Re: kdb: add rdmsr and wrmsr commands for i386

2007-05-17 Thread Bernardo Innocenti
Keith Owens wrote:

 Before using MSR, you must first check that the cpu supports the
 instruction, rd/wrmsr cause an oops on 486 or earlier.  Also using an
 invalid msr number causes an oops, so use rd/wrmsr_safe().

I didn't bother implementing those checks because kdb recovers
nicely from GPF anyway.  It's the valid MSR writes that could
cause unrecoveable problems! :)

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


Re: modprobe msr fail!

2007-05-17 Thread Mitch Bradley
CONFIG_X86_MSR=y

means that MSR support is compiled into the kernel, not as a module. So
you do not need to do modprobe. The functionality is built-in. There
will be a file named /dev/cpu/0/msr.


In image 386, the config was:

CONFIG_X86_MSR=m

so with that build, you did need to do the modprobe. Before you do the
modprobe, the /dev/cpu/0/msr file will not exist, but it will appear
after the modprobe.

[EMAIL PROTECTED] wrote:

 Hi,

 I am Luna.

 I want to modprobe msr on OLPC. (I use MIT image 426 on B2)

 But it is fail. The error message is “FATAL: Module msr not found.”

 To look at config file, I find

 CONFIG_X86_MSR=y

 Someone can tell me what happen.

 I had returned MIT image 386. The command really works.

 Thanks.

 

 ___
 Devel mailing list
 Devel@laptop.org
 http://mailman.laptop.org/mailman/listinfo/devel
   
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


modprobe msr fail!

2007-05-17 Thread Luna.Huang
Hi,

I am Luna.

I want to modprobe msr on OLPC. (I use MIT image 426 on B2)

But it is fail. The error message is “FATAL: Module msr not found.”

To look at config file, I find  

CONFIG_X86_MSR=y

 

Someone can tell me what happen.

I had returned MIT image 386. The command really works.

Thanks.

 

___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


RE: modprobe msr fail!

2007-05-17 Thread Luna.Huang
Sorry, 
I misunderstand the mean.
The functionality has been built-in, and I get right msr value.
Thank you.
Luna

-Original Message-
From: Mitch Bradley [mailto:[EMAIL PROTECTED] 
Sent: Thursday, May 17, 2007 3:10 PM
To: Luna Huang (�S��玲)
Cc: devel@laptop.org
Subject: Re: modprobe msr fail!

CONFIG_X86_MSR=y

means that MSR support is compiled into the kernel, not as a module. So
you do not need to do modprobe. The functionality is built-in. There
will be a file named /dev/cpu/0/msr.


In image 386, the config was:

CONFIG_X86_MSR=m

so with that build, you did need to do the modprobe. Before you do the
modprobe, the /dev/cpu/0/msr file will not exist, but it will appear
after the modprobe.

[EMAIL PROTECTED] wrote:

 Hi,

 I am Luna.

 I want to modprobe msr on OLPC. (I use MIT image 426 on B2)

 But it is fail. The error message is “FATAL: Module msr not found.”

 To look at config file, I find

 CONFIG_X86_MSR=y

 Someone can tell me what happen.

 I had returned MIT image 386. The command really works.

 Thanks.

 

 ___
 Devel mailing list
 Devel@laptop.org
 http://mailman.laptop.org/mailman/listinfo/devel
   
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


Re: [Trac #1425] Verify that camera activities close the camera device when not in use.

2007-05-17 Thread Bert Freudenberg
On May 17, 2007, at 14:26 , Zarro Boogs per Child wrote:

 #1425: Verify that camera activities close the camera device when  
 not in use.
 - 
 +--
   Reporter:  jg  |   Owner:  cjb
   Type:  task|  Status:  new
   Priority:  high|   Milestone:  BTest-4
  Component:  distro  | Version:
 Resolution:  |Keywords:  power
   Verified:  0   |
 - 
 +--
 Comment (by tomeu):

  Yesterday we added an 'active' property to  
 sugar.activity.Activity. The
  shell will set it to True or False when the user switches between
  activites. You can listen to changes on the 'notify::active' signal.

  Is this enough?

