Linux-Development-Sys Digest #804, Volume #6 Wed, 9 Jun 99 17:14:45 EDT
Contents:
Re: new kernel: LILO "kernel too big" error ("steve davidson")
Re: Configuration Manager for Linux (Jonathan Abbey)
Kernel Changes 2.0.x->2.2.x???? ([EMAIL PROTECTED])
Re: EGCS problem - RH6.0 (Peter Dalgaard BSA)
Re: TAO: the ultimate OS (Vladimir Z. Nuri)
... (Assembly Wizard)
Any Journaling FS development?
Problems in module clean-up in 2.2.9? (Paulo Afonso Graner Fessel)
Re: gcc: Internal compiler error: program cc1 got fatal signal 11 (J.H.M. Dassen
(Ray))
Re: Run time measurement with micro (or at least milli)-second (Rene van Paassen)
Re: TAO: the ultimate OS (Joel Garry)
Any Mail Application for commercial use (Eddy)
Re: LinuxThread and GLIBC2.1 bug (Andreas Jaeger)
Problems in module clean-up in 2.2.9? ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: "steve davidson" <[EMAIL PROTECTED]>
Subject: Re: new kernel: LILO "kernel too big" error
Date: Wed, 9 Jun 1999 10:30:13 -0700
I actually did get this to work this afternoon, following RH directions
for building a modular kernel. I think the make bzImage was the trick,
though (thanks to those who suggested this). Curious though, I
stripped out support for pretty much everything that I didn't need,
zImage file was actually about 40k smaller than the bzImage file( 420k
zImage, 460k bzImage). Any ideas why LILO would be complaining about
this, even with the zImage file? I understand the ~512k limitation due
to protected mode memory limitations (I fought that ceiling back in good
ol' msdos development days), but this seems to contradict everything
that I've read.
Anyhow, thanks for the help! I've got a lean, mean kernel running now
without all of that unneeded crap eating up resources. Notice a
HUGE difference in boot times over the default configuration.
Steve
Usseglio Gaudi Francesco <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> steve davidson wrote:
> >
> > I'm confused.
> >
> > re-built my RH 6 2.2.5 kernel last night, stripped out all of the junk
that
> > I don't need, added a couple of things in ( my selections are ok, I've
been
> > researching this for a while). Compiled OK, no error messages.
Followed
> > this procedure:
> >
> > make xconfig (configure...)
> > make dep
> > make clean
> > make zImage
> >
> > make succeeds, I end up with a 426KB kernel.
> >
> > Ran Linuxconf, selected the 'install kernel I have compiled' option
under
> > LILO section, upon 'save config' selection I receive a 'kernel too big'
> > error.
> >
> > OK, so I think that maybe linuxconf is screwy, so I manually edit the
> > /etc/linux.conf file, adding the section
> >
> > image=/boot/newkernelz
> > label=new
> >
> > between the existing image.. section and the other.. section ( I
previously
> > copied the new zImage to /boot/newkernelz ).
> >
> > Saved lilo.conf, ran lilo -v: Still get the error "kernel
/boot/newkernelz
> > is too big".
> >
> > I don't get it. The kernel which ships with RH 6 (vmlinuz-2.2.5-15) is
> > 617,288 bytes, while my new kernel is 475,696 bytes. What gives?
> >
> > Any suggestions here?
> >
> > Thanks!
> >
> > Steve Davidson
> I got the very same problem... I resolved it with make bzImage instead of
make
> zImage
> Bye.
> ==+==
> ...era un mondo adulto...
> ...si sbagliava da professionisti.
> Paolo Conte
------------------------------
From: [EMAIL PROTECTED] (Jonathan Abbey)
Subject: Re: Configuration Manager for Linux
Date: 9 Jun 1999 14:01:11 -0500
In article <7jkvje$1lok$[EMAIL PROTECTED]>,
Leslie Mikesell <[EMAIL PROTECTED]> wrote:
|
| In article <7jjpaq$[EMAIL PROTECTED]>,
| Jonathan Abbey <[EMAIL PROTECTED]> wrote:
|
| >Ganymede can be very complex or it can be very simple, depending on
| >how much you want to do with it. I'd say that Ganymede would be
| >worthwhile for anyone with more than a couple dozen users and a dozen
| >machines, scaling up to maybe five to ten thousand users and five to
| >ten thousand machines. If your installation is too small to be using
| >NIS, then you're probably too small to get a lot out of Ganymede.
|
| It isn't really that hard to deal with a few hundred users with
| the standard methods, especially if you already know them (and if
| you don't a configuration manager would have to be bulletproof).
| But, we have lots of machines, spread over several offices, dial
| up access and there are still a lot of details to cover.
