netnice-kernels — a ML for netnice kernel module and its porting

Re: [Netnice-kernels] Linux Kernel

[Scott B. <sco...@ve...>]

Hi Kartikey,

Thanks for the new patch, I have applied it and am running it now.
I did have to disable vif_fork and the changes in ip_output.c in
order to run the kernel here, I also needed to manually update
kernel/exit.c.

Regards


Scott B

Re: [Netnice-kernels] Linux Kernel

[Takashi O. <ta...@cs...>]

thanks kartikey,

and it's good to know that you feel okay now.

but, where's the file...? i think it is not attached.


-- taka

p.s.

i have to apologize to you. today, i realized that a guy i asked to
send you the additional supplies did not send them to you. i'm
asking him to ship everything asap. so, if you have changed your
mailing address, please let me know. sorry...


kartikey bhatt wrote:
> 
> hi taka,
> 
> i am feeling alright now. thanks.
> i dont know what is the problem with
> online cvs repository. but here i have
> regenerated the patch using typical -urN
> option, bzipped it and sending it again.
> to apply just go to the fresh 2.6.7 directory
> and issue comman bzip2 -dc netnice-linux.diff.bz2 | patch -p1 --dry-run.
> 
> regards,
> kartikey

Re: [Netnice-kernels] Linux Kernel

[kartikey b. <kar...@ho...>]

hi taka,

i am feeling alright now. thanks.
i dont know what is the problem with
online cvs repository. but here i have
regenerated the patch using typical -urN
option, bzipped it and sending it again.
to apply just go to the fresh 2.6.7 directory
and issue comman bzip2 -dc netnice-linux.diff.bz2 | patch -p1 --dry-run.

regards,
kartikey

>From: Takashi Okumura <ta...@cs...>
>To: kartikey bhatt <kar...@ho...>
>CC: sco...@ve...,net...@li...
>Subject: Re: [Netnice-kernels] Linux Kernel
>Date: Thu, 07 Oct 2004 17:12:04 -0400
>
>hi kartikey!
>
>thanks for the patch. how're you feeling now??
>
>i don't know why the code on the CVS isn't the latest one,
>but, that is okay.
>
>i also tried the patch, and got the same error as Scott did.
>
> > patch: **** malformed patch at line 14: };
>
>what option do i need to give to the patch command??
>i'm using patch 2.5.4, and i've tried -p0, -p1, -p2, -p3
>options in the target directory, but, they didn't work.
>
>any thoughts?
>
>thanks!
>
>-- taka
>
>kartikey bhatt wrote:
> >
> > hello,
> >
> > may be here comes the answer to all your
> > questions. this is the patch that i worked
> > on and where everything is in place. it
> > compiles neatly without any modifications.
> > just apply it against fresh 2.6.7 tree.
> > and from make menuconfig select
> > drivers->networking options->Netnice for Linux.
> >
> > regards
> > --kartikey
> >
> > >From: Takashi Okumura <ta...@cs...>
> > >To: Scott Brumbaugh <sco...@ve...>
> > >CC: net...@li...
> > >Subject: Re: [Netnice-kernels] Linux Kernel
> > >Date: Wed, 06 Oct 2004 13:59:18 -0400
> > >
> > >hello,
> > >
> > >Scott Brumbaugh wrote:
> > > >
> > > > > thanks for the message. i realized that vif_free() and some of
> > > > > other functions are still incomplete. i don't know why it doesn't
> > > > > compile, since i've checked that it compiled (in india!)...
> > > > >
> > > >
> > > > >From the CVS the files vif_input.c, vif_output.c, and vif_subr.c 
>will
> > > > pass through the compiler and an object file will be produced.
> > > > However, the conditional at the top:
> > > >
> > > > #ifdef CONFIG_NETNICE
> > > >
> > > > will prevent any of the contained functions from being compiled.  So
> > > > the object file produced is really not of any use because it 
>contains
> > > > no executable code.  The definition of CONFIG_NETNICE is obtained by
> > > > including linux/config.h and the #ifdef CONFIG_NETNICE must occur
> > > > _after_ linux/config.h has been included.  When this is done and the
> > > > functions in the files actually pass through the compiler the errors
> > > > occur.
> > > >
> > > > As for socket.c some of the variables passed into the functions 
>within
> > > > the sections delimited by CONFIG_NETNICE use variables that are not
> > > > found anywhere in the file.
> > >
> > >okay. i took a look at the files, and now i understand. the last case
> > >is the easiest; it is using a pointer, "struct proc *p", although linux
> > >doesn't have that variable in the socket accept function. maybe you're
> > >right. but, still i think it is relatively easy to fix...
> > >
> > > > Also, the Kconfig (kernel configuration) setup is not complete and I
> > > > needed to add this so I could compile.
> > > >
> > > > If this worked before I imagine that there is a later version 
>missing
> > > > from your CVS.
> > >
> > >i checked that the files on the CVS is the ones i got from kartikey.
> > >
> > >kartikey, do you have newer files? or, any comments??
> > >
> > >thanks,
> > >
> > >-- taka
> > >
> > >
> > >
> > >-------------------------------------------------------
> > >This SF.net email is sponsored by: IT Product Guide on 
>ITManagersJournal
> > >Use IT products in your business? Tell us what you think of them. Give 
>us
> > >Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out 
>more
> > >http://productguide.itmanagersjournal.com/guidepromo.tmpl
> > >_______________________________________________
> > >Netnice-kernels mailing list
> > >Net...@li...
> > >https://lists.sourceforge.net/lists/listinfo/netnice-kernels
> >
> > _________________________________________________________________
> > All the news that matters. All the gossip from home.
> > http://www.msn.co.in/NRI/ Specially for NRIs!
> >
> >   
>--------------------------------------------------------------------------------
> >
> >                             Name: netnice-linux.diff
> >    netnice-linux.diff       Type: Plain Text (text/plain)
> >                         Encoding: 8bit
>

