[9fans] third edition, installation experiences

[9fans] third edition, installation experiences

[Fco. J. Ballesteros]

Just to let you know that Plan 9 runs just fine on a Dell Inspiron
7k.

BTW, anyone with a Lucent wavelan driver? If none, I'm willing to
write one myself. But would prefer to reuse :-)

--
    ()    ascii ribbon campaign - against html mail
    /\                          - against microsoft attachments

Re: [9fans] third edition, installation experiences

[kajri jain]

in addition to scott's useful list,
i think if you have to add a ctlr to vgadb,
the modified vgadb doesn't make it to the hard disk.


....

> * If you told the install-floppy-maker that you have ethernet,
> but then you use the local media to load the distribution, the ip
> information never gets configured, which makes ndb/dns pause a while
> before timing out.  Don't panic if that happens.
>
> * It looks like glenda's home directory has a mode 0555 tmp directory
> in it, which should be 0777.

Re: [9fans] third edition, installation experiences

[Lyndon Nerenberg]

>Just to let you know that Plan 9 runs just fine on a Dell Inspiron
>7k.

Did you have to modify vgadb? When I tried to install it didn't recognise
the video chip. I added an entry to vgadb that forced the installation
to recognize the Rage LT Pro, but the result was trash on the screen.

Which display settings did you use when generating the floppy?

--lyndon

>What about all the people who hoarded tonnes of spam in their bunkers?

I hoard spam on my hard drive.  When I heard about the coming
Y2K worries, I downloaded a lifetime supply from the net.
			-- Charlie Gibbs in alt.folklore.computers

Re: [9fans] third edition, installation experiences

[Fco. J. Ballesteros]

>>>>> "Lyndon" == Lyndon Nerenberg <lyndon@messagingdirect.com> writes:

    >> Just to let you know that Plan 9 runs just fine on a Dell
    >> Inspiron 7k.

    Lyndon> Did you have to modify vgadb? When I tried to install it

No. I just lied, I just said I have a `versalx' (1024x768) even though I
don't think I got one on the i7k :-)

BTW, be very careful to really cold-boot your machine or you'll get
a funny rio display.

--
    ()    ascii ribbon campaign - against html mail
    /\                          - against microsoft attachments

Re: [9fans] third edition, installation experiences

[Lyndon Nerenberg]

    >>> Just to let you know that Plan 9 runs just fine on a Dell
    >>> Inspiron 7k.

    Lyndon> Did you have to modify vgadb? When I tried to install it

    > No. I just lied, I just said I have a `versalx' (1024x768) even though I
    > don't think I got one on the i7k :-)

Okay, I tried this, but still no joy. This time the screen is readable
enough that I can see the mouse pointer, but there's still lots of
junk, and the machine freezes up solid. (I did the cold-start from
power-down as you suggested.)

Could I get you to send me the plan9.ini, vgadb and vgainfo.txt from your
boot floppy? I'm beginning to think we might have different revisions
of the LT Pro chipset (or else something is different between our
hardware configurations).

Here's my vgainfo.txt from the latest attempt:

