Xen initrd

Xen initrd

[Kurt Garloff]

[-- Attachment #1: Type: text/plain, Size: 311 bytes --]

Hi,

docu says that a second module config line in the grub/menu.lst file
would load the initrd. This fails for me.
Is the docu correct and it should work?

PS: If this is the wrong list to ask such questions, please advice.
-- 
Kurt Garloff                   <kurt@garloff.de>             [Koeln, DE]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Xen initrd

[Ian Pratt]

> docu says that a second module config line in the grub/menu.lst file
> would load the initrd. This fails for me.
> Is the docu correct and it should work?

It certainly works for all the versions of grub I've ever tried
it on. Having an 'initrd' line doesn't.

Odd.

Ian


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Perry Couprie]

Hoi,

i am having problem starting a xen kernel with Adaptec Raid card,
is it possible to use the dpt-i2o card under xen.
I am using kernel 2.6.

RAID bus controller: Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

Greeting Perry.


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Martin Maney]

On Tue, Nov 30, 2004 at 03:48:09PM +0100, Perry Couprie wrote:
> i am having problem starting a xen kernel with Adaptec Raid card,
> is it possible to use the dpt-i2o card under xen.
> I am using kernel 2.6.

If you're using the easy binary installer, perhaps it simply wasn't
built with support for that controller?  I ran into that on an old
Proliant - the panic was perfectly explained when I realized it didn't
have the 53c8xx driver built in, and a suitably adjusted build fixed it
right up.

Speaking of rebuilding xen0/xenU kernels, what's the suggested
procedure to get to the point where one can "make *config" to make
changes?  I may have overlooked it, but what I ended up doing, almost
accidentally, was interrupting the xen0 build (I must have pushed the
wrong button, but my notes are unclear at this point), then clean and
*config and after that it worked very nicely indeed.  But slowly, which
reminds me to ask if the Xen build has a toggle for using both of the
CPUs - on a single PII/450 the pair of kernels took around 50 minutes.

-- 
The phenomenon of financial excess associated with
promising novel technologies is a recurring feature
of the last two centuries.  -- Andrew Odlyzko



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Jan Kundrát]

Martin Maney wrote:
> But slowly, which
> reminds me to ask if the Xen build has a toggle for using both of the
> CPUs - on a single PII/450 the pair of kernels took around 50 minutes.

`make -j3 vmlinuz`?
I read somewhere that the number should be number_of_cpus + 1.





-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Brian Wolfe]

Yes, this creates the optimal CPU load. at any one time two fo the three
should always be compiling while the third is doing reads/writes to the
filesystem.

More than #cpu+1 and you end up with procs waiting for cpu time. Less
and yoiu leave CPU time idle.

On a UP machine doing j > 1 actually slows things down due to scheduling
conflicts.

On Wed, 2004-12-01 at 16:21 +0100, Jan Kundrát wrote:
> Martin Maney wrote:
> > But slowly, which
> > reminds me to ask if the Xen build has a toggle for using both of the
> > CPUs - on a single PII/450 the pair of kernels took around 50 minutes.
>
> `make -j3 vmlinuz`?
> I read somewhere that the number should be number_of_cpus + 1.
>
>
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel
--



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Martin Maney]

On Wed, Dec 01, 2004 at 10:31:17AM -0600, Brian Wolfe wrote:
> Yes, this creates the optimal CPU load. at any one time two fo the three
> should always be compiling while the third is doing reads/writes to the
> filesystem.

> On Wed, 2004-12-01 at 16:21 +0100, Jan Kundr?t wrote:
> > `make -j3 vmlinuz`?
> > I read somewhere that the number should be number_of_cpus + 1.

Right.  Sorry, I hadn't set out to ask and the question came out a
little ambiguous.  What I *should* have asked was whether the Xen build
as a whole, or even just the "make kernels" part, was known to work
with a parallelized build.  I can't recall what long-ago experience
made me think it might be a problem.