_________________________________________________________________
1 Paisa Auction with no reserve price. 
http://ads2.baazee.com/cgi-bin/banners/redirect.pl?id=2563 Steal a Deal - 
Nokia phones.

Re: [Netnice-kernels] Linux Kernel

[Scott B. <sco...@ve...>]

On Thu, 7 Oct 2004, Takashi Okumura wrote:

> hi kartikey!
> 
> thanks for the patch. how're you feeling now?? 
> 
> i don't know why the code on the CVS isn't the latest one,
> but, that is okay.
> 
> i also tried the patch, and got the same error as Scott did.
> 
> > patch: **** malformed patch at line 14: };
> 
> what option do i need to give to the patch command??
> i'm using patch 2.5.4, and i've tried -p0, -p1, -p2, -p3
> options in the target directory, but, they didn't work.
> 
> any thoughts?
> 
> thanks!
> 
> -- taka
> 

Hi,

I am pretty sure that the patch file attachment was corrupted by an
email program.  The error message about the malformed patch on line 14
occurs because there is a space character missing from the start of
the line.  I reformatted the first hunk, adding the spaces where they
were needed and it applied ok.  Notice also that the hunk headers have
line breaks added to them.

Here is a cut and paste of the area of the patch file that causes the
error message and a reformatted version that applies. The original
patch looks like so, notice hat the '}' character is directly under
the '+' above it.  so it looks like this:


 +
 +#ifdef CONFIG_NETNICE
 +      caddr_t                 so_vifnet;
 +#endif
 };

 struct vm_area_struct;

This should look like the following in the unified diff format. Notice
that there is a blank space before the '}':

 +
 +#ifdef CONFIG_NETNICE
 +      caddr_t                 so_vifnet;
 +#endif
  };

  struct vm_area_struct;

I think Some email programs will remove a single blank space at the
start of a line. Patch needs these blank spaces.


Thanks,


Scott Brumbaugh

Re: [Netnice-kernels] Linux Kernel

[Takashi O. <ta...@cs...>]

hi kartikey!

thanks for the patch. how're you feeling now?? 

i don't know why the code on the CVS isn't the latest one,
but, that is okay.

i also tried the patch, and got the same error as Scott did.

> patch: **** malformed patch at line 14: };

what option do i need to give to the patch command??
i'm using patch 2.5.4, and i've tried -p0, -p1, -p2, -p3
options in the target directory, but, they didn't work.

any thoughts?

thanks!

-- taka

kartikey bhatt wrote:
> 
> hello,
> 
> may be here comes the answer to all your
> questions. this is the patch that i worked
> on and where everything is in place. it
> compiles neatly without any modifications.
> just apply it against fresh 2.6.7 tree.
> and from make menuconfig select
> drivers->networking options->Netnice for Linux.
> 
> regards
> --kartikey
> 
> >From: Takashi Okumura <ta...@cs...>
> >To: Scott Brumbaugh <sco...@ve...>
> >CC: net...@li...
> >Subject: Re: [Netnice-kernels] Linux Kernel
> >Date: Wed, 06 Oct 2004 13:59:18 -0400
> >
> >hello,
> >
> >Scott Brumbaugh wrote:
> > >
> > > > thanks for the message. i realized that vif_free() and some of
> > > > other functions are still incomplete. i don't know why it doesn't
> > > > compile, since i've checked that it compiled (in india!)...
> > > >
> > >
> > > >From the CVS the files vif_input.c, vif_output.c, and vif_subr.c will
> > > pass through the compiler and an object file will be produced.
> > > However, the conditional at the top:
> > >
> > > #ifdef CONFIG_NETNICE
> > >
> > > will prevent any of the contained functions from being compiled.  So
> > > the object file produced is really not of any use because it contains
> > > no executable code.  The definition of CONFIG_NETNICE is obtained by
> > > including linux/config.h and the #ifdef CONFIG_NETNICE must occur
> > > _after_ linux/config.h has been included.  When this is done and the
> > > functions in the files actually pass through the compiler the errors
> > > occur.
> > >
> > > As for socket.c some of the variables passed into the functions within
> > > the sections delimited by CONFIG_NETNICE use variables that are not
> > > found anywhere in the file.
> >
> >okay. i took a look at the files, and now i understand. the last case
> >is the easiest; it is using a pointer, "struct proc *p", although linux
> >doesn't have that variable in the socket accept function. maybe you're
> >right. but, still i think it is relatively easy to fix...
> >
> > > Also, the Kconfig (kernel configuration) setup is not complete and I
> > > needed to add this so I could compile.
> > >
> > > If this worked before I imagine that there is a later version missing
> > > from your CVS.
> >
> >i checked that the files on the CVS is the ones i got from kartikey.
> >
> >kartikey, do you have newer files? or, any comments??
> >
> >thanks,
> >
> >-- taka
> >
> >
> >
> >-------------------------------------------------------
> >This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
> >Use IT products in your business? Tell us what you think of them. Give us
> >Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
> >http://productguide.itmanagersjournal.com/guidepromo.tmpl
> >_______________________________________________
> >Netnice-kernels mailing list
> >Net...@li...
> >https://lists.sourceforge.net/lists/listinfo/netnice-kernels
> 
> _________________________________________________________________
> All the news that matters. All the gossip from home.
> http://www.msn.co.in/NRI/ Specially for NRIs!
> 
>   --------------------------------------------------------------------------------
> 
>                             Name: netnice-linux.diff
>    netnice-linux.diff       Type: Plain Text (text/plain)
>                         Encoding: 8bit

