cbfs fix

[Myles Watson]

This fixes the boot for CBFS with the s2895.  As an added bonus it
still works on serengeti :)

The problem was that the master cbfs record was located at the end of
the flash and overwrote anything that was there.  Luckily for most
targets this wasn't a problem, but for ck804 or mcp55-based machines
it overwrote the romstrap.

The fix is simple:
1. Put the master cbfs record above the bootblock instead of on it.

There are many improvements that can be made to cbfs, but for now I
think this improves the situation substantially.

Signed-off-by: Myles Watson

Thanks,
Myles

[Peter Stuge]

Myles Watson wrote:
> Put the master cbfs record above the bootblock instead of on it.
>
> Signed-off-by: Myles Watson

Acked-by: Peter Stuge
> /* Replace the LAR header */

Copypaste. Yum!

//Peter

[ron minnich]

Another fine job by the Monster Fix-it Team!

Signed-off-by: Ronald G. Minnich

http://www.truveo.com/Monster-Fixit-Team/id/1650801844

[ron minnich]

I'm tired.

Acked-by: Ronald G. Minnich

On Fri, Apr 24, 2009 at 4:02 PM, ron minnich  wrote:
> Another fine job by the Monster Fix-it Team!
>
> Signed-off-by: Ronald G. Minnich
>
> http://www.truveo.com/Monster-Fixit-Team/id/1650801844

[Peter Stuge]

ron minnich wrote:
> I'm tired.

Beer and Sesame Street at work.. Your job rocks! :p

//Peter

Newbie source commit question

[Lim, Vincent]

Hi All:

I would like to add the family10h Rev C0-C2 support to coreboot and I
just wanted to make sure I've got the procedure right.

Here are my simple chances:

Index: src/cpu/amd/model_10xxx/model_10xxx_init.c
===================================================================
--- src/cpu/amd/model_10xxx/model_10xxx_init.c  (revision 4206)
+++ src/cpu/amd/model_10xxx/model_10xxx_init.c  (working copy)
@@ -142,6 +142,10 @@
        { X86_VENDOR_AMD, 0x100f2A },
    { X86_VENDOR_AMD, 0x100f22 },
    { X86_VENDOR_AMD, 0x100f23 },
+       { X86_VENDOR_AMD, 0x100f40 },           /* RB-C0 */
+       { X86_VENDOR_AMD, 0x100F42 },           /* RB-C2 */ 
+       { X86_VENDOR_AMD, 0x100F52 },           /* BL-C2 */ 
+       { X86_VENDOR_AMD, 0x100F62 },           /* DA-C2 */ 
        { 0, 0 },
     };

Since I checkout the source as developer, I just have to do a 'svn ci
.../model_10xxx_init.c', and someone will ack my commit ?

Please let me know this is the right process...

Best Regards,

Vincent Lim
SimNow Team
Performance CoE
Central Engineering
T 512.602.1618
F 512.602.7745

Newbie source commit question

[Patrick Georgi]

Am Samstag 25 April 2009 04:54:59 schrieb Lim, Vincent:
> Hi All:
>
> I would like to add the family10h Rev C0-C2 support to coreboot and I
> just wanted to make sure I've got the procedure right.
>
> Here are my simple chances:
> Index: src/cpu/amd/model_10xxx/model_10xxx_init.c
>
> --- src/cpu/amd/model_10xxx/model_10xxx_init.c  (revision 4206)
> +++ src/cpu/amd/model_10xxx/model_10xxx_init.c  (working copy)
> @@ -142,6 +142,10 @@
>         { X86_VENDOR_AMD, 0x100f2A },
>    { X86_VENDOR_AMD, 0x100f22 },
>    { X86_VENDOR_AMD, 0x100f23 },
>   { X86_VENDOR_AMD, 0x100f40 },           /* RB-C0 */
>   { X86_VENDOR_AMD, 0x100F42 },           /* RB-C2 */
>   { X86_VENDOR_AMD, 0x100F52 },           /* BL-C2 */
>   { X86_VENDOR_AMD, 0x100F62 },           /* DA-C2 */
>   { 0, 0 },
> };
>
> Since I checkout the source as developer, I just have to do a 'svn ci
> .../model_10xxx_init.c', and someone will ack my commit ?

