Linux-Development-Sys Digest #950, Volume #6 Sun, 11 Jul 99 22:14:10 EDT
Contents:
Re: "DbgPrint" in Linux? (Michael Meissner)
Re: Memory Management Bug (Stefan Proels)
Hello! ([EMAIL PROTECTED])
SOCK_PACKET (Bayee)
Re: driver for AMCC S5933 (Albrecht Dre�)
SCSI device/partition limit? (Cory Papenfuss)
Re: when will Linux support > 2GB file size??? (Adam Jacobs)
ACAD for Linux ... ([EMAIL PROTECTED])
ACAD for Linux ... ([EMAIL PROTECTED])
Re: MICROSOFT LINUX DISTRIBUTION (Captain Insano)
SkyOS ("Szeleney Robert")
Re: SOCK_PACKET (Andi Kleen)
Init fails to load (Aaron Tomb)
Re: CD-ROM File Time Bug ("Terry D. Boldt")
Kernel version 2.3.9+ ("Zachary Kuznia")
----------------------------------------------------------------------------
Subject: Re: "DbgPrint" in Linux?
From: Michael Meissner <[EMAIL PROTECTED]>
Date: 11 Jul 1999 11:23:25 -0400
[EMAIL PROTECTED] (Yung-Hsiang Lu) writes:
> Does anybody know how to do this (redirect printk to the serial line)?
> I plan to do some performance study about a hard disk. Therefore,
> logging on a hard disk doesn't seem a good way because it will pollute
> the data.
Read the syslog(2), syslog(3), and syslog.conf(5) man pages. In particular,
each syslog entry has a facility (auth, crron, daemon, kern, etc.) and a
priority (debug, info, notice, warning, error, crit, alert, panic) that you can
use to decide where the messages go. For your case, use a priority of debug to
go to a fifo or tty, and info or something lower to go to a disk file. In the
kernel then, use the debug priority for the messsages you want.
--
Michael Meissner, Cygnus Solutions
PMB 198, 174 Littleton Road #3, Westford, Massachusetts 01886
email: [EMAIL PROTECTED] phone: 978-486-9304 fax: 978-692-4482
------------------------------
From: Stefan Proels <[EMAIL PROTECTED]>
Subject: Re: Memory Management Bug
Date: Sun, 11 Jul 1999 06:56:19 +0200
Sean Walton wrote:
>
> There should be a FAQ on this: it is a common *and*avoidable* programming
> problem. Linux uses a "lazy allocation" scheme: a program can allocate as
> much memory as is wants, but the actual physical allocation does not occur
> until written to. So, it's actually possible to allocate 2GB (4GB on some
> patched systems) where only 128MB of VM is really available.
> Is this a defect? No, it was literally designed this way. [BTW, your
> system may not be locked up at all--it's possibly just thrashing.] There
No, it was completely locked up. No disk I/O. It didn't even allow me to
switch to another console.
> has been many, many discussions regarding this, and many have asserted that
> it is a major, fundamental defect in Linux. Torvalds says that it is
> supposed to work that way. That's that.
> Here are some rationales:
> -Many programs simply allocate a bunch of space and never use it. You
> may say that this is poor programming, but consider memory management, it's
> actually quite system-friendly: fragmentation doesn't happen as readily.
> -Memory is a resource that should appear inexhaustable. No program
> should have to worry about "heap overflow" errors on *nix systems.
> -Any program that tries to allocate the world deserves to die (of
> course, should it kill the OS or other apps?). Why are you trying to
> allocate all of memory/swap? If you really want to allocate the world, go
> buy the RAM/disk-space and load it up. Or, you can query the system what
> resources are available and allocate based on that info.
I have no way argued that it's a Bad Thing to handle allocation this
way. I just don't think that it's a Good Thing to enable every ordinary
user to crash the system. As I said I ran into this problem because of a
bug in a program of mine. Yes, this was my fault, but that shouldn't
crash a the system. If no space is left to page out a frame it should
kill the process in question. I realize that this isn't perfect because
this could hit every process, not just the one allocating the world. A
better strategy might be to kill the process which has allocated most.
setrlimit doesn't fix the problem. You still could have multiple
processes
(users) allocating like crazy.
Stefan
------------------------------
From: [EMAIL PROTECTED]
Subject: Hello!
Date: Sun, 11 Jul 1999 19:25:45 +0400
Reply-To: [EMAIL PROTECTED]
Dear friends,
I hope that you'll spare some time to browse through this site:
http://www.bbe.8m.com/index.html
Regards,
Yasser Thosip.