Re: [Netnice-kernels] Linux Kernel

[Scott B. <sco...@ve...>]

On Thu, 7 Oct 2004, kartikey bhatt wrote:

> hello,
> 
> may be here comes the answer to all your
> questions. this is the patch that i worked
> on and where everything is in place. it
> compiles neatly without any modifications.
> just apply it against fresh 2.6.7 tree.
> and from make menuconfig select
> drivers->networking options->Netnice for Linux.
> 
> 
> regards
> --kartikey
> 

Hi,

Thanks for the patch.  According to your note the patch
compiles, but does it run and does netnice function?

Also, I am having difficulty applying this patch, perhaps
hotmail has mangled the attachment.  If you are able to
apply the patch on your end could show me the command that
you use:

[scottb@tamarack linux-2.6.7]$ pwd
/home/scottb/netnice/linux-2.6.7
[scottb@tamarack linux-2.6.7]$ patch --verbose -p2 --dry-run < ../netnice-linux.diff
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|diff -uprN ../home/cloud9/linux/include/linux/net.h
|./linux-patched/include/linux/net.h
|--- ../home/cloud9/linux/include/linux/net.h   2004-09-26 01:39:02.000000000
|+1000
|+++ ./linux-patched/include/linux/net.h        2004-10-07 14:07:05.000000000 +1000
--------------------------
Patching file include/linux/net.h using Plan A...
patch: **** malformed patch at line 14: };

[scottb@tamarack linux-2.6.7]$


Thanks,

Scott Brumbaugh

Re: [Netnice-kernels] Linux Kernel

[kartikey b. <kar...@ho...>]

hello,

may be here comes the answer to all your
questions. this is the patch that i worked
on and where everything is in place. it
compiles neatly without any modifications.
just apply it against fresh 2.6.7 tree.
and from make menuconfig select
drivers->networking options->Netnice for Linux.


regards
--kartikey

>From: Takashi Okumura <ta...@cs...>
>To: Scott Brumbaugh <sco...@ve...>
>CC: net...@li...
>Subject: Re: [Netnice-kernels] Linux Kernel
>Date: Wed, 06 Oct 2004 13:59:18 -0400
>
>hello,
>
>Scott Brumbaugh wrote:
> >
> > > thanks for the message. i realized that vif_free() and some of
> > > other functions are still incomplete. i don't know why it doesn't
> > > compile, since i've checked that it compiled (in india!)...
> > >
> >
> > >From the CVS the files vif_input.c, vif_output.c, and vif_subr.c will
> > pass through the compiler and an object file will be produced.
> > However, the conditional at the top:
> >
> > #ifdef CONFIG_NETNICE
> >
> > will prevent any of the contained functions from being compiled.  So
> > the object file produced is really not of any use because it contains
> > no executable code.  The definition of CONFIG_NETNICE is obtained by
> > including linux/config.h and the #ifdef CONFIG_NETNICE must occur
> > _after_ linux/config.h has been included.  When this is done and the
> > functions in the files actually pass through the compiler the errors
> > occur.
> >
> > As for socket.c some of the variables passed into the functions within
> > the sections delimited by CONFIG_NETNICE use variables that are not
> > found anywhere in the file.
>
>okay. i took a look at the files, and now i understand. the last case
>is the easiest; it is using a pointer, "struct proc *p", although linux
>doesn't have that variable in the socket accept function. maybe you're
>right. but, still i think it is relatively easy to fix...
>
> > Also, the Kconfig (kernel configuration) setup is not complete and I
> > needed to add this so I could compile.
> >
> > If this worked before I imagine that there is a later version missing
> > from your CVS.
>
>i checked that the files on the CVS is the ones i got from kartikey.
>
>kartikey, do you have newer files? or, any comments??
>
>thanks,
>
>-- taka
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
>Use IT products in your business? Tell us what you think of them. Give us
>Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
>http://productguide.itmanagersjournal.com/guidepromo.tmpl
>_______________________________________________
>Netnice-kernels mailing list
>Net...@li...
>https://lists.sourceforge.net/lists/listinfo/netnice-kernels

_________________________________________________________________
All the news that matters. All the gossip from home. 
http://www.msn.co.in/NRI/ Specially for NRIs!