First, you'll have to provide a Signed-off-by  (see 
http://www.coreboot.org/Development_Guidelines#Sign-off_Procedure, especially 
the "Developer's Certificate of Origin". Basically, it states that you're 
allowed to release this change.

Then, someone reviews it, and - if good enough - ACKs it. I know what this 
change does, and know that it's fine, so I can do it:

Acked-by: Patrick Georgi

Now you can "svn ci" that change, with a commit message stating
1) what that commit does
2) your Signed-off-by line
3) my (or someone elses) Acked-By line(s)

That way, we have a record that you state that you are allowed to release that 
change under the terms of coreboot (in this case: GPLv2), that you employer 
approves (if he has any say), and that someone (in this case: I) approved of 
that change.

Btw, does this mean that the microcode for those CPUs will be released, too?

Regards,
Patrick Georgi

Newbie source commit question

[Lim, Vincent]

Hi Patrick:

Thanks for the explanation. For formality, I will do the sign-off here:

Index: src/cpu/amd/model_10xxx/model_10xxx_init.c
===================================================================
--- src/cpu/amd/model_10xxx/model_10xxx_init.c  (revision 4206)
+++ src/cpu/amd/model_10xxx/model_10xxx_init.c  (working copy)
@@ -142,6 +142,10 @@
        { X86_VENDOR_AMD, 0x100f2A },
    { X86_VENDOR_AMD, 0x100f22 },
    { X86_VENDOR_AMD, 0x100f23 },
+       { X86_VENDOR_AMD, 0x100f40 },           /* RB-C0 */
+       { X86_VENDOR_AMD, 0x100F42 },           /* RB-C2 */ 
+       { X86_VENDOR_AMD, 0x100F52 },           /* BL-C2 */ 
+       { X86_VENDOR_AMD, 0x100F62 },           /* DA-C2 */ 
        { 0, 0 },
     };

add the family10h Rev C0-C2 support to coreboot.

Signed-off-by: Vincent Lim [e-mail]

Best Regards,

Vincent Lim
SimNow Team
Performance CoE
Central Engineering
T 512.602.1618
F 512.602.7745

-----Original Message-----
From: Patrick Georgi [mailto:[e-mail]] 
Sent: Friday, April 24, 2009 11:46 PM
To: [e-mail]
Cc: Lim, Vincent
Subject: Re: [coreboot] Newbie source commit question

Am Samstag 25 April 2009 04:54:59 schrieb Lim, Vincent:
> Hi All:
>
> I would like to add the family10h Rev C0-C2 support to coreboot and I
> just wanted to make sure I've got the procedure right.
>
> Here are my simple chances:
> Index: src/cpu/amd/model_10xxx/model_10xxx_init.c
>
> --- src/cpu/amd/model_10xxx/model_10xxx_init.c  (revision 4206)
> +++ src/cpu/amd/model_10xxx/model_10xxx_init.c  (working copy)
> @@ -142,6 +142,10 @@
>         { X86_VENDOR_AMD, 0x100f2A },
>    { X86_VENDOR_AMD, 0x100f22 },
>    { X86_VENDOR_AMD, 0x100f23 },
>   { X86_VENDOR_AMD, 0x100f40 },           /* RB-C0 */
>   { X86_VENDOR_AMD, 0x100F42 },           /* RB-C2 */
>   { X86_VENDOR_AMD, 0x100F52 },           /* BL-C2 */
>   { X86_VENDOR_AMD, 0x100F62 },           /* DA-C2 */
>   { 0, 0 },
> };
>
> Since I checkout the source as developer, I just have to do a 'svn ci
> .../model_10xxx_init.c', and someone will ack my commit ?

First, you'll have to provide a Signed-off-by  (see 
http://www.coreboot.org/Development_Guidelines#Sign-off_Procedure,
especially 
the "Developer's Certificate of Origin". Basically, it states that
you're 
allowed to release this change.

Then, someone reviews it, and - if good enough - ACKs it. I know what
this 
change does, and know that it's fine, so I can do it:

Acked-by: Patrick Georgi

Now you can "svn ci" that change, with a commit message stating
1) what that commit does
2) your Signed-off-by line
3) my (or someone elses) Acked-By line(s)

That way, we have a record that you state that you are allowed to
release that 
change under the terms of coreboot (in this case: GPLv2), that you
employer 
approves (if he has any say), and that someone (in this case: I)
approved of 
that change.

