i have a new amd600-based system that's has some
interesting behavior. if i pxe boot (using ether the old
or new 9load) without any extra options, the machine
boots fine. however, if i add *nomp=1, the machine
manages to load the kernel, but hangs right after printing
the user/kernel memory
has anyone been successful with pcie-based nvidia controllers?
i have tried a geforce 7200gs (10de/01d3) pcie-based card and
a 10de/0242 builtin. no joy with either, but pretty much the
same symptom. the screen is repeated 4/6/8 time horizontally
and compressed to a few inches high vertically.
// ...is the idea that you would only want to cpu(1) to another
// machine because of the services or network topology you
// (as a human) want rather than just to find more grunt.
Certainly originally cpu(1) was frequently used simply to get to a
bigger machine. That's the way many of the
Hi,
I'm a newbie, so probably it is a stupid question for you.
I've been trying to mount a server on mountpoint, but everytime I get
the following message:
mount: mount /n/inst: fossil authCheck: auth protocol not finished
What does it mean?
Is it an error of mine?
I've searched in the
Note that there are, in fact, two *nbiosload=1 lines and two
*noahciload=1 lines.
why?
- erik
That's the same question I'm asking :-)
you can delete the extra copies. you may also delete the
blank lines.
- erik
get a bigger hard drive. 160g hard drives can be had for $45.
there was a time when 100gb was considered infinite storage.
- erik
So, the decision in the linux virtualization world is to make all
paravirtual devices look like ... drum roll ... PCI devices.
That's really tragic. Virtualisation of devices would be such a good
opportunity to invent a nice simple interface abstracting away from
all the hardware
The opportunity still exists -- only one driver needs to implement
their numeric hack - 9p. Then the rest can be based off of that.
Unfortunately, evolution just comes slow and painful.
-eric
are 9p mesages really the right vehicle for this? 9p messages
provides a serialized and in
are 9p mesages really the right vehicle for this? 9p messages
provides a serialized and in a standard byte order. this requires
byte reordering (on intel) and copying. but are these really needed?
the guest and host are on the same platform, so the guest can
pass pointers to the host.
it looks to me that sdata.c only prints this when i/o times out.
since you're sending a non-dma command, the most likely problem
is a missed interrupt.
that's not very helpful, is it.
is there any chance your original problem was a real i/o error?
- erik
I'm trying to run the most recent
PBS1... Plan 9 from Bell Labs
ELCR: 0C00
Initial probe, to find plan9.ini... dev A0 port 1F0 config 85C0
capabilities 0F00 mwdma 0407
pcirouting 8086/27DF at pin 1 irq 10
dev A0 port 18D0 config 045A capabilities 2F00 mwdma 0007 udma 203F
LLBA sectors 117231408
ahci0 port 0xd4544400: hba
Exactly: they are an aesthetic convenience compared with multiple
echo invocations
but so are quotes compared with multiple backslashes.
though it is annoying they can't be used in fn definitions.
That's a bug which could be fixed ...
i was thinking about this topic again the other
You forgot `-quotes, which required \-quoting to nest in sh.
i knew i was forgetting something. that whole reparsing thing
always bothered me.
- erik
So, what's the rationale behind using Channels?
Is passing by reference not recommended by Plan 9 design?
D
channels (the libthread variety) may pass by reference or value.
so passing by reference or value is orthagonal to the question
of why channels.
for example, if you have a couple of
Well, yeah, I get that. I think the question I need to
ask is why use Channels at all?
What is the benefit of using Channels when threads can't
be bound to a specific CPU in a multi-core system? Aren't
Channels just a wrapper for locking a mutex and setting
a variable or adding to a queue?
* erik quanstrom ([EMAIL PROTECTED]) wrote:
but it doesn't work if you use curly braces
Of course it does; the syntax is somewhat unexpected, though:
for(i in 1 2 3){
cat !
