etherboot-developers — Discussion of developer issues

Re: [Etherboot-developers] sgmltools

[<Gee...@xs...>]

On Sat, Jul 19, 2003 at 11:07:36PM +1000, Ken Yap wrote:
> >While playing with the documentation I noticed that some entities like 
> >_ appear literally in the output. I am using
> >
> >SGMLtools-Lite version 3.0.3
> >
> >Is this a known problem, or do I need another version of sgmltools?
> 
> I'm using sgmltools-lite 3.0.2. I see what you mean by the _ in
> the sgml files. This probably is due to no definition for this entity.
> The SGML files were created by automatic conversion of the ones on 5.0
> so it's quite possible the process is not perfect. For example there are
> still some unclosed tags.  Maybe _ should be called something
> else, &underscore;?  Anybody know the official name of _?
> 

underscore


IIRC it not a resevered character in SGML/XML,
so you can s/_/_/.


Geert Stappers

Re: [Etherboot-developers] Developer documentation

[<ke...@et...>]

>I just started to update the chapter "Writing an Etherboot Driver" for 5.2. 
>Could some expert please confirm or correct the following statements?
>
>- The reset routine in mydriver.c is gone

Yes, it seems that Eric made _disable the externally visible routine now.

>- For a new driver there is no need to edit any file besides the new .c file 
>in src/drivers/net and it will be automatically added to the build process.

Yes, genrules will grab the IDs (declared with the new macros) from the
.c file and create the Roms file that's included by arch/*/Makefile

>Tim, I think it would be nice to include your porting notes as well. I'll 
>also add some notes about keeping the etherboot and linux driver in sync.

And don't forget to add yourselves to the authors list.

Re: [Etherboot-developers] sgmltools

[<ke...@et...>]

>While playing with the documentation I noticed that some entities like 
>_ appear literally in the output. I am using
>
>SGMLtools-Lite version 3.0.3
>
>Is this a known problem, or do I need another version of sgmltools?

I'm using sgmltools-lite 3.0.2. I see what you mean by the _ in
the sgml files. This probably is due to no definition for this entity.
The SGML files were created by automatic conversion of the ones on 5.0
so it's quite possible the process is not perfect. For example there are
still some unclosed tags.  Maybe _ should be called something
else, &underscore;?  Anybody know the official name of _?

[Etherboot-developers] sgmltools

[Georg B. <gb...@us...>]

While playing with the documentation I noticed that some entities like 
_ appear literally in the output. I am using

SGMLtools-Lite version 3.0.3

Is this a known problem, or do I need another version of sgmltools?


Georg

[Etherboot-developers] Developer documentation

[Georg B. <gb...@us...>]

I just started to update the chapter "Writing an Etherboot Driver" for 5.2. 
Could some expert please confirm or correct the following statements?

- The reset routine in mydriver.c is gone
- For a new driver there is no need to edit any file besides the new .c file 
in src/drivers/net and it will be automatically added to the build process.


Tim, I think it would be nice to include your porting notes as well. I'll 
also add some notes about keeping the etherboot and linux driver in sync.

Georg

RE: [Etherboot-developers] tlan driver sucessfully boots ltsp

[Timothy L. <tl...@ro...>]

> It's a revision 48. It doesn't send any packets unfortunately. Also
one
> would expect a reply from the DHCP server if that were so, but none of
> that either. This is the command line I used:
> 

I believe that I may have the 48 as well.  I must check.  However, it
seems that there is a difference in the 35 and 48.  In the 35, it does
not look like an EOC is ever set in the tx status.  In the 48 it is.

I have managed to get both cards working but not with the same code and
I have yet to figure out how to integrate both sections of code so that
it works correctly.

A job for tomorrow night...

Tim

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003
 

Re: [Etherboot-developers] tlan driver sucessfully boots ltsp

[<ke...@et...>]

>> So autonegotiation works now. Unfortunately it doesn't send any
>packets
>> to the wire, according to tcpdump. It prints a . and then TX Time
>> Out..TRANSMIT: OC2326 is busy (Head=206304 Tail=1)
>
>If you check tcpdump I believe you will find that it does send one
>packet.  Which chip revision do you have (Its printed to the screen)?  I
>have a 32 that works fine but the 42? that I borrowed today does what
>you describe.  I am reviewing...

