options MCLSHIFT=12
in order to avoid a copy when a mbuf cluster has to be sent to a domU
(which is the common case when the packet comes from a physical network
interface).
receive system:
- on the receive side, attach the mapped buffer as external storage
instead of copying it. As a mapped buffer may not live much longer, we
have to deal with the fact that one page of buffer may containt several
packets, and it's not possible to map them several times. Use a hashed list
to keep track of mapped pages, and use reference counters.
- on the transmit side, when MCLBYTES == PAGE_SIZE, give away the mbuf
cluster page when possible instead of copying it. Keep a pool of physical
pages to map in place of the page we give away. When copying, use a
pool_cache(9) to manage copy buffers (use mclpool_cache when
MCLBYTES == PAGE_SIZE, otherwise use a private pool/pool_cache) instead
of a local list. This should reduce the number of hypercalls and MMU
operations in the copy case as well.
send queue. This give upper layer an opportunity to queue up all available
packets before starting to process them. This reduce the number of interrupt
generated on the backend, and the time spent doing hypercalls in a significant
way.
The first test, SHADOW_INSTANCE, includes an example of a shadowed
instance: 'child' only attaches to attribute hook, therefore the instance
'child* at parentii?' is shadowed by 'child* at hook?'.
However, that configuration file is still valid, and therefore must be
accepted.
since both pool_get() and pool_put() can call wakeup().
instead, allocate the struct sadata_upcall before taking
sched_lock in mi_switch() and free it after releasing sched_lock.
clean up some modularity warts by adding a callback to
struct sadata_upcall for freeing sa_arg.
split the single list of pool cache groups into three lists:
completely full, partially full, and completely empty.
use LIST instead of TAILQ where appropriate.
o Rework do_kill_orphans() to use that information and mark explicitely
orphaned devices (i.e., the ones whose missing ancestor has been
negated)
o Make a distinction between erroneous orphans and explicit orphans.
Error out on the former, ignore the later (but print a warning when -v
is used)
Yes, now config(1) will actually stop if you comment out a parent. That
should help people still hoping adjustkernel is relevant these days :)
Tested on Krups and a SPARCbook 3GX
Works only with wscons kernels, currently there's official support on Krups
with igsfb and more or less official support for SPARCbooks with pnozz and
SPARCstations with cgsix.
alignment architectures
fix ETHER_ALIGN to 2 (same value as on FreeBSD) - appears VGE_FIXUP_RX
code cuts part of packet otherwise; also add comment about it's purpose
PR: 31323 by Murata Shuuichirou
