[ECOS] wavelan driver

[ECOS] wavelan driver

[Andrea Acquaviva - DEIS-micrel]

Hi,

I'm in the process of writing an ecos driver for the
wavelan card, so I need some info about the tcp/ip
stack.

In particular, looking at the driver_doc document, I
don't find where is the correct way to pass the IP
source and destination address to the card.

For example, what is the data structure that handle this
info in the HRDWR_send function?

Thanks,
Andrea.

Re: [ECOS] wavelan driver

[Jonathan Larmour]

"Andrea Acquaviva - DEIS-micrel (93785)" wrote:
> 
> Hi,
> 
> I'm in the process of writing an ecos driver for the
> wavelan card, so I need some info about the tcp/ip
> stack.
> 
> In particular, looking at the driver_doc document, I
> don't find where is the correct way to pass the IP
> source and destination address to the card.

The stack expects to just be able to fill that in and send it as "data". If
you need to pass an address to the card as well as the data, you would
probably need to extract the addresses back out from the packet you've been
passed to send.

But FYI anyway, the header creation gets done in ip_output.c

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine
Come to the Red Hat TechWorld open source conference in Brussels!
Keynotes, techie talks and exhibitions    http://www.redhat-techworld.com/

Re: [ECOS] wavelan driver

[Andrea Acquaviva - DEIS-micrel]

Quoting Jonathan Larmour <jlarmour@redhat.com>:

> "Andrea Acquaviva - DEIS-micrel (93785)" wrote:
> > 
> > Hi,
> > 
> > I'm in the process of writing an ecos driver for the
> > wavelan card, so I need some info about the tcp/ip
> > stack.
> > 
> > In particular, looking at the driver_doc document, I
> > don't find where is the correct way to pass the IP
> > source and destination address to the card.
> 
> The stack expects to just be able to fill that in and
send it as "data". If
> you need to pass an address to the card as well as the
data, you would
> probably need to extract the addresses back out from
the packet you've been
> passed to send.
> 
> But FYI anyway, the header creation gets done in
ip_output.c
> 
> Jifl
> -- 
> Red Hat, Rustat House, Clifton Road, Cambridge, UK.
Tel: +44 (1223) 271062
> Maybe this world is another planet's Hell -Aldous
Huxley || Opinions==mine
> Come to the Red Hat TechWorld open source conference
in Brussels!
> Keynotes, techie talks and exhibitions   
http://www.redhat-techworld.com/
> 

Yes, the address is passed through the sg_list as a
normal data.
The problem is that I need to pull this address
to program the appropriate registers in the card, so I
need to know where it is located.
How can I find that? By looking at ip_output.c?

Andrea.

 

Re: [ECOS] wavelan driver

[Andrea Acquaviva - DEIS-micrel]

Quoting Jonathan Larmour <jlarmour@redhat.com>:

> "Andrea Acquaviva - DEIS-micrel (93785)" wrote:
> > 
> > Hi,
> > 
> > I'm in the process of writing an ecos driver for the
> > wavelan card, so I need some info about the tcp/ip
> > stack.
> > 
> > In particular, looking at the driver_doc document, I
> > don't find where is the correct way to pass the IP
> > source and destination address to the card.
> 
> The stack expects to just be able to fill that in and
send it as "data". If
> you need to pass an address to the card as well as the
data, you would
> probably need to extract the addresses back out from
the packet you've been
> passed to send.
> 
> But FYI anyway, the header creation gets done in
ip_output.c
> 
> Jifl
> -- 
> Red Hat, Rustat House, Clifton Road, Cambridge, UK.
Tel: +44 (1223) 271062
> Maybe this world is another planet's Hell -Aldous
Huxley || Opinions==mine
> Come to the Red Hat TechWorld open source conference
in Brussels!
> Keynotes, techie talks and exhibitions   
http://www.redhat-techworld.com/
> 

Now I realized that I really don't need to pull the IP
address from the data, so I have not more problems about
that.

Thank you.

Re: [ECOS] wavelan driver

[Andrea Acquaviva - DEIS-micrel]

Quoting Jonathan Larmour <jlarmour@redhat.com>:

> "Andrea Acquaviva - DEIS-micrel (93785)" wrote:
> > 
> > Hi,
> > 
> > I'm in the process of writing an ecos driver for the
> > wavelan card, so I need some info about the tcp/ip
> > stack.
> > 
> > In particular, looking at the driver_doc document, I
> > don't find where is the correct way to pass the IP
> > source and destination address to the card.
> 
> The stack expects to just be able to fill that in and
send it as "data". If
> you need to pass an address to the card as well as the
data, you would
> probably need to extract the addresses back out from
the packet you've been
> passed to send.
> 
> But FYI anyway, the header creation gets done in
ip_output.c
> 
> Jifl
> -- 
> Red Hat, Rustat House, Clifton Road, Cambridge, UK.
Tel: +44 (1223) 271062
> Maybe this world is another planet's Hell -Aldous
Huxley || Opinions==mine
> Come to the Red Hat TechWorld open source conference
in Brussels!
> Keynotes, techie talks and exhibitions   
http://www.redhat-techworld.com/
> 

Hi, I've a new question about the driver.

In eth_drv.c:

1- the eth_drv_recv function requires packet data to
be   passed to it in order to fill the sg_list, so this
list will be filled with the whole packet except the
headers (802.11 and 802.3), is this correct? 
If not, what is the content of the sg_list? In addition,
because of this function is expecting only a 802.3
header, I have to pass the packet lenght minus the
802.11 header size as "total_len" argument? 

2- the eth_drv_send function fills the sg_list and calls
the HRDWR_send. What is the content of this list? More
precisely, where can I find the src and dst MAC
addresses? (I need to fill the 802.11 header with this
address)

Thank you,
Andrea.

[ECOS] wavelan driver

[Andrea Acquaviva - DEIS-micrel]

Hi,
I have another question about the wavelan driver.
Now my driver is becaming working because I can receive
and send packet. The problem is now that the net tests
don't work, so I think there is something wrong in the 
interface to the hardware independent driver of ecos:

Here are some simple questions:
- total_len argument of recv and HRDWR_send must be the
total packet len (data+ethernet header)?

- Do I need to change the endianness of the ethernet
type field before to pass it to the card? (in the frame
specs it must be big endian)

Thanks for your help,
Andrea