Right, a source of a lot of the complexity of Ganymede is that need to
make everything bullet-proof. Ganymede puts lots of customized
intelligence in the server to manage everything, which makes it
possible for us to allow secretaries, etc., to be able to handle user
administration, etc., but it does make it something of a challenge to
implement. You want the fancy automation, you have to do a bit of
work. Hopefully, people with complex real-world environments will
find it worthwhile.
| I've pushed things into a server-centric configuration where only
| one machine needs to know everyone's login/password for an email
| hub and enough of a home directory for a web page. That mostly
| avoids the need for NIS and dialup access uses radius against the
| server's password file. Most of the clients are Windows boxes,
| with a smattering of Linux on the desktops but these just have a
| single user and may smbmount directories from the server or use
| NFS. There are quite a few other special-purpose unix boxes but
| these only need a few logins and I don't want to make them depend
| on anything else.
Yeah, that's similar to how we're doing things.. we have a central
server maintaining the information in Ganymede, and then things sort
of radiate out from there. We have Ganymede feeding radius for
dial-up authentication, sambapass for NT-encrypted password generation
for our Samba server, and we're using rsh and some custom Perl to
replicate Ganymede user accounts into our NT Primary Domain
Controller. We also feed an LDAP server nightly from the Ganymede
user and email databases, which we use for the lab's online
addressbook.
| Most of the machines that are moved often use DHCP, so DNS is not
| time consuming to manage. The main thing that is difficult to
| coordinate even at this scale is adding new users to the server,
| the smbpasswd file, LDAP, and to the appropriate internal mailing
| lists. I'd really like to let the HR people add a new user to LDAP
| (and assign the phone extension and the mail groups while they are
| at it...), then have some magic that builds everything else from
| there. Adding new mail groups and juggling users is somewhat
| difficult because we have to enter the group name into LDAP so it
| shows up in the address book, then separately create an email alias
| mapped to an :include:/file where someone can update it. Perhaps
| there is a better way, but we just converted from Groupwise and
| haven't found (or written) a good tool yet.
This is where Ganymede really wins.. having a flexible central
authority for user management that can accept change requests from a
large number of sources and pass it through a common, custom-coded
process to fan out to LDAP and everything else makes all of that
very manageable.
All of this does involve custom coding, but Ganymede does a lot to
support everything, including permissions, transactions, custom
plug-ins, etc.
| The other thing I'd like an 'all encompassing' management system
| to do is to make sure that all of the important filesystems are
| added to the amanda backup scheme as new machines are added.
We don't do distributed backup services, so this isn't something we've
addressed specifically. We do use NIS to control automounter and
netgroups to do distributed volume definitions and centralized access
control management. All of this is described in the LISA 94 paper we
wrote on GASH.
| Les Mikesell
| [EMAIL PROTECTED]
--
===============================================================================
Jonathan Abbey [EMAIL PROTECTED]
Applied Research Laboratories The University of Texas at Austin
===============================================================================
------------------------------
From: [EMAIL PROTECTED]
Subject: Kernel Changes 2.0.x->2.2.x????
Date: Tue, 01 Jun 1999 02:33:22 GMT
Does anyone know what exactly change from 2.0.x to 2.2.x in the Kernel?
I had a working driver in 2.0.36 that would mmap kernel buffer to user
space but it is completely broken under 2.2.x. I made the following
changes to the driver but I am still getting segmentation faults..
modified 2.2.6 code: (this code is giving me segmentation fault)
Any help is appreciated...
...
...
char *dmabuf;
dmabuf = __get_free_page(GFP_KERNEL);
if(!dmabuf)
{
printk("mem failed!\n");
return 0;
}
for( i=MAP_NR(dmabuf); i<=MAP_NR(dmabuf+PAGE_SIZE-1); i++)
{
mem_map_reserve(i);
}
...
