public inbox for linuxppc-dev@ozlabs.org 
 help / color / mirror / Atom feed
* questions
@ 2008-05-26  0:32 Kevin Diggs
  2008-05-25 21:31 ` questions Brad Boyer
  0 siblings, 1 reply; 3+ messages in thread
From: Kevin Diggs @ 2008-05-26  0:32 UTC (permalink / raw)
  To: linuxppc-dev

Hi,

	In idle_6xx.S one finds instructions like:

	lis   r4,powersave_nap@ha
	lwz   r4,powersave_nap@l(r4)

Can someone explain what this is doing? Presumably the first is loading 
an address and the second a value. What do the '@ha' and '@l' do?

	Also, is there any performance difference between:

	lbz rD,d(rA)
	lhz rD,d(rA)
	lwz rD,d(rA)

	While I'm wasting your time, I picked up an ADB infrared wireless 
keyboard. I think it works. But not under Linux. Should it?

	And is there any reason to prefer one over the other for doing byte 
swapping:

	lwbrx rD,rA,rB
	stwbrx rS,rA,rB

kevin

^ permalink raw reply	[flat|nested] 3+ messages in thread
* questions
@ 1999-12-23  3:08 Ted Merrill
  0 siblings, 0 replies; 3+ messages in thread
From: Ted Merrill @ 1999-12-23  3:08 UTC (permalink / raw)
  To: linuxppc-embedded



First a general comment:

There were some good comments about board support packages in the
mail archive for this list.
Yes, i realize board support packages are a pain, however:
The linux kernel source is going to explode if it has to support 
directly all of the board variations that will have linux ports.
Statements like "switch(machine_type)" won't cut it.
A more achievable goal is to have the kernel support directly
all of the common cpu chips, bus controllers and the like,
and configure the kernel with a combination of ifs and #ifs controlled
by configuration variables of fairly fine grain, selectable by
a board support configuration file that lives at a distance from
the rest of the code.

Now for the specific question of interest to me:

For the sort of projects that my company is working on,
vxWorks is a better fit than Linux because we don't really need
separate process/kernel address spaces.
What we're doing on the main processor isn't really that complicated
but does require IP networking... thus vxWorks or Linux.
We can (i hope) put up with this unwanted feature of Linux 
in order to get the good stuff, principally the publicly developed
networking software and debugging tools.
I'm thinking the doable thing with Linux is to have only a single
process (filling in for the init process) with multiple threads.
Sounds like the "sash" program could be the beginning of the model for this.
The single process approach combined with static linking
blows away the whole problem of stripping down the C library to fit...
I was trying to make a list of processes running on my pc
that might be needed (and so would be a problem).
We won't have a local disk and we won't be serving very much
but we probably will want to serve both telnet (to a debugging shell, not sh)
and snmp for debug and management purposes, thus we may need inetd ...
this might be some amount of work to turn into library functions.
I'm pretty clueless how snmpd can be reasonably used in an embedded
system were some of the data is in a process and some in the kernel...
i'll have to read up more.
Can anyone share any experience with this?

Thanks
Ted Merrill
ArrayComm, Inc.


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2008-05-26  2:44 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-05-26  0:32 questions Kevin Diggs
2008-05-25 21:31 ` questions Brad Boyer
  -- strict thread matches above, loose matches on Subject: below --
1999-12-23  3:08 questions Ted Merrill

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox