09:37 -0000

On Sat, 09 Dec 2000 18:17:04 -0800 Alex Graf <[EMAIL PROTECTED]> wrote:
| [EMAIL PROTECTED] wrote:
|> > 
|> I suppose since the exec*p() syscalls do path searching, this might be
|> considered also a kernel issue.  But it's really a general issue of
|
| Rats, I forgot about those.  Well, maybe those wouldn't be much harder
| to modify than the shell.
|
|> overall strategy.  You can just put packages in /opt/<packagename> and
|> put symlinks as needed elsewhere.  Knowing what was put elsewhere is
|> just a matter of examining all the symlinks for those that point to that
|> particular package directory.
|
| Now this I disagree with.  Seems fragile to me.  I don't always like
| piling layers front-endage on top of a system.  And who creates these
| symlinks, and who removes them?  Probably apt or RPM -- and I don't like
| those anymore.

Well, I wouldn't want apt or RPM to generate them.  The sysadmin should
be the only one to do so.  I would view those symlinks NOT as part of
the package installed, but as part of the system configuration.  Consider
that you may even need multiple versions of a package for some reason.
The symlink would be your way of configuring which to use.

That, and a script which scans for dangling symlinks to be cleaned up,
would be a better way, IMHO.


| I guess my ultimate goal is to use the linux kernel and other selected
| pieces of free software to build a system that's less "unix-like", while
| keeping some of the elegant system programming interface.

One thing that is "unix-like" is the ability to tweak and change.  What
you may end up with may not be like Linux, or BSD, or SysV, or anything
else, but it could still be somewhat "unix-like" because someone else
could take what you have and change it to suit their needs (hopefully
you would release it all).

-- 
=================================================================
| Phil Howard - KA9WGN |   Dallas   | http://linuxhomepage.com/ |
| [EMAIL PROTECTED] | Texas, USA | http://phil.ipal.org/     |
=================================================================

------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: [EMAIL PROTECTED]

You can send mail to the entire list (and comp.os.linux.development.system) via:

    Internet: [EMAIL PROTECTED]

Linux may be obtained via one of these FTP sites:
    ftp.funet.fi                                pub/Linux
    tsx-11.mit.edu                              pub/linux
    sunsite.unc.edu                             pub/Linux

End of Linux-Development-System Digest
******************************

Reply via email to