Re: [Haifux] Lecture Slides are Ready (finally)
Orr Dunkelman <[EMAIL PROTECTED]> writes: > I know they are not perfect (feel free to send comments), > I know they are about month late, > > I know that... > > But enjoy: > > http://www.haifux.org/lectures/83 -- Qmail & SMTP. > > -- > Orr Dunkelman, > [EMAIL PROTECTED] > > "Any human thing supposed to be complete, must for that reason infallibly > be faulty" -- Herman Melville, Moby Dick. > > Spammers: http://vipe.technion.ac.il/~orrd/spam.html > > -- > Haifa Linux Club Mailing List (http://www.haifux.org) > To unsub send an empty message to [EMAIL PROTECTED] > > > Slides 2 and 3 of the PDF seem to be identical. -- "Cut your own wood and it will warm you twice" Regards, Yoni Rabkin Katzenell -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] gcc -pg, but no gmon.out
(what does haifux got to do with lkml anyway? :P~~~ we deserve our own alias!) On Tue, 27 Jan 2004, Shachar Tal wrote: > I have a somewhat complex piece of code that needs profiling. It is > composed of numerous Makefiles. I added "-pg -g" to the compiler > parameters, but when the resulting binary runs, gmon.out is not created. did you also pass '-pg' to the link command? or, if you use ld directly, did you tell it to use /lib/gcrt.o and -lc_p? 'pinfo gprof', then read about 'compiling', i've been reading from that page as i type... -- guy "For world domination - press 1, or dial 0, and please hold, for the creator." -- nob o. dy -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] gcc -pg, but no gmon.out
On Tue, Jan 27, 2004, Shachar Tal wrote about "[Haifux] gcc -pg, but no gmon.out": > Hi fellas, > > I have a somewhat complex piece of code that needs profiling. It is > composed of numerous Makefiles. I added "-pg -g" to the compiler > parameters, but when the resulting binary runs, gmon.out is not created. > > When I create a simpler project (Makefile and code file set) to test this > against, gmon.out is created correctly. What you may have done wrong is to forget to add "-pg" to the *linking* phase. It is necessary, but not enough, to add "-pg" to the compilation phase, but you must make sure that the final linking is done with -pg as well, otherwise your program isn't compiled with the right bootstrap code that actually generates the gmon.out. Check to see that the "cc -o progname object1.o object2.o ..." line in the "make" output has the "-pg" parameter too. Also, as Orna pointed out, gmon.out is only created when the program exits. You must therefore make sure it indeed exits, and not, for example, killed with an uncaught signal. P.S. To your linking stage, you will most likely want to also add "-lc_p". This will allow you to profile the standard C library calls as well. Without this, if you do a lot of library calls or IO you'll may have a lot of unexplained time in the profiling results. -- Nadav Har'El| Tuesday, Jan 27 2004, 5 Shevat 5764 [EMAIL PROTECTED] |- Phone: +972-53-790466, ICQ 13349191 |This space is for sale - inquire inside. http://nadav.harel.org.il | -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] gcc -pg, but no gmon.out
On Tue, 27 Jan 2004, Shachar Tal wrote: > Hi fellas, > > I have a somewhat complex piece of code that needs profiling. It is > composed of numerous Makefiles. I added "-pg -g" to the compiler > parameters, but when the resulting binary runs, gmon.out is not created. > > When I create a simpler project (Makefile and code file set) to test this > against, gmon.out is created correctly. > > Any thoughts? > Does your big program end cleanly? I remember I had problems on irregular exits. Also note that static functions cannot be profiled - hence I use a macro instead of the word static, so that I can replace it at once when I want to profile. Orna. -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
[Haifux] gcc -pg, but no gmon.out
Hi fellas, I have a somewhat complex piece of code that needs profiling. It is composed of numerous Makefiles. I added "-pg -g" to the compiler parameters, but when the resulting binary runs, gmon.out is not created. When I create a simpler project (Makefile and code file set) to test this against, gmon.out is created correctly. Any thoughts? -- Bye, Shachar. -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
[Haifux] Lecture Slides are Ready (finally)
I know they are not perfect (feel free to send comments), I know they are about month late, I know that... But enjoy: http://www.haifux.org/lectures/83 -- Qmail & SMTP. -- Orr Dunkelman, [EMAIL PROTECTED] "Any human thing supposed to be complete, must for that reason infallibly be faulty" -- Herman Melville, Moby Dick. Spammers: http://vipe.technion.ac.il/~orrd/spam.html -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] Need help ASAP - Till Tuesday 12:00