Re: [Netnice-kernels] Linux Kernel

[Takashi O. <ta...@cs...>]

hello,

Scott Brumbaugh wrote:
>
> > thanks for the message. i realized that vif_free() and some of
> > other functions are still incomplete. i don't know why it doesn't
> > compile, since i've checked that it compiled (in india!)...
> >
> 
> >From the CVS the files vif_input.c, vif_output.c, and vif_subr.c will
> pass through the compiler and an object file will be produced.
> However, the conditional at the top:
> 
> #ifdef CONFIG_NETNICE
> 
> will prevent any of the contained functions from being compiled.  So
> the object file produced is really not of any use because it contains
> no executable code.  The definition of CONFIG_NETNICE is obtained by
> including linux/config.h and the #ifdef CONFIG_NETNICE must occur
> _after_ linux/config.h has been included.  When this is done and the
> functions in the files actually pass through the compiler the errors
> occur.
> 
> As for socket.c some of the variables passed into the functions within
> the sections delimited by CONFIG_NETNICE use variables that are not
> found anywhere in the file.

okay. i took a look at the files, and now i understand. the last case
is the easiest; it is using a pointer, "struct proc *p", although linux
doesn't have that variable in the socket accept function. maybe you're
right. but, still i think it is relatively easy to fix...

> Also, the Kconfig (kernel configuration) setup is not complete and I
> needed to add this so I could compile.
> 
> If this worked before I imagine that there is a later version missing
> from your CVS.

i checked that the files on the CVS is the ones i got from kartikey.

kartikey, do you have newer files? or, any comments??

thanks,

-- taka

[Netnice-kernels] Linux Kernel

[Scott B. <sco...@ve...>]

On Tue, 5 Oct 2004, Takashi Okumura wrote:

> hi,
> 
> thanks for the message. i realized that vif_free() and some of
> other functions are still incomplete. i don't know why it doesn't
> compile, since i've checked that it compiled (in india!)...
> 

From the CVS the files vif_input.c, vif_output.c, and vif_subr.c will
pass through the compiler and an object file will be produced.
However, the conditional at the top:

#ifdef CONFIG_NETNICE

will prevent any of the contained functions from being compiled.  So
the object file produced is really not of any use because it contains
no executable code.  The definition of CONFIG_NETNICE is obtained by
including linux/config.h and the #ifdef CONFIG_NETNICE must occur
_after_ linux/config.h has been included.  When this is done and the
functions in the files actually pass through the compiler the errors
occur.

As for socket.c some of the variables passed into the functions within
the sections delimited by CONFIG_NETNICE use variables that are not
found anywhere in the file.

Also, the Kconfig (kernel configuration) setup is not complete and I
needed to add this so I could compile.

If this worked before I imagine that there is a later version missing
from your CVS.



Regards,


Scott Brumbaugh

[Netnice-kernels] Re: Linux kernel compilation

[Takashi O. <ta...@cs...>]

quick comment.

i've been thinking that netnice267 compiles. kartikey, any comments??


-- taka

Scott Brumbaugh wrote:
> 
> Hi Takashi,
> 
> I wanted to let you know where I am in the development.  In order to
> start the port I need to be able to run a kernel that can at least
> install a root vif on each real network interface.  I found that the
> 2.6.7 kernel in CVS with the tag netnice267 will not compile or link.
> I can tell that the internals have not been completed.
> 
> Nevertheless, I think all I need to get started is a working
> vif_attach that can be called when the kernel installs a real
> interface.  In working toward that I have spent today hacking on the
> cvs version and getting it to compile and link.
> 
> I don't think I need the the modified sk_buff structures yet nor do I
> need process structures or vif_fork.  I think all I need at this point
> is a working vif_attach (called vif_register in the linux cvs) and
> vif_alloc to start nnfs.
> 
> I have attached a context diff of the changes that I have made to
> netnice267 to get it to compile and link.  I will try to get it to run
> with vif_attach early next week and add a fs/nnfs directory and then
> submit that to your cvs.
> 
> Thanks,
> 
> Scott Brumbaugh
> 
>   --------------------------------------------------------------------------------
> 
>                        Name: nn.dif.txt
>    nn.dif.txt          Type: Plain Text (TEXT/plain)
>                    Encoding: BASE64
>                 Description: nndiff

[Netnice-kernels] new member

[Scott B. <sco...@ve...>]

Hi,

I am working on the Linux VFS based control interface to netnice.
Basically implementing NNFS that appears in FreeBSD5 in Linux.  I have
done a lot of professional Linux work in the, past mainly in user
space but I have also done a couple kernel modules and I am looking
forward to getting a lot of good experience in the kernel with this
project. I expect the Linux nnfs implementation to take about a month
to complete.

I live in California and have a small business web site where I
advertise locally at http://www.inyotech.com


Regards,


Scott B

[Netnice-kernels] Linux kernel compilation

[Scott B. <sco...@ve...>]

Hi Takashi,

I wanted to let you know where I am in the development.  In order to
start the port I need to be able to run a kernel that can at least
install a root vif on each real network interface.  I found that the
2.6.7 kernel in CVS with the tag netnice267 will not compile or link.
I can tell that the internals have not been completed.