}
fu
!
fu
fu
fu
Martin
On 11/7/07, erik quanstrom [EMAIL PROTECTED] wrote:
// wiring data to processors would make more sense.
is what's meant here wiring (data+proc) to processors would make more sense?
no, if you're using threads (procs), by definition mutiple actors are
chewing on the same data
Michaelian Ennis wrote:
I'd really like to use this font in plan9. Is there a guide out there of
what is required?
http://www.levien.com/type/myfonts/inconsolata.html
Get the freetype port from http://mirtchovski.com/p9/freetype/
Convert the font to ttf and then use ttf2subf.
there is a
cool :-) will it be a separate program or part of *c?
cfront works by turning c++ input into plan c output,
in much the same way a conventional c compiler turns
c input into assembly output. (well they often run the
assembler behind your back.)
- erik
I like here documents, I think they look better than multi-line echoes though
it is annoying they can't be used in fn definitions.
i was thinking about this topic again the other day during an
idle moment. here documents are part of the parsing routine.
but they don't end up in rc's machine
However, I'm also thinking longer term - towards the FastOS project
where users aren't going to necessarily want to fire up a whole
'environment' such as Inferno or Drawterm just to execute an
application on the cluster -- particularly if that application isn't
graphical.
i think i don't
We've typically configured IP for our one standalone auth server
from /cfg/$sysname/cpurc; I want to know if I've understood why
you're doing it with plan9.ini bootargs instead: Is the advantage
that you get the variables set in /net/ndb (fs, auth, etc) that I
don't get when I call
On 11/1/07, erik quanstrom [EMAIL PROTECTED] wrote:
i think it makes sense to use the medium changes (not connections, if
possible) to determine the version. the marvell and aoe driver
consider a device changed if the serial# or number of sectors
change.that is something most io clients
further, this behavior only holds for /dev/sdXX/data and a few other files,
like
/dev/fd?^(diskc ctl). i don't see how this could break anything.
could you give a senario that would be broken in this case?
i can't find any programs that use the qid.version for anything except
to infer that
um, i don't think i said that it was new behaviour. i did say that i thought
it
was an abuse of the semantics, and i still think so. to me the devfloppy
code you point to looks like it was just using the Chan as a
convenient place to store
i think you should complain to presoto or rob. it
i'm sorry? the server is the one *generating* the qid, surely?
for instance, in the case of cdfs, the version number comes from the
scsi(2) library function, and doesn't have any necessary correspondence with
qid.vers as found in the data file served by cdfs. so strictly speaking,
the server
try doing ls -lt /mail/fs/mbox.
i wish this worked!
i do, too. (i made it work in my version.) but that's really beside the point.
fileservers can't be forced to maintain all of the stat fields.
they are free to lie as the wish.
- erik
/n/sources/plan9/sys/src/9/port/devsd.c:177
???
i'm probably being stupid here, but all i see there is devsd dealing
with its own data structures, with no external qids involved at all.
an SDunit doesn't correspond with any file that devsd uses, as far as i can
see.
there is no race.
The same thing as grep(1)
On Nov 5, 2007, at 4:26 PM, andrey mirtchovski wrote:
What do you think does the Look command do?
no. grep finds regular expressions. look finds strings.
- erik
the qid.vers does get exported.
indeed it does. but if clients aren't expected to make use of it,
why *is* it exported? i guess i'm arguing that the notions of device
medium version (a.k.a. nchanges) and file version are two
separate things, and it doesn't really make sense to present
the
I use in my home network local DNS+DHCP server and my fs/auth/cpu
Plan9 server gets statically configured IP address. So my network
isn't defined in my Plan9 ndb database.
First I configure my ethernet device (ip/ipconfig) and then dns
service (ndb/dns -r).
I am the beginner with Plan9 and
i'm getting this error wheb trying to boot plan9 using qemu.
i'm using qemu 0.9.0 and i've downloaded CD Image of plan9 from here:
http://plan9.bell-labs.com/plan9/download.html in the date: nov-2-2007
the error is here: http://i4.photobucket.com/albums/y111/Adrielbr/qemu.png
if you cannot
there is a bad order in initialization of network in /rc/bin/cpurc:
ndb/dns -r
ip/ipconfig
When I boot with these settings I get message:
ndb/dns: can't read my ip address
I suggest to swap these commands.
on plan 9, ndb, not dns is generally required to resolve ip
yes. i believe this bug was fixed. i had this problem, too.
