[9fans] more bitsy stuff

[9fans] more bitsy stuff

[john]

So it turns out that I can build a kernel that will boot Axel's old
paqdisk without a problem, but if I use my own kernel with my own
freshly created paqdisk, I get the crash below:


root is from (paq)[paq]:

paq...paqfs: sysfatal reading block: /dev/flash/ramdisk:

version...panic: boot process died: sys: write on closed pipe pc=0x00006c74

panic: boktrace /kernel/path c0010d28 c011714c c01171a9

ot procc01170f8=c0010d24 ess diec01170fc=c0070aec d: sys:c0117104=c003afe4  write c0117128=c003b044 on clos

ed pipe c0117140=c003b2a8 pc=0x00c0117144=c0010d24 006c74
c0117148=c003b2ac
c0117150=c006d790

c01171b0=c008d610 c01171b4=c008d3bc c01171bc=c008f200 c01171c8=c0078f94

c01171d4=c008d14c c01171d8=c008e37c c01171dc=c008d510 c01171e0=c008d14c

c01171e8=c008d3f8 c01171f4=c008d3bc c01171fc=c008f200 c0117208=c0078f94

c011721c=c008f6c0 c011723c=c006022c c0117248=c006add0 c0117250=c006add8

c011725c=c006d750 c0117274=c00108f8 c0117280=c00790b8 c0117284=c0079594

c01172ac=c0070fc0 c01172b4=c00108dc c01172c8=c0010418 c01172e8=c0008248

c01172f0=c0079594

it's a wonderful day to die

This makes me think that there was perhaps a change in mkpaqfs along
the line?  To look at the error, it seems that there's something wrong
with my ramdisk image; any insights, anyone?  Who has been maintaining
the bitsy stuff?


John

[9fans] ARMv6 compiler ? (was: more bitsy stuff)

[prem]

Hi All,
I am thinking of porting Plan9 to latest ARM cores including v6 and v7
(in my upcoming holidays). I think the compiler is old (pre-v5) though
i am not sure.
could anyone tell me if 5c and friends support v6. if it doesn't how
difficult it is to add it?
Also I have installed latest inferno sources from google svn (on
Linux). is it possible to (cross-)compile plan9 sources on Linux ?

thanks
/Prem

On Mar 14, 2:20 am, j...@csplan9.rit.edu wrote:
> So it turns out that I can build a kernel that will boot Axel's old
> paqdisk without a problem, but if I use my own kernel with my own
> freshly created paqdisk, I get the crash below:
>
> root is from (paq)[paq]:
>
> paq...paqfs: sysfatal reading block: /dev/flash/ramdisk:
>
> version...panic: boot process died: sys: write on closed pipe pc=0x00006c74
>
> panic: boktrace /kernel/path c0010d28 c011714c c01171a9
>
> ot procc01170f8=c0010d24 ess diec01170fc=c0070aec d: sys:c0117104=c003afe4  write c0117128=c003b044 on clos
>
> ed pipe c0117140=c003b2a8 pc=0x00c0117144=c0010d24 006c74
> c0117148=c003b2ac
> c0117150=c006d790
>
> c01171b0=c008d610 c01171b4=c008d3bc c01171bc=c008f200 c01171c8=c0078f94
>
> c01171d4=c008d14c c01171d8=c008e37c c01171dc=c008d510 c01171e0=c008d14c
>
> c01171e8=c008d3f8 c01171f4=c008d3bc c01171fc=c008f200 c0117208=c0078f94
>
> c011721c=c008f6c0 c011723c=c006022c c0117248=c006add0 c0117250=c006add8
>
> c011725c=c006d750 c0117274=c00108f8 c0117280=c00790b8 c0117284=c0079594
>
> c01172ac=c0070fc0 c01172b4=c00108dc c01172c8=c0010418 c01172e8=c0008248
>
> c01172f0=c0079594
>
> it's a wonderful day to die
>
> This makes me think that there was perhaps a change in mkpaqfs along
> the line?  To look at the error, it seems that there's something wrong
> with my ramdisk image; any insights, anyone?  Who has been maintaining
> the bitsy stuff?
>
> John

Re: [9fans] ARMv6 compiler ? (was: more bitsy stuff)

