Linux-Development-Sys Digest #235, Volume #6 Fri, 8 Jan 99 02:14:15 EST
Contents:
Re: blocksize / file write speed anomaly (Jerry Dinardo)
Re: blocksize / file write speed anomaly (Jerry Dinardo)
Re: linux in embedded system (mlw)
Re: WDM Emulator, anyone? (Jens Kristian S�gaard)
Re: disheartened gnome developer (Navindra Umanee)
Re: reboot problem with adaptec 2940uw (aic7881rev1) (Johan Kullstam)
Re: Why I'm dumping Linux, going back to Windblows (Johan Kullstam)
Re: Ultra-DMA safe in 2.2.0-pre? (ian)
Re: disheartened gnome developer (Christopher B. Browne)
Re: WDM Emulator, anyone? (Tony Hoyle)
Re: Glibc2.0.7 where is it ? (Allin Cottrell)
----------------------------------------------------------------------------
From: Jerry Dinardo <[EMAIL PROTECTED]>
Subject: Re: blocksize / file write speed anomaly
Date: Thu, 07 Jan 1999 10:01:46 -0500
I did not remove the file before my random writes because i wanted to simulate DB
creation and DB update. However , I changed my program to do the random writes on a new
file.
STEP 1
1. create file
2. write 20000 2K blocks sequentially
3. close
9 seconds
STEP 2
4. open existing
5. 10000 2K blocks on 4K boundary (block 0,2,4,6 ...)
6. close
102 seconds
STEP3
7. RECREATE file
8. 10000 2K blocks on 4K boudary (block 0,2,4,6 ...)
9 close
7 seconds
The procesing in STEP 2 and STEP 3 is identical except the file is new in step 3 and it
runs 14 times faster.
I also noticed that the update rate is about the disk RPM rate.
This leads me to believe that linux is reading the 2K block before writing it. I know
that linux would have to read block if it did not start on sector boudary or it was not
an even multiple of sector size. However , I thought the magic number was 512 and I
used
2048 just to be safe.
Does anyone know under what conditions linux will read block before writing it??????
* I needed to increase the file size to get the same results on machines with more
memory. I have been testing on 32M and 64M machines.
thanks jerry
------------------------------
From: Jerry Dinardo <[EMAIL PROTECTED]>
Subject: Re: blocksize / file write speed anomaly
Date: Thu, 07 Jan 1999 10:06:20 -0500
This is a multi-part message in MIME format.
==============8AB842C22C9192A3E4C13B3F
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I did not remove the file before my random writes because i wanted to simulate DB
creation and DB update. However , I changed my program to do the random writes on a new
file.
STEP 1
1. create file
2. write 20000 2K blocks sequentially
3. close
9 seconds
STEP 2
4. open existing
5. 10000 2K blocks on 4K boundary (block 0,2,4,6 ...)
6. close
102 seconds
STEP3
7. RECREATE file
8. 10000 2K blocks on 4K boudary (block 0,2,4,6 ...)
9 close
7 seconds
The procesing in STEP 2 and STEP 3 is identical except the file is new in step 3 and it
runs 14 times faster.
I also noticed that the update rate is about the disk RPM rate.
This leads me to believe that linux is reading the 2K block before writing it. I know
that linux would have to read block if it did not start on sector boudary or it was not
an even multiple of sector size. However , I thought the magic number was 512 and I
used
2048 just to be safe.
Does anyone know under what conditions linux will read block before writing it??????
* I needed to increase the file size to get the same results on machines with more
memory. I have been testing on 32M and 64M machines.
thanks jerry
==============8AB842C22C9192A3E4C13B3F
Content-Type: application/x-unknown-content-type-cfile; name="Jpd.c"
Content-Transfer-Encoding: base64
Content-Disposition: inline; filename="Jpd.c"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==============8AB842C22C9192A3E4C13B3F==
------------------------------
From: mlw <[EMAIL PROTECTED]>
Subject: Re: linux in embedded system
Date: Fri, 08 Jan 1999 04:21:54 +0000
Eric Hegstrom wrote:
>
> Well the real-time Linux project might be a place to start:
>
> http://luz.cs.nmt.edu/~rtlinux/
>
> hope this helps some.
> Cheers,
> Eric
>
> ggyy wrote:
> >
> > Does anyone know that how to use Linux in embedded system? I means to burn
> > the linux kernel binaries into ROM and then load into RAM after booting. My
> > friend is working on Linux and FreeBSD for DVD player and would like to put
> > the kernel binaries into the ROM. Hope anyone outthere could help. Thank
> > you.
