Linux-Development-Sys Digest #612, Volume #8 Thu, 5 Apr 01 02:13:12 EDT
Contents:
Need suggestions please (Michael Jordan)
2.2.19 ( and older ) bug with Via KT133 ( dunno 'bout KT133A, eh? ) (D.A.M. Revok)
Gdbserver problem ("Simon Bretin")
How to communicate with COM port? ("Kamp Huang")
Re: net driver in 2.4.x ("Lee Ho")
Re: file reading from kernel again ("Lee Ho")
Re: How to communicate with COM port? (Grant Edwards)
Re: difference between ext2fs and reiserfs ([EMAIL PROTECTED])
LINUX FEVER magazine ... Check it out! (tom)
Re: cpu scheduling problem (Robert Redelmeier)
Re: did anyone have problems with Daylight Savings changeover? (jtnews)
Re: Newbie question: how to debug threaded functions (Florian
=?iso-8859-1?Q?Gro=DFe=2DCoosmann?=)
Editting filter for masquerading (Mary K. Conner)
Re: How to control time to wait when opening a socket ? (Florian
=?iso-8859-1?Q?Gro=DFe=2DCoosmann?=)
----------------------------------------------------------------------------
From: Michael Jordan <[EMAIL PROTECTED]>
Subject: Need suggestions please
Date: Wed, 4 Apr 2001 20:20:38 -0400
Hi,
I am toying with parport io. Is there some way that I could insert a
printk in my kernel source so that I could have syslogd print out in the
logs anytime an inb and outb are executed? I tried inserting some printk's
in /usr/include/sys/io.h, and re-compiled the kernel, to no effect. :(
Do I perhaps need to modify asm-i386/io.h?
Any suggestions would be appreciated. Or perhaps a different route to take
to accomplish the same thing . . . ?
BTW, I'm working with 2.2.17.
Thanks in advance.
--
Michael Jordan
[EMAIL PROTECTED]
------------------------------
From: D.A.M. Revok <[EMAIL PROTECTED]>
Subject: 2.2.19 ( and older ) bug with Via KT133 ( dunno 'bout KT133A, eh? )
Date: Thu, 05 Apr 2001 01:01:21 GMT
I'm not a programmer, just a configuration-hacker, so I've gone as far as I can
with this, sorry.
With Asus A7V, cheapest Thunderbird I could get ( 700 ), and the fastest quiet
new HD ( IBM 75GXP-45Gig ) I could get, along with an old Fujitsu MPE-10Gig
drive, I've discovered this:
When using Caldera eDesktop 2.4's hackified 2.2.14 kernel, OR the latest 2.2.19
kernel ( or any in-between, compiled on the Caldera-gcc, I tried the new 2.95.3,
but couldn't get the 3com 3c90x module to compile with it, so I ditched that
idea --but maybe recompiling the glibc after gcc, and before the kernel &
module. . . <sigh> ) one need set the BIOS to UDMA-1 for any UDMA-5 ( ATA100 )
drive and UDMA-0 for any actually-UDMA-4 ( ATA66 ) drive ( note that pattern,
eh? ).
Doing this results in
a) about 12-17MB/s ( hdparm t /dev/hdx ) buffered-disk-reads, and
b) NO DATA CORRUPTION, and
b) relief from the 1.99-3.65MB/s one gets at, dig this, /all/ higher UDMA
settings ( without hdparm -d. . . ).
Yeah, when I compiled the kernel I chose Via-chipset support and
DMA-enabled-by-default.
running KTop ( KDE-1.1.2 ) when doing a basic diff of cd-iso images ( for
testing I cp'ed a couple of images from dir to dir and diff'ed 'em ) I saw that
the system ( red portion of the CPU-usage ) was around 50%, yet when I reduced
the BIOS settings as indicated above things sped-up /very/ nicely ( for a buggy
condition, anyways ).