It's a revision 48. It doesn't send any packets unfortunately. Also one
would expect a reply from the DHCP server if that were so, but none of
that either. This is the command line I used:

$ tcpdump -lenx -s 1500 port bootps or port bootpc

RE: [Etherboot-developers] tlan driver sucessfully boots ltsp

[Timothy L. <tl...@ro...>]

> So autonegotiation works now. Unfortunately it doesn't send any
packets
> to the wire, according to tcpdump. It prints a . and then TX Time
> Out..TRANSMIT: OC2326 is busy (Head=206304 Tail=1)
> 

If you check tcpdump I believe you will find that it does send one
packet.  Which chip revision do you have (Its printed to the screen)?  I
have a 32 that works fine but the 42? that I borrowed today does what
you describe.  I am reviewing...

Tim

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003
 

Re: [Etherboot-developers] tlan driver sucessfully boots ltsp

[<ke...@et...>]

>I just committed a version of the tlan driver that can successfully load
>and boot the ltsp kernel.
>
>Now I need some testers and a way to work around the lack of an
>interrupt timer to remove the recursive-ness that Ken mentioned from the
>auto-negotiation.  

Well what do you know, it was the idiot at this end.

I had recently moved the cables and hadn't noticed that cable wasn't
plugged into the switch. Also I was looking at the wrong column of
lights at the switch.

So autonegotiation works now. Unfortunately it doesn't send any packets
to the wire, according to tcpdump. It prints a . and then TX Time
Out..TRANSMIT: OC2326 is busy (Head=206304 Tail=1)

I did a cvs update from the 5.1 tree but is it possible my version is
still not the latest? What I have are:

$ ls -l tlan.?
-rw-r--r--    1 ken      users       50912 2003-07-17 17:59 tlan.c
-rw-r--r--    1 ken      users       14738 2003-07-17 17:59 tlan.h

Re: *** GMX Spamverdacht *** Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) -FIXED CODE

[Anselm M. H. <an...@ho...>]

Hello Ken,

Thursday, July 17, 2003, 4:03:11 PM, you wrote:

> On further examination, I'm not happy with the way you have done it
> using recursion. Each recursive call of nfs consumes about 400 bytes on
> the stack. If you convert the recursion to iteration, you don't need to
> make recursion static and you don't need to waste 400 bytes of stack per
> level of indirection.

OK, I didn't get to my terminal till 1:30 AM and have to leave this
morning at 8:30, so no etherboot coding for me before monday.
It's not too hard, but it will have to wait.
Just to lower your expectations :-/

Have a nice weekend,
 Anselm

Re: *** GMX Spamverdacht *** Re: [Etherboot-developers] Loading kernel over NFS fails(solved!)-FIXED CODE

[Anselm M. H. <an...@ho...>]

> Hint: I won't be shocked if you use a goto instead of turning the loop
> inside out and risk breaking the existing algorithm.

My programming practice lector would kill you for suggesting a goto in C,
but that was what I had planned in the right place. I decided otherwise
more or less because of this doctrine of not using goto.

ttyul8r,

Anselm

Re: [Etherboot-developers] Loading kernel over NFS fails (solved!)-FIXED CODE

[<ke...@et...>]

>> the stack. If you convert the recursion to iteration, you don't need to
>> make recursion static and you don't need to waste 400 bytes of stack per
>> level of indirection.
>
>OK, that's an answer I can live with.
>I'll look into it later this afternoon, of course you are right.

Hint: I won't be shocked if you use a goto instead of turning the loop
inside out and risk breaking the existing algorithm.

Re: *** GMX Spamverdacht *** Re: [Etherboot-developers] Loading kernel over NFS fails (solved!)-FIXED CODE

[Anselm M. H. <an...@ho...>]

> On further examination, I'm not happy with the way you have done it
> using recursion. Each recursive call of nfs consumes about 400 bytes on
> the stack. If you convert the recursion to iteration, you don't need to
> make recursion static and you don't need to waste 400 bytes of stack per
> level of indirection.

OK, that's an answer I can live with.
I'll look into it later this afternoon, of course you are right.

Greetings,
Anselm

Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) -FIXED CODE

[<ke...@et...>]

>I'd not like that one because of something I'm thinking about right now.

On further examination, I'm not happy with the way you have done it
using recursion. Each recursive call of nfs consumes about 400 bytes on
the stack. If you convert the recursion to iteration, you don't need to
make recursion static and you don't need to waste 400 bytes of stack per
level of indirection.

Re: *** GMX Spamverdacht *** Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) -FIXED CODE

[Anselm M. H. <an...@ho...>]

> Two other ways of doing it:
>
> If there is initialisation code for NFS, make it a file scope static and
> set it to 0 there.
>
> Otherwise make it a global scope variable and set it in main() to 0 when
> starting up.