9804261509.aa06201
gives a different error (Invalid address) than unknown user.
any local name with an invalid character will do this.
i think you need to update /mail/lib/validatesender. here's
my fix:
#!/bin/rc
mail -x $1 |
For someone like me who have not gotten used to the entire
plan 9 ideas yet. Something like this would be very useful.
But I think this type of documentation should be placed in the
wiki not in /sys/doc. Once you have learned the system, you only
need the man files as references or to get
all the /bin/patch/* files are just shell scripts, so it's easy to see
what's going on. which stage gives you permission errors? in a default
setup, you'll have to be in group sys, as that's who owns all the
sources.
if you can't be in sys for whatever reason, you'll have a somewhat
more
The new 9load works on all the machines I was able to test it on,
including booting from USB flash disks and SATA disks, in AHCI and
non-AHCI modes. The only way it's going to work for more people is if
the community at large uses it and helps me to understand why it
doesn't work on
to be a bit picky, the qid.version doesn't indicate the status of the device,
it indicates how many times the media have changed.
it doesn't make sense for a process to blithly continue writing to the
new medium without getting an error.
i agree with that, and for read-only devices, using
i haven't thought this through, but perhaps this would
be an easier problem if we didn't change 9p, but we changed
the model of the caching server.
the current proposals assume that the caching servers don't
know that other caches are out there. alternatively,
the caching servers participate in
Any maliciousness on the part of clients in this scheme would result
in (possibly one-time only) temporary denial of service to users
sharing a file; such users are not usually maliciously inclined.
it doesn't take malice. 1 faulty client will do.
- erik
The problem is that the clients with higher latencies badly need to
be able to cache. And the ones with better latencies can afford not
caching :)
the irony is, the higher the latency, the greater the cost of syncronization.
- erik
sorry i didn't see that email for a bit. i'm at home working on
loading the dump (up to 25 feb 2006). and i got distracted taking a look
at what i needed to implement the last bit of functionality on my list.
- erik
There's a funny obsession in this discussion with optimal performance
in the least common scenarios.
but surely that's absolutely typical of most such discussions in computing
and perhaps sports.
oddly, it's the same with spam.
- erik
there appear to be typos in your links
- erik
Hello there 9fans mailing list,
I recently burned/downloaded several copies of the (Live)CD (to make sure I
got a good burn), but when trying to boot from it, I always managed to get
an error just before the StartUp Menu was supposed to appear.
Oh. I apologize for that:
http://home.earthlink.net/~eldanen/pastes/dmesg
http://home.earthlink.net/~eldanen/pastes/plan9output
i don't see any error. (the pcirouting whine is harmless.)
does it just hang?
i'm pretty sure that plan 9 will not recognize your hard
drive, as it is plugged
By the way, yes it does hang. I plugged the HDD into Sata #1 (no #0 in the
BIOS.) I then inserted the plan9 CD, and it hung at the same point. What
next?
i guess bios doesn't know how to count. ;-)
geoff may have better ideas than i, but you might try disabling the
floppy drive with bios.
Hehe. There is no floppy drive. But thanks for all the help you've given
me.
exactly. disable it in bios.
- erik
Dear List,
I am getting the above message when trying to change to another user
on my system. What does it mean?
Regards,
Christian
does cpu -u $newuser work?
this is a little bit of a guess, but the key suspect seems to be
port/devcap.c:229: error(Eshort);
i.e.
what does /sys/log/auth say?
Nothing related to this event. I get a suspicious warning every day:
bob Oct 30 15:37:10 keyfs starting warnings: 47274196 471c6237
But I am not sure if this is related...
my really wild guess is that you changed the hostowner and didn't
add him to
Maybe also broken: passwd(1). What happens if you use it to change
your hostowner's password?
you also need to do auth/wrkeys if you are going to change the hostowner's
password. if you use a ken fs, you will also need to set the password on the
fs. not sure about fossil.
i haven't found it
What about this change:
1058,1059c1058,1059
syslog(0, auth, keyfs starting warnings: %lux %lux,
now, lastwarning);
---
syslog(0, auth, keyfs starting warnings at %s, last
was %s,
hurumph. Don't know where i got that from - I tried to base my v9fs
cacheing stuff on cfs.c, but I don't see any qid.version==0 checks
there. Then I thought perhaps it was from a conversation wtih Russ
when I was doing cacheing in v9fs -- but on searching my gmail he was
saying that it
Why don´t add a QTCTL bit to Qid.type?
It would mean this file does not behave like a regular file, do not cache and
handle with care).
why are the current namespace conventions insufficient?
/mnt, /net, and /dev hold most, if not all, of the special files.
- erik
The dynamic nature of namespace works against such conventions.
Besides it would be nice to have a mechanism that could work in other
systems that use 9p. File servers should be able to convey whether a
file is cache-able or not.
i'm not sure i follow this argument. plan 9 namespaces are
Sure - however, there is a case for loose caches as well. For example,
lots of remote file data is essentially read-only, or at the very
worst its updated very infrequently. Brucee had
i might be speaking out of school. but i worry about the qualifiers
essentially and very infrequently.
2. I live in America, where the government was stupid enough to change what
day Daylight Savings Time changes fall on. Unfortunately, release 4 was
released back in 2002 - before the new law came into play. What will I have
to do to fix the daylight savings time rules? This is a particular
this timezone should cover the dst trial period. let me know
if there are problems. if there are none, i will submit it.
to try it out, copy the attachment to /adm/timezone/Australia_West
then
cp /adm/timezone/Australia_West
cat /adm/timezone/Australia_West/env/timezone
this timezone should cover the dst trial period. let me know
if there are problems. if there are none, i will submit it.
Huh, thanks! Lots of interest in timezones on the list it seems.
At first when I tried it date started reporting GMT, but it works
fine after I added a newline to the
I don't know how many of you see this stuff, but ...
http://tinyurl.com/234l92
Our industry is into this every little thing is different mode it seems.
ron
If that were true... what's with XML everywhere?
this isn't slashdot. you're supposed to read the article
before commenting.
-
If that were true... what's with XML everywhere?
this isn't slashdot. you're supposed to read the article
before commenting.
- erik
what the article outlines is a meta standard that makes no
decision on authentication for pluggable storage. it just
outlines a layer of goo ontop of
Yeah. I just watch this stuff go by, in continuous flow. I'm seeing it
now with NFS. It's just more and more crud piled higher and higher.
The perceived solution to problems with software is to add more
software, in each case, which is why (e.g.) Linux kernel is growing
*exponentially*.
I
We still fix some bugs (a bug in grap(1) was found this morning), but
I don't know about the system in general.
the whole system is maintained. you can look for yourself at the changes
at
http://plan9.bell-labs.com/sources/plan9/sys/src/
- erik
$ traceroute plan9.bell-labs.com
traceroute: unknown host plan9.bell-labs.com
nameserver problems.
ndb/dnsdebug plan9.bell-labs.com, or on linux
dig plan9.bell-labs.com (which likely won't work) then
dig +trace @a.root-servers.net plan9.bell-labs.com
- erik
in that case, one should build a sandbox, climb into it and import the
fs. the potential damage is contained. maybe 9fs should have an
option to do that.
What if the trojan broke out of that sandbox? Or knows how to
import other parts of the namespace into its process? Namespaces
on
How? When I put
path=(/bin . /usr/pietro/bin)
at the end of my /usr/pietro/profile, rc resets it.
first, rc doesn't read that file, it reads $home/lib/profile.
second, the preferred way of adding $home/bin/rc and
$home/bin/$cputype to your path are by binding them
like so
and one fewer account on sources. there's a check on that
sort of behavior.
- erik
1) rc: the value of $path is (. /bin). It is a classic case not to
have . as the first directory when searching for programs - it allows
Trojan horses to form.
if you're the only one using your
How about forking off a server process that lets me execute arbitrary
commands as you?
How about placing trojan processes in your person bin directory?
How about subtly corrupting all of the writable data in your filesystem?
How about setting up a spam bot on your machine? Using your
BSP is not really sufficient for scaling or building all applications.
It's a great hammer, but not all the apps are nails.
ron
could you elaborate or give a pointer explaining why
bsp is insufficient?