main->snarf
vga->snarf
mach64xx->snarf
vga->dump
vga misc             E7
vga feature          00
vga sequencer        03 00 03 00 02
vga crt              73 4F 4F 97 53 84 B4 1F - 00 4F 0D 0E 00 00 06 90
                     91 87 8F 28 1F 8F B5 A3 - FF
vga graphics         00 00 00 00 00 10 0E 00 - FF
vga attribute        00 01 02 03 04 05 14 07 - 38 39 3A 3B 3C 3D 3E 3F
                     0C 00 0F 08 00
vga vm a b           16777216        0
vga vmz               8388608
vga apz               8388608
vga linear                  1
vga->attr: 0xC00F4=MACH64
mach64xx->dump
mach64xx pci 2fce0 regtable pciregs io e800 portio
mach64xx ccru 300
mach64xx HTotalDisp          004F0073
mach64xx HSyncStrtWid        00040053
mach64xx VTotalDisp          018F01B4
mach64xx VSyncStrtWid        00070191
mach64xx VlineCrntVline      008B03FF
mach64xx OffPitch            0A000000
mach64xx IntCntl             C0000134
mach64xx CrtcGenCntl         024B0200
mach64xx OvrClr              B517DDD5
mach64xx OvrWidLR            00000000
mach64xx OvrWidTB            00000000
mach64xx CurClr0             7FACF009
mach64xx CurClr1             CC093460
mach64xx CurOffset           00000000
mach64xx CurHVposn           00000000
mach64xx CurHVoff            00000000
mach64xx ScratchReg0         04100400
mach64xx ScratchReg1         00000000
mach64xx ClockCntl           00AD0000
mach64xx BusCntl             7333A101
mach64xx MemCntl             10751A7B
mach64xx ExtMemCntl          64000C01
mach64xx MemVgaWpSel         00010000
mach64xx MemVgaRpSel         00010000
mach64xx DacRegs             00FF3440
mach64xx DacCntl             8701200A
mach64xx GenTestCntl         00000000
mach64xx ConfigCntl          00003F42
mach64xx ConfigChipId        DC004C42
mach64xx ConfigStat0         00A40095
mach64xx ConfigStat1         00000000
mach64xx DspConfig           004709D9
mach64xx DspOnOff            04F00714
mach64xx DpBkgdClr           FFFFFFFF
mach64xx DpChainMsk          FFFFFFFF
mach64xx DpFrgdClr           FFFFFFFF
mach64xx DpMix               FFFFFFFF
mach64xx DpPixWidth          FFFFFFFF
mach64xx DpSrc               FFFFFFFF
mach64xx DpWriteMsk          FFFFFFFF
mach64xx LcdIndex            00000405
mach64xx LcdData             80000000
mach64xx PLL         AD D5 40 64 D9 03 C0 D4 - D4 D4 00 F1 A6 1B 00 00
                     06 A8 80 00 18 A3 CC 10 - 08 00 00 00 00 00 00 02
mach64xx LCD ConfigPanel     0F030274
mach64xx LCD GenCntl         406D24D6
mach64xx LCD DstnCntl        00000000
mach64xx LCD HfbPitchAddr    00000F00
mach64xx LCD HorzStretch     C0000A00
mach64xx LCD VertStretch     80000000
mach64xx LCD ExtVertStretch  0057FF00
mach64xx LCD LtGio           00007000
mach64xx LCD PowerMngmnt     0200040B
mach64xx LCD ZvgPio          02000000
mach64xx VCLK0	31619314
mach64xx VCLK1	31619314
mach64xx VCLK2	31619314
mach64xx VCLK3	0

rom table offset 10A
freq table offset 8F6
memclk 100000000
ref_freq 29500000
ref_divider 64
min_freq 9840000
max_freq 236000000
pd 0 value 4 (|4)
post = 3
mach64xx pixel clock = 65140000

