Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on Microsoft Virtual PC

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on Microsoft Virtual PC

[Sasidhar Kasturi]

erik quanstrom wrote:
>> Hello,
>>        I tried to install plan 9 on virtual pc and succedded. But,
>> once i boot through plan 9  errors are coming.
>> Attached the screen shot of the error.
>>
>
> the plan 9 ide driver apparently disagrees with virtual pc.
> this isn't a fossil or venti problem.
>
> unfortunately, i have no access to virtual pc so the rest of this
> email is a thought experiment ....
>
> command 30
> data f09534b0 limit f0964eb0 dlen 8192 status 0 error 0
> lba 210592 -> 210592, count 16 -> 16 (16)
>  0x00 0x04 0xac 0x36 0x03 0xe0 0x58
> 0x40: e306 0x42: c0000 x48: 00
> 0x4a 0000
>
> so
>     the command is write sectors (Cws; 0x30)
>     lba 0x0336a0 (according to the driver)
>
> the registers say:
>
> 0    1    2    3    4    5    6    As
>     0x00    0x04    0xac    0x36    0x03    0xe0    0x58
>     error    count    lba0    lba1    lba2    stat
>
> at lba 0x3036ac (4 sectors to go) status Bsy|Df|Drdy.
>
> i would think you're issuing a command before the device
> is ready for you.  this patch may help.
>
> minooka; diff -c /sys/src/9/pc/sdata.c
> /n/dump/2008/0908/sys/src/9/pc/sdata.c
> /sys/src/9/pc/sdata.c:1418,1424 -
> /n/dump/2008/0908/sys/src/9/pc/sdata.c:1418,1424
>       ctlr = drive->ctlr;
>       cmdport = ctlr->cmdport;
>       ctlport = ctlr->ctlport;
> -     if(ataready(cmdport, ctlport, drive->dev, Bsy|Drq, Drdy,
> 101*1000) < 0)
> +     if(ataready(cmdport, ctlport, drive->dev, Bsy|Drq, 0, 101*1000)
> < 0)
>           return -1;
>         ilock(ctlr);
>
> - erik
>
>
>

Hello Erik,
      Thank you for that.  But, I don't have the source code. Can you
please point me to the place where i can download the source of plan 9.
Also, how to rebuild the code once i get it. If there is any step by
step tutorial sort of thing, please let me know.

Thanking you,
Sasidhar.

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on Microsoft Virtual PC

[grai]

> can you please point me to the place where i can download the source of plan 9. Also, how
> to rebuild the code once i get it.

If you want to try Erik's patch, you could install plan 9 on real
hardware or under a different virtual machine like qemu or vmware, and
follow the instructions at
http://plan9.bell-labs.com/wiki/plan9/Compiling_kernels/

Or just keep using the other VM and give up on Microsoft Virtual PC
for the moment.

grai

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on

[erik quanstrom]

>> can you please point me to the place where i can download the source of plan 9. Also, how
>> to rebuild the code once i get it.
>
> If you want to try Erik's patch, you could install plan 9 on real
> hardware or under a different virtual machine like qemu or vmware, and
> follow the instructions at
> http://plan9.bell-labs.com/wiki/plan9/Compiling_kernels/
>
> Or just keep using the other VM and give up on Microsoft Virtual PC
> for the moment.

this is a repeat of a private email; i didn't think
this was going to be generally interesting.

option 1:
assuming the live cd boots, there is a test kernel on sources
in my contrib area which can be put into 9fat without booting
from the virtual hard drive.

	open a new window in the installer.  i'll assume that /dev/sdC0
	is your virtual hard drive.

	; cd /dev/sdC0
	; disk/fdisk -p data>ctl ; disk/prep -p plan9>ctl
	; 9fat:
	; cd /n/9fat:
	; 9fs sources
	; cp /n/sources/contrib/quanstro/9pcf.gz 9pcide.gz
	now edit plan9.ini.  add a second bootfile line like:
	bootfile=sdC0!9fat!9pcide.gz

now reboot and hope. :-)

(for those following along at home, the sdata.c i'm working from
is in /n/sources/contrib/quanstro/sdata.c.)

option 2:
install onto aoe storage.  vblade
	http://sourceforge.net/project/showfiles.php?group_id=130453&package_id=143790
runs on linux, if you have a linux machine.  or perhaps you could
run it on virtual pc.  (there's a version for plan 9, too.)
you'll need to put these lines into your plan9.ini

	# assuming that your vblade is running as 42.0
	# and you are using ether0 and ether1 to serve aoe:

	aoeif=ether0 ether1
	aoedev=f!#æ/aoe/42.0

(in any event, there are many ways to try to get this working.
e.g. if you had a seperate machine as a fileserver, you could boot
from it and be more free to experiment.)

dispite the gratitous ad hominem, i'm sorry that this is difficult.
unfortunately, i don't have a windows box or virtual pc.  thus
i can't properly test the change, even though it seems required
by the ata spec.

even if i did, it's presumptious to assume that the folks at the
labs have the time to review changes and incorporate them.

- erik

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on

[grai]

> in any event, there are many ways to try to get this working.

(I meant to include something in my email about a clumsy method like
re-packing plan9.iso with the modified kernel.)

