[9fans] rdbfs(4) woes

[9fans] rdbfs(4) woes

[Steven Stallion]

9fans,

I dusted off rdbfs(4) today to dig into a kernel issue and ended up
running into what looks like the same issue bwc reported back in 2005:

barstow% import cons /mnt/consoles
barstow% rdbfs -d -t /386/9pccpu /mnt/consoles/target
attach /mnt/consoles/target
barstow% acid -k 1 /386/9pccpu
/386/9pccpu:386 plan 9 boot image
/sys/lib/acid/port
/sys/lib/acid/386
acid: lstk()
got Tread tag 1190 fid 768 offset 56 count 4: here
goes...r00000038...error interrupted
r00000038...error interrupted
r00000038...error interrupted
r00000038...error interrupted
r00000038...error interrupted

Other relevant bits of info:

barstow% 9fat:
barstow% grep console'=' /n/9/plan9.ini
console=0

barstow% import cons /lib/ndb /n/cons
barstow% ndb/query -f /n/cons/consoledb console target
console=target port=7 dev=/dev/eia8

barstow% import cons '#t' /n/t
barstow% cat /n/t/eia8status
b9600 c0 d1 e0 l8 m0 pn r1 s1 i0
dev(8) type(0) framing(4) overruns(0) berr(0) serr(0) cts dsr

All of this was done post ^t ^t D on the target's console.

Any ideas?

Steve

Re: [9fans] rdbfs(4) woes

[Steven Stallion]

Bump, though admittedly this is not as much fan as complaining about
rc's shortcomings.

On Tue, Aug 28, 2012 at 7:07 PM, Steven Stallion <sstallion@gmail.com> wrote:
> 9fans,
>
> I dusted off rdbfs(4) today to dig into a kernel issue and ended up
> running into what looks like the same issue bwc reported back in 2005:
>
> barstow% import cons /mnt/consoles
> barstow% rdbfs -d -t /386/9pccpu /mnt/consoles/target
> attach /mnt/consoles/target
> barstow% acid -k 1 /386/9pccpu
> /386/9pccpu:386 plan 9 boot image
> /sys/lib/acid/port
> /sys/lib/acid/386
> acid: lstk()
> got Tread tag 1190 fid 768 offset 56 count 4: here
> goes...r00000038...error interrupted
> r00000038...error interrupted
> r00000038...error interrupted
> r00000038...error interrupted
> r00000038...error interrupted
>
> Other relevant bits of info:
>
> barstow% 9fat:
> barstow% grep console'=' /n/9/plan9.ini
> console=0
>
> barstow% import cons /lib/ndb /n/cons
> barstow% ndb/query -f /n/cons/consoledb console target
> console=target port=7 dev=/dev/eia8
>
> barstow% import cons '#t' /n/t
> barstow% cat /n/t/eia8status
> b9600 c0 d1 e0 l8 m0 pn r1 s1 i0
> dev(8) type(0) framing(4) overruns(0) berr(0) serr(0) cts dsr
>
> All of this was done post ^t ^t D on the target's console.
>
> Any ideas?
>
> Steve

Re: [9fans] rdbfs(4) woes

[Rudolf Sykora]

On 2 September 2012 20:49, Steven Stallion <sstallion@gmail.com> wrote:
> Bump, though admittedly this is not as much fan as complaining about
> rc's shortcomings.

Unfortunately, I believe there are just no men to help you. At least I
can't. And the knowledge in the community seems to be slowly
disappearing. Here, sometimes, you "stand there alone".

Ruda

Re: [9fans] rdbfs(4) woes

[hiro]

for non mac os related questions come to irc.

Re: [9fans] rdbfs(4) woes

[erik quanstrom]

On Sun Sep  2 14:52:18 EDT 2012, sstallion@gmail.com wrote:
> Bump, though admittedly this is not as much fan as complaining about
> rc's shortcomings.
>
> On Tue, Aug 28, 2012 at 7:07 PM, Steven Stallion <sstallion@gmail.com> wrote:
> > 9fans,
> >
> > I dusted off rdbfs(4) today to dig into a kernel issue and ended up
> > running into what looks like the same issue bwc reported back in 2005:
> >
> > barstow% import cons /mnt/consoles
> > barstow% rdbfs -d -t /386/9pccpu /mnt/consoles/target
> > attach /mnt/consoles/target
> > barstow% acid -k 1 /386/9pccpu
> > /386/9pccpu:386 plan 9 boot image
> > /sys/lib/acid/port
> > /sys/lib/acid/386
> > acid: lstk()
> > got Tread tag 1190 fid 768 offset 56 count 4: here
> > goes...r00000038...error interrupted
> > r00000038...error interrupted
> > r00000038...error interrupted

i've never used rdb.  on quick inspection, it looks like
you're going to have to set consuart = nil in talkrdb().
but that's just a guess.  obviously rdb appears a one-way
trip.

other than that, i've only got trite ideas.  personally,
i would ratrace acid and rdbfs (from another window), and see exactly
what system calls are generating which notes.

you may also wish to temporarly remove consolefs from
the mix, and use stk() or a simplier command like regs()
or *PC instead of lstk().  but again, sorry for the trite
advice.

good luck.

- erik