vmf 65000000 vmdf 75000000 vf1 0 vbw 0
vga->init
mach64xx->init
rom table offset 10A
freq table offset 8F6
memclk 10000
memclk 10000... x 12.307692...t 12... xprec 4...fifosz 393.846154...fprec 9...prec 4...afifosz 32...fifooff 380.000000...pfc 7...rcc 9...fifoon 42.000000...
dbdumpmode
type=versalx, size=1024x768x8
frequency=65000000
x=1024 (0x400), y=768 (0x300), z=8 (0x8)
ht=1328 (0x530), shb=1096 (0x448), ehb=1232 (0x4D0)
shs=1072 (0x430), ehs=1232 (0x4D0)
vt=806 (0x326), vrs=771 (0x303), vre=777 (0x309)
hsync=45, vsync=45, interlace=0
vga->dump
vga flag             Fdump|Finit|Fsnarf
vga misc             E3
vga feature          00
vga sequencer        03 01 0F 00 0A
vga crt              A1 7F 88 9A 86 1A324 FD - 00 60 00 00 00 00 00 00
                    303 292FF 80 60303304 A3 -7FF
vga graphics         00 00 00 00 00 50 05 0F - FF
vga attribute        00 01 02 03 04 05 06 07 - 08 09 0A 0B 0C 0D 0E 0F
                     41 FF 0F 00 00
vga clock[0] f       31619314
vga clock[0] d i m          0        0       -       64
vga clock[0] n p q r      212        3       -        0        0
vga vm a b           16777216        0
vga vmz               8388608
vga apz               8388608
vga linear                  1
vga->attr: 0xC00F4=MACH64
mach64xx->dump
mach64xx flag        Ulinear|Uenhanced|Fdump|Finit|Fsnarf
mach64xx pci 2fce0 regtable pciregs io e800 portio
mach64xx ccru 300
mach64xx HTotalDisp          007F00A5
mach64xx HSyncStrtWid        00340085
mach64xx VTotalDisp          02FF0325
mach64xx VSyncStrtWid        00260302
mach64xx VlineCrntVline      008B03FF
mach64xx OffPitch            20000000
mach64xx IntCntl             00000000
mach64xx CrtcGenCntl         030B0200
mach64xx OvrClr              00000000
mach64xx OvrWidLR            00000000
mach64xx OvrWidTB            00000000
mach64xx CurClr0             7FACF009
mach64xx CurClr1             CC093460
mach64xx CurOffset           00000000
mach64xx CurHVposn           00000000
mach64xx CurHVoff            00000000
mach64xx ScratchReg0         04100400
mach64xx ScratchReg1         00000000
mach64xx ClockCntl           00000002
mach64xx BusCntl             7333A101
mach64xx MemCntl             10751A7B
mach64xx ExtMemCntl          64000C01
mach64xx MemVgaWpSel         00010000
mach64xx MemVgaRpSel         00010000
mach64xx DacRegs             00FF3440
mach64xx DacCntl             8701200A
mach64xx GenTestCntl         00000000
mach64xx ConfigCntl          00000000
mach64xx ConfigChipId        DC004C42
mach64xx ConfigStat0         00A40095
mach64xx ConfigStat1         00000000
mach64xx DspConfig           004A0627
mach64xx DspOnOff            00A805F0
mach64xx DpBkgdClr           FFFFFFFF
mach64xx DpChainMsk          FFFFFFFF
mach64xx DpFrgdClr           FFFFFFFF
mach64xx DpMix               FFFFFFFF
mach64xx DpPixWidth          00020202
mach64xx DpSrc               FFFFFFFF
mach64xx DpWriteMsk          FFFFFFFF
mach64xx LcdIndex            00000405
mach64xx LcdData             80000000
mach64xx PLL         AD D5 40 64 D9 03 C0 D4 - D4 D4 00 F1 A6 1B 00 00
                     06 A8 80 00 18 A3 CC 10 - 08 00 00 00 00 00 00 02
mach64xx LCD ConfigPanel     0F034274
mach64xx LCD GenCntl         406D20D2
mach64xx LCD DstnCntl        00000000
mach64xx LCD HfbPitchAddr    00000F00
mach64xx LCD HorzStretch     00000A00
mach64xx LCD VertStretch     00000000
mach64xx LCD ExtVertStretch  0017FB00
mach64xx LCD LtGio           00007000
mach64xx LCD PowerMngmnt     0200040B
mach64xx LCD ZvgPio          02000000
mach64xx VCLK0	31619314
mach64xx VCLK1	31619314
mach64xx VCLK2	31619314
mach64xx VCLK3	0

rom table offset 10A
freq table offset 8F6
memclk 100000000
ref_freq 29500000
ref_divider 64
min_freq 9840000
max_freq 236000000
pd 0 value 4 (|4)
post = 3
mach64xx pixel clock = 65140000

main->exits


--lyndon

In the force if Yoda's so strong, construct a sentence with words in
the proper order then why can't he?