Btw, does this mean that the microcode for those CPUs will be released,
too?

Regards,
Patrick Georgi

Newbie source commit question

[Lim, Vincent]

Hi Patrick:

Regarding micro-patch, I do not have the complete story yet. Bao Zheng
and Marc Jones added ucode support for Rev B (DR) and Rev C (RB/BL/DA)
but I have not tested it out on a actual HW. I am not sure anyone has
done it ? If I cannot get boards to test these revs, I probably will try
it on SimNow. Although the patches will be just 'pass-thru' and have no
effect on the Simulator, at least I can verify that this mechanism
works.

I will find out more by next week.

Best Regards,

Vincent Lim
SimNow Team
Performance CoE
Central Engineering
T 512.602.1618
F 512.602.7745

-----Original Message-----
From: Patrick Georgi [mailto:[e-mail]] 
Sent: Friday, April 24, 2009 11:46 PM
To: [e-mail]
Cc: Lim, Vincent
Subject: Re: [coreboot] Newbie source commit question

Am Samstag 25 April 2009 04:54:59 schrieb Lim, Vincent:
> Hi All:
>
> I would like to add the family10h Rev C0-C2 support to coreboot and I
> just wanted to make sure I've got the procedure right.
>
> Here are my simple chances:
> Index: src/cpu/amd/model_10xxx/model_10xxx_init.c
>
> --- src/cpu/amd/model_10xxx/model_10xxx_init.c  (revision 4206)
> +++ src/cpu/amd/model_10xxx/model_10xxx_init.c  (working copy)
> @@ -142,6 +142,10 @@
>         { X86_VENDOR_AMD, 0x100f2A },
>    { X86_VENDOR_AMD, 0x100f22 },
>    { X86_VENDOR_AMD, 0x100f23 },
>   { X86_VENDOR_AMD, 0x100f40 },           /* RB-C0 */
>   { X86_VENDOR_AMD, 0x100F42 },           /* RB-C2 */
>   { X86_VENDOR_AMD, 0x100F52 },           /* BL-C2 */
>   { X86_VENDOR_AMD, 0x100F62 },           /* DA-C2 */
>   { 0, 0 },
> };
>
> Since I checkout the source as developer, I just have to do a 'svn ci
> .../model_10xxx_init.c', and someone will ack my commit ?

First, you'll have to provide a Signed-off-by  (see 
http://www.coreboot.org/Development_Guidelines#Sign-off_Procedure,
especially 
the "Developer's Certificate of Origin". Basically, it states that
you're 
allowed to release this change.

Then, someone reviews it, and - if good enough - ACKs it. I know what
this 
change does, and know that it's fine, so I can do it:

Acked-by: Patrick Georgi

Now you can "svn ci" that change, with a commit message stating
1) what that commit does
2) your Signed-off-by line
3) my (or someone elses) Acked-By line(s)

That way, we have a record that you state that you are allowed to
release that 
change under the terms of coreboot (in this case: GPLv2), that you
employer 
approves (if he has any say), and that someone (in this case: I)
approved of 
that change.

Btw, does this mean that the microcode for those CPUs will be released,
too?

Regards,
Patrick Georgi

Newbie source commit question

[Peter Stuge]

Hi Vincent,

Lim, Vincent wrote:
> I would like to add the family10h Rev C0-C2 support to coreboot and

Great!
> I just wanted to make sure I've got the procedure right.

As you may know by now, this is how it goes:

1. Work in local development tree
2. Generate patch file with svn diff
3. Send email to the list with log message, maybe some extra
   explanations, Signed-off-by: and the patch file.
4. Wait for Acked-by: or discussion.
5. If Acked-by: then commit with log message, which ends with
   Signed-off-by: and all Acked-by: received in 4.

In 3. some like to include the patch contents in the email body, some
attach the file, some do both. I like to use attachments because then
I don't have to worry about my mail program reformatting the patch in
case anyone wants to apply it to their development trees.

When using attachments, please check that the MIME type for the
attachment is set to text/plain (often automatic if the file name
ends with .patch) to allow easier reading for all list recipients.

Others can improve upon your patch, and then send a new patch with
their Signed-off-by: added to yours. In that case both developers who
signed off usually watch the discussion or the first can even ack the
new changes of the second. This is not very common, but has happened.

