cgpt supports GptNextKernelEntry() and GptUpdateKernelEntry()

cgpt supports GptNextKernelEntry() and GptUpdateKernelEntry()

[Bill Richardson, 2010-05-05 15:04:16]

We need to change the bits used for the attribute flags to conform to the UEFI
spec. Otherwise LGTM.

http://codereview.chromium.org/1922004/diff/1/4
File src/platform/vboot_reference/cgptlib/cgpt_internal.h (right):

http://codereview.chromium.org/1922004/diff/1/4#newcode52
src/platform/vboot_reference/cgptlib/cgpt_internal.h:52: *     63  -- do not
automounting
We can only use bits 48-63 for our purposes. Bits 0-47 are reserved for use by
the UEFI spec (section 5.3.3, table 16).

The Chromium OS Drive Map document lists the bit fields to use.

[Louis, 2010-05-05 17:16:10]

Thanks, Bill. I've refined my code. I am waiting for Gaurav's review before
submit.

http://codereview.chromium.org/1922004/diff/1/4
File src/platform/vboot_reference/cgptlib/cgpt_internal.h (right):

http://codereview.chromium.org/1922004/diff/1/4#newcode52
src/platform/vboot_reference/cgptlib/cgpt_internal.h:52: *     63  -- do not
automounting
My bad. I read that then forgot that. :-(  Revised now.
On 2010/05/05 15:04:16, Bill Richardson wrote:
> We can only use bits 48-63 for our purposes. Bits 0-47 are reserved for use by
> the UEFI spec (section 5.3.3, table 16).
> 
> The Chromium OS Drive Map document lists the bit fields to use.

[Bill Richardson, 2010-05-05 17:50:51]

You may have a long wait. Gaurav has gone to get married, so if he doesn't
reply tonight he probably won't for a couple of weeks. Just check it in -
we're the only ones using it, and we'll be making many more changes before
Gaurav returns.


On Wed, May 5, 2010 at 10:16 AM, <yjlou@chromium.org> wrote:

> Thanks, Bill. I've refined my code. I am waiting for Gaurav's review before
> submit.
>
>
> http://codereview.chromium.org/1922004/diff/1/4
> File src/platform/vboot_reference/cgptlib/cgpt_internal.h (right):
>
> http://codereview.chromium.org/1922004/diff/1/4#newcode52
> src/platform/vboot_reference/cgptlib/cgpt_internal.h:52: *     63  -- do
> not automounting
> My bad. I read that then forgot that. :-(  Revised now.
> On 2010/05/05 15:04:16, Bill Richardson wrote:
>
>> We can only use bits 48-63 for our purposes. Bits 0-47 are reserved
>>
> for use by
>
>> the UEFI spec (section 5.3.3, table 16).
>>
>
>  The Chromium OS Drive Map document lists the bit fields to use.
>>
>
> http://codereview.chromium.org/1922004/show
>

[gauravsh, 2010-05-05 18:01:55]

Hmm, did someone say my name? LGTM too.

(on proper vacation starting next week. meanwhile, expect intermittent flakiness
in responsiveness).

[Louis, 2010-05-05 18:18:12]

hmm... then I will push my CL first. But I haven't tested it on
'build_autotest.sh firmware_VbootCrypto' yet because it takes too long to do
that. Make sure we won't break the continuous build when you push your CL.

On Wed, May 5, 2010 at 11:01 AM, <gauravsh@chromium.org> wrote:

> Hmm, did someone say my name? LGTM too.
>
> (on proper vacation starting next week. meanwhile, expect intermittent
> flakiness
> in responsiveness).
>
>
> http://codereview.chromium.org/1922004/show
>

[gauravsh, 2010-05-05 18:26:36]

On Wed, May 5, 2010 at 11:17 AM, Louis Yung-Chieh Lo <yjlou@chromium.org> wrote:
> hmm... then I will push my CL first. But I haven't tested it on
> 'build_autotest.sh firmware_VbootCrypto' yet because it takes too long to do
> that. Make sure we won't break the continuous build when you push your CL.

You should run that before you push the CL. If you just build
firmware_VbootCrypto, it shouldn't take that long except for the first
time, when other autotest dependencies will be emerged too.

There's no way to be sure that the build won't break - the ARM
cross-compiler has some subtle differences from X86 in terms of how it
flags warnings.

>
> On Wed, May 5, 2010 at 11:01 AM, <gauravsh@chromium.org> wrote:
>>
>> Hmm, did someone say my name? LGTM too.
>>
>> (on proper vacation starting next week. meanwhile, expect intermittent
>> flakiness
>> in responsiveness).
>>
>> http://codereview.chromium.org/1922004/show
>
>



-- 
-g

[Louis, 2010-05-05 18:40:51]

ahhh... I've submitted the CL. But I am running the firmware_VbootCrypto
right now. will send another CL if I encounter error.