[Charles Forsyth]

>is it possible to (cross-)compile plan9 sources on Linux ?

i did that for Solaris years ago. the compilers are the easy bit; you
also need to modify mkfiles to add a root variable.  (it was 8 years ago
and i can't remember many details.)  for Linux, now, you're probably
better off if you can get THNX going, which makes plan 9 run as a guest under Linux.
better still is just to build or borrow a cheap machine and run Plan 9 on it.

i'll need to look at the chart, if i can find one, that gives the differences between the
various arm stages.  until the most recent revisions with FP i didn't think there was much
of interest to add to the compiler, from C's point of view.  (the compiler does generate the old FP.)

Re: [9fans] ARM compiler

[erik quanstrom]

speaking of arm:

does the compiler emit conditional instructions
like subgt?

- erik

Re: [9fans] ARM compiler

[Richard Miller]

> speaking of arm:
>
> does the compiler emit conditional instructions
> like subgt?

term% cat t.c
int
f(int x)
{
	int y = 0;
	if (x > 0)
		y -= 1;
	return y;
}
term% 5c -S t.c
	TEXT	f+0(SB),0,$4
	MOVW	$0,R2
	CMP	$0,R0,
	SUB.GT	$1,R2,R2
	MOVW	R2,R0
	RET	,
	RET	,
	END	,

Re: [9fans] ARM compiler

[Rob Pike]

Spoilsport.

-rob

On Tue, Mar 18, 2008 at 8:46 AM, Richard Miller <9fans@hamnavoe.com> wrote:
> > speaking of arm:
>  >
>  > does the compiler emit conditional instructions
>  > like subgt?
>
>  term% cat t.c
>  int
>  f(int x)
>  {
>         int y = 0;
>         if (x > 0)
>                 y -= 1;
>         return y;
>  }
>  term% 5c -S t.c
>         TEXT    f+0(SB),0,$4
>         MOVW    $0,R2
>         CMP     $0,R0,
>         SUB.GT  $1,R2,R2
>         MOVW    R2,R0
>         RET     ,
>         RET     ,
>         END     ,
>
>
>

Re: [9fans] ARM compiler

[Bruce Ellis]

rtfc(1) is on the rtfm(1) page.

grep of /sys/src/cmd/5c/*.c may have been enough.

brucee

On Tue, Mar 18, 2008 at 9:28 AM, Rob Pike <robpike@gmail.com> wrote:
> Spoilsport.
>
> -rob
>
>
> On Tue, Mar 18, 2008 at 8:46 AM, Richard Miller <9fans@hamnavoe.com> wrote:
> > > speaking of arm:
> >  >
> >  > does the compiler emit conditional instructions
> >  > like subgt?
> >
> >  term% cat t.c
> >  int
> >  f(int x)
> >  {
> >         int y = 0;
> >         if (x > 0)
> >                 y -= 1;
> >         return y;
> >  }
> >  term% 5c -S t.c
> >         TEXT    f+0(SB),0,$4
> >         MOVW    $0,R2
> >         CMP     $0,R0,
> >         SUB.GT  $1,R2,R2
> >         MOVW    R2,R0
> >         RET     ,
> >         RET     ,
> >         END     ,
> >
> >
> >
>
>

Re: [9fans] ARM compiler

[erik quanstrom]

> rtfc(1) is on the rtfm(1) page.
>
> grep of /sys/src/cmd/5c/*.c may have been enough.
>
> brucee

despite your usual lack of information your jab is stale
and obvious.

it is interesting that your standards for appropriate 9fans
questions are in no way met by your retorts.

- erik

Re: [9fans] ARM compiler

[Bruce Ellis]

information was precise. ignore it.

brucee

On Tue, Mar 18, 2008 at 4:07 PM, erik quanstrom <quanstro@quanstro.net> wrote:
> > rtfc(1) is on the rtfm(1) page.
> >
> > grep of /sys/src/cmd/5c/*.c may have been enough.
> >
> > brucee
>
> despite your usual lack of information your jab is stale
> and obvious.
>
> it is interesting that your standards for appropriate 9fans
> questions are in no way met by your retorts.
>
> - erik
>
>
>