[9fans] 9atom installation report on Acer C710 Chromebook

[Ramakrishnan Muthukrishnan]

Hi,

I have an Acer C710 Chromebook
()
with chromeos bootloader replaced with seabios+coreboot from John Lewis
(). The
machine used to work fairly well with this bootloader (except some
glitches like non-working suspend/resume) with Debian GNU/Linux. The one
I have has Celeron 1007u processor.

This morning, I decided to install 9atom. The installation went quite
well but towards the end, while it is compiling amd64 executables, I got
a panic and a freeze.

panic: cpu0: cclose 0xf01c45f8
dumpstack: disabled
cpu0: exiting
cpu1: exiting

Will be be happy to help debug and get 9atom to run fine on this
machine. It has got a 16GB SSD and 2GB RAM. I chose fossil+venti while
formatting the disk. The installation process was the very smooth except
for the above errors.

Ramakrishnan

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[erik quanstrom]

> This morning, I decided to install 9atom. The installation went quite
> well but towards the end, while it is compiling amd64 executables, I got
> a panic and a freeze.
> 
> panic: cpu0: cclose 0xf01c45f8
> dumpstack: disabled
> cpu0: exiting
> cpu1: exiting

well, that's not good.  i will take a look.

- erik

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[erik quanstrom]

On Sat Feb 14 09:40:55 PST 2015, quans...@quanstro.net wrote:
> > This morning, I decided to install 9atom. The installation went quite
> > well but towards the end, while it is compiling amd64 executables, I got
> > a panic and a freeze.
> > 
> > panic: cpu0: cclose 0xf01c45f8
> > dumpstack: disabled
> > cpu0: exiting
> > cpu1: exiting
> 

for those following along at home, the panic is here

if(c->flag&CFREE)
panic("cclose %#p", getcallerpc(&c));

and called from

acid; src(0xf01c45f8)
/sys/src/nix/port/chan.c:240
 235return;
 236decref(&npath);
 237free(p->s);
 238for(i=0; imlen; i++)
 239if(p->mtpt[i])
>240cclose(p->mtpt[i]);
 241free(p->mtpt);
 242free(p);
 243}
 244
 245/*

it would be interesting to know if a second try at installation
also fails in the same way.  if it does not, then it will look a lot
like a race in chan.c.  if it fails in the same way, we really can't
infer anything.

- erik

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[Ramakrishnan Muthukrishnan]

On Sat, Feb 14, 2015, at 11:22 PM, erik quanstrom wrote:
> On Sat Feb 14 09:40:55 PST 2015, quans...@quanstro.net wrote:
> > > This morning, I decided to install 9atom. The installation went quite
> > > well but towards the end, while it is compiling amd64 executables, I got
> > > a panic and a freeze.
> > > 
> > > panic: cpu0: cclose 0xf01c45f8
> > > dumpstack: disabled
> > > cpu0: exiting
> > > cpu1: exiting
> > 
> 
> for those following along at home, the panic is here
> 
>   if(c->flag&CFREE)
>   panic("cclose %#p", getcallerpc(&c));
> 
> and called from
> 
> acid; src(0xf01c45f8)
> /sys/src/nix/port/chan.c:240
>  235 return;
>  236 decref(&npath);
>  237 free(p->s);
>  238 for(i=0; imlen; i++)
>  239 if(p->mtpt[i])
> >240  cclose(p->mtpt[i]);
>  241 free(p->mtpt);
>  242 free(p);
>  243 }
>  244 
>  245 /*
> 
> it would be interesting to know if a second try at installation
> also fails in the same way.  if it does not, then it will look a lot
> like a race in chan.c.  if it fails in the same way, we really can't
> infer anything.

I see a crash elsewhere now. Tried installation 3 times again and all
crashed at different places. 

One of them is:

rc: note: sys: trap: fault write addr=0x7ee0 pc=0x20d790
rc 9024: suicide: sys: trap: ...

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[Ramakrishnan Muthukrishnan]