Nevertheless, I think all I need to get started is a working
vif_attach that can be called when the kernel installs a real
interface.  In working toward that I have spent today hacking on the
cvs version and getting it to compile and link.

I don't think I need the the modified sk_buff structures yet nor do I
need process structures or vif_fork.  I think all I need at this point
is a working vif_attach (called vif_register in the linux cvs) and
vif_alloc to start nnfs.

I have attached a context diff of the changes that I have made to
netnice267 to get it to compile and link.  I will try to get it to run
with vif_attach early next week and add a fs/nnfs directory and then
submit that to your cvs.



Thanks,


Scott Brumbaugh

Re: [Netnice-kernels] NNFS was: Re: Sourceforge Account

[Takashi O. <ta...@cs...>]

Hi Scott,

thanks for the update info.

Scott Brumbaugh wrote:
>
> Been using the Boot CD and experimenting with the fundamentals of the nnfs
> filesystem using the /proc/network/lo0/vif1, setting a symlink in
> /proc/curproc/sockets and using telnet to localhost.

sounds good. although we're going to add a library layer, libnetnice, on
top of the low level procfs-like API, it is definitely helpful to get
the feel. you may find our research paper helpful, also, to  understand
its behavior and the internal algorithm.

T. Okumura and D. Mosse', "Virtualizing Network I/O on End-host Operating
System: Operating System Support for Network Control and Resource Protection",
IEEE Transactions on Computers, Oct 2004. 

> I think I should proceed with the mkdir/rmdir operation first, I think
> this is relatively straight forward and maybe a small piece of the
> control file implementation.  Next should come the implementation of
> the <pid>/sockets directory, the code here will come from the existing
> procfs code.  After that would be the symlink operation between the
> socket and the and the vif, I suspect that this may be the most
> difficult.

the approach sounds reasonable. please let us know if you need further help.

> I will be going out of town tomorrow and will return Tuesday.  Looking
> forward to a conference call next week when I return.

alright. have a good trip :-)


-- taka

[Netnice-kernels] A new member

[Takashi O. <ta...@cs...>]

dear netnice hackers,

as you might have realized, a new member, Mr. Scott Brumbaugh, has joined us
for the porting of the procfs API. he's running a consultation business
(right?) in California. i would appreciate your help, when he needs some.

Scott, can you add any missing information, or more preferably, send a
self-introduction message to the list?

thanks!

-- taka

Re: [Netnice-kernels] NNFS was: Re: Sourceforge Account

[Scott B. <sco...@ve...>]

Hi Takashi,

> 
> my point is this: please use the CDROM system, to learn how it works. it
> is a working specification, you need to use for your linux-NNFS. but, to
> study how it is implemented on FreeBSD, please use the FreeBSD5/NNFS
> implementation, by checking out the source from our CVS repository.
> otherwise, you'll have a hard time to understand the messy implementation.
> (i think you're on the right track about this, right? since you're mentioning
> nnfs_vncache_alloc()... just in case.)
> 
> http://cvs.sourceforge.net/viewcvs.py/netnice/FreeBSD5/?only_with_tag=netnice53
> 
> thanks!!
> 

Thanks, I was working from the FreeBSD5 branch with the netnice521 tag.  I 
have since done cvs update -r netnice53 and will reference these files in 
the future.  

Been using the Boot CD and experimenting with the fundamentals of the nnfs
filesystem using the /proc/network/lo0/vif1, setting a symlink in
/proc/curproc/sockets and using telnet to localhost.

I am getting more comfortable with the internals. The VFS interface is
different under Linux than FreeBSD in that Linux VFS is more OO.  On
FreeBSD NNFS there are a lot of functions with switch statements
branching on the type of nnfs node (root vif, vif, virtual file,
etc...) .  On Linux VFS the different inodes (vnodes on BSD) each
contain function pointers which differ depending on the inode type.  I
think this will make the code smaller in the end.  

I think I should proceed with the mkdir/rmdir operation first, I think
this is relatively straight forward and maybe a small piece of the
control file implementation.  Next should come the implementation of
the <pid>/sockets directory, the code here will come from the existing
procfs code.  After that would be the symlink operation between the
socket and the and the vif, I suspect that this may be the most
difficult.

I will be going out of town tomorrow and will return Tuesday.  Looking
forward to a conference call next week when I return.


Thanks,


Scott Brumbaugh

Re: [Netnice-kernels] NNFS was: Re: Sourceforge Account

[Takashi O. <ta...@cs...>]

hi scott,

Scott Brumbaugh wrote:
> 
> Hi Takashi,
> 
> On Tue, 28 Sep 2004, Takashi Okumura wrote:
> 
> > hi scott,
> >
> > i still don't understand your question. but, i think it is a good
> > idea to try our bootable CD to know how it works. this is definitely
> > a better specification of the API, better than the simplified description
> > of the page you mentioned.
> 
> Thanks for the pointer, I had not actually run netnice. Now, I have
> downloaded the ISO and run it excersizing netnice by doing telnet
> localhost and examining in /proc/network and /proc/<pid>/sockets/.
> Before I was working from the docs and from the FreeBSD5 source.
> 
> > but, i don't think it is a right way, to create a file system which
> > contains just the directories under /proc/network of FreeBSD and
> > mount it under /proc/network of linux-procfs. we need to setup a lot
> > of links between process-dependent directories under /proc/PIDs and
> > VIF directories under /proc/network, and thus, they cannot be separated
> > that way, i believe.
> 
> Considering further and with your reply, I can see the point.  The
> code that implements the interaction between the /proc/<pid>/sockets
> entries and the vif interfaces under /proc/network/<interface>/<vif>
> needs to be in the same kernel module.

this is correct, i believe.

