Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.apple2.programmer > #1698 > unrolled thread

Marina, now with Berkeley sockets

Back to article view | Back to comp.sys.apple2.programmer

Contents

  Marina, now with Berkeley sockets D Finnigan <dog_cow@macgui.com> - 2015-06-16 20:54 +0000
    Re: Marina, now with Berkeley sockets David Schmidt <schmidtd@my-deja.com> - 2015-06-16 17:32 -0400
      Re: Marina, now with Berkeley sockets D Finnigan <dog_cow@macgui.com> - 2015-06-17 21:28 +0000
      Re: Marina, now with Berkeley sockets D Finnigan <dog_cow@macgui.com> - 2015-06-23 23:26 +0000
    Re: Marina, now with Berkeley sockets D Finnigan <dog_cow@macgui.com> - 2015-06-17 21:23 +0000
      Re: Marina, now with Berkeley sockets Michael J. Mahon <mjmahon@aol.com> - 2015-06-17 17:01 -0500
        Re: Marina, now with Berkeley sockets D Finnigan <dog_cow@macgui.com> - 2015-06-18 23:18 +0000
          Re: Marina, now with Berkeley sockets Michael J. Mahon <mjmahon@aol.com> - 2015-06-18 19:39 -0500
    Re: Marina, now with Berkeley sockets Your Name <YourName@YourISP.com> - 2015-06-18 09:45 +1200

#1698 — Marina, now with Berkeley sockets

The June 2015 edition of Marina is here, featuring a Berkeley sockets API.
It's not complete, but what's done does work. Read more on what's new:
http://macgui.com/news/article.php?t=391

Or just visit the web site in my sig.

-- 
]DF$
The Marina IP stack for Apple II--
http://marina.a2hq.com/

[toc] | [next] | [standalone]

#1699

On 6/16/2015 4:54 PM, D Finnigan wrote:
> The June 2015 edition of Marina is here, featuring a Berkeley sockets API.

Glad to see the source code.  In its current form, how much memory 
footprint is required (understanding the send buffer is variable - so 
let's just say 1492 for now)?

[toc] | [prev] | [next] | [standalone]

#1701

David Schmidt wrote:
> On 6/16/2015 4:54 PM, D Finnigan wrote:
>> The June 2015 edition of Marina is here, featuring a Berkeley sockets
>> API.
> 
> Glad to see the source code.  In its current form, how much memory 
> footprint is required (understanding the send buffer is variable - so 
> let's just say 1492 for now)?
>

Code is a little over 9,000 bytes, but significant space is taken by the
DHCP, DNS, and ARP subsystems that all use dedicated buffers for send
operations. Incoming buffer should be 1514 bytes, and outgoing buffer is
1500 because the protocol headers are stored separate from the payload (see
ETH.S on disk 2). In future, I will probably have to introduce support for
multiple receive buffers in order to support simultaneous TCP and UDP
sockets.

-- 
]DF$
The Marina IP stack for Apple II--
http://marina.a2hq.com/

[toc] | [prev] | [next] | [standalone]

#1721

David Schmidt wrote:
> On 6/16/2015 4:54 PM, D Finnigan wrote:
>> The June 2015 edition of Marina is here, featuring a Berkeley sockets
>> API.
> 
> Glad to see the source code.  In its current form, how much memory 
> footprint is required

Regarding some of the modules:

 These two could possibly be eliminated to save some RAM if you don't need
 the functionality:

DNS - 1,346 bytes
DHCP - 1,707 bytes

Marina's DHCP is a rather generous implementation of the RFC, so you could
even eliminate code within that module to get a bare-bones implementation.
You could also just jettison it once you've gotten a lease, and overwrite it
with some other code or data.

-- 
]DF$
The Marina IP stack for Apple II--
http://marina.a2hq.com/

[toc] | [prev] | [next] | [standalone]

#1700

Antoine Vignau wrote:
> +1
>

Thanks. There is still so much left to do! :-0

But I suppose I should have left instructions for changing the Uthernet slot
number. Here they are:

Boot source disk 1 and insert disk 2 into drive 2. In Merlin Pro, press D to
switch to drive 2, then load ETH. Right at the top you will see all the
EQUates for the slot addresses. Change these as needed.

Then at the main menu, Save, then Write to text file. Switch to drive 1.
Load MARINA, then ASM, then switch back to Drive 1 again and press O to save
the binary to disk, and you should be set to go!

-- 
]DF$
The Marina IP stack for Apple II--
http://marina.a2hq.com/

[toc] | [prev] | [next] | [standalone]

#1703

D Finnigan <dog_cow@macgui.com> wrote:
> Antoine Vignau wrote:
>> +1
>> 
> 
> Thanks. There is still so much left to do! :-0
> 
> But I suppose I should have left instructions for changing the Uthernet slot
> number. Here they are:
> 
> Boot source disk 1 and insert disk 2 into drive 2. In Merlin Pro, press D to
> switch to drive 2, then load ETH. Right at the top you will see all the
> EQUates for the slot addresses. Change these as needed.
> 
> Then at the main menu, Save, then Write to text file. Switch to drive 1.
> Load MARINA, then ASM, then switch back to Drive 1 again and press O to save
> the binary to disk, and you should be set to go!

Merlin allows you to request input during assembly. 

You could request the slot number, then use that to compute all the equates
at assembly time. 
-- 
-michael - NadaNet 3.1 and AppleCrate II: http://home.comcast.net/~mjmahon

[toc] | [prev] | [next] | [standalone]

#1705

Michael J. Mahon wrote:
> 
> Merlin allows you to request input during assembly. 
> 
> You could request the slot number, then use that to compute all the
> equates
> at assembly time. 
>

Will do. I also need to finish putting conditional assembly opcodes around
all my debug code. But I feared if I didn't release now it wouldn't get out
this month and I wanted a June release. :-0

-- 
]DF$
The Marina IP stack for Apple II--
http://marina.a2hq.com/

[toc] | [prev] | [next] | [standalone]

#1706

D Finnigan <dog_cow@macgui.com> wrote:
> Michael J. Mahon wrote:
>> 
>> Merlin allows you to request input during assembly. 
>> 
>> You could request the slot number, then use that to compute all the
>> equates
>> at assembly time. 
>> 
> 
> Will do. I also need to finish putting conditional assembly opcodes around
> all my debug code. But I feared if I didn't release now it wouldn't get out
> this month and I wanted a June release. :-0

Macros can be very helpful for debugging output/logging and assertions.
They permit easy changes to debugging methods and "levels". 
-- 
-michael - NadaNet 3.1 and AppleCrate II: http://home.comcast.net/~mjmahon

[toc] | [prev] | [next] | [standalone]

#1702

In article <dog_cow-1434488118@macgui.com>, D Finnigan
<dog_cow@macgui.com> wrote:
> 
> Marina, now with Berkeley sockets

Well Marina had better give Berkeley his sockets back, or the police
may have to be called.  ;-)

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.apple2.programmer

csiph-web