> >
> > With Best Regards,
> > Cheng-Ta Wu,
>
> --
> Eric Hegstrom tele: 520-617-0072 x402
> Sonoran Scanners, Inc. email: [EMAIL PROTECTED]
There are a couple ways to do this. The dreaded Disk On Chip guys have a
disk driver for their nvram disk chip, between 40 and 100 meg.
You could format a rom as a ram disk image and boot off it. Use an
additional ram diskas tmp space.
You could write a little boot loader and store a compressed ram disk
image in rom.
It is all up to you.
How much rom you got?
--
Mohawk Software
Windows 95, Windows NT, UNIX, Linux. Applications, drivers, support.
Visit the Mohawk Software website: www.mohawksoft.com
------------------------------
From: [EMAIL PROTECTED] (Jens Kristian S�gaard)
Crossposted-To: comp.os.ms-windows.programmer.nt.kernel-mode
Subject: Re: WDM Emulator, anyone?
Date: 08 Jan 1999 04:38:49 +0100
mlw <[EMAIL PROTECTED]> writes:
> I'm a few days in to it and the group has grown to four. I will be
> working on dividing up the various areas of the kernel this week.
I hope the project succeedes. I would like to help, but as I haven't
got any experience with NT internals ( nor Linux really ) I probably
can't... ;-) ( unless you have some programming that can be done
without too much knowledge of the system ).
> BTW: What's up with this?
> /packages/wine-990103/miscemu/main.c:106: undefined reference to
> `Callout'
It seems the generated Makefile from 990103 is a bit erraneous.
Just edit the top-level Makefile; go to the line that builds wine and
wine.sym.
Change the line that starts with $(CC) to:
$(CC) -o wine /usr/X11/lib/libXxf86dga.a memory/local.o
windows/queue.o controls/menu.o ole/storage.o misc/callback.o $(EMUOBJS) $(ALT_LINK)
$(LDOPTIONS) $(X_LIBS) $(XLIB) $(LIBS)
( there should be no line-breaks ofcourse )
You ofcourse should include libXxf86dga.a if you haven't got a DGA
enabled XFree X server.
--
Jens Kristian S�gaard,
[EMAIL PROTECTED]
------------------------------
From: Navindra Umanee <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: disheartened gnome developer
Date: 7 Jan 1999 05:41:24 GMT
[EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> If you don't want a flame war, Navindra, than I suggest you 1) don't take what
> people say out of context 2) don't distort the facts.
Whoopsie, did I do that.
> I clearly emphisized in my previous post that I did not mean to discourage
> anyone from developing for KDE/QT. I merely said developers should understand
> the implications of the license.
Agreed. I'd add that people should also understand the implications
of GPL'ing their work.
>> > I don't care what Eric Raymond or Richard Stallman says. As I see it: open
>> > patch != open source.
>>
>> Give us a break. The patch clause has been removed for a long time
>> now. We've already had this flame war.
>
> Not true. Go read the license http://www.troll.no/qpl/ clause #3
>
> "3. You may make modifications to the Software and distribute your
> modifications, in a form that is separate from the Software, such as patches."
> ^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> In other words, you can't fork the development, or even put your own version
Sure you can fork the development.
> on a CVS repository and have yourself and others commit patches to on it.
Okay, here I admit I'm not sure what "in a form that is separate from
the Software" means. Considering patches though, you can put the
patches themselves in CVS alongside the original tree (you can
generate diffs from the original easily enough from a local PRCS/CVS
tree).
> Take a look at the KDE CVS repository. Where's the toolkit?? Harmony?? What's
> that doing there??
Huh? Don't be so disingenuous.
>> > inheritance/polymorphism scheme is different from what's traditionally
>> > taught in school. Hence there's a bit of a learning curve for most people,
>> > but once you've learned it's no harder than c++.
>>
>> It also doesn't make GTK+ any easier than Qt does it?
>
> I never said it did. Did I??????????? Go back and read the thread and get the
> *context* of what I said. Please don't take what I say out of context
No, it was meant to be an unrelated comment, not a personal attack or
anything.
-N.
--
"These download files are in Microsoft Word 6.0 format. After unzipping,
these files can be viewed in any text editor, including all versions of
Microsoft Word, WordPad, and Microsoft Word Viewer." [Microsoft quote]
< http://www.cs.mcgill.ca/~navindra/editors/ >
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Subject: Re: reboot problem with adaptec 2940uw (aic7881rev1)
Date: 07 Jan 1999 10:20:27 -0500
> Johan Kullstam wrote:
> > i've got a problem with rebooting a computer with adaptec 2940uw
> > (aic7881rev1).
> >
> > the first time i boot up, everything goes well. the scsi card finds
> > my harddrive at id 0 lun 0. the kernel boots. happy happy joy joy.