> I was originally thinking that
> /proc/<pid>/sockets could be created dynamically from within the nnfs
> kernel module (I think linux procfs may have this feature) but it
> would probably be difficult and error prone.

so, it's the source of the confusion. BSD procfs do not have such a
dynamic feature. we need to hardcode everything...

the approach of Linux-procfs sounds reasonable, because it's easier to
extend the feature (support more directories and functionality). but,
since we do not need to make it too generic, i think BSD approach is
better, in respect to its management overhead. we need to think about
our approach...

> I think the linux /proc/<pid> implementation is mostly contained in
> the functions at the end of the file linux/fs/proc/base.c -
> proc_pid_readdir, and proc_task_readdir.  I will need to investigate
> further but at this point does not look too complex to extract and
> port to nnfs.

this sounds good. i think it is preferable to support them. but, please
keep in mind that it is still an open problem, how much of the linux-procfs
functionality we need to support in NNFS.


> > i guess that the confusion is mostly coming from the difference of
> > organizations between FreeBSD-procfs and linux-procfs. and, maybe
> > it is a good idea to have a tele-conference to avoid further confusion.
> > we have a polycom system in my department, and i use MSN messenger
> > also (ta...@ho...).
> >
> 
> I would prefer a teleconference to IM initially.  For efficiency I
> would suggest we try next week maybe Wednesday or Thursday as I will
> be out of town this weekend and I want to study study up on the system
> some more.

okay.

> As it is, I was about ready to move into implementing the mkdir
> operation on the virtual directories under /network.  I am studying
> the functions vif_alloc and nnfs_vncache_alloc as I think these are
> the most important here, I am thinking they should port with few
> changes.

btw, as you might have realized, procfs implementation of netnice API is
the most horrible part of the system. because of this, i've totally
re-written the code for FreeBSD 5/netnice (which is still a pre-alpha
version), as NNFS. the code works exactly the same way as the procfs of the
CDROM netnice (based on FreeBSD4.9), but, the whole system is not functional.

my point is this: please use the CDROM system, to learn how it works. it
is a working specification, you need to use for your linux-NNFS. but, to
study how it is implemented on FreeBSD, please use the FreeBSD5/NNFS
implementation, by checking out the source from our CVS repository.
otherwise, you'll have a hard time to understand the messy implementation.
(i think you're on the right track about this, right? since you're mentioning
nnfs_vncache_alloc()... just in case.)

http://cvs.sourceforge.net/viewcvs.py/netnice/FreeBSD5/?only_with_tag=netnice53

thanks!!


-- taka

Re: [Netnice-kernels] NNFS was: Re: Sourceforge Account

[Scott B. <sco...@ve...>]

Hi Takashi,

On Tue, 28 Sep 2004, Takashi Okumura wrote:

> hi scott,
> 
> i still don't understand your question. but, i think it is a good
> idea to try our bootable CD to know how it works. this is definitely
> a better specification of the API, better than the simplified description
> of the page you mentioned.

Thanks for the pointer, I had not actually run netnice. Now, I have
downloaded the ISO and run it excersizing netnice by doing telnet
localhost and examining in /proc/network and /proc/<pid>/sockets/.
Before I was working from the docs and from the FreeBSD5 source.

> but, i don't think it is a right way, to create a file system which
> contains just the directories under /proc/network of FreeBSD and
> mount it under /proc/network of linux-procfs. we need to setup a lot
> of links between process-dependent directories under /proc/PIDs and
> VIF directories under /proc/network, and thus, they cannot be separated
> that way, i believe.

Considering further and with your reply, I can see the point.  The
code that implements the interaction between the /proc/<pid>/sockets
entries and the vif interfaces under /proc/network/<interface>/<vif>
needs to be in the same kernel module.  I was originally thinking that
/proc/<pid>/sockets could be created dynamically from within the nnfs
kernel module (I think linux procfs may have this feature) but it
would probably be difficult and error prone.

I think the linux /proc/<pid> implementation is mostly contained in
the functions at the end of the file linux/fs/proc/base.c -
proc_pid_readdir, and proc_task_readdir.  I will need to investigate
further but at this point does not look too complex to extract and
port to nnfs.

> i guess that the confusion is mostly coming from the difference of
> organizations between FreeBSD-procfs and linux-procfs. and, maybe
> it is a good idea to have a tele-conference to avoid further confusion.
> we have a polycom system in my department, and i use MSN messenger
> also (ta...@ho...).
> 

I would prefer a teleconference to IM initially.  For efficiency I
would suggest we try next week maybe Wednesday or Thursday as I will
be out of town this weekend and I want to study study up on the system
some more.