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on

[Federico G. Benavento]

relax

On Sat, Oct 11, 2008 at 4:43 PM, Eris Discordia
<eris.discordia@gmail.com> wrote:
>> dispite the gratitous ad hominem, i'm sorry that this is difficult.
>
> Don't take it seriously, erik. Even if you do, what's wrong with being a
> cold, insensitive, unsympathetic computer wizard (that's what high IQ and
> low EQ means)? There are a number of them out there, in very nice positions,
> exalted by many admirers from around the world. It's a good way of being.
>
> --On Saturday, October 11, 2008 10:43 AM -0400 erik quanstrom
> <quanstro@quanstro.net> wrote:
>
>>>> can you please point me to the place where i can download the source of
>>>> plan 9. Also, how to rebuild the code once i get it.
>>>
>>> If you want to try Erik's patch, you could install plan 9 on real
>>> hardware or under a different virtual machine like qemu or vmware, and
>>> follow the instructions at
>>> http://plan9.bell-labs.com/wiki/plan9/Compiling_kernels/
>>>
>>> Or just keep using the other VM and give up on Microsoft Virtual PC
>>> for the moment.
>>
>> this is a repeat of a private email; i didn't think
>> this was going to be generally interesting.
>>
>> option 1:
>> assuming the live cd boots, there is a test kernel on sources
>> in my contrib area which can be put into 9fat without booting
>> from the virtual hard drive.
>>
>>        open a new window in the installer.  i'll assume that /dev/sdC0
>>        is your virtual hard drive.
>>
>>        ; cd /dev/sdC0
>>        ; disk/fdisk -p data>ctl ; disk/prep -p plan9>ctl
>>        ; 9fat:
>>        ; cd /n/9fat:
>>        ; 9fs sources
>>        ; cp /n/sources/contrib/quanstro/9pcf.gz 9pcide.gz
>>        now edit plan9.ini.  add a second bootfile line like:
>>        bootfile=sdC0!9fat!9pcide.gz
>>
>> now reboot and hope. :-)
>>
>> (for those following along at home, the sdata.c i'm working from
>> is in /n/sources/contrib/quanstro/sdata.c.)
>>
>> option 2:
>> install onto aoe storage.  vblade
>>
>>  http://sourceforge.net/project/showfiles.php?group_id=130453&package_id=
>> 143790 runs on linux, if you have a linux machine.  or perhaps you could
>> run it on virtual pc.  (there's a version for plan 9, too.)
>> you'll need to put these lines into your plan9.ini
>>
>>        # assuming that your vblade is running as 42.0
>>        # and you are using ether0 and ether1 to serve aoe:
>>
>>        aoeif=ether0 ether1
>>        aoedev=f!#æ/aoe/42.0
>>
>> (in any event, there are many ways to try to get this working.
>> e.g. if you had a seperate machine as a fileserver, you could boot
>> from it and be more free to experiment.)
>>
>> dispite the gratitous ad hominem, i'm sorry that this is difficult.
>> unfortunately, i don't have a windows box or virtual pc.  thus
>> i can't properly test the change, even though it seems required
>> by the ata spec.
>>
>> even if i did, it's presumptious to assume that the folks at the
>> labs have the time to review changes and incorporate them.
>>
>> - erik
>>
>>
>
>



-- 
Federico G. Benavento

Re: [9fans] fossil:diskWriteRawFailed error with Plan 9 on

[Eris Discordia]

> dispite the gratitous ad hominem, i'm sorry that this is difficult.

Don't take it seriously, erik. Even if you do, what's wrong with being a 
cold, insensitive, unsympathetic computer wizard (that's what high IQ and 
low EQ means)? There are a number of them out there, in very nice 
positions, exalted by many admirers from around the world. It's a good way 
of being.

--On Saturday, October 11, 2008 10:43 AM -0400 erik quanstrom 
<quanstro@quanstro.net> wrote:

>>> can you please point me to the place where i can download the source of
>>> plan 9. Also, how to rebuild the code once i get it.
>>
>> If you want to try Erik's patch, you could install plan 9 on real
>> hardware or under a different virtual machine like qemu or vmware, and
>> follow the instructions at
>> http://plan9.bell-labs.com/wiki/plan9/Compiling_kernels/
>>
>> Or just keep using the other VM and give up on Microsoft Virtual PC
>> for the moment.
>
> this is a repeat of a private email; i didn't think
> this was going to be generally interesting.
>
> option 1:
> assuming the live cd boots, there is a test kernel on sources
> in my contrib area which can be put into 9fat without booting
> from the virtual hard drive.
>
> 	open a new window in the installer.  i'll assume that /dev/sdC0
> 	is your virtual hard drive.
>
> 	; cd /dev/sdC0
> 	; disk/fdisk -p data>ctl ; disk/prep -p plan9>ctl
> 	; 9fat:
> 	; cd /n/9fat:
> 	; 9fs sources
> 	; cp /n/sources/contrib/quanstro/9pcf.gz 9pcide.gz
> 	now edit plan9.ini.  add a second bootfile line like:
> 	bootfile=sdC0!9fat!9pcide.gz
>
> now reboot and hope. :-)
>
> (for those following along at home, the sdata.c i'm working from
> is in /n/sources/contrib/quanstro/sdata.c.)
>
> option 2:
> install onto aoe storage.  vblade
> 	http://sourceforge.net/project/showfiles.php?group_id=130453&package_id=
> 143790 runs on linux, if you have a linux machine.  or perhaps you could
> run it on virtual pc.  (there's a version for plan 9, too.)
> you'll need to put these lines into your plan9.ini
>
> 	# assuming that your vblade is running as 42.0
> 	# and you are using ether0 and ether1 to serve aoe:
>
> 	aoeif=ether0 ether1
> 	aoedev=f!#æ/aoe/42.0
>
> (in any event, there are many ways to try to get this working.
> e.g. if you had a seperate machine as a fileserver, you could boot
> from it and be more free to experiment.)
>
> dispite the gratitous ad hominem, i'm sorry that this is difficult.
> unfortunately, i don't have a windows box or virtual pc.  thus
> i can't properly test the change, even though it seems required
> by the ata spec.
>
> even if i did, it's presumptious to assume that the folks at the
> labs have the time to review changes and incorporate them.
>
> - erik
>
>