***********
------------------------------
Date: Mon, 12 Jul 1999 00:11:06 +0800
From: Bayee <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.development.apps
Subject: SOCK_PACKET
Hi all,
I am currently doing a project that need to convert tcp frame into
some propiatry network
and vice versa, like this,
Other Host <---- TCP/IP -----> (eth0) Linux Gateway
(ppp0/eth1) <----- >Propairty Network
Can I bind SOCK_PACKET to eth0 to read/write IP packet ?
Regards,
Bayee
------------------------------
Crossposted-To:
comp.os.linux.hardware,comp.os.linux.setup,comp.os.linux.misc,comp.os.linux.development.apps
From: [EMAIL PROTECTED] (Albrecht Dre�)
Subject: Re: driver for AMCC S5933
Date: Tue, 6 Jul 1999 06:32:58 GMT
[Posted and mailed]
> Yes,
> A nice gentleman from Italy shared some development code he had worked
> on for the s5933. It seems to work pretty darn well.
> His name is Andrea Cisternino <[EMAIL PROTECTED]>
[snip]
>> Is there driver for AMCC S5933 (Linux Red Hat for PC Pentium)?
>> If there is, as it to teceive?
I recently wrote a driver for a Hunt Engineering DSP card which also uses the
S5933. You will find it on the metalab archive (aka sunsite)
http://metalab.unc.edu/pub/Linux/kernel/misc-cards/
under the name hepc3-0.9.0.tar.gz. I am working on a new version with support
for 2.2 kernels and LinuxPPC, which is almost finished except for the docs.
If you are interested, please send me a mail, and I will send you the "pre-
release"...
Hope this helps, Albrecht,
--
+-----------------------------------------------------------------------------+
| Dr.-Ing. Albrecht Dre\ss ---- |
| Max-Planck-Institut f\"ur Radioastronomie |\ / /o o\ |
| Abteilung f\"ur Infrarot-Interferometrie | \ / | / | |
| Auf dem H\"ugel 69 | \ | \ ---/ |
| D-53121 Bonn (Germany) ------------+------+------------------- |
| | / | |
| Phone (+49) 228 525 319 | / / |
| Fax (+49) 228 525 411 |/ / |
| Mail [EMAIL PROTECTED] |
+-------------- electrical engineers do it with less resistance --------------+
------------------------------
From: Cory Papenfuss <[EMAIL PROTECTED]>
Subject: SCSI device/partition limit?
Date: Sun, 11 Jul 1999 12:50:50 -0500
Hey... I've got a fairly unique problem with scsi. I've got an
Adaptec 2940 OEM with three drives as the main system (sda=swap, /, /usr;
sdb=/home; sdc=/usr/local). This works fine and is statically compiled
into the kernel.
The problems is the RAID I've come up with on the second SCSI
controller. It's an NCR53c825, 50-pin narrow card. It's basically filled
up. DDS-1 tape drive at ID0, and then 6 1GB drives, with the first two in
a striped RAID0 and the last four in another. I'm using the ncr53c8xx
driver as a kernel module. Here's the problem:
When the module loads, it can see partitions on sdd-sdi (ID1-6 on NCR
controller) when the modules loads, but does not register all the
partitions in /proc/partitions. It can only see through sdg1. Also, when
I unload and load the module again, it loses two drive letters -- as
though the previous load didn't give them back (i.e. sdd-i becomes sdf-k
the next time). I've tried the 53c7,8xx driver also, which gives back the
letters, but still doesn't see them all.
One more bit of info. Each of the individual software RAIDs works
by itself. So, two devices or four devices is fine, but when all six
disks are on the bus, it can't register enough devices.
Any clues?
Thanks,
-Cory
FYI, here's some relevent dumps:
Module load:
ncr53c8xx: at PCI bus 0, device 8, function 0
ncr53c8xx: 53c825 detected
ncr53c825-0: rev=0x02, base=0xffecef00, io_port=0x7800, irq=7
ncr53c825-0: ID 7, Fast-10, Parity Checking
ncr53c825-0: restart (scsi reset).
scsi1 : ncr53c8xx - version 3.2
scsi : 2 hosts.
