On (08/14/07 18:20), Peter Memishian wrote:
>
> * I'd expect "autoneg" to be no/yes rather than 0/1. Not sure
> what the numeric values for pause and rem_fault mean, but I'd
> prefer words there too, if possible.
Raymond's reworking the pause definition, but..
my understanding of the ieee802.3(5) man page for rem_fault
for adv/peeradv rows a value of "i" maps to mii_rem_fault[i] where
static char *mii_rem_fault[] = {
"good", "offline", "link fail", "autoneg fail"
};
for capable, i can be 0/1 indicating no/yes (not capable/capable)
current value of rem_fault is "good" iff both adv/peeradv rows are "good",
"possibly bad" otherwise.
so if my understanding is right, then the output of dladm show-ether
now comes out to (some additional bug fixes were encountered for the
SPEED column since yesterday)
LINK PARAM STATE AUTONEG SPEED DUPLEX PAUSE
REM_FAULT
bge0 current up on 1G full 1
good
capable -- on 1G,100M,10M full,half 1
yes
adv -- on 1G full,half 1
good
peeradv -- on 1G full 1
good
bge1 current up on 1G full 1
good
capable -- on 1G,100M,10M full,half 1
yes
adv -- on 1G full 1
good
peeradv -- on 1G full 1
good
And, after doing something like
# ./dladm set-linkprop -p adv_autoneg_cap=0 bge1
we'd have
# ./dladm show-ether -x
LINK PARAM STATE AUTONEG SPEED DUPLEX PAUSE
REM_FAULT
bge0 current up on 1G full 1
good
capable -- on 1G,100M,10M full,half 1
yes
adv -- on 1G full,half 1
good
peeradv -- on 1G full 1
good
bge1 current up on 1G full 1
good
capable -- on 1G,100M,10M full,half 1
yes
adv -- off 1G full 1
good
peeradv -- on 1G full 1
good
does that make sense?
--Sowmini
