Alireza Dabagh [MS] wrote:
> BTW, although sending a packet from a device/protocol edge that is in
low
> power state is not desired, it is not fatal.
Thanks Ali! This was the missing piece of information.
> "Eric Brueggemann" <eric.brueggemann@gmail.com> wrote in message
> news:cd7gp2$an0@odah37.prod.google.com...
> > 1. Does NDIS serialize PNP events with respect to send/receive
paths?
> > 2. If not, is it a requirement that a packet does not leave the
driver
> > from an edge that is in low power?
> >
> > If the answers are 1)no and 2)yes, as I suspect they are, then I
> > believe we have a bug in the PASSTHRU driver.
So as I now understand, the answer to 2 is no and, hence, there is no
bug.
The confusion came from the apparent effort in PASSTHRU to guarantee no
packets leave low-power edges. So, when I saw a violation of this, I
assumed it was a bug. IMHO, a clarifying comment in the code would be
helpful to future newcomers.
Thanks,
Eric
Alireza Dabagh [MS] wrote:
> Assuming the answers to the questions are yes and no, what is the
bug? are
> you referring to lack of protection for MPDeviceState and
PTDeviceState
> variables?
>
> BTW, although sending a packet from a device/protocol edge that is in
low
> power state is not desired, it is not fatal.
>
> -ali
>
> --
> This posting is provided "AS IS" with no warranties, and confers no
rights.
>
> "Eric Brueggemann" <eric.brueggemann@gmail.com> wrote in message
> news:cd7gp2$an0@odah37.prod.google.com...
> > Thanks Maxim. I think I found the thread you're referring to. There
> > certainly was good information there. I now know, among other
things,
> > that MPHalt can be called without PtUnbind ever being called.
> >
> > Unfortunately, however, it didn't address my issue. I believe my
> > question boils down to this:
> > 1. Does NDIS serialize PNP events with respect to send/receive
paths?
> > 2. If not, is it a requirement that a packet does not leave the
driver
> > from an edge that is in low power?
> >
> > If the answers are 1)no and 2)yes, as I suspect they are, then I
> > believe we have a bug in the PASSTHRU driver.
> >
> > As always, I am very grateful for any insight anybody may have on
this.
> >
> > Cheers,
> >
> > Eric
> >
> >