Re: [9fans] third edition, installation experiences

[Douglas A. Gwyn]

jmk@plan9.bell-labs.com wrote:
>         >  i8259enable: irq 7 shared but not level
>         >  intrenable: couldn't enable irq 7, tbdf 0xffffffff for lpt
> The reason for the message is that some other driver has registered as
> wanting to use that IRQ already and ...

Unfortunately I'm getting a similar conflict with IRQ 3, which in
my system is assigned for exclusive use of the 3C509B (I had no
real choice as there are no spare IRQs on my system and I even
had to sacrifice a serial port to get that one).  I assume this
is because the installation/resulting Plan 9 systems are assuming
the existence of 2 serial ports; how do I disable the second one
(in PLAN9.INI presumably)?

Re: [9fans] third edition, installation experiences

[Scott Schwartz]

jmk:
| There's a misunderstanding in the numerology, IRQ 7 has a base vector
| added to it of (in this case) 32, so you should look in '#P/irqalloc'
| for something using vector 39 (the output actually shows you which
| IRQ maps to which vector, important in SMP machines).

Ah, right.  The actual collision was with sb16, documented as using irq7.

Re: [9fans] third edition, installation experiences

[Scott Schwartz]

rsc:
> See if cdfs(4) is any better than the kernel.
> Every once in a while it is.

One has to compile that from source; it's not installed in /bin in the
distribution.  But it does seem to work.  Thanks for the pointer.

Re: [9fans] third edition, installation experiences

[jmk]

	>The other day I said something about not being able to
	>assign an irq.  The actual error message was:
	>
	>  i8259enable: irq 7 shared but not level
	>  intrenable: couldn't enable irq 7, tbdf 0xffffffff for lpt
	>
	>According to '#P/irqalloc',  irq 7 is assigned to mathemu.

There's a misunderstanding in the numerology, IRQ 7 has a base vector
added to it of (in this case) 32, so you should look in '#P/irqalloc'
for something using vector 39 (the output actually shows you which
IRQ maps to which vector, important in SMP machines).

The reason for the message is that some other driver has registered as
wanting to use that IRQ already and, as this is an 8259 PIC, you can't
share interrupts unless the corresponding bit is set in the Edge/Level
Control Register (ELCR) and apparently it isn't.

--jim

Re: [9fans] third edition, installation experiences

[presotto]

	I also had my first crash.  I echo-ed "kill" to the control
	file of a process '#IOtcpack".  Instant reboot.

You shouldn't be able to kill that process, I'll fix that.  It's a kernel
process that can leave bad pointers lying around if you terminate it.

Re: [9fans] third edition, installation experiences

[rob pike]

	/dev/draw seems slower than /dev/bitblt.  I wonder if the difference is
	from not doing runtime code generation.

Perhaps.  Draw() is also a much more general operator, which causes it to be
slower in the worst case.  On the other hand, we (well, Russ) worked on getting
draw() to talk to the graphics accelerator on some of the cards, and there are
simple examples like catting /dev/words where the new system is *much*
faster (factors of 10 or more) than even bitblt was.  I hope folks will
connect the code to the hardware on more graphics cards so the speedups
will be more widely available.

On the issue of run-time code generation, we abandoned it because it was
too hard to write portable code to handle the synchronization of instruction
and data caches.  The problem is not processors so much as systems; the
details of cache management in different architectures are often delegated
to motherboard makers who choose widely varying ways to handle cache
flushing and make it very hard for a portable program to divine what to do.
If you're just on PCs, it's not so hard, but we use a lot of non-PC machines.
Also, it's possible that the rise of JITs will help this situation, but if the
VGA industry is anything to go by...

	Plumbing is... interesting.  I normally like to have several editors
	running, keeping different work in different contexts.  The new system
	makes that difficult, because when one editor opens a file, the
	others want to open it too.

There was a long discussion about how and whether to handle fan-out in
the plumber.  What we did reflects our style of use.  If it doesn't suit your
needs, you have the source...  Also of course it's mostly controlled by
configuration files, so you might be able to try another arrangement just
by using different plumbing rules.  (The fan-out isn't explicit in the
configuration, but I suppose on reflection it could be made to be.) I've
also done tricks where I've rebound things to rearrange plumbing ports
for temporary purposes; that might work well for you.