As it is, I was about ready to move into implementing the mkdir
operation on the virtual directories under /network.  I am studying
the functions vif_alloc and nnfs_vncache_alloc as I think these are
the most important here, I am thinking they should port with few
changes.



Thanks,

Scott Brumbaugh

Re: [Netnice-kernels] NNFS was: Re: Sourceforge Account

[Takashi O. <ta...@cs...>]

hi scott,

i still don't understand your question. but, i think it is a good
idea to try our bootable CD to know how it works. this is definitely
a better specification of the API, better than the simplified description
of the page you mentioned.

	http://www.netnice.org/download.cgi

i personally think the problem of NNFS on Linux is that we may not need
all of the features of linux-procfs (or, contents under /proc). this is
because, although i'm thinking to mount NNFS at /proc and thinking to provide
capability of linux-procfs, it might not be the best solution in respect
to the amount of coding we need. we may simply mount nnfs at /network
or somewhere else, not sticking to the procfs compatibility. so, that
is a open problem, and i'm happy to discuss the issue.

but, i don't think it is a right way, to create a file system which
contains just the directories under /proc/network of FreeBSD and
mount it under /proc/network of linux-procfs. we need to setup a lot
of links between process-dependent directories under /proc/PIDs and
VIF directories under /proc/network, and thus, they cannot be separated
that way, i believe.

i guess that the confusion is mostly coming from the difference of
organizations between FreeBSD-procfs and linux-procfs. and, maybe
it is a good idea to have a tele-conference to avoid further confusion.
we have a polycom system in my department, and i use MSN messenger
also (ta...@ho...).

i'm sorry that i'm not answering to all of questions, but, i believe
that it is a good starting point to try the working system first,
and then, to think how we realize a similar feature on Linux.

hope this helps, and i would apologize if you've already tried the
FreeBSD/Netnice.


-- taka

Scott Brumbaugh wrote:
>
> There is an illustration and text description under the heading 'File
> System Representation' .  Mounted directly under /proc is the nnfs
> network/ directory which holds the VIF interfaces themselves.  However
> under /proc are directories corresponding to the PIDs of running
> processes and under them are sockets/ directories that contain links
> to the VIF interfaces under network/.
> 
> I am new to this project so my understanding may be wrong, but to
> create this structure on linux my idea is to create the nnfs network/
> directory and mount it and then modify the existing /proc adding a
> sockets/ virtual directory under each PID and under each sockets/
> directory establish links to the VIF interface under network/ .
> 
> >From the text description on the web page I read:
> 
>     A directory under /proc/network/ represents a VIF. ...
> 
> Then, in the next paragraph:
> 
>     Directories under /proc/[pid]/ are process entries, as usual. We
>     added a sockets directory under each process entry. ...
> 
> On Linux, /proc/[pid] exists and it has been my intention to figure
> out if a sockets directory can be added under each one.  Now maybe
> this will not be possible due to the design of linux procfs but at
> this point I have not seen that it is not possible to create this
> sockets directory.
> 
> Now, from your response, I am reasoning that your approach here would
> be to copy the code from linux procfs that exposes the PIDS and put
> that code into the new nnfs thus creating a directory structure under
> the new nnfs that contains both the VIFS and the PIDs that are using
> the VIFS.  This structure is different from the illustration and the
> high level description of the API as far as I can tell.  However, I am
> just starting out here and trying to get up to speed quickly so I am
> perhaps rushing my understanding.

[Netnice-kernels] NNFS was: Re: Sourceforge Account

[Scott B. <sco...@ve...>]

On Tue, 28 Sep 2004, Takashi Okumura wrote:

> hi,
> 
> thanks for the message. i feel better by a message like that, because
> i can sense your progress :-)
> 
> > I will need to see if I can simply modify the existing /proc in order
> > to support the socket files without having to add process ids to nnfs.
> 
> i actually do not understand your question. you may reuse code from 
> linux-procfs, as much as you want, as log as your NNFS follows the
> specification of NNFS.
> 
> you may send the question to the kernel ML, because others may have
> good answers.
> 
> thanks!
> 
> 
> -- taka
> 

Hi Takashi,

When I ask this question I am referring to the API description on the
web page at http://www.netnice.org/pukiwiki-e.php?API%20overview .
There is an illustration and text description under the heading 'File
System Representation' .  Mounted directly under /proc is the nnfs
network/ directory which holds the VIF interfaces themselves.  However
under /proc are directories corresponding to the PIDs of running
processes and under them are sockets/ directories that contain links
to the VIF interfaces under network/.

I am new to this project so my understanding may be wrong, but to
create this structure on linux my idea is to create the nnfs network/
directory and mount it and then modify the existing /proc adding a
sockets/ virtual directory under each PID and under each sockets/
directory establish links to the VIF interface under network/ .

From the text description on the web page I read:

    A directory under /proc/network/ represents a VIF. ...

Then, in the next paragraph:

    Directories under /proc/[pid]/ are process entries, as usual. We
    added a sockets directory under each process entry. ...
 
On Linux, /proc/[pid] exists and it has been my intention to figure
out if a sockets directory can be added under each one.  Now maybe
this will not be possible due to the design of linux procfs but at
this point I have not seen that it is not possible to create this
sockets directory.