> >
> > the fun ends when i go to reboot. i type shutdown -r now but now the
> > adaptec card prints a line about being itself and hangs. it does not
> > go on to print stuff about configuration by typing Cntl-A. it does
> > not find any disks at id 0 and lun 0.
> >
> > hitting the reset button does not help. the machine hangs at the same
> > place. i must power-down and then re-power-up. what is going on
> > here?
Joop Marijne <[EMAIL PROTECTED]> writes:
> Had same problem here to, with one of my newer servers the problem
> disappeared. I think that was a newer card doesn't have this problem
> anymore.
> The older server was a P166 from 1995. My newer PII from 1998 was
> fine.
did you use the same scsi controller on the new system or did you
change everything? maybe it's a bios problem on my motherboard. can
the scsi controller bios be upgraded?
--
johan kullstam
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: 07 Jan 1999 10:31:14 -0500
[EMAIL PROTECTED] (Chris Hanson) writes:
> I'm also - surprise, surprise -- a software engineer. I write
> applications and other software in C & C++ for a living. I know what
> I'm doing on Unix, Windows, and MacOS. I prefer the Macintosh both as
> a user and developer platform, however, because of its interface
> consistency (and consequent ease of use), and because other Macintosh
> developers are of a like philosophy.
>
> Frankly, I'm insulted that you equate "able to deal with complexity"
> and "wants to deal with complexity all the time in everything." Just
> because what I do for a living is complicated doesn't mean I want to
> go out of my way to make it moreso. (That's why I use CodeWarrior and
> not MPW for development. Its project manager frees me from having to
> write and maintain makefiles, track dependencies, use a separate
> source code control system, etc., and gives me all sorts of extra
> features -- like a class browser -- to boot. Sure, I could learn all
> sorts of tools, and keep track of some stuff in my head, but I'm more
> productive with my toolset.)
>
> What frustrates me is that Linux *doesn't* have to be hard to use or
> administer. Look at what NeXT did with NEXTSTEP and OpenStep, look at
> what Apple did with A/UX and what they're doing with MacOS X Server,
> look at Irix, look at BeOS (which while it isn't Unix is still
> extremely Unix-like). All of them are relatively easy to use and
> administer, because they take the idea of consistency of interface and
> the philosophy of ease of use to heart.
>
> Fortunately, upon further exploration of the Linux world, there are at
> least some developers who get it, and many of them are hard at work.
> I'm looking forward to the day when firing up a terminal window on my
> Linux box is as rare as it is on NEXTSTEP or BeOS. Hopefully it'll
> come soon.
i agree with everything you say about consistency being important. on
the other hand, i think that command-line interfaces are also good.
what is wrong with the unix command-lines is that they are
*inconsistent* not that you have to type a lot. command lines offer a
different user interface than menus. sometimes mouse pointing makes
sense. sometimes typing a command makes sense.
it's all the wacky one letter options changing from app to app (-r is
recursive for rm, -R is recursive for ls). every program feeling that
it's its god-given right to introduce its very own slightly
incompatible version of regexp syntax. the way find seems to have
been done by someone from a non-unix planet and the way it only does
90% of the job leaving you to do things like
find <dir> -type f -name '*~' -print | sed 's!.*~!mv \1 ~/.trash!' | sh
--
johan kullstam
------------------------------
From: ian <[EMAIL PROTECTED]>
Subject: Re: Ultra-DMA safe in 2.2.0-pre?
Date: Fri, 08 Jan 1999 00:58:37 -0500
what types of problems. I have a P200 MMX running on a VIA MVP3 chipset.
I had alot of trouble getting Linux to stay setup and running fine.
something different went wrong from X not starting except with XDM to
other probs.
it may have been related to UDMA, but then again I have had troubles
like that with Slakware before and maybe the ReHat I just bought will do
the trick.
I think that 2.0.35 supports UDMA but I am not sure
Andreas Spengler wrote:
>
> Hi folks,
>
> I recently switched my system to RH5.2 and got a new harddisk with
> U-DMA.
> I am using this on a VIA Board with MVP3 CS. Using kernel 2.1.131 I got
> regular
> ext2fs-errors using the DMA feature and the support for the VIA bridge
> with my new HD.
>
> Is this corrected in 2.2.0-whatever ... ???
>
> Thanks for any answers,
>
> Andreas Spengler
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: disheartened gnome developer
Date: 8 Jan 1999 06:14:06 GMT
Reply-To: [EMAIL PROTECTED]
On Thu, 7 Jan 1999 20:30:23 -0800, jedi <[EMAIL PROTECTED]> posted:
>On 7 Jan 1999 17:37:50 -0800, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>>In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] says...