On Sun, Feb 15, 2015, at 11:43 AM, Ramakrishnan Muthukrishnan wrote:
> On Sat, Feb 14, 2015, at 11:22 PM, erik quanstrom wrote:
> > On Sat Feb 14 09:40:55 PST 2015, quans...@quanstro.net wrote:
> > > > This morning, I decided to install 9atom. The installation went quite
> > > > well but towards the end, while it is compiling amd64 executables, I got
> > > > a panic and a freeze.
> > > > 
> > > > panic: cpu0: cclose 0xf01c45f8
> > > > dumpstack: disabled
> > > > cpu0: exiting
> > > > cpu1: exiting
> > > 
> > 
> > for those following along at home, the panic is here
> > 
> > if(c->flag&CFREE)
> > panic("cclose %#p", getcallerpc(&c));
> > 
> > and called from
> > 
> > acid; src(0xf01c45f8)
> > /sys/src/nix/port/chan.c:240
> >  235 return;
> >  236 decref(&npath);
> >  237 free(p->s);
> >  238 for(i=0; imlen; i++)
> >  239 if(p->mtpt[i])
> > >240cclose(p->mtpt[i]);
> >  241 free(p->mtpt);
> >  242 free(p);
> >  243 }
> >  244 
> >  245 /*
> > 
> > it would be interesting to know if a second try at installation
> > also fails in the same way.  if it does not, then it will look a lot
> > like a race in chan.c.  if it fails in the same way, we really can't
> > infer anything.
> 
> I see a crash elsewhere now. Tried installation 3 times again and all
> crashed at different places. 
> 
> One of them is:
> 
> rc: note: sys: trap: fault write addr=0x7ee0 pc=0x20d790
> rc 9024: suicide: sys: trap: ...

Perhaps I should add that I tried the 386 install as well and that
crashed randomly as well. 

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[erik quanstrom]

> > I see a crash elsewhere now. Tried installation 3 times again and all
> > crashed at different places. 
> > 
> > One of them is:
> > 
> > rc: note: sys: trap: fault write addr=0x7ee0 pc=0x20d790
> > rc 9024: suicide: sys: trap: ...
> 
> Perhaps I should add that I tried the 386 install as well and that
> crashed randomly as well. 

that's quite wierd.  either (a) we're corrupting memory on a regular
basis, and your machine has an unfortunate memory layout that
gets hit by this issue, or (b) there actually is something special about
your machine.

* is the internal disk presenting an ahci interface?  

* have you tried memtest86.  i expect it to pass, but it's always good
to be sure.

* can you point me at a specs page?

- erik

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[Ramakrishnan Muthukrishnan]

On Mon, Feb 23, 2015, at 07:39 PM, erik quanstrom wrote:
> > > I see a crash elsewhere now. Tried installation 3 times again and all
> > > crashed at different places. 
> > > 
> > > One of them is:
> > > 
> > > rc: note: sys: trap: fault write addr=0x7ee0 pc=0x20d790
> > > rc 9024: suicide: sys: trap: ...
> > 
> > Perhaps I should add that I tried the 386 install as well and that
> > crashed randomly as well. 
> 
> that's quite wierd.  either (a) we're corrupting memory on a regular
> basis, and your machine has an unfortunate memory layout that
> gets hit by this issue, or (b) there actually is something special about
> your machine.
> 
> * is the internal disk presenting an ahci interface?  

I don't know. I will check today evening.

> * have you tried memtest86.  i expect it to pass, but it's always good
> to be sure.

No, I haven't. I will do that too tonight.

> * can you point me at a specs page?



The spec of my machine is this:

- CPU - Celeron 1007u
- 16GB SSD
- 2 GB RAM
(rest as per the above page)

I modified the boot rom to have coreboot + seabios payload so that I can
install a non-chromeos system. It used to happily run Debian but was
locking up occasionally. Perhaps it has a bad RAM or a faulty DDR
initialization in the coreboot.

Re: [9fans] 9atom installation report on Acer C710 Chromebook

[Ramakrishnan Muthukrishnan]

On Tue, Feb 24, 2015, at 09:50 AM, Ramakrishnan Muthukrishnan wrote:
> On Mon, Feb 23, 2015, at 07:39 PM, erik quanstrom wrote:
> > > > I see a crash elsewhere now. Tried installation 3 times again and all
> > > > crashed at different places. 
> > > > 
> > > > One of them is:
> > > > 
> > > > rc: note: sys: trap: fault write addr=0x7ee0 pc=0x20d790
> > > > rc 9024: suicide: sys: trap: ...
> > > 
> > > Perhaps I should add that I tried the 386 install as well and that
> > > crashed randomly as well. 
> > 
> > that's quite wierd.  either (a) we're corrupting memory on a regular
> > basis, and your machine has an unfortunate memory layout that
> > gets hit by this issue, or (b) there actually is something special about
> > your machine.
> > 
> > * is the internal disk presenting an ahci interface?  
> 
> I don't know. I will check today evening.
> 
> > * have you tried memtest86.  i expect it to pass, but it's always good
> > to be sure.
> 
> No, I haven't. I will do that too tonight.

Ran memtest86 (booted off a USB stick running grml) and it didn't show
any memory errors.

> > * can you point me at a specs page?
> 
> 
> 
> The spec of my machine is this:
> 
> - CPU - Celeron 1007u
> - 16GB SSD
> - 2 GB RAM
> (rest as per the above page)
> 
> I modified the boot rom to have coreboot + seabios payload so that I can
> install a non-chromeos system. It used to happily run Debian but was
> locking up occasionally. Perhaps it has a bad RAM or a faulty DDR
> initialization in the coreboot.