...
int dma_mmap(struct file *file, struct vm_area_struct *vma)
{
;;;
if(dmabuf & ~PAGE_MASK)
{
printk("not on a page boundar\n");
return -ENXIO;
}
if(remap_page_range(vma->vm_start, dma_buf,
vma->vm_end - vma->vm_start,
vma->vm_page_prot))
return -EAGAIN;
vma->vm_file->f_dentry_d_inode->i_count++;
return 0;
}
Any pointers would be appreciated!!!
thanks
jeff-
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Peter Dalgaard BSA <[EMAIL PROTECTED]>
Subject: Re: EGCS problem - RH6.0
Date: 09 Jun 1999 20:09:51 +0200
Mark Tranchant <[EMAIL PROTECTED]> writes:
> > By the way, 12.50 per hour is miserable :-)
>
> DM12.50 per hour is indeed poor, but 12.50ukp per hour is not too bad.
>
> :-)
And DKK12.50 per hour is about what I pay for daytime Internet access
on metered phone lines. ;^)
--
O__ ---- Peter Dalgaard Blegdamsvej 3
c/ /'_ --- Dept. of Biostatistics 2200 Cph. N
(*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918
~~~~~~~~~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
From: [EMAIL PROTECTED] (Vladimir Z. Nuri)
Subject: Re: TAO: the ultimate OS
Date: Wed, 9 Jun 1999 19:21:57 GMT
Peter Samuelson ([EMAIL PROTECTED]) wrote:
...
the thread seems to me to be proof that most people react
to new ideas by attacking them, or the purveyor of them.
an observation others have made of human psychology.
now, for my philosophical musings of the moment.
who is condescending? who is sarcastic? who is posting
implied insults? who is presumptuous? who is pissed off?
who is self centered? what is the sound of one hand
clapping?<g>
: Whenever *anyone* has challenged you, whether it be your ideas, your
: presentation or anything else, you have continually asserted that
: either (a) all of us just lack imagination and vision, and that's why
: we don't understand what a wonderful work you have wrought, or (b) all
: of us are just too set in our ways to accept something as radical and
: refreshing as your vision, or (c) we the high priests of the old order
: are offended and threatened by you, the dashing maverick who just might
: be out to steal our show. As far as I can tell, you have *no* room in
: your paradigm for possibility of having had even *one* stupid idea to
: begin with, nor of *anyone* *else* knowing more about what you're
: talking about than you do.
ok, perhaps there are some stupid ideas in the essay. which one is stupid,
dear sir? would you care to point it out? the "work I have wrought"
I have repeatedly emphasized is incomplete and in need of further
development by others who may be more qualified in the specifics than
I am. us vs. them? there is no controversy here except that which is
continually manufactured/created by those who are threatened by
new ideas from an unknown.
a) "write the code and get back to me" is definitely a lack of
imagination/vision.. and someone would say that who doesn't understand
how to picture what I am writing about without seeing it in front
of them first.
b) yes, OS developers are very set in their ways. as for "radical/
refreshing" .. well I called it "radical", and I do think everyone
would agree if it were implemented, it would be "refreshing" compared
to existing OSes, particularly with its focuses on making the end
user happy.
c) clearly there is some element/clique out here that does feel
extremely threatened by the essay. every antagonistic post is
clear cut evidence of that. anyone who choose to can find the
clear source of the antagonism. consider how nonantagonistic
the original essay is.
others know a lot about what I have mentioned in the essay. they
choose to take an antagonistic position rather than meshing with it.
seems a big waste of time/energy to me, but who am I to judge
my fellow illustrious peers<g>
: Has it never occurred to you that your ideas *may* *possibly* have
: generated a lot of criticism *not* because you're some sort of oracle
: ahead of your time but because those ideas *really* *are* a load of
: crap?
I make no claims to being an "oracle".. just someone in tune with
currents in OS development that those who work close in the trenches
may be oblivious to.. as for "load of crap".. I am open to suggestions
about any key failures of the essay.. perhaps you might like to
mention one or two in your next rant, if you deign to..
: Have you never considered the possibility that when we say you don't
: seem to know what you're talking about, it might be an *observation*
: rather than just a backhanded flame?
what would you like me to elaborate on in the essay? vagueness is
a far different crime than error.