Vendor: ARCHIVE Model: Python 25501-XXX Rev: 296A
Type: Sequential-Access ANSI SCSI revision: 01
Detected scsi tape st0 at scsi1, channel 0, id 0, lun 0
Vendor: QUANTUM Model: EMPIRE_1080S Rev: 1242
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdd at scsi1, channel 0, id 1, lun 0
Vendor: QUANTUM Model: EMPIRE_1080S Rev: 1242
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sde at scsi1, channel 0, id 2, lun 0
Vendor: MICROP Model: 2210-09MQ1001901 Rev: HQ30
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdf at scsi1, channel 0, id 3, lun 0
Vendor: CONNER Model: CFP1060S 1.05GB Rev: 2035
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdg at scsi1, channel 0, id 4, lun 0
Vendor: SEAGATE Model: ST11200N Rev: 9682
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdh at scsi1, channel 0, id 5, lun 0
ncr53c825-0-<6,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
Vendor: IBM Model: 0662S12 !O Rev: 2 23
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdi at scsi1, channel 0, id 6, lun 0
ncr53c825-0-<1,0>: tagged command queue depth set to 8
ncr53c825-0-<2,0>: tagged command queue depth set to 8
ncr53c825-0-<3,0>: tagged command queue depth set to 8
ncr53c825-0-<4,0>: tagged command queue depth set to 8
ncr53c825-0-<5,0>: tagged command queue depth set to 8
ncr53c825-0-<6,0>: tagged command queue depth set to 8
ncr53c825-0-<1,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
SCSI device sdd: hdwr sector= 512 bytes. Sectors= 2109376 [1029 MB] [1.0
GB]
sdd: sdd1
ncr53c825-0-<2,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
SCSI device sde: hdwr sector= 512 bytes. Sectors= 2109376 [1029 MB] [1.0
GB]
sde: sde1
ncr53c825-0-<3,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
SCSI device sdf: hdwr sector= 512 bytes. Sectors= 2065250 [1008 MB] [1.0
GB]
sdf: sdf1
ncr53c825-0-<4,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
SCSI device sdg: hdwr sector= 512 bytes. Sectors= 2074880 [1013 MB] [1.0
GB]
sdg: sdg1
cat /proc/partitions
major minor #blocks name
8 0 1037440 sda
8 1 66544 sda1
8 2 103424 sda2
8 3 867328 sda3
8 16 1025430 sdb
8 17 1024586 sdb1
8 32 533127 sdc
8 33 532464 sdc1
8 48 1054688 sdd
8 49 1054598 sdd1
8 64 1054688 sde
8 65 1054598 sde1
8 80 1032625 sdf
8 81 1032176 sdf1
8 96 1037440 sdg
8 97 1037296 sdg1
------------------------------
From: [EMAIL PROTECTED] (Adam Jacobs)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: when will Linux support > 2GB file size???
Date: 11 Jul 1999 19:35:27 GMT
In article <7m8qtt$[EMAIL PROTECTED]>,
Byron A Jeff <[EMAIL PROTECTED]> wrote:
>...
>Simply put most apps don't need 2GB+ files...
>...
>So BTW why exactly do you need 2GB+ files?
I for one need huge -- O(1G) -- files to store the data acquired from long
multichannel electrophysiology experiments. Acquiring data (8-bit samples,
say) at 10 KHz on 64 channels for 6 hours can generate 13G of raw data,
In reality the file sizes I'm using right now are more like several hundred
megs, because I process the data during acquisition instead of saving raw
samples, But there are numerous reasons to want to save the raw data
as well. In a previous setup I routinely had file sizes approaching 1G.
With four times as many channels of data in my present setup, sizes over
2G are quite easy to contemplate. Saving the data per se is not difficult
(could use raw partition, or multiple files, or whatever) but it would be
convenient to be able to use the filesystem transparently.
I'd think high-definition video storage/editing would be a more mainstream
application for similar file sizes...
adam
------------------------------
From: [EMAIL PROTECTED]
Subject: ACAD for Linux ...
Date: Sat,10 Jul 1999 19:21:33+2000
Look at this web page to learn more about ACAD for Linux !!!
http://209.218.86.64/linux.html
If you know someone who can be interested in that please tell him
or her about this web page.
Thank you for your time.
------------------------------
From: [EMAIL PROTECTED]
Subject: ACAD for Linux ...
Date: Sat,10 Jul 1999 19:21:33+2000
Look at this web page to learn more about ACAD for Linux !!!
http://209.218.86.64/linux.html
If you know someone who can be interested in that please tell him
or her about this web page.
Thank you for your time.
------------------------------
From: [EMAIL PROTECTED] (Captain Insano)
Subject: Re: MICROSOFT LINUX DISTRIBUTION
Date: 11 Jul 1999 19:27:11 GMT
In article <[EMAIL PROTECTED]>,
Samuel Brown <[EMAIL PROTECTED]> wrote:
>I had a someone tell me that Microsoft will sell their own linux
>distribution. Is this true?
>
>They said it would have IE 5 and EXPLORER as the window manager
>and the setup program would be really simple.
>It will have word and excel 2000 also.
>Are they taking over LINUX?
>
I don't think MS has plans for Linux just yet, although I'm sure
someone there has given it a little thought. However, if they did
produce a Linux version of Office 2000, I would not be surprised at
all if it shipped with IE as the desktop/window manager.
It could be that your friend over heard someones theory about how
MS would enter the Linux market. This may just happen if and when
Linux becomes a viable portion of the market. I half-expect to see
something like this happen myself (e.g. MS produces IE and Office
for the Mac, and an IE for Unix systems).