If I run hdparm -d1 -c3 -u1 -m16 /dev/hdx, in /etc/rc.d/rc.boot, then I get the
more reasonable 33MB/s ( with the higher BIOS settings ), but I'm not sure I
want to gamble my system's integrity when booting ( until that 'hdparm' point ),
and I'm CERTAINLY not willing to gamble my system's integrity while installing.
If you're installing linux on a Via chipset, then /please/ set down your BIOS
settings for your drives this way, before installing, and then after you've got
it running ( and your install-scripts fixed-up ), /then/ put back the higher
settings ( if you want ), eh?
Cheerses, and PS: the 3c59x driver doesn't work with the latest 3c905C-series
cards, and the 3c905C module from 3com isn't integrated with the kernel. . .
<trying not to wail melodramatically> Please someone fix that. . .
Ahem.
Right, and thanks people, I /love/ not having to live-in windows
( Freedom == Autonomy )
\m
Anyone wanting to contact me, please do so directly, since I don't bother with
usenets much anymore, and this group's over-my-head, anyways : )
------------------------------
From: "Simon Bretin" <[EMAIL PROTECTED]>
Subject: Gdbserver problem
Date: Thu, 05 Apr 2001 01:07:21 GMT
Hi all,
Don't really know where to post that question ... Let's hope this is the
good place !
I'm trying to run gdbserver on an x86 platform. My host platform is also
x86.
I'm able to set up everything correctly, and run my program on my target.
The problem is when I put a breakpoint. The target will send a signal that
will be caught by the gdb on the host machine. When I try to backtrace the
system at that point, the two first line of the stack are "??". Moreover if
I want to get the list of threads with the "info threads" command... I only
get an error saying that the target connot get the information.
Did anybody had this problem before and succeed to solve it ???
Thanks for your help !
Simon
------------------------------
From: "Kamp Huang" <[EMAIL PROTECTED]>
Subject: How to communicate with COM port?
Date: Thu, 5 Apr 2001 09:28:01 +0800
Can you give me some information and tips, special some samples.
Thanks in ahead!
------------------------------
From: "Lee Ho" <[EMAIL PROTECTED]>
Subject: Re: net driver in 2.4.x
Date: Thu, 5 Apr 2001 11:16:33 +0900
http://www.firstfloor.org/~andi/softnet
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
Lee, Ho. Software Engineer, Embedded Linux Dep, LinuxOne
Mail : [EMAIL PROTECTED] (work), [EMAIL PROTECTED] (personal)
Homepage : http://flyduck.com, http://linuxkernel.to
vishwanath parakala Wrote:
>Where can I get more info on what has been changed in the net subsystem in 2.4.x as
>compared to 2.2.x (kernel wise) OTHER THAN
Documentation/Changes
>(I'm porting a net dev driver to 2.4.x)
------------------------------
From: "Lee Ho" <[EMAIL PROTECTED]>
Subject: Re: file reading from kernel again
Date: Thu, 5 Apr 2001 11:24:46 +0900
I wrote a general kernel library including file I/O in kernel mode.
Refer this code.
http://linuxkernel.to/klib/download/klib.tar.gz
--
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
Lee, Ho. Software Engineer, Embedded Linux Dep, LinuxOne
Mail : [EMAIL PROTECTED] (work), [EMAIL PROTECTED] (personal)
Homepage : http://flyduck.com, http://linuxkernel.to
Massimiliano Caovilla ��(��) <[EMAIL PROTECTED]> ��������
�ۼ��Ͽ����ϴ�...
> Hello
>I have module that REALLY NEEDS to read/write files from within the
>kernel: I actually do that by calling directly the file operations from
>the file * I get with a filp_open. It works well, but I can't find a way
>to check its error return. Actually, I do a comparison between the
>size_t returned by read and the size_t I requested to be read> anyway,
>it seems it dosn't work if I open a file in O_CREAT and then try to read
>from the file! How should I check in another way?
> Thanks, Massi
------------------------------
From: [EMAIL PROTECTED] (Grant Edwards)
Subject: Re: How to communicate with COM port?