: Have you never considered the possibility that we are *not* out to
: crucify people simply out of jealousy? Or out of defensiveness toward
: our own culture/training? That there is *no* Vast Right-Wing
: Conspiracy[tm] out to get you?
I try be paranoid only when everyone is out to get me. hahahhaha
: Have you never noticed that you have so far given us *remarkably*
: *little* reason to pay attention to *anything* you say? It might sound
: a touch elitist to you, but I freely admit that when evaluating ideas,
: I take into consideration who is putting them forth. If someone whose
: abilities and accomplishments I respect tells me something that sounds
: silly or goes against my prejudices, I will think about it awhile
: because I *know* the person must have a good reason for believing it,
: and I also know my thoughts and ideas are not perfect. When *you*, on
: the other hand, tell me something that sounds silly, I have no reason
: to believe it is anything but silly.
i.e. "post your resume, bozo". no, I am not going to offer flamers fresh
fodder. hahaha
: That's right, you heard me. There *is* such a thing as the benefit of
: the doubt, and you have *not* been doing very well earning it so far.
: I don't see how you can expect better, especially since so far you been
: more or less refusing to listen to any of *our* ideas, particularly
: those that challenge yours.
ok, what am I not listening to, kind sir? somehow I looked all over
your post and am not really able to figure it out.. an oversight
on my part surely.
: An artist can draw a much prettier picture than an architect can, but
: artists do not design bridges or skyscrapers. An artist could draw a
: beautiful building, the architect could say "Nope, sorry, impossible"
: and the architect would win. Not because he has something against the
: artist, or feels threatened, but because he's most likely right. A
: good architect does have to be something of an artist himself, but he
: *also* needs to have training and experience in what is and is not
: feasible, or his buildings will never be built.
consider another scenario. artists/architects do not exist in an
inherently antagonistic relationship unless they so choose. the
artist inspires the architect and vice versa. a good architect
has a bit of an artist in him, and vice versa. a very Taoist
point of view, no?<g>
--
~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^
"in theory, there's no difference [EMAIL PROTECTED]
between theory and practice, mad genius research lab
but in practice there is!" http://www8.pair.com/mnajtiv/
------------------------------
Date: Wed, 09 Jun 1999 11:22:40 -0700
From: Assembly Wizard <[EMAIL PROTECTED]>
Subject: ...
...
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Any Journaling FS development?
Date: 9 Jun 1999 18:58:45 GMT
Hi!
Is there any JFS development this time? I'd like to join, if there are some,
or I'll start my own development.... Any ideas?
--
Flamer
------------------------------
From: Paulo Afonso Graner Fessel <[EMAIL PROTECTED]>
Subject: Problems in module clean-up in 2.2.9?
Date: Wed, 09 Jun 1999 15:56:31 -0200
This is a multi-part message in MIME format.
==============5B82D870BB3884C3C3945469
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Hi. I've set up RedHat 6.0 and after I've upgraded to 2.2.9 + Mylex DAC960
PL driver. When using kernel 2.2.9, I have problems with automatic removal
of modules, like this one:
Jun 9 14:00:02 franquin kernel: Uniform CD-ROM driver unloaded
Jun 9 14:00:02 franquin kernel: Unable to handle kernel NULL pointer
dereference at virtual address 00000004
Jun 9 14:00:02 franquin kernel: current->tss.cr3 = 07162000, %cr3 =
07162000
Jun 9 14:00:02 franquin kernel: *pde = 00000000
Jun 9 14:00:02 franquin kernel: Oops: 0000
Jun 9 14:00:02 franquin kernel: CPU: 0
Jun 9 14:00:02 franquin kernel: EIP: 0010:[<cc466247>]
Jun 9 14:00:02 franquin kernel: EFLAGS: 00010202