I personally like the Windows desktop and wouldn't mind seeing it
implemented for Linux. Windows user-friendly GUI with Linux's
reliability might make a winning combination.
And if Linux were to support the MS COM/OLE architecture, it just
might become a useful operating system.
--
Mathematics and Computer Science Student.
------------------------------
From: "Szeleney Robert" <[EMAIL PROTECTED]>
Subject: SkyOS
Date: Sun, 11 Jul 1999 22:32:46 +0200
Hi!!!
Have you tested SkyOS already???
Please tell me what you think!!
Thanks, Cu, Bertl!!!
http://skyos.ics.at
------------------------------
From: Andi Kleen <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.development.apps
Subject: Re: SOCK_PACKET
Date: 11 Jul 1999 21:46:20 +0200
Bayee <[EMAIL PROTECTED]> writes:
> Hi all,
>
> I am currently doing a project that need to convert tcp frame into
> some propiatry network
> and vice versa, like this,
>
> Other Host <---- TCP/IP -----> (eth0) Linux Gateway
> (ppp0/eth1) <----- >Propairty Network
>
> Can I bind SOCK_PACKET to eth0 to read/write IP packet ?
Yes, although a SOCK_RAW socket would work in linux[1] too (and would be easier
for you because it is a bit higher level). Get man-pages-1.24 and read
raw(7) and packet(7)
-Andi
[1] Not on most BSD derived stacks.
--
This is like TV. I don't like TV.
------------------------------
From: Aaron Tomb <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,redhat.kernel.general
Subject: Init fails to load
Date: 11 Jul 1999 20:30:56 GMT
Hi,
I'm having a bizzare problem.
I have a new system with A DFI K6XV3+ MB, an AMD K6-2/400 CPU, 128MB ram,
Award BIOS from early June.
Whenever I start up the system for the first time, the kernel boots (using
LILO) and gets as far as "Freeing unused kernel memory, 44k freed", and
then stops. It never runs init.
If I reset the computer then, it works fine, although I have to hit the
hardware reset, Ctrl+Alt+Del doesn't work. It also works if I hit reset
before Linux even starts to load, while the BIOS is still doing its stuff.
It seems like it's a hardware problem in a sense, but it's very consistent.
The hardware isn't flaky, it just has some mistake in it? I think it's
something that could be worked around in the kernel, but I don't know quite
how.
Just for additional information, the BIOS sometimes doesn't detect all of
the memory. I've never had any problem with corruption of any sort. I think
the RAM is fine. It just sometimes doesn't all appear. I also think I need
to shorten my IDE cables for UDMA to work.
Oh, yes, and it's kernel 2.2.9-27, with the Mandrake distribution.
Thanks,
Aaron-
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
Date: Sun, 11 Jul 1999 19:19:49 -0500
From: "Terry D. Boldt" <[EMAIL PROTECTED]>
Subject: Re: CD-ROM File Time Bug
First let me thank all who have replied. Your efforts are appreciated. I am
still toying with this - situation - and haven't fully figured out the
problems on all of the platforms as yet.
I reset my hardware clock from EST to UTC (Red Hat advises leaving the h/w
clock on local time since on multi-boot machines the other OSs usually
expect local time on the h/w clock). Ran the program '/usr/sbin/timeconfig'
as Red Hat recommends to set something (they never specify what is set or
how the OS recognises to what the h/w clock is set) to signal Linux that
h/w clock now set to UTC and rebooted. The time under KDE is now off by one
hour - seems that KDE doesn't understand DST. However, the 'ls' listed
filetimes are more 'reasonable' and doing 'TZ= ls' displays filetimes that
are commensurate with the EDT to UTC correction. Neither filetime agrees
with the 'ls' output included by the CD manufacturer, but then the
manufacturer doesn't state whether they displayed local or UTC times - so
who knows what is right or consistent or anything else in this
situation.??? Also none of the filetimes agree with the DOS or OS/2
filetimes. So I have four sets of filetimes:
1) Linux translated to local
2) Linux left in UTC - I guess (TZ='')
3) The manufacterers filetimes (timezone setting unknown)
4) DOS and OS/2 filetimes
and all four sets differ.
Anybody know how to get KDE to recognize DST??
------------------------------
From: "Zachary Kuznia" <[EMAIL PROTECTED]>
Subject: Kernel version 2.3.9+
Date: Sun, 11 Jul 1999 21:35:59 -0400
In the 2.3.9 and 2.3.10 kernels I have found an error compiling the fat
module used to mount fat16 and fat32 drives. It finds an unresolved
reference to the function update_vm_cache. I have checked the source and
there hasn't been anything added or changed to the fat code, so would anyone
know which source file has/had the update_vm_cache function?
Zachary Kuznia
[EMAIL PROTECTED]
------------------------------
** 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
******************************