>>
>>>
>>> Troll Tech, whatever you or I might think of them, are
>>> still just a company. Idealism and wishful thinking
>>> won't alter that or change the constraints that companies
>>> operate under.
>>
>>so, are you saying there should be no companies who work for profit??
>>
>>if eveything is free, companies will close, and there will be no jobs.
>>
>>is that what you want? no jobs?
>
> You do understand the distinction between an application,
> an API or the OS do you not? I would imagine not since you're
> a Windows cheerleader.
Nor does it seem possible for people to comprehend the basic fact that
free software is not free of cost.
Thus, the purported "freeness" of Internet Explorer has to have some
serious conditions sitting behind it.
There Ain't No Such Thing As A Free Lunch.
>>this is the free and great capitalist country we live in, where companies
>>are created to make wealth for its stock holders, to employ people and
>>to spread the American dream far and wide.
>
> That's why one should not trust a corporation as far as
> one could throw it's board or it's shareholders. Like
> most groups of humans, such groups often resemble a lynch
> mob with the CEO and board as potential tree ornaments.
Interestingly, it seems to be more "free and capitalist" when things
are being controlled by relatively small minorities of investors; the
IRS seems to be acting in ways to discourage the proliferation of
small businesses.
>>I for sure do not mind, and very happy to pay for my software products,
>>becuase I know when I do that, I am helping a company make some profit,
>>which in turn give people money to buy something I might make one day, and
>>therefor we all prosper and gain wealth and happeniss.
>>
>>This is why I buy these Linux distrubtions from Fry's where I know the prices
>>are higher and where I know I can get Linux for free if I wanted to. But I
>>dont. I want to pay for it.
>>
>>Look live the American dream.
>
> The american dream is not despotism. Therefore, we should
> not tolerate such things in our economies or encourage them
> to be formed.
>
> I've spent considerably more on Linux software, commercial
> or otherwise, so spare me pseudo-capitalist tripe.
Large corporations often behave as virtual despots; the legal
wrangling of MSFT and many of the other members of the software
industry look more like feudalism rather than capitalism...
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
[EMAIL PROTECTED] - "What have you contributed to Linux today?..."
------------------------------
From: [EMAIL PROTECTED] (Tony Hoyle)
Crossposted-To: comp.os.ms-windows.programmer.nt.kernel-mode
Subject: Re: WDM Emulator, anyone?
Date: Thu, 07 Jan 1999 17:26:23 GMT
On Thu, 07 Jan 1999 18:06:58 +0900, SONE Takeshi <[EMAIL PROTECTED]>
wrote:
>Andy Lutomirski wrote:
>> You do NOT need the NT system call interface. At all. This is
>> undocumented, and Mark Russinovich is the only person who uses it.
>
>I know...
>I just thought that if you develop NT driver interface, you can share
>the effort for developing syscall interface for applications to run.
>Kernel mode NT emulator should be more compatible than user mode one
>(Wine), because it can run all *real* (M$) DLLs.
>I suppose even real WIN32K.SYS can run in kernel mode, it is much like
>drivers.
>
I always thought Wine was rather a lot of effort to get NT stuff
running... Provided you can write hal.dll and ntdll.dll it should be
possible to run the entire Winnt system on top of that. The really
hard work would be getting the device drivers to interact properly
with the Linux kernel.
It also doesn't help that the NT kernel API is almost entirely
undocumented. Still, with enough people working on it it should be
possible to get something going.
Tony
btw. 'dumpbin /imports win32k.sys' doesn't look too frightening, it's
just a lot of work working out the parameters to the functions.
====================================================================================
If Bill Gates had a dime for every time a Windows box crashed...
.... Oh, wait a minute, he already does.
====================================================================================
[EMAIL PROTECTED] http://sale.netfusion.co.uk
====================================================================================
------------------------------
From: Allin Cottrell <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Glibc2.0.7 where is it ?
Date: Thu, 07 Jan 1999 12:05:06 -0500
Matt wrote:
>
> Help,
>
> I need to upgrade to glibc2.0.7 I have upgraded to 2.0.6 but
> now I need glib2.0.7 does anyone know where I can get a download
> of the locale, crypt and other files I need (like the 2.0.6 ones ?
ftp://alpha.gnu.org/pub/gnu
Latest 2.0.7 is pre6. Get the locale, linuxthreads, and crypt
add-on files that have version numbers as close to 2.0.7pre6 as
possible, counting from below. (Some of the files have not
changed and are not numbered 2.0.7pre6).
--
Allin Cottrell
Department of Economics
Wake Forest University, NC
------------------------------
** 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
******************************
