December 2003 - Legacy - lists.gtalug.org

Compiling the 2.6 kernel
by fraser-Txk5XLRqZ6CsTnJN9+BGXg＠public.gmane.org 26 Dec '03

26 Dec '03

On December 26, 2003 11:51 am, Fraser Campbell wrote: > > doesn't ash (or any shell) require a kernel underneath? So why don't I > > just toss initrd and boot straight into the kernel I was going to use > > anyway? > > Debian kernels are modular. To answer your question, if you wish to do away with initrd then I don't see why you cannot. Just compile IDE/SCSI/whatever support directly into the kernel and you should be fine without an initrd. Have you looked into Debian's make-kpkg tool, it is trivial to compile your own Debian kernels (i.e. dpkg installable kernels) using it. -- Fraser Campbell <fraser-Txk5XLRqZ6CsTnJN9+BGXg at public.gmane.org> http://www.wehave.net/ Georgetown, Ontario, Canada Debian GNU/Linux -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Compiling the 2.6 kernel
by fraser-Txk5XLRqZ6CsTnJN9+BGXg＠public.gmane.org 26 Dec '03

26 Dec '03

On December 26, 2003 12:30 am, pking123-rieW9WUcm8FFJ04o6PK0Fg at public.gmane.org wrote: > doesn't ash (or any shell) require a kernel underneath? So why don't I just > toss initrd and boot straight into the kernel I was going to use anyway? Debian kernels are modular. Almost everything is a module (even basic IDE support). I'm far from an expert on initrd but, I believe, the initrd is required in these circumstances (where you need a module to mount your root filesystem). If your root filesystem is on IDE then you need the ide_disk module loaded before you can mount the root filesystem, once that's mounted then additional modules that you need can simply be modprobed /lib/modules on your root filesystem. Another case of needing an initrd would be if you have your root filesystem on a software raid device (well some hardware raids as well probably). LVM is likely another case where you'd require initrd, although having / on lvm is not recommended. > This is what I am accustomed to doing in just about every other compilation > I have done.Why the extra layer of bureaucracy? By using modules you can support a maximum range of hardware without having every single driver compiled into the kernel. > As for the modules, how do I know which modules to put in there? I > recognise the "benefit" of initrd is to selectively install modules in a > way that does not cause conflicts with other modules. So, I guess that > means I can't include all of them. :-) There are a series of files and directories under /etc/mkinitrd/ that let you specify additional modules to load, programs to add to the initrd, filesystem to use, scripts to run, etc. mkinitrd is just a (very configurable) shell script so you can read through it to figure out how the Debian folk decide what goes in, I've never looked at it in detail. > Nevertheless, I have created the beginnings of my own initrd, but am at a > loss as to the modules, and how to write modules.conf. You probably should not manually modify modules.conf (man update-modules), if you're talking about modules.conf in the initrd then I'm not too sure ... hopefully covered in the various manpages. -- Fraser Campbell <fraser-Txk5XLRqZ6CsTnJN9+BGXg at public.gmane.org> http://www.wehave.net/ Georgetown, Ontario, Canada Debian GNU/Linux -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Waaaay offtopic
by james.knott-bJEeYj9oJeDQT0dZR+AlfA＠public.gmane.org 26 Dec '03

26 Dec '03

cbbrowne-HInyCGIudOg at public.gmane.org wrote: >>We don't have to argue about the importance of vitamins and minerals, >>do we? > > > If you're lacking sufficient quantities, that's certainly cause for > problems. > > But adding more may not do anything more profound than making one's pee > very expensive. And some vitamins in excess are harmfull. -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Waaaay offtopic
by cbbrowne-HInyCGIudOg＠public.gmane.org 26 Dec '03

26 Dec '03

> We don't have to argue about the importance of vitamins and minerals, > do we? If you're lacking sufficient quantities, that's certainly cause for problems. But adding more may not do anything more profound than making one's pee very expensive. -- "cbbrowne","@","acm.org" http://www.ntlug.org/~cbbrowne/nonrdbms.html Is the surface of a planet the right place for an expanding technological civilization? -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Formatting in C++ (fwd)
by jingsu-26n5VD7DAF2Tm46uYYfjYg＠public.gmane.org 26 Dec '03

26 Dec '03

My appologies... I double checked and realized I copied the code incorrectly. Heh. Should've used cut&paste instead of re-typing it. The line should read: std::cout << setw(2) << std::hex << (unsigned it)buf[i]; -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Formatting in C++
by wildberger-iRg7kjdsKiH3fQ9qLvQP4Q＠public.gmane.org 26 Dec '03

26 Dec '03

On December 25, 2003 09:41 pm, Jing Su wrote: > I tried a short example, which seemed to work... snippet below. The only > catch I found was I had to cast the number. I think the operator > overload is sending the number into the character interpretor instead of > the numeric interpretor. > > std::cout << setw(2) << (unsigned int)std::hex << buf[i]; > This gives me a long string of numbers starting with 134515590134515590....... The number sequense in buf[] is 01 00 00 00 5A FF 00 F0 C3 E2 This prints fine with printf( "%.2X ",buf[i]); I also tried a few other suggestins, given by by Peter Peres but also with negative success. It seems that the good old printf() is still the best to use. It just bugs me not being able to do it with cout. John -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