Now, from your response, I am reasoning that your approach here would
be to copy the code from linux procfs that exposes the PIDS and put
that code into the new nnfs thus creating a directory structure under
the new nnfs that contains both the VIFS and the PIDs that are using
the VIFS.  This structure is different from the illustration and the
high level description of the API as far as I can tell.  However, I am
just starting out here and trying to get up to speed quickly so I am
perhaps rushing my understanding.

I am CC'ing sending this reply to netnice-kernels so perhaps someone
there can set me straight.


Thanks,


Scott Brumbaugh

Re: [Netnice-kernels] OpenBSD status

[Takashi O. <ta...@cs...>]

hi,

just a status update.

fujita-kun has finished most of the procfs debugging, and left pittsburgh
for japan. he will start debugging vif_*.c shortly, and will release
an alpha version, by the end of the month.

i would be really happy if you post a short description of your status,
to the ML. kartikey, please let me know your status and future schedule.

thanks!

-- taka

Takashi Okumura wrote:
> 
> just a status update, on behalf of Fujita-kun.
> 
> Fujita-kun is finishing /proc. i hope that we release OpenBSD/Netnice alpha
> by the end of the week.
> 
> i would be really happy if you post a short description of your status,
> to the ML.
> 
> thanks!
> 
> -- taka
> 
> Takashi Okumura wrote:
> >
> > kuroki-san, and fujita-kun; i would be happy if you update the page
> > to share your plan with others interested in your code.
> >
> > and, if possible, please let me know (reply to the mailing list) about
> > the status of the porting. i think fujita-kun will release the alpha
> > version for OpenBSD next week :-)
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Netnice-kernels mailing list
> Net...@li...
> https://lists.sourceforge.net/lists/listinfo/netnice-kernels

[Netnice-kernels] OpenBSD status

[Takashi O. <ta...@cs...>]

just a status update, on behalf of Fujita-kun.

Fujita-kun is finishing /proc. i hope that we release OpenBSD/Netnice alpha
by the end of the week.

i would be really happy if you post a short description of your status,
to the ML.

thanks!

-- taka

Takashi Okumura wrote:
> 
> kuroki-san, and fujita-kun; i would be happy if you update the page
> to share your plan with others interested in your code.
> 
> and, if possible, please let me know (reply to the mailing list) about
> the status of the porting. i think fujita-kun will release the alpha
> version for OpenBSD next week :-)

Re: [Netnice-kernels] Homepage Renewal

[Takashi O. <ta...@cs...>]

thanks kartikey for the updating.

kuroki-san, and fujita-kun; i would be happy if you update the page
to share your plan with others interested in your code.

and, if possible, please let me know (reply to the mailing list) about
the status of the porting. i think fujita-kun will release the alpha
version for OpenBSD next week :-)


thanks!

-- taka

Takashi Okumura wrote:
> 
> Dear netnice kernel developers,
> 
> i have just renewed our project home page. in this update, i have entirely
> rewritten the contents using the wiki technology, and now, you can create and
> maintain your own page. i have setup simple template pages for you. so,
> please put your porting schedule and other related information, such as
> request for testers, and use the page for communication with your audiences.
> to edit a page, click on an edit icon at the bottom of the page.
> 
>         OpenBSD         http://www.netnice.org/pukiwiki-e.php?OpenBSD
>         NetBSD          http://www.netnice.org/pukiwiki-e.php?NetBSD
>         Linux           http://www.netnice.org/pukiwiki-e.php?Linux
> 
> thank you for your cooperation and contributions to the project.
> 
> thanks!!
> 
> -- taka
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by OSTG. Have you noticed the changes on
> Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
> one more big change to announce. We are now OSTG- Open Source Technology
> Group. Come see the changes on the new OSTG site. www.ostg.com
> _______________________________________________
> Netnice-kernels mailing list
> Net...@li...
> https://lists.sourceforge.net/lists/listinfo/netnice-kernels

[Netnice-kernels] Homepage Renewal

[Takashi O. <ta...@cs...>]

Dear netnice kernel developers,

i have just renewed our project home page. in this update, i have entirely
rewritten the contents using the wiki technology, and now, you can create and
maintain your own page. i have setup simple template pages for you. so,
please put your porting schedule and other related information, such as
request for testers, and use the page for communication with your audiences.
to edit a page, click on an edit icon at the bottom of the page.

	OpenBSD		http://www.netnice.org/pukiwiki-e.php?OpenBSD
	NetBSD		http://www.netnice.org/pukiwiki-e.php?NetBSD
	Linux		http://www.netnice.org/pukiwiki-e.php?Linux

thank you for your cooperation and contributions to the project.

thanks!!

-- taka

Re: [Netnice-kernels] Hi Linux

[Takashi O. <ta...@cs...>]

kartikey bhatt wrote:
> 
> Hi Linux
> You are just an inch away from final netnice patch

really!?


-- taka

[Netnice-kernels] Hi Linux

[kartikey b. <kar...@ho...>]

Hi Linux
You are just an inch away from final netnice patch

_________________________________________________________________
Claim your Citibank Ready Cash today. http://go.msnserver.com/IN/52041.asp 
It’s fast, easy and affordable.