I'd not like that one because of something I'm thinking about right now.
I'd like a menu system I started working on recently to retrieve
additional data (e.g. background bitmap files, font files...) via tftp; or
NFS at user's discretion.
For me, a callback function in etherboot like
retrieve_file ( char * url, (void *)(block_management_function) would be
the ideal solution.
This would even open the possibility for e.g. harddisk restorage via
multicast tftp... ideas open. In any case, NFS would be called from
probably more than two points in etherboot code. Setting the recursion
variable only inside that function was the safest way to go.

Anselm

Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) - FIXED CODE

[<ke...@et...>]

>Setting recursion to 0 at all exit-points of NFS is probably not so
>nice, but should guarantee this bug won't bite us.
>If you agree, I'll commit; the change is in my local codebase yet.

Two other ways of doing it:

If there is initialisation code for NFS, make it a file scope static and
set it to 0 there.

Otherwise make it a global scope variable and set it in main() to 0 when
starting up.

Re: *** GMX Spamverdacht *** Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) - FIXED CODE

[Anselm M. H. <an...@ho...>]

Hello Ken,

> There's an obscure bug in your code waiting to bite someday. You declare
> "recursion" as a static int set to 0. Problem is if Etherboot restarts,
> this will not get reset to 0 and sometime later, when NFS access through
> symlinks is retried, the limit will be exceeded, even though it's a
> fresh attempt to load the file.

Setting recursion to 0 at all exit-points of NFS is probably not so
nice, but should guarantee this bug won't bite us.
If you agree, I'll commit; the change is in my local codebase yet.

Best regards,
 Anselm Martin Hoffmeister
 Stockholm Projekt Computer-Service
 <an...@ho...>

-- 
Disclaimer - These opiini^H^H damn! ^H^H ^Q ^[ .. :w :q :wq :wq! ^d X^?
exit X Q ^C ^c ^? :quitbye  CtrlAltDel  ~~q :~q logout save/quit :!QUIT
^[zz ^[ZZZZZZ ^H  man vi ^@ ^L ^[c ^# ^E ^X ^I ^T ? help helpquit ^D ^d
man help ^C exit ?Quit ?q CtrlShftDel "Hey, what does this button d..."

Re: [Etherboot-developers] Loading kernel over NFS fails (solved!) - FIXED CODE

[<ke...@et...>]

>Quick google'ing gave mention of 5 or 15 as symlinks depth, I think Linus
>didn't agree to upping the limit to 15 finally.
>I just introduced a constant NFS_MAXLINKDEPTH into CVS (in
>etherboot.h), set to 16 to go for sure. At least one thing we
>definitively are "better" than the Linux kernel ;-)

Anselm,

There's an obscure bug in your code waiting to bite someday. You declare
"recursion" as a static int set to 0. Problem is if Etherboot restarts,
this will not get reset to 0 and sometime later, when NFS access through
symlinks is retried, the limit will be exceeded, even though it's a
fresh attempt to load the file.

Re: [Etherboot-developers] Linux World Expo UK; 2-4 September

[<ebi...@ln...>]

ja...@Mc... writes:


> But, I see on the linux.conf website, they rate themselves up
> there with the Ottawa Linux Symposium.  That takes place next
> week, and I WILL be there :)

Then perhaps we can run into each other.

> > where half the acronyms being tossed around casually hurt my brain.
> 
> Yeah, hurts my brain too.  But, I figure eventually, some of it
> will finally make some sense.


Hmm, imppossible to comprehend acronyms.  Sounds like a challenge.

Eric

[Etherboot-developers] pci scan bug fix commited

[<ebi...@ln...>]

I have just commited a bug fix where the pci driver matching code was
attempting to setup pci devices with class == 0.

There is another bug in there if two drivers happen to match a pci
device by class, and neither of them works.  In that case the drivers
drop into an infinite loop.

Eric

RE: [Etherboot-developers] tlan driver sucessfully boots ltsp

[Timothy L. <tl...@ro...>]

> It waited for the ready signal and then nothing more appeared on the
> screen, I didn't snoop on the wire to see if it sent out a request.
> Maybe it did finish the auto-negotiation and was in the poll loop.
Maybe
> silly me forgot to edit the DHCP config file. I'm hopeful that it will
> work this weekend.

I believe that it prints the speed as the last thing.  Let me verify
that...