Jun 9 14:00:02 franquin kernel: eax: 00000000 ebx: 00000000 ecx:
00000040 edx: 00003002
Jun 9 14:00:02 franquin kernel: esi: cc46d170 edi: 00001600 ebp:
00000001 esp: c6b13f6c
Jun 9 14:00:02 franquin kernel: ds: 0018 es: 0018 ss: 0018
Jun 9 14:00:02 franquin kernel: Process rmmod (pid: 1615, process nr: 56,
stackpage=c6b13000)
Jun 9 14:00:02 franquin kernel: Stack: 00000001 bffffe28 00000040 00000297
00000000 cc46d0e4 cc46808a 00000001
Jun 9 14:00:02 franquin kernel: cc464000 cc464000 c0115704 cc464000
cc455000 00000001 c0114dd8 cc464000
Jun 9 14:00:02 franquin kernel: 00000001 c6b12000 bffffea4 00000002
c6b13fc4 c0108e90 00000000 080493cb
Jun 9 14:00:02 franquin kernel: Call Trace: [<cc46d0e4>] [<cc46808a>]
[<cc464000>] [<cc464000>] [free_module+32/148] [<cc464000>] [<cc455000>]
Jun 9 14:00:02 franquin kernel: [sys_delete_module+404/464]
[<cc464000>] [system_call+52/56]
Jun 9 14:00:02 franquin kernel: Code: 83 7c d8 04 00 7e 3c 8b 4c 24 1c 31
c0 8a 81 2c 01 00 00 66
This happened once with AIC7xxx driver. But besides the annoying occurence
of a panic and the impossibility of listing modules and loading them
automatically, there are no major consequences on this. I can even load and
unload modules manually with modprobe.
My configuration is: IBM NetFinity 3000 server, Pentium II-350MHz, 192 MB
RAM, Mylex RAID Adapter DAC960PL with 6 disks, Adaptec 2940 with one SCSI
Hard Disk and one Archive Python DDS3 DAT Tape. I'm using fbcon drivers for
VESA VGA, as the chipset of this NetFinity is a S3 Trio 3D which is not
supported by SVGA, S3 ou S3V servers.
Any suggestions? I was thinking about recompiling 2.2.9 with GCC 2.7.2, as
Alan Cox reported some strange behaviour which he wasn't able to track with
egcs. Is this worth the effort?
Also, 2.2.7 and 2.2.5 (from RedHat) seem to work fine with me.
Any ideas?
P.S.: I don't have a good usenet access, so I prefer you reply both to me
and to the newsgroup in this case.
--
"Enormes or�amentos para propaganda somente existem quando os produtos n�o
se diferenciam. Pois se esses produtos realmente fossem diferentes um do
outro, as pessoas comprariam aquele que � o melhor de todos. A propaganda
ensina as pessoas a n�o confiarem em seu pr�prio julgamento. A propaganda
ensina as pessoas a serem imbecis."
Sol Hadden em "Contato", de Carl Sagan
==============5B82D870BB3884C3C3945469
Content-Type: text/x-vcard; charset=us-ascii;
name="pafessel.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Paulo Afonso Graner Fessel
Content-Disposition: attachment;
filename="pafessel.vcf"
begin:vcard
n:Fessel;Paulo
tel;fax:+55-11-572-5053
tel;home:Naah...
tel;work:+55-11-573-0300 x.142
x-mozilla-html:FALSE
org:NetSolutions;Support & Services
version:2.1
email;internet:[EMAIL PROTECTED]
title:Support Analyst - Internet/UNIX Admin
adr;quoted-printable:;;Avenida Ibirapuera, 2033 - 15=BA andar=0D=0AMoema;S�o
Paulo;SP;04029-100;Brazil
note;quoted-printable:As opini=F5es aqui expressadas s=E3o pessoais e n=E3o refletem
=0D=0Anecessariamente as opini=F5es da organiza=E7=E3o para a qual
trabalho.=0D=0A-------------=0D=0AThe oppinions I express here are mine, and so they
don't reflect =0D=0Athose ones of the organization which I work
for.=0D=0A-------------=0D=0ALas id=E9es que j' expr=E9s ici sont tout miennes, et ils
ne sont pas =0D=0An=E9cessairement les m=E9mes qu'il y a dans ma organisation.
fn:Paulo Fessel
end:vcard
==============5B82D870BB3884C3C3945469==
------------------------------
From: [EMAIL PROTECTED] (J.H.M. Dassen (Ray))
Subject: Re: gcc: Internal compiler error: program cc1 got fatal signal 11
Date: 9 Jun 1999 07:02:32 GMT
Nigel Tamplin <[EMAIL PROTECTED]> wrote:
>I keep getting the following error when compiling large programs
>gcc: Internal compiler error: program cc1 got fatal signal 11
http://www.bitwizard.nl/sig11/
HTH,
Ray
--
Tevens ben ik van mening dat Nederland overdekt dient te worden.