- erik
it's possible for the first page not to render until you go forward one page
and back to the first page. this is a new problem introduced with the
page cache code.
- erik
1) rc: the value of $path is (. /bin). It is a classic case not to
have . as the first directory when searching for programs - it allows
Trojan horses to form.
if you're the only one using your system, how could this be a problem?
but assuming you have multiple users on your system, how
thanks.
- erik
BSP is essentially all about the inner loop. In this loop, you do
the work, and, at the bottom of the loop, you tell everyone what you
have done.
So you are either computing or communicating. Which means, on your
$100M computer, that you are using about $50M of it over
i refreshed my copy on sources (/n/sources/contrib/quanstro/src/myfs).
if used with /n/sources/contrib/quanstro/9loadaoe, it will recognize
the results of the e820 scan. my machine reports,
found 1 e820 memory banks 3326MB+13056MB
topofmem = cfc0
cpuiddx bfebfbff
pse:
i refreshed my copy on sources (/n/sources/contrib/quanstro/src/myfs).
if used with /n/sources/contrib/quanstro/9loadaoe, it will recognize
the results of the e820 scan. my machine reports,
found 1 e820 memory banks 3326MB+13056MB
topofmem = cfc0
cpuiddx bfebfbff
pse:
So, those with experience with threading implementations on weird
real-time or embedded operating systems:
Have you ever ran into a thread implementation where two threads
could *not* directly access each other's .bss (or equiv)/heap?
i.e. have you ever encountered a scenario where
under plan 9, don't we call those processes?
that doesn't sound like any imbedded environment i've worked in.
it's getting mighty fancy to have one set of pagetables. ;-)
You're cute, but I'm trying to be serious.
It's an odd question, I know.
well although i was being funny, the point
term% grep cpu /dev/kmesg
cpu0: 998MHz CentaurHauls Via C3 Eden-N (cpuid: AX 0x069A DX 0x381BA3F)
cpu1: 997MHz CentaurHauls Via C3 Eden-N (cpuid: AX 0x069A DX 0x381BA3F)
term% grep type /dev/vgactl
type vesa
do you know why vesa works with via multiprocessors?
i haven't seen it work with
mv /rc/bin/service/tcp567 /rc/bin/service/!tcp567
Also, there was no /386/9pccpuf kernel; I compiled my own, but had the
same issue as Antonin that it was too big, so I cut out all the video
drivers and all the network drivers but one. Now it boots, but I get
the following message every few
So, first off, I've been looking (again) at some of the stuff about
the blit and gnot terminals, and I'm wondering--how widespread
were they? The commercialized versions, were many sold? They
look really cool and I'd like to find one to play with... but is my only
chance to drive to Bell Labs
yeah, it would save some effort for me, I am trying to buy new
hardware and a list would be helpful.
i think that any intel pcie-based gigabit nic is supported, including
82563 motherboard solutions. intel 63xxesb (enterprise south
bridge)
often
On Mon Oct 22 10:41:26 EDT 2007, [EMAIL PROTECTED] wrote:
- Some SATA chipsets are now supported thanks to Erik, but you still
cannot install
plan9 to SATA devices because 9load has not been recompiled to do so
(yet).
this is not correct. any sata chipset that will emulate the
I did specify the authserver (it's the same machine). Nat shouldn't be a
problem, because all machines in question are connected through a single
hub. It's just that cpu works, drawterm from the system next to it doesn't.
Conversely, drawterm to mordor, which lies behind a NAT'ing router,
http://www.amazon.com/Intel-PILA8460C3BLK-Desktop-Network-Adapter/dp/B000957HZW/ref=sr_1_1/103-5585202-5631020?ie=UTF8s=electronicsqid=1192976576sr=8-1
this card is $30 and does gigabit. you might find it cheeper at newegg.
notes are not compatable with threaded applications.
notes are not intended for general communications. what
is your intended application? you may want send/recv instead.
notes aren't substantially weaker. they solve a different
problem. from notify(2)
When a process raises an
Current rc uses temporal file in cooking here doc.
However I prefer that entire contents of the here doc is in memory.
Now memory is cheap and I never write so much contents in here doc.
Speed and readability is much important for me.