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003
 

Re: [Etherboot-developers] tlan driver sucessfully boots ltsp

[<ke...@et...>]

>I just committed a version of the tlan driver that can successfully load
>and boot the ltsp kernel.

Wonderful. Good onya. If you do a backport to 5.0 I can release 5.0.11,
and put 5.0 into maintenance mode (de facto anyway), and then
concentrate on getting 5.2 out.

>Now I need some testers and a way to work around the lack of an
>interrupt timer to remove the recursive-ness that Ken mentioned from the
>auto-negotiation.  

Will give it a whirl this weekend. What you want to do is wait for a few
seconds to see if the link can be established, otherwise print a
message, abort and let Etherboot continue, and hopefully the user will
see the message and deal with the link issue.  The Linux driver uses
timers because the OS has to do other things while waiting.

>Ken you mentioned issues with your card not getting past the
>auto-negotiation stage.  Can you provide any additional details?  

It waited for the ready signal and then nothing more appeared on the
screen, I didn't snoop on the wire to see if it sent out a request.
Maybe it did finish the auto-negotiation and was in the poll loop. Maybe
silly me forgot to edit the DHCP config file. I'm hopeful that it will
work this weekend.

[Etherboot-developers] tlan driver sucessfully boots ltsp

[Timothy L. <tl...@ro...>]

Hi All

I just committed a version of the tlan driver that can successfully load
and boot the ltsp kernel.

Now I need some testers and a way to work around the lack of an
interrupt timer to remove the recursive-ness that Ken mentioned from the
auto-negotiation.  

Ken you mentioned issues with your card not getting past the
auto-negotiation stage.  Can you provide any additional details?  

At this point I still need to clean up the code.  It works but it isn't
pretty!

Regards

Tim

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003
 

Re: [Etherboot-developers] Linux World Expo UK; 2-4 September

[<ja...@Mc...>]

On Wed, 16 Jul 2003, Peter Lister wrote:

> > I've signed up for a booth at the UK Linux World, but the show
> > has been postponed until probably February.  It had previously
> > been scheduled for early September.
> > 
> > So, there should be lots of time to arrange an Etherboot booth,
> > assuming they still have room available.
> 
> Note that the venue is the Birmingham NEC, NOT in London, despite what
> http://www.linuxworld.com/story/33815.htm might say. Amazingly enough,
> not everything in Britain is in London... I'm in Oxford, which is an
> hour away from the NEC by train. I'd be happy to help (e.g. lug a banner
> up there), since I'll go anyway, but I don't know anything about booking
> shows. Looking at the web page there is apparently a .org pavilion, but
> no details.
> 
> BTW, the NEC is next to Birmingham Airport, which has quite a few cheap
> European flights, but intercontental travellers still have to come up
> from Heathrow or Gatwick.

The NEC is also next to the train station, which also makes travel
pretty easy.

I was there last month for the LinuxUser & Developer conference,
and from Detroit, I flew to Amsterdam, and then to Birmingham, which
is a bit wierd, but still a pretty good trip.

FYI, both Amsterdam and Birmingham airports have wireless access :)

Jim McQuillan
ja...@Lt...

Re[2]: [Etherboot-developers] Linux World Expo UK; 2-4 September

[Anselm M. H. <an...@ho...>]

Hello Peter,

> Note that the venue is the Birmingham NEC, NOT in London, despite what
> http://www.linuxworld.com/story/33815.htm might say. Amazingly enough,
> not everything in Britain is in London... I'm in Oxford, which is an
> hour away from the NEC by train. I'd be happy to help (e.g. lug a banner
> up there), since I'll go anyway, but I don't know anything about booking
> shows. Looking at the web page there is apparently a .org pavilion, but
> no details.

> BTW, the NEC is next to Birmingham Airport, which has quite a few cheap
> European flights, but intercontental travellers still have to come up
> from Heathrow or Gatwick.

So count me in, I live halfway between Cologne and Frankfurt Intl'
Airport (35 min train each), certainly there will be some cheap travel
occasions. I hope they'll release a fix date soon.

Best regards,
 Anselm Martin Hoffmeister
 Stockholm Projekt Computer-Service
 <an...@ho...>

-- 
Merke: Nicht das OS macht dich zu einem interessanteren Gespraechs-
partner, sondern das, was du darueber weisst. Und die Toleranz macht
dich dann noch zu einem liebenswerten Gespraechspartner.
(Buelent Caliskan in de.org.ccc)