-- 
The people who are serious in the open source world
don't really think that making money is an evil process.
Not even Richard Stallman.  Where you get that from is your
cliche 12-year old on Slashdot who doesn't understand it. -- Bruce Perens



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Ian Pratt]

> i am having problem starting a xen kernel with Adaptec Raid card,
> is it possible to use the dpt-i2o card under xen.
> I am using kernel 2.6.
> 
> RAID bus controller: Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

I think there's been at least one other report of this card not
working. It's most likely due to a latent bug in the driver,
perhaps confusion over bus vs physical addresses (you can get
away with this on native x86). 

If you enable the driver in the config and build it with tracing
etc then it'll probably be fairly easy to fix. Eyeballing the
code might even do it.

Ian


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen and Adaptec (formerly DPT) SmartRAID V Controller (rev 01)

[Keir Fraser]

 > 
> > i am having problem starting a xen kernel with Adaptec Raid card,
> > is it possible to use the dpt-i2o card under xen.
> > I am using kernel 2.6.
> > 
> > RAID bus controller: Adaptec (formerly DPT) SmartRAID V Controller (rev 01)
> 
> I think there's been at least one other report of this card not
> working. It's most likely due to a latent bug in the driver,
> perhaps confusion over bus vs physical addresses (you can get
> away with this on native x86). 
> 
> If you enable the driver in the config and build it with tracing
> etc then it'll probably be fairly easy to fix. Eyeballing the
> code might even do it.
> 
> Ian

There's one use of virt_to_phys() in the driver
(drivers/scsi/dpt_i20.c). You might want to try changing that to
virt_to_bus(), and report back.

 -- Keir


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Xen initrd

[Kurt Garloff]

[-- Attachment #1: Type: text/plain, Size: 729 bytes --]

Hi Ian,

On Tue, Nov 30, 2004 at 01:47:47PM +0000, Ian Pratt wrote:
> Kurt wrote:
> > docu says that a second module config line in the grub/menu.lst file
> > would load the initrd. This fails for me.
> > Is the docu correct and it should work?
> 
> It certainly works for all the versions of grub I've ever tried
> it on. Having an 'initrd' line doesn't.
> 
> Odd.

Actually it does work if I feed uncompressed initrd images to grub.
It does not seem to like gzipped ones. 

Regards,
-- 
Kurt Garloff                   <kurt@garloff.de>             [Koeln, DE]
Physics:Plasma modeling <garloff@plasimo.phys.tue.nl> [TU Eindhoven, NL]
Linux: SUSE Labs (Director)    <garloff@suse.de>            [Novell Inc]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Xen initrd

[Kurt Garloff]

[-- Attachment #1: Type: text/plain, Size: 694 bytes --]

Hi,

On Fri, Dec 10, 2004 at 09:20:55AM +0100, Kurt Garloff wrote:
> Actually it does work if I feed uncompressed initrd images to grub.
> It does not seem to like gzipped ones. 

To document the solved mystery:
Our mkinitrd does attach the bootsplash screen to the gzipped initrd
image. grub decompressor gets confused and bails out.

Uncompressing throws away the trailing garbage (words from gunzip).
It still works after recompressing ...

Regards,
-- 
Kurt Garloff                   <kurt@garloff.de>             [Koeln, DE]
Physics:Plasma modeling <garloff@plasimo.phys.tue.nl> [TU Eindhoven, NL]
Linux: SUSE Labs (Director)    <garloff@suse.de>            [Novell Inc]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Xen initrd

[Rik van Riel]

On Tue, 30 Nov 2004, Kurt Garloff wrote:

> docu says that a second module config line in the grub/menu.lst file
> would load the initrd. This fails for me.
> Is the docu correct and it should work?

On my laptop it looks like this, and works...

title Fedora Core (2.6.9-1.1006_FC4.rielxen0)
         root (hd0,0)
         kernel /xen-2.6.9.gz dom0_mem=180000
         module /vmlinuz-2.6.9-1.1006_FC4.rielxen0 ro root=LABEL=/ debug 3 xencons=off
         module /initrd-2.6.9-1.1006_FC4.rielxen0.img


-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/