Date: Thu, 05 Apr 2001 02:52:50 GMT
On Thu, 5 Apr 2001 09:28:01 +0800, Kamp Huang <[EMAIL PROTECTED]> wrote:
>Can you give me some information and tips, special some samples.
Learn how to crosspost. Answered in another group.
--
Grant Edwards grante Yow! .. I wonder if I
at ought to tell them about my
visi.com PREVIOUS LIFE as a COMPLETE
STRANGER?
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: difference between ext2fs and reiserfs
Crossposted-To: comp.os.linux.misc
Date: Thu, 05 Apr 2001 02:58:22 GMT
[EMAIL PROTECTED] (Rod Smith) writes:
> In article <UaJy6.124985$[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] writes:
> > [EMAIL PROTECTED] (Rod Smith) writes:
> >> In article <w3my6.122397$[EMAIL PROTECTED]>,
> >> [EMAIL PROTECTED] writes:
> >> >
> >> >> Unfortunately, both ReiserFS and ext3fs suffer from the same file-
> >> >> and partition-size limits as ext2fs. These are starting to become
> >> >> issues for some people -- particularly the 4GB file-size
> >> >> limit. Therefore, unless those limits are raised, neither of these
> >> >> filesystems will really do as more than a stop-gap measure.
> >> >
> >> > Hum? "The same file-size limits as ext2fs" would indicate something
> >> > slightly over 4TB.
> >> >
> >> > With 4KB blocks, the limit on filesize is 4TB + 4GB + 4MB + 12 * 4KB.
> >>
> >
> > Nope. Check this out, from /usr/src/linux/include/linux/ext2_fs.h:
> >
> > Nope. Check this out.
> >
> > [cbbrowne@chvatal:testdeb] dd if=/dev/zero of=5g bs=1G count=5
> > 5+0 records in
> > 5+0 records out
> > [cbbrowne@chvatal:testdeb] ls -l 5*
> > -rw-r--r-- 1 root root 5368709120 Apr 4 11:59 5g
>
> I've reproduced your demonstration with a 2.4.0 kernel; HOWEVER, when I
> rebooted into a 2.2.17 kernel, the system reported the file size as
> being 4,294,967,295 bytes -- one byte short of 4GB. (I didn't bother
> trying to CREATE the file under the 2.2.17 kernel.) I therefore suspect
> that one of two things is going on:
>
> 1) The 4GB limit I referred to was real in ext2fs as implemented by the
> 2.2.x kernel (with VFS imposing its own 2GB file size limit atop
> that). Presumably ext2fs was extended in the 2.3.x tree.
> 2) Something about the 2.2.17 kernel caused buggy file size reporting.
> As a practical matter, there's not much difference between the two;
> the bottom line is that a "stock" 2.2.x kernel can't handle more
> than 2GB because of VFS limits, but a 2.4.x kernel can handle more
> than 4GB on ext2fs. In any event, there's still a *LOT* of obsolete
> documentation floating around out there -- much of it (including a
> paper on the official ext2fs SourceForge site,
> http://e2fsprogs.sourceforge.net/ext2intro.html) refers to a 2GB
> ext2fs file size limit. I guess I'll have to start being careful
> about kernel version numbers when referring to ext2fs file size
> limits....
This is likely to be something in flux over the next little while;
getting full support for big files requires:
a) That the filesystems support the big size;
b) That VFS support the big size [not true until late 2.2.x, _maybe_,
certainly late breaking 2.3.x and 2.4.x do...];
c) That GLIBC support the big size;
d) That applications be compiled with the GLIBC support for "big
size."
Note also that the dearth support for big files on ext2 was only true
for 32 bit architectures; it has worked on Alpha/UltraSPARC for a
while now.
The next time the major distributions "churn" through another version,
the 32 bit limitation will pretty likely evaporate away.
Mind you, NFS2 only supports exporting 32 bit file sizes, so NFS will
cause this limitation to still be visible here and there for a while
yet...