Slava Shklyar wrote: 1. init reads a location of needed files from paths.h The SHELL is defined there. Maybe you can change the default shell only for single user in the init.c. Search for ':[sS1]:' regexp in this file ... This may be a bit too late, but I would simply change the "/bin/sh" string that appears in the binary /sbin/init file. Use khexedit for that, for example. This is quick, dirty, may not work, but most probably will. Be sure to have a backup copy of the original /sbin/init, and also make sure that you have some kind of rescue disc to boot from (knoppix and alike) in the not-so-unlikely event that your computer will refuse to boot. Eli -- Web: http://www.billauer.co.il -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] ANN: Lectures: Writing device drivers in Linux and Windoze
On Tuesday 27 January 2004 02:56, Muli Ben-Yehuda wrote: > The first draft of the slides for this one are now available at >http://www.mulix.org/lectures/intro_to_linux_device_drivers/intro_linux_device_drivers.pdf. Very good. Small remarks/questions: - I assume they learn the actual build procedure elsewhere. If not, than at least some pointers are needed. - The idea to patch the core for hooking into timer interrupts is very good for learning (the power of open source...). However, because the interrupt you talk about is (by incident) the timer, people may get the wrong idea that in Linux patching is needed for timing... (which is a very common task). You may want to mention the existance of the simple timer_* API. - Some mention (maybe verbally) about kernel versioning (which is used for their exercise?) -- Understanding the Linux kernel API's are a moving target (intentionally...) - Are they getting (maybe in their lectures) some orderly view about: * Different locking schemes? (spinlocks, reader-writers, semaphores,...) * Different schemes for defering work? (softirqs, tasklet API, work-queues, timer API's) Bye, -- Oron Peled Voice/Fax: +972-4-8228492 [EMAIL PROTECTED] http://www.actcom.co.il/~oron Linux lasts longer! -- "Kim J. Brand" <[EMAIL PROTECTED]> -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
[Haifux] Re: Acks on logo votes
On Tue, 27 Jan 2004, Shlomi Fish wrote: > > Hi! > > I prefer logo 4 to logo 6 to the rest. > Oops! Replied to all by accident - sorry. Regards, Shlomi Fish > Regards, > > Shlomi Fish > > > > -- > Shlomi Fish[EMAIL PROTECTED] > Home Page: http://t2.technion.ac.il/~shlomif/ > > You are banished! You are banished! You are banished! > > Hey? I'm just kidding! > > -- > Haifa Linux Club Mailing List (http://www.haifux.org) > To unsub send an empty message to [EMAIL PROTECTED] > > -- Shlomi Fish[EMAIL PROTECTED] Home Page: http://t2.technion.ac.il/~shlomif/ You are banished! You are banished! You are banished! Hey? I'm just kidding! -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] ANN: Lectures: Writing device drivers in Linux and Windoze
On Tue, Jan 27, 2004 at 09:30:30AM +0200, Oron Peled wrote: > On Tuesday 27 January 2004 02:56, Muli Ben-Yehuda wrote: > > The first draft of the slides for this one are now available at > >http://www.mulix.org/lectures/intro_to_linux_device_drivers/intro_linux_device_drivers.pdf. > > Very good. Small remarks/questions: > - I assume they learn the actual build procedure elsewhere. If not, than at > least some pointers are needed. I think they do, but I'll mention it verbally. With 2.6, it's really trivial - a one line change in the Makefile and a two line adition to Kconfig. > - The idea to patch the core for hooking into timer interrupts is very > good for learning (the power of open source...). However, because > the interrupt you talk about is (by incident) the timer, people may > get the wrong idea that in Linux patching is needed for timing... > (which is a very common task). You may want to mention the existance > of the simple timer_* API. I will, definitely. The reason for patching into the timer interrupt is to show deferring work from the top half to the bottom half context. timers expire in the botom half, so it wouldn't make much sense to defer work there. > - Some mention (maybe verbally) about kernel versioning (which is used > for their exercise?) -- Understanding the Linux kernel API's are > a moving target (intentionally...) I'm not sure this is something I should get into... I'll keep it in mind. > - Are they getting (maybe in their lectures) some orderly view about: > * Different locking schemes? (spinlocks, reader-writers, > semaphores,...) I think so. I was asked to keep this part relatively simple. > * Different schemes for defering work? (softirqs, tasklet API, > work-queues, timer API's) Definitely. The lecture slides are available on the web, for the curious. Very interesting reading, the instructors did a swell job. Cheers, Muli -- Muli Ben-Yehuda http://www.mulix.org | http://mulix.livejournal.com/ "the nucleus of linux oscillates my world" - [EMAIL PROTECTED] signature.asc Description: Digital signature
[Haifux] Re: Acks on logo votes
Hi! I prefer logo 4 to logo 6 to the rest. Regards, Shlomi Fish -- Shlomi Fish[EMAIL PROTECTED] Home Page: http://t2.technion.ac.il/~shlomif/ You are banished! You are banished! You are banished! Hey? I'm just kidding! -- Haifa Linux Club Mailing List (http://www.haifux.org) To unsub send an empty message to [EMAIL PROTECTED]
Re: [Haifux] Need help ASAP - Till Tuesday 12:00
1.