Failing that, B is a shell script that could easily be adapted to use other
information to direct the file to a port other than edit, or to add
attributes that different editors could check.

-rob

Re: [9fans] third edition, installation experiences

[Russ Cox]

  My cdrom is recognized (cat sdD0/ctl looks sensible), but
  the data file is all zeros.

See if cdfs(4) is any better than the kernel.
Every once in a while it is.

  /dev/draw seems slower than /dev/bitblt.  I wonder if the difference is
  from not doing runtime code generation.

Yes, that's much of it; the code already has many
special checks to handle the common cases, but
the general case is indeed slower than it ought
to be.  I have a runtime code generator started (and stopped,
for the time being), but mostly as an experiment to see what
might be gained.  I expect that the problems in keeping the
instruction and data caches up to date on non-x86
and non-Intel x86 chips will make it not practical,
but maybe the win will be enormous.  I don't know.

  Plumbing is... interesting.  I normally like to have several editors
  running, keeping different work in different contexts.  The new system
  makes that difficult, because when one editor opens a file, the
  others want to open it too.

Run plumber in the window before starting the editor,
and it will be in its own ``plumbing space''
if you will.

Russ

Re: [9fans] third edition, installation experiences

[Scott Schwartz]

A few more first impressions:

The other day I said something about not being able to
assign an irq.  The actual error message was:

  i8259enable: irq 7 shared but not level
  intrenable: couldn't enable irq 7, tbdf 0xffffffff for lpt

According to '#P/irqalloc',  irq 7 is assigned to mathemu.

I also had my first crash.  I echo-ed "kill" to the control
file of a process '#IOtcpack".  Instant reboot.

My cdrom is recognized (cat sdD0/ctl looks sensible), but
the data file is all zeros.

/dev/draw seems slower than /dev/bitblt.  I wonder if the difference is
from not doing runtime code generation.

Plumbing is... interesting.  I normally like to have several editors
running, keeping different work in different contexts.  The new system
makes that difficult, because when one editor opens a file, the
others want to open it too.

Re: [9fans] third edition, installation experiences

[jmk]

Fco. J. Ballesteros <nemo@gsyc.escet.urjc.es> wrote:
	Just to let you know that Plan 9 runs just fine on a Dell Inspiron
	7k.

	BTW, anyone with a Lucent wavelan driver? If none, I'm willing to
	write one myself. But would prefer to reuse :-)

thanks for the feedback.

there was a driver for the old wavelan cards which we didn't include
as we were unsure if we were allowed to release it. there's a summer
student just started on writing a driver for the new 802.11 cards,
but a parallel effort on your part wouldn't be a bad thing.

--jim

[9fans] third edition, installation experiences

[Scott]

My notes are at home, but here are a few comments from memory.

* The installation instructions didn't say so, but the disk
partitioning programs accept '?' as a command to get some help.  That's
useful to learn that you can use 't' to change a fallow partition to
type PLAN9 (a fallow partition conveniently available, since I had been
planning to install Linux there for the last five years but somehow
never got around to it.)

* When the kernel boots, it says something about 128MB swap.  In the
second edition, the penalty for having less backing store allocated
than the kernel expected was a panic the first time one of those page
frames was written out.  Is that still the case?  The installation
instructions suggest that it should be harmless.

* The system complained about not being able to assign an irq to the
printer port.

* The vga stuff works fine with a #9 771 card.

* The installer is willing to try reading from an ext2 partition, but I
had mixed results with that.  (Ok, linux was recently installed on a
second disk.)  The mini-shell was unwilling to cd into /home, for
unknown reasons.  Putting plan9.9gz into / worked around that.

* If you told the install-floppy-maker that you have ethernet,
but then you use the local media to load the distribution, the ip
information never gets configured, which makes ndb/dns pause a while
before timing out.  Don't panic if that happens.

* It looks like glenda's home directory has a mode 0555 tmp directory
in it, which should be 0777.