On Wed, May 5, 2010 at 11:26 AM, Gaurav Shah <gauravsh@chromium.org> wrote:

> On Wed, May 5, 2010 at 11:17 AM, Louis Yung-Chieh Lo <yjlou@chromium.org>
> wrote:
> > hmm... then I will push my CL first. But I haven't tested it on
> > 'build_autotest.sh firmware_VbootCrypto' yet because it takes too long to
> do
> > that. Make sure we won't break the continuous build when you push your
> CL.
>
> You should run that before you push the CL. If you just build
> firmware_VbootCrypto, it shouldn't take that long except for the first
> time, when other autotest dependencies will be emerged too.
>
> There's no way to be sure that the build won't break - the ARM
> cross-compiler has some subtle differences from X86 in terms of how it
> flags warnings.
>
> >
> > On Wed, May 5, 2010 at 11:01 AM, <gauravsh@chromium.org> wrote:
> >>
> >> Hmm, did someone say my name? LGTM too.
> >>
> >> (on proper vacation starting next week. meanwhile, expect intermittent
> >> flakiness
> >> in responsiveness).
> >>
> >> http://codereview.chromium.org/1922004/show
> >
> >
>
>
>
> --
> -g
>

[Louis, 2010-05-13 10:19:04]

by running following commands, the cgptlib can be compiled successfully,
which means it won't break the continuous build:

  x86:
    ./build_packages --board=x86-generic --withautotest --jobs=16 &&
    ./build_autotest.sh --board="x86-generic" firmware_VbootCrypto &&
  arm:
    ./build_packages --board=arm-generic --withautotest --jobs=16 &&
    ./build_autotest.sh --board="arm-generic" firmware_VbootCrypto

Let me know if it is not enough.

On Thu, May 6, 2010 at 2:40 AM, Louis Yung-Chieh Lo <yjlou@chromium.org>wrote:

> ahhh... I've submitted the CL. But I am running the firmware_VbootCrypto
> right now. will send another CL if I encounter error.
>
>
> On Wed, May 5, 2010 at 11:26 AM, Gaurav Shah <gauravsh@chromium.org>wrote:
>
>> On Wed, May 5, 2010 at 11:17 AM, Louis Yung-Chieh Lo <yjlou@chromium.org>
>> wrote:
>> > hmm... then I will push my CL first. But I haven't tested it on
>> > 'build_autotest.sh firmware_VbootCrypto' yet because it takes too long
>> to do
>> > that. Make sure we won't break the continuous build when you push your
>> CL.
>>
>> You should run that before you push the CL. If you just build
>> firmware_VbootCrypto, it shouldn't take that long except for the first
>> time, when other autotest dependencies will be emerged too.
>>
>> There's no way to be sure that the build won't break - the ARM
>> cross-compiler has some subtle differences from X86 in terms of how it
>> flags warnings.
>>
>> >
>> > On Wed, May 5, 2010 at 11:01 AM, <gauravsh@chromium.org> wrote:
>> >>
>> >> Hmm, did someone say my name? LGTM too.
>> >>
>> >> (on proper vacation starting next week. meanwhile, expect intermittent
>> >> flakiness
>> >> in responsiveness).
>> >>
>> >> http://codereview.chromium.org/1922004/show
>> >
>> >
>>
>>
>>
>> --
>> -g
>>
>
>

[Louis, 2010-05-18 06:39:58]

Hi Gaurav,

Just a remind that my cl <http://codereview.chromium.org/2082007/show>has
modified the Makefiles.

On Thu, May 13, 2010 at 6:18 PM, Louis Yung-Chieh Lo <yjlou@chromium.org>wrote:

> by running following commands, the cgptlib can be compiled successfully,
> which means it won't break the continuous build:
>
>   x86:
>     ./build_packages --board=x86-generic --withautotest --jobs=16 &&
>     ./build_autotest.sh --board="x86-generic" firmware_VbootCrypto &&
>   arm:
>     ./build_packages --board=arm-generic --withautotest --jobs=16 &&
>     ./build_autotest.sh --board="arm-generic" firmware_VbootCrypto
>
> Let me know if it is not enough.
>
> On Thu, May 6, 2010 at 2:40 AM, Louis Yung-Chieh Lo <yjlou@chromium.org>wrote:
>
>> ahhh... I've submitted the CL. But I am running the firmware_VbootCrypto
>> right now. will send another CL if I encounter error.
>>
>>
>> On Wed, May 5, 2010 at 11:26 AM, Gaurav Shah <gauravsh@chromium.org>wrote:
>>
>>> On Wed, May 5, 2010 at 11:17 AM, Louis Yung-Chieh Lo <yjlou@chromium.org>
>>> wrote:
>>> > hmm... then I will push my CL first. But I haven't tested it on
>>> > 'build_autotest.sh firmware_VbootCrypto' yet because it takes too long
>>> to do
>>> > that. Make sure we won't break the continuous build when you push your
>>> CL.
>>>
>>> You should run that before you push the CL. If you just build
>>> firmware_VbootCrypto, it shouldn't take that long except for the first
>>> time, when other autotest dependencies will be emerged too.
>>>
>>> There's no way to be sure that the build won't break - the ARM
>>> cross-compiler has some subtle differences from X86 in terms of how it
>>> flags warnings.
>>>
>>> >
>>> > On Wed, May 5, 2010 at 11:01 AM, <gauravsh@chromium.org> wrote:
>>> >>
>>> >> Hmm, did someone say my name? LGTM too.
>>> >>
>>> >> (on proper vacation starting next week. meanwhile, expect intermittent
>>> >> flakiness
>>> >> in responsiveness).
>>> >>
>>> >> http://codereview.chromium.org/1922004/show
>>> >
>>> >
>>>
>>>
>>>
>>> --
>>> -g
>>>
>>
>>
>