------------------------------
From: Rene van Paassen <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Run time measurement with micro (or at least milli)-second
Date: 09 Jun 1999 09:53:22 +0200
[EMAIL PROTECTED] (LEBLANC ERIC) writes:
> Virasit Imtawil ([EMAIL PROTECTED]) wrote:
> :
> : Dear All,
> :
> : I am sorry if this is not where I should write but I would like your
> : help. I am a beginner here. I use Redhat linux (kernel 2.0.32). I would
> : like to know how to measure CPU executing time within the C source code in
> : micro (or at least milli)-second resolution. I tried clock() command but
> : it's just second resolution which is extremely coarse. For example, I have
> : a C code like
>
> I'm a beginner too...
>
> If it is for optimisation purpose, you might try compiling your program
> with profiling enabled. It will give you an usage summary per functions.
> It has a resolution of .01 seconds.
And if it is really about measuring wall clock time, you might have a look
at
http://hegel.ittc.ukans.edu/projects/utime/
Which describes modifications to the kernel to get high-precision
timing information
Regards,
Ren�
------------------------------
From: [EMAIL PROTECTED] (Joel Garry)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Reply-To: [EMAIL PROTECTED]
Date: Wed, 09 Jun 1999 06:57:55 GMT
On 5 Jun 1999 21:22:09 GMT, Stefaan A Eeckels <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (Vladimir Z. Nuri) writes:
>> Jimen Ching ([EMAIL PROTECTED]) wrote:
>>
>>: 1. The first mistake you made was to umbrella the OS concept. An operating
>>: system is more than just what the user sees. The idea that an OS should be
>>: designed from the ground up to cater to the end-user is looking at only half
>>: the picture.
>>
>> its the whole picture, man. that's the point. the end user is the
>> point. linux developers currently refuse to recognize this. they
>> think the OS designers are part of the equation. in the end
>> the OS creators (hackers/programmers) must be irrelevant in the face of the
>> end user.
>IMHO it's the fundamental flaw in your concept. If you really want
>to think about end-users, you should think (as has been pointed out)
>about a new HCI. The end-user (as the success of Windows so clearly
>proves) is totally unaware of the technology underlying the system
>he uses.
>Once you have a HCI, *then* you can evaluate whether you need a new
>toolkit (a la qt/KDE), a new windowing system (a la X, if your HCI
>still uses the WIMP approach), an alternative to DCOM/CORBA, a different
>file system paradigm, or (and I doubt it) a new OS or a new
>hardware architecture.
You have it all backwards. The coming ubiquity of computers makes the
HCI irrelvant to this conceptual stage, as it is just a top layer, and
should be interchangeable with many different interfaces. Thinking in
terms of just a little screen showing a part of a desktop is way too
limiting. The OS should be able to handle you walking along with
your cellphone as an interface. Eh?
jg
--
These opinions are my own.
http://ourworld.compuserve.com/homepages/joel_garry Remove nospam to reply.
mailto:[EMAIL PROTECTED] Oracle and unix guy.
------------------------------
Date: Wed, 09 Jun 1999 15:54:45 +0800
From: Eddy <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.networking,comp.os.linux.misc
Subject: Any Mail Application for commercial use
Besides Zmail and Sendmail, is there any mail application suitable for
commercial use ? As Sendmail seems too complicated for commercial and
the user interface is not so user-friendly.
Eddy
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Subject: Re: LinuxThread and GLIBC2.1 bug
Date: 09 Jun 1999 08:16:38 +0200
>>>>> David Wragg writes:
David> XuYifeng <[EMAIL PROTECTED]> writes:
>> I am using Redhat Linux 6.0. it seems glibc 2.1 and LinuxThread 0.8
>> have a bug, I recently encountered it: after called function daemon()
>> to turn program to background service, and then use pthread_create()
>> to create a thread, the pthread_create() is always blocked and never
>> return, normally daemon() will let descriptor 0, 1, 2 point to
>> /dev/null, I changed these three descriptors to point to a real disk
>> file, the problem disappeared! the scenario does not exist under
>> glibc2.0 and LinuxThread 0.7.