--
(reverse (concatenate 'string "ac.notelrac.teneerf@" "454aa"))
http://vip.hex.net/~cbbrowne/resume.html
"What this list needs is a good five-dollar plasma weapon."
--paraphrased from `/usr/bin/fortune`
------------------------------
Date: Wed, 04 Apr 2001 23:33:26 -0400
From: tom <[EMAIL PROTECTED]>
Subject: LINUX FEVER magazine ... Check it out!
http://linuxfever.hypermart.net
Become part of the fever today!
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
Date: Wed, 04 Apr 2001 23:56:52 -0500
From: Robert Redelmeier <[EMAIL PROTECTED]>
Subject: Re: cpu scheduling problem
[EMAIL PROTECTED] wrote:
>
> Robert Redelmeier <[EMAIL PROTECTED]> writes:
> > �ڵ��� wrote:
> > >
> > > If you look into the source code, you can find the fact that when a new
> > > process enters runqueue, it is located at the tail of the queue.
> >
> > That makes me think: What would happen if new processes
> > were put _at the head_ of the run queue???
> >
> > Of course, daemonic spawners would cause trouble and
> > unfairly hog resources.
> >
> > But the run-of-the-mill well-behaved app might see
> > better response. I suspect most threads either block
> > on IO or terminate very quickly. No real purpose
> > it making it wait all through the queue. Better
> > response at a slighly lower performance for CPU
> > hogs that eat up their slice.
> >
> > Comments?
>
> I suspect you won't likely get a whole lot of improved response out of
> your strategy; it should really only provide any substantive boost the
> very first time the program gets a "slice" of time.
Oh, I agree it isn't likely to be a big boost. But
if the thread is short, at least it would get into
the IO queues quicker. And a quick thread might
terminate in the first slice.
Think of a proggie that launches a thread to do something
like an arp call. It would come back very quickly or block
rather than wait through the whole queue.
Big threads, especially background worker CPU intensive
threads wouldn't benefit at all.
-- Robert
------------------------------
From: jtnews <[EMAIL PROTECTED]>
Subject: Re: did anyone have problems with Daylight Savings changeover?
Date: Thu, 05 Apr 2001 05:45:24 GMT
I've had the exact problem
you described, with the clock being
advanced twice in one of my systems
with Windows and Linux in a dual boot
config.
I've set all my hw clocks to UTC for
all my pure Linux machines.
I'll just have to work on making
the other one a pure linux machine
too.
"H�kon Alstadheim" wrote:
>
> Rick Ace <[EMAIL PROTECTED]> writes:
>
> > "H�kon Alstadheim" wrote:
> >
> > > Set your hardware clock to UTC, and change the appropriate config
> > > things. If you don't want to do that, live with it, or boot windows
> > > once after every change (i.e. once in fall and once in spring) . It
> > > will adjust your hardware clock.
> >
> > Hmmm. Windows might be good for something after all :oP
>
> Tongue in cheek and all that, but just to keep the record straight:
> windows is wrong on insisting that the RTC^H^H^H Hardware Clock be in
> local time. Linux can actually work with both UTC and local time in
> the hardware clock.
>
> As we all know local time changes an hour in each direction every
> year. Also computers have been known to move. When these things happen
> we do *not* want the date a file was last changed to be affected.
> System time should be UTC.
>
> The windows way works, sort of, if you only run one OS. If linux did
> the same as windows , you might end up with a *TWO* hour adjustment
> after having booted windows and then Linux (depending on how the phase
> of the moon affects windows ability to set the DST bit in CMOS). :/.
> You *can* actually get linux to do the same thing as windows. Just put
> something like "hwclock --systohc" in your halt script, and create a
> check in your boot script to see whether daylight saving changed while
> the machine was off. Add or subtract an hour accordingly with hwclock
> --date=. This boot script should of course connect to the net and
> download the latest laws on daylight saving, and parse out the start
> and end dates, but that's a different story.