Kenji Arisawa
the main argument i can see for having
I think it is considerably easier to create here documents on the fly (eg.
awk output) than correctly quoted strings.
that's the beauty of rc-style quoting. a simple sed script is enough
to stuff '. it's easy to do this on the fly. your awk script needn't
know. i wrote a bundle-like
when i try using this
“ 201cleft double quotation mark
” 201dright double quotation mark
the postscript looks fine on plan 9 but not elsewhere.
if i used `` and '' instead and use /sys/doc/cleanup, the postscript
doesn't look right either.
what's the secret?
- erik
Did you download the unicode postscript fonts? The invocation of
aux/download is somewhat intricate, but in the list archives is a message
from Russ(?) showing how it's done. I've got bitten before because of these
German umlauts we're using here.
thanks for the pointer. this is what worked
Got this part fixed, at least in part. I just set up a standalone CPU
server here at home, so I just run drawterm to there, import
my directory from the distant server, and play away. I guess if I had
this new CPU/auth server boot from the distant server's fossil, I wouldn't
even have to
ps: limiting that number from below is left as an exercise to the reader :)
my feynmann, we can further limit this number from above by noting
that we likely do not have 10^8 processors/person. 100 seems like
a reasonable upper bound. that would leave us with a mere 10^11
processors tops.
:-)
is this a google interview question? i love those NOT!
assume a microprocessor is 1 cm³ on average. the slabs of silicone
used to make such processors are 1 m³ in volume, so one of those can
make 1 000 000 (10⁶) microprocessors. to make a single slab of
silicone 453kg of raw
I have 64MB on the box. However, the reason of the fault was out of
memory condition.
I was wondered that there was no swap configured in cpurc (termrc do have
this).
Copying lines which configure swap from termrc to cpurc have solved the
problem.
Thanks to everybody,
Yaroslav.
it
I have 64MB on the box. However, the reason of the fault was out of
memory condition.
I was wondered that there was no swap configured in cpurc (termrc do have
this).
Copying lines which configure swap from termrc to cpurc have solved the
problem.
Thanks to everybody,
Yaroslav.
i
i wonder if the kernel is too big for the 4MB of temporary page
table mappings that l.s sets up. the kernel itself decompresses
to ~3MB and starts at 1MB. there's not much room for the
kernel pagetable..
(i had that problem just over the weekend, but with ken's kernel.
switching to 4MB pages
page says 'converting from troff to postscript' but then never
actually displays anything. I'm sure there's something basic
I'm forgetting here--can somebody help me figure out what's
wrong?
the file descriptor is never closed, thus page never stops to render
what it's got.
what you want is
cd /sys/src/9/pc
mk 'CONF=pc' install
mk 'CONF=pcf' install
mk 'CONF=pccpuf' install
9fat:
cd /386
cp 9load 9pcf 9pccpuf /n/9fat
# reboot
Just a guess,
add a mk clean between each mk of a kernel, the mkfile doesn't always clean
up
On 10/15/07, erik quanstrom [EMAIL PROTECTED] wrote:
i wonder if the kernel is too big for the 4MB of temporary page
table mappings that l.s sets up. the kernel itself decompresses
to ~3MB and starts at 1MB. there's not much room for the
kernel pagetable..
(i had that problem just
Yes, I know I can make a custom kernel.
But I assume that pccpuf config file from the distribution CD will work.
Without any corrections. Custom kernel should be second step.
regardless of what should be, you have verified that it does not work.
we're officially onto step two.
- erik
It looks like Erik gets the prize. How did the kernels get to be so big?
venti? (ha!)
- erik
Is the usable size of kernel limited?
I use NOT gziped variation of kernel, just as is written in Compiling
kernels (Plan 9 wiki).
Am I wrong?
Antonin
yes. l.s sets up a temporary page table with 4MB mapped. the whole
uncompressed kernel plus any memory needed for the kernel page table
i had a need the other day to do an e820 scan from
a kernel without the plan 9 realmode business. since
switching back to realmode seems a bit problematic,
and i have no need for realmode in the stuff i'm doing,
i wanted to do it before the switch to protected mode,
which happens in 9load. so i
301 - 400 of 1350 matches
Mail list logo