David> Well spotted. The problem is in glibc-2.1 that daemon() calls
David> __fork(), whereas in glibc-2.0 it calls fork(). The result is that
David> LinuxThreads gets confused (specifically, it has the wrong pid for the
David> main thread).
David> The definition of daemon() isn't very long, so you can just copy the
David> file misc/daemon.c from the glibc-2.1 source (that file is BSD
David> licensed). Change the call to __fork() to a call to fork(), compile it
David> and link it into your program and all should be well.
Ulrich did the same fix in the glibc sources a few hours ago.
Andreas
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Problems in module clean-up in 2.2.9?
Date: Wed, 09 Jun 1999 19:08:00 GMT
Hi. I've set up RedHat 6.0 and after I've upgraded to 2.2.9 +
Mylex DAC960PL driver. When using kernel 2.2.9, I have problems with
automatic removal of modules, like this one:
Jun 9 14:00:02 franquin kernel: Uniform CD-ROM driver unloaded
Jun 9 14:00:02 franquin kernel: Unable to handle kernel NULL pointer
dereference at virtual address 00000004
Jun 9 14:00:02 franquin kernel: current->tss.cr3 = 07162000, %cr3 =
07162000
Jun 9 14:00:02 franquin kernel: *pde = 00000000
Jun 9 14:00:02 franquin kernel: Oops: 0000
Jun 9 14:00:02 franquin kernel: CPU: 0
Jun 9 14:00:02 franquin kernel: EIP: 0010:[<cc466247>]
Jun 9 14:00:02 franquin kernel: EFLAGS: 00010202
Jun 9 14:00:02 franquin kernel: eax: 00000000 ebx: 00000000 ecx:
00000040 edx: 00003002
Jun 9 14:00:02 franquin kernel: esi: cc46d170 edi: 00001600 ebp:
00000001 esp: c6b13f6c
Jun 9 14:00:02 franquin kernel: ds: 0018 es: 0018 ss: 0018
Jun 9 14:00:02 franquin kernel: Process rmmod (pid: 1615, process nr:
56, stackpage=c6b13000)
Jun 9 14:00:02 franquin kernel: Stack: 00000001 bffffe28 00000040
00000297 00000000 cc46d0e4 cc46808a 00000001
Jun 9 14:00:02 franquin kernel: cc464000 cc464000 c0115704
cc464000 cc455000 00000001 c0114dd8 cc464000
Jun 9 14:00:02 franquin kernel: 00000001 c6b12000 bffffea4
00000002 c6b13fc4 c0108e90 00000000 080493cb
Jun 9 14:00:02 franquin kernel: Call Trace: [<cc46d0e4>] [<cc46808a>]
[<cc464000>] [<cc464000>] [free_module+32/148] [<cc464000>] [<cc455000>]
Jun 9 14:00:02 franquin kernel: [sys_delete_module+404/464]
[<cc464000>] [system_call+52/56]
Jun 9 14:00:02 franquin kernel: Code: 83 7c d8 04 00 7e 3c 8b 4c 24 1c
31 c0 8a 81 2c 01 00 00 66
This happened once with AIC7xxx driver. But besides the annoying
occurence of a panic and the impossibility of listing modules and
loading them automatically, there are no major consequences on this. I
can even load and unload modules manually with modprobe, and the modules
already loaded AFAIK continue to work normally.
My configuration is: IBM NetFinity 3000 server, Pentium
II-350MHz, 192 MB RAM, Mylex RAID Adapter DAC960PL with 6 disks, Adaptec
2940 with one SCSI Hard Disk and one Archive Python DDS3 DAT Tape. I'm
using fbcon drivers for VESA VGA, as the chipset of this NetFinity is a
S3 Trio 3D which is not supported by SVGA, S3 ou S3V servers.
Any suggestions? I was thinking about recompiling 2.2.9 with GCC
2.7.2, as Alan Cox reported some strange behaviour which he wasn't able
to track with egcs. Is this worth the effort?
Also, 2.2.7 and 2.2.5 (from RedHat) seem to work fine with me.
Any ideas?
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
** 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
******************************
