If I run “nettrap”, the command line created for my 82557 comes out wrong.
“nettrap” gives me:
Net.ether82557 -p1000 -i9 -I0 -l1
For some reason, it looks like Net.ether82557 has different parameter option
letters than the other Net.* drivers. Comparing the output of “use
Net.ether905” and “use Net.ether82557”, I was able to construct the
following command line:
Net.ether82557 -a1000 -i9 -p0 -l1
Which actually loaded the driver just fine. This bug appears to be a
problem even during installation, where I can watch it waiting for the
device to become available.
As a side note, what is the best way to wait for the driver to initialize?
Right now I just “sleep 5” after running the driver so that “Socklet” won’t
execute in “tcpip.1” before the drive has finished initializing. (And, btw,
this method should be included in the stock sysinit.1 file…)
Thanks!
For what it’s worth, the only option I have ever had to use
with these cards is “-l” when there is more than one card or
on one occasion, the -s setting for a flaky hub.
The new TCP/IP suite has an if_up utility that is smarter than a
sleep - I don’t know when it was made available.
Richard
Kees Cook wrote:
If I run “nettrap”, the command line created for my 82557 comes out wrong.
“nettrap” gives me:
Net.ether82557 -p1000 -i9 -I0 -l1
For some reason, it looks like Net.ether82557 has different parameter option
letters than the other Net.* drivers. Comparing the output of “use
Net.ether905” and “use Net.ether82557”, I was able to construct the
following command line:
Net.ether82557 -a1000 -i9 -p0 -l1
Which actually loaded the driver just fine. This bug appears to be a
problem even during installation, where I can watch it waiting for the
device to become available.
As a side note, what is the best way to wait for the driver to initialize?
Right now I just “sleep 5” after running the driver so that “Socklet” won’t
execute in “tcpip.1” before the drive has finished initializing. (And, btw,
this method should be included in the stock sysinit.1 file…)
Thanks!
Yeah, the only reason I was bringing it up was to help the QNX folks fix
their install scripts. Basically, if you have an 82557/8/9, and you install
QNX4, you can’t use your net card until you dig around in the scripts for a
day. 
“Richard R. Kramer” <rrkramer@kramer-smilko.com> wrote in message
news:3AC90D90.DD22C95B@kramer-smilko.com…
For what it’s worth, the only option I have ever had to use
with these cards is “-l” when there is more than one card or
on one occasion, the -s setting for a flaky hub.
The new TCP/IP suite has an if_up utility that is smarter than a
sleep - I don’t know when it was made available.
Richard
Kees Cook wrote:
If I run “nettrap”, the command line created for my 82557 comes out
wrong.
“nettrap” gives me:
Net.ether82557 -p1000 -i9 -I0 -l1
For some reason, it looks like Net.ether82557 has different parameter
option
letters than the other Net.* drivers. Comparing the output of “use
Net.ether905” and “use Net.ether82557”, I was able to construct the
following command line:
Net.ether82557 -a1000 -i9 -p0 -l1
Which actually loaded the driver just fine. This bug appears to be a
problem even during installation, where I can watch it waiting for the
device to become available.
As a side note, what is the best way to wait for the driver to
initialize?
Right now I just “sleep 5” after running the driver so that “Socklet”
won’t
execute in “tcpip.1” before the drive has finished initializing. (And,
btw,
this method should be included in the stock sysinit.1 file…)
Thanks!