PPP Session disconnect

Now I could have misunderstood this, but I’ve got a few questions regar=
ding
the usefulness of this:

  1. What do you expect to happen if the other end sends a packet to the
    standby machine?
  2. Doesn’t a PPP connection have a “watchdog” packet (idle-packet or so=
    me
    such) that needs to be sent every so often to make sure that the other =
    end
    is still there?
  3. Assuming that the above problems have workarounds, when the system i=
    s
    restored, what do you think will happen when the other end starts resen=
    ding
    all the packets that it sent when the system was “sleeping”?

[N.B. I’m no expert at PPP, but I’ve worked with other network solution=
s
using custom-built protocol’s]


Mats

xxxxx@lists.osr.com wrote on 09/01/2004 04:26:56 PM:

Hi All,

We has a problem with Windows PPP stack. The dial up is terminating
it’s session once the system enters into the standby mode.We want to
maintain the dialup connection (that is PPP session) even in the
standy mode. Is there any way to do that.

All your suggestions are welcomed.

Thanks,
Cyril

[image removed] — Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=3D256 You are currently
subscribed to ntdev as: unknown lmsubst tag argument: ‘’ To
unsubscribe send a blank email to xxxxx@lists.osr.com
ForwardSourceID:NT0000280E=

Ok, I see what you’re trying to do. Unfortunately, I’m not entirely sure
how you would solve this, I’m not an expert on Serial/Network Drivers (nor
really an expert on any other drivers, I’ve only been working with Windows
drivers for 18 months + a few weeks at a previous job modifying an existing
driver).

But here’s my (maybe incorrect) understanding of how things work in standby
mode: There are no active windows drivers for anything. The chipset in the
machine is responsible for detecting “Wake-up” signals, and then waking the
processor from standby. The actual standby operation is performed in the
BIOS or by executing ACPI pseudo-code, and once in Standby, the processor
need to be awakened by hardware signalling. Once it’s awakened, it will
require a fair bit of excercising to get back to proper Windows operations
(it may seem immediate to you and me, but to the hardware, it’s quite a
long time).

You may be able to set the hardware to wake up on serial port activity or
interrupts on PCI-bus to indicate that the modem received something, but
there will at that point not be any OS there to handle the interrupts and
no capability of receiving packets.

You do realize that a PPP-packet is collected over the serial port, and
there is (usually) not enough buffer in the serial port to actually get a
full packet into the system without getting some processing done. And I
suspect you may loose the start of the packet before the processor has
awakened. Maybe not, but I think it’s quite likely.

Again, I don’t know how PPP works, but I guess if you don’t send responses
to the other side’s “Echo” request, the other end will terminate the
connection…

I think the only solution in this case is to say “Don’t go into standby
when connected to the network”.

But I’m only guessing from my knowledge of protocols in general and my
understanding of power management from the processor and chipset hardware
standpoint.

Someone that understands how windows works and how these type of drivers
work may have better information.


Mats

xxxxx@lists.osr.com wrote on 09/03/2004 08:08:57 AM:

Hi Mats,

Thanks for replying.

  1. We are expecting the system to wake up when it receives the
    packet from the otherend.(question number three is solved i guess)

  2. When we are in standby we will not send any echo request packet
    to the other side.

Any suggestions / Comments.

Thanks,
Cyril

On Wed, 01 Sep 2004 Mats PETERSSON wrote :
>
>
>
>
>
>Now I could have misunderstood this, but I’ve got a few questions
regarding
>the usefulness of this:
>1. What do you expect to happen if the other end sends a packet to the
>standby machine?
>2. Doesn’t a PPP connection have a “watchdog” packet (idle-packet or
some
>such) that needs to be sent every so often to make sure that the other
end
>is still there?
>3. Assuming that the above problems have workarounds, when the system is
>restored, what do you think will happen when the other end starts
resending
>all the packets that it sent when the system was “sleeping”?
>
>[N.B. I’m no expert at PPP, but I’ve worked with other network solutions
>using custom-built protocol’s]
>
>–
>Mats
>
>xxxxx@lists.osr.com wrote on 09/01/2004 04:26:56 PM:
>
> > Hi All,
> >
> > We has a problem with Windows PPP stack. The dial up is terminating
> > it’s session once the system enters into the standby mode.We want to
> > maintain the dialup connection (that is PPP session) even in the
> > standy mode. Is there any way to do that.
> >
> > All your suggestions are welcomed.
> >
> > Thanks,
> > Cyril
> >
> >
> > [image removed] — Questions? First check the Kernel Driver FAQ at
> > http://www.osronline.com/article.cfm?id=256 You are currently
> > subscribed to ntdev as: unknown lmsubst tag argument: ‘’ To
> > unsubscribe send a blank email to xxxxx@lists.osr.com
> > ForwardSourceID:NT0000280E
>
>
>—
>Questions? First check the Kernel Driver FAQ at http://www.
osronline.com/article.cfm?id=256
>
>You are currently subscribed to ntdev as: unknown lmsubst tag argument:
‘’
>To unsubscribe send a blank email to xxxxx@lists.osr.com

[image removed] — Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256 You are currently
subscribed to ntdev as: unknown lmsubst tag argument: ‘’ To
unsubscribe send a blank email to xxxxx@lists.osr.com
ForwardSourceID:NT00002A7E

The power chip on the mobo usually has some “WAKE on” gates.
They are, in turn, attached to some “wake” wires on the peripherals like
the PCI PME.

If you want the sleeping machine to be woken from Ethernet, then you need
a) the Ethernet card to generate PME and b) the power chip to wake the machine
from PME.

To switch both states on, there must be a wait-wake handler for the
Ethernet, and also the wait-wake handler in the ACPI.SYS which controls the
mobo. This is all the mechanics.

Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com