C programming question
by wildberger-iRg7kjdsKiH3fQ9qLvQP4Q＠public.gmane.org 26 Dec '03

26 Dec '03

On December 26, 2003 05:29 am, Peter L. Peres wrote: > od -j 20 -N100 -tx1 /dev/mem > Peter Thanks Peter, od -j0 -N100 -tx1 /dev/mem ==> 0000000 01 00 00 00 5a ff 00 f0 c3 e2 00 f0 5a ff 00 f0 0000020 5a ff 00 f0 54 ff 00 f0 57 ff 00 f0 5a ff 00 f0 0000040 a5 fe 00 f0 87 e9 00 f0 7c 00 00 f0 7c 00 00 f0 0000060 7c 00 00 f0 7c 00 00 f0 57 ef 00 f0 7c 00 00 f0 0000100 14 00 00 c0 4d f8 00 f0 41 f8 00 f0 24 e0 00 f0 0000120 39 e7 00 f0 59 f8 00 f0 2e e8 00 f0 d2 ef 00 f0 0000140 53 ff 00 f0 0000144 This does the trick as far as the numbers go, but the location indicators are still not right. They should be in incrementsc of 10 and not 20 because they count in hex the number of bytes. John -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Formatting in C++
by plp-ysDPMY98cNQDDBjDh4tngg＠public.gmane.org 26 Dec '03

26 Dec '03

On Thu, 25 Dec 2003, John Wildberger wrote: > How can I use cout <<... to perform the equivalent to > printf( " %.2X ",buf[i] ); > when buf[] was declared as uint8_t buf[] ; > and the content of buf[i] are hex numbers like 0x5e, or 0xfd, etc ? > I think that I would have to include #include <iomanip> ,but I am at a loss > how to accomplish the formatting. cout << uint8 << buf[i]; // example only, uint8 may not work on your c++ Look up the correct syntax for the uint8 there are a dozen options. Peter -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

C programming question
by plp-ysDPMY98cNQDDBjDh4tngg＠public.gmane.org 26 Dec '03

26 Dec '03

On Thu, 25 Dec 2003, John Wildberger wrote: > ==> At offset 20 : 54 FF 00 F0 57 FF 00 F0 5A FF > > > > > od -j 20 -N100 -t x2 /dev/mem ( I modified this from 2 to 100 > 0000140 ff53 f000 > 0000144 > > This might be shorter, but very confusing. Note the numbering for each line. > It is double of what it should be.(2 times hex) od -j 20 -N100 -tx1 /dev/mem ;-) Peter -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0

Compiling the 2.6 kernel
by pking123-rieW9WUcm8FFJ04o6PK0Fg＠public.gmane.org 25 Dec '03

25 Dec '03

Date sent: Fri, 26 Dec 2003 01:47:23 -0500 From: Jing Su <jingsu-26n5VD7DAF2Tm46uYYfjYg at public.gmane.org> To: tlug-lxSQFCZeNF4 at public.gmane.org Subject: Re: [TLUG]: Compiling the 2.6 kernel Send reply to: tlug-lxSQFCZeNF4 at public.gmane.org > > Any suggestions or readings would be of help. > > I too decided to play with 2.6 for the day, though I got frustrated and > stopped. Anyone know what the deal with the new modules is? Based on > some browsing of newsgroups via google, it seems that a new module loader > is needed? (I'm using the one shipped with RH8). Anyways, I went and got > the latest/greatest non-development version of the module loader. > modules.conf has been replaced with modprobe.conf, and modprobe.devfs. > Haven't quite figured out these things yet. > > Anyways, I've been getting module dependency errors and they refuse to > load when I boot the new kernel.... still investigating it though, to > figure out what's going wrong. Cover the obvious ... Did you do a "make modules" then a "make modules_install"? The kernel I installed keeps complaining about an unrecognised block (0,0) on the boot partition, and does a kernel panic. When I load the default kernel, it does not have the problem. However, this time I am trying to do this without initrd (I compiled the kernel with initrd turned off, and do not have an initrd line in lilo.conf for that kernel). > > As for initrd, I found that it's mostly used for scsi systems. Atleast > it's true in my case. I needed to pack up the the scsi board module into > the initrd image. After a while, I found the whole process of initrd > creation to be annoying, and just compiled the scsi driver into the kernel > itself, and did a regular boot without initrd. > > -Jing > -- > The Toronto Linux Users Group. Meetings: http://tlug.ss.org > TLUG requests: Linux topics, No HTML, wrap text below 80 columns > How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml ========================================================= Paul King http://www3.sympatico.ca/pking123/ -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

1 0