init reads a location of needed files from paths.h
The SHELL is defined there. Maybe you can change the
default shell only for single user in the init.c. Search for ':[sS1]:'
regexp in this file ...
2.
# This spec file was generated by Gideon
# Please report any problem to KDevelop Team <[EMAIL PROTECTED]>
# Thanks to Matthias Saou for his explanations on
http://freshrpms.net/docs/fight.html
Name: appName
Version: 0.1
Release: release
Vendor: Vendor
Copyright: License
Summary: Summary
Group: groupe
Packager: Packager
BuildRoot: %{_tmppath}/%{name}-root
Source: appName-0.1.tar.gz
%description
Description
%prep
%setup
CFLAGS="$RPM_OPT_FLAGS" CXXFLAGS="$RPM_OPT_FLAGS" ./configure \
--target=i386
%build
%configure
make
%install
rm -rf %{buildroot}
%makeinstall
%clean
rm -rf %{buildroot}
%post -p /sbin/ldconfig
%postun -p /sbin/ldconfig
%files
%defattr(-, root, root)
%doc AUTHORS COPYING ChangeLog NEWS README TODO
%{_bindir}/*
%{_libdir}/*.so.*
%{_datadir}/%{name}
%{_mandir}/man8/*
%changelog
guy keren wrote:
> On Mon, 26 Jan 2004, Raichshtain Gilad wrote:
>
>
>
>> We're doing a project in OS. The project goal is to create a new
>> shell (from scrtch) and replace the root shell with this new
>> shell.
>>
>> We need help with the following issues:
>>
>>
>
> and you think someone will "jump for your command" and help you,
> when you've less then 24 hours to get this thing done?
>
> my first advice is: ask for a delay in delivering the project.
> you're bound not to make it on time.
>
>
>
>> 1. One of the project goals is to be able to run the shell as the
>> default shell at Single Mode. Two questions: Which file do I have
>> to update in order to make it the default shell at Single Mode ?
>> And what kind of initiations should I do in order for it to work
>> ? Anything else that might effect our execution in Single mode?
>>
>>
>
> assuming that your shell is not compatible with bourne shell, you
> cannot just replace the normall shell (/bin/sh) - you will be
> breaking down your entire system (e.g. the shell scripts, which
> handle launching of the system, will stop working).
>
> i am not sure how init knows which shell to run as the default
> shell for single-user mode. perhaps it uses the shell defined for
> user 'root' - try changing that iwth the 'chsh' command. do note
> that you'd better make sure your shell actually works, and that you
> have a currently open root-shell, before doing this. then try to
> login as root, see if this works. if not - you'll be able to use
> the still open root shell to change root to using the previous
> shell again (you did check which shell this is before making the
> switch - didn't you?).
>
> otherwise, you'll have to read the source code of 'init', in order
> to find which shell it uses for single-user mode. if this is a
> hrd-coded value, you're out of luck (unless you install a modified
> 'init' binary on your system)
>
>
>
>> 2. We have to package our project using RPM. Does anyone have any
>> simple sample for spec file that we can use?
>>
>>
>
> did you try searching for one on the internet? if not, now is the
> time...
>
>
>
--
Haifa Linux Club Mailing List (http://www.haifux.org)
To unsub send an empty message to [EMAIL PROTECTED]