For trivial things such as reformatting, wording changes, and more,
it is fine to self-ack, ie. skip 2. 3. and 4. and instead commit
directly with Signed-off-by: and Acked-by: yourself.

Hope this helps!

//Peter

Newbie source commit question II

[Lim, Vincent]

Hi All:

Thanks for details process, I shall follow in the future commit.

I ran into username problem during commit:

Authentication realm: svn://coreboot.org:3690 LinuxBIOS Password for
'vlim': 
Authentication realm: svn://coreboot.org:3690 LinuxBIOS
Username: VincentLim
Password for 'VincentLim': 
Authentication realm: svn://coreboot.org:3690 LinuxBIOS
Username: VincentLim
Password for 'VincentLim': 
svn: Commit failed (details follow):
svn: Authentication error from server: Username not found
svn: Your commit message was left in a temporary file:
svn:
'/proj/simnow-data3/users/vlim/coreboot/coreboot/coreboot-v2/src/cpu/amd
/svn-commit.3.tmp'
beemer:/proj/simnow-data3/users/vlim/coreboot/coreboot/coreboot-v2/src/c
pu/amd/m

I checked out using command "svn co
svn://[e-mail]/repos/trunk/coreboot-v2

Maybe I do not have set-up/permission to check-in code yet ?

Best Regards,

Vincent Lim
SimNow Team
Performance CoE
Central Engineering
T 512.602.1618
F 512.602.7745

-----Original Message-----
From: [e-mail]
[mailto:[e-mail]] On Behalf Of Peter Stuge
Sent: Saturday, April 25, 2009 8:01 AM
To: [e-mail]
Subject: Re: [coreboot] Newbie source commit question

Hi Vincent,

Lim, Vincent wrote:
> I would like to add the family10h Rev C0-C2 support to coreboot and

Great!
> I just wanted to make sure I've got the procedure right.

As you may know by now, this is how it goes:

1. Work in local development tree
2. Generate patch file with svn diff
3. Send email to the list with log message, maybe some extra
   explanations, Signed-off-by: and the patch file.
4. Wait for Acked-by: or discussion.
5. If Acked-by: then commit with log message, which ends with
   Signed-off-by: and all Acked-by: received in 4.

In 3. some like to include the patch contents in the email body, some
attach the file, some do both. I like to use attachments because then
I don't have to worry about my mail program reformatting the patch in
case anyone wants to apply it to their development trees.

When using attachments, please check that the MIME type for the
attachment is set to text/plain (often automatic if the file name
ends with .patch) to allow easier reading for all list recipients.

Others can improve upon your patch, and then send a new patch with
their Signed-off-by: added to yours. In that case both developers who
signed off usually watch the discussion or the first can even ack the
new changes of the second. This is not very common, but has happened.

For trivial things such as reformatting, wording changes, and more,
it is fine to self-ack, ie. skip 2. 3. and 4. and instead commit
directly with Signed-off-by: and Acked-by: yourself.

Hope this helps!

//Peter

Newbie source commit question II

[Peter Stuge]

Lim, Vincent wrote:
> I checked out using command "svn co
> svn://[e-mail]/repos/trunk/coreboot-v2
>
> Maybe I do not have set-up/permission to check-in code yet ?

All other account names have lowercase letters only. Please try again
using vincentlim@ and see if that works better.

(Note that the username is only required when committing, that's why
checkout works even if the username is not correct.)

//Peter

Newbie source commit question II

[Carl-Daniel Hailfinger]

Hi Vincent,

let me explain a few things about our revision control usage:

We use centralized subversion, so you have to get an account from Stefan
before you can commit. I have added Stefan to this message. (Usually,
developers can apply for an account after they have submitted a few
patches and the patches were accepted. Only very few developers on this
list have accounts with commit permissions.)
Everything you commit will directly appear in the public tree and you
can't change the commit message later. So if you want to send a patch
for review, do _not_ commit it, but run svn diff. The output of svn diff
can be mailed to the list for review. After you get an Acked-by from
someone on the list, you can commit.

Regards,
Carl-Daniel

[Myles Watson]

> Acked-by: Ronald G. Minnich 
> Acked-by: Peter Stuge

Rev. 4209.

Thanks,
Myles