I assume this is a DBus signal? Could you add that to the API spec?

http://wiki.laptop.org/go/Activity_Factory_DBus_API

- Bert -


___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


Re: [Trac #1520] Entering text in Paint needs an extra 'enter'

2007-05-17 Thread Walter Bender
Manu is working on it... making steady progress and working with Eben
and the mt of mtpaint

On 5/17/07, Zarro Boogs per Child [EMAIL PROTECTED] wrote:
 #1520: Entering text in Paint needs an extra 'enter'
 --+-
   Reporter:  Carla|   Owner:  walter
   Type:  defect   |  Status:  new
   Priority:  normal   |   Milestone:  Trial-2
  Component:  pixel paint  | Version:  Build 385
 Resolution:   |Keywords:
   Verified:  0|
 --+-
 Changes (by jg):

   * milestone:  Untriaged = Trial-2

 Comment:

  Walter, do we have someone who is supposed to sugarize paint?

 --
 Ticket URL: http://dev.laptop.org/ticket/1520#comment:1
 One Laptop Per Child http://laptop.org/



-- 
Walter Bender
One Laptop per Child
http://laptop.org
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


Re: owner id in .sugar/default/

2007-05-17 Thread Holger Levsen
Hi,

On Thursday 17 May 2007 04:04, Ivan Krstić wrote:
 Hal Murray wrote:
  That feels like the tip of a security iceberg.  Somebody has to be able
  to authorize access to data on the server without the appropriate key,
  including getting the key.
  I don't think that's anything new from the computer security standpoint. 
  You have to trust your sysadmin.  The interesting part for OLPC will be
  bringing the local sysadmins up to speed on security.

I agree.

 Correct. I explained this to people in today's security meeting: the
 school server maintains a UUID - child identity mapping. Backups are
 identified as belonging to a particular UUID. A teacher can log into the
 school server and use a graphical interface to reassign existing backups
 for a particular UUID to another UUID by modifying the mapping. This
 covers laptop destruction or exchange for any reason.

Yup. But it would also be nice, if the pupils can ask the server for their 
backups, without going via the teacher. For that, an access key on the laptop 
would be needed. (So that it's not possible to request someone elses backup.)

In case the laptop breaks or is stolen, the backup should be accessable via 
the teacher. (And a new laptop key needs to be genereated.)

So IMHO the backup has to be stored encrypted twice: once with a school key, 
and once with a laptop key (kids key). And it would surely be nice, if the 
laptop keys survives reflashing the laptop.

 Once the kids are old enough that they're worried about the teacher
 using a spare XO to invade their privacy,

I dont think the teachers are the (biggest) security threat here. Random 
strangers on the other side of the street are more worrysome IMHO. (As we all 
know, strangers with candy... ;-)

 For more details, see P_DOCUMENT_BACKUP and P_PASSWORD in
 http://wiki.laptop.org/go/OLPC_Bitfrost .

Neither P_DOCUMENT_BACKUP nor P_PASSWORD seem complete to me. Also it says 
that http://wiki.laptop.org/go/Bitfrost is the authoritive version, which is 
much less specific. Is there a process to finalize the document and make it 
binding?

I joined the security list today.


regards,
Holger


pgpdfQPRkeyQK.pgp
Description: PGP signature
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel


Re: owner id in .sugar/default/

2007-05-17 Thread Ivan Krstić
Holger Levsen wrote:
 Yup. But it would also be nice, if the pupils can ask the server for their 
 backups, without going via the teacher. For that, an access key on the laptop 
 would be needed. (So that it's not possible to request someone elses backup.)

I don't think you understood my original mail. Backups are retrieved
from the server by virtue of a laptop's UUID -- not a cryptographic key
-- so the *only* instance where you have to ask a teacher to obtain a
backup is when your laptop was destroyed.

 In case the laptop breaks or is stolen, the backup should be accessable via 
 the teacher. (And a new laptop key needs to be genereated.)

See above.

 So IMHO the backup has to be stored encrypted twice

Unnecessary.

 I dont think the teachers are the (biggest) security threat here. Random 
 strangers on the other side of the street are more worrysome IMHO.

We're talking about *backup* security here, not security in general.

 Neither P_DOCUMENT_BACKUP nor P_PASSWORD seem complete to me.

P_DOCUMENT_BACKUP describes the general principle; P_PASSWORD isn't
complete because it's not completely clear how that'll work yet, and
it's not a hard target for the ship date unless we have some extra time
left over to implement it.

 Is there a process to finalize the document and make it binding?

I'm working on finishing some sections, but what's there now is
authoritative.

-- 
Ivan Krstić [EMAIL PROTECTED] | GPG: 0x147C722D
___
Devel mailing list
Devel@laptop.org
http://mailman.laptop.org/mailman/listinfo/devel