>
> P.S: If you have N installations on a given machine, DST adjustment
> will obviously be N hours, when you've booted each OS once. Try
> dual-booting two windows installations and see.
>
> --
> H�kon Alstadheim, Montreal, Quebec, Canada
------------------------------
From: Florian =?iso-8859-1?Q?Gro=DFe=2DCoosmann?= <[EMAIL PROTECTED]>
Subject: Re: Newbie question: how to debug threaded functions
Date: Thu, 05 Apr 2001 07:58:03 +0200
Hi,
"Thomas A. Anderson" wrote:
> I recently run into a problem in which the GDB compiler
> cannot see what is happening inside the threaded function. Debugging info
> just points to the line just before the thread starts to execute. I would
> like to see what is happening inside my threaded code.
>
> What tools are there to help me with this problem? Do I have to get a new
> GDB version that supports debugging of multithreaded programs? Are there any
> utilities that are specifically designed for this?
>
> my devel system:
> -------------------------
> GNU gdb 4.18
> gcc version egcs-2.91.66
> Slackware 7.0 with kernel 2.4.3
> -------------------------
I tried gdb 4.18 about 1.5 years ago und run into the same problems. There is
a patch for 4.18 but it's better to fetch gdb 5.00 and try it. Read the
included manual and FAQ for multi-threading problems. I still have some
problems with gdb 5.00 and glibc 2.2 but they are minor ones and I can
debug.
Cheers, Florian
------------------------------
From: [EMAIL PROTECTED] (Mary K. Conner)
Subject: Editting filter for masquerading
Date: Thu, 05 Apr 2001 05:57:53 GMT
I have a networked app that I need to run through a Linux box that
does masquerading. The app transmits its own IP number and port,
which the other half of the app on the other side is supposed to
connect to. This obviously will not work with masquerading. I do not
have source code for the app so fixing things that way is not an
option. What I want to do is install a filter that will recognize and
edit the packet with the IP/port info to send the other side app to a
redirected port on the Linux box. I'm comfortable programming under
Linux, but I have never done anything like try to hook a kernel
service like this before.
Is there such a thing in existance already? It would save me a lot of
work. If not, any suggestions for how to go about doing this,
especially avoiding security risks, would be appreciated.
------------------------------
From: Florian =?iso-8859-1?Q?Gro=DFe=2DCoosmann?= <[EMAIL PROTECTED]>
Subject: Re: How to control time to wait when opening a socket ?
Date: Thu, 05 Apr 2001 08:07:27 +0200
Hi,
[EMAIL PROTECTED] wrote:
> =
> How to control time to wait when opening a socket ?
> =
> Do you know how to control the timeout of the "socket" connect
> function. I think it is possible to do with the sysctl executable.
You mean the connect() ?
> Better yet, do you know the proper syntax for the C function "sysctl"
> which will control this,
Just do a simple system("sysctl do_what_i_want"); and read the man page
of sysctl IF YOU REALLY WANT TO BREAK YOUR SYSTEM.
It's normally a bad idea to change systemwide variables. I may be wrong
but it sounds that you like to adjust the timeout for a specific
application only. In this case open the socket by socket() and
put it in the asynchroneous mode. Then you have the ability to =
select() or poll() (poll is faster under Linux) against the socket
after a connect() call. Quotation from the man page of connect()
of the error code in this case:
EINPROGRESS
The socket is non-blocking and the connection can=AD
not be completed immediately. It is possible to
select(2) or poll(2) for completion by selecting
the socket for writing. After select indicates
writability, use getsockopt(2) to read the SO_ERROR
option at level SOL_SOCKET to determine whether
connect completed successfully (SO_ERROR is zero)
or unsuccessfully (SO_ERROR is one of the usual
error codes listed here, explaining the reason for
the failure).
Use this.
Cheers, Florian
------------------------------
** 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 by posting to the
comp.os.linux.development.system newsgroup.
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
******************************
