[meta-freescale] Updates to meta-fsl-arm (master) in 2014-11-05

[Otavio Salvador]

Hello Andreas,

(adding Lauren on Cc)

On Thu, Nov 6, 2014 at 9:58 AM, Andreas Müller
<schnitzeltony at googlemail.com> wrote:
> On Wed, Nov 5, 2014 at 9:38 PM, Otavio Salvador <otavio at ossystems.com.br> wrote:
>> commit 8e5e871079bdaf7659171102f008bad28c728eed
>> Author: Lauren Post <lauren.post at freescale.com>
>> Date:   Fri Oct 31 10:17:16 2014 -0500
>>
>>     linux-imx: Upgrade to 3.10.17-1.0.2 patch release
>>
>>     Changes to the following areas.
>>
>>     - pfuse 100
>>     - power management
>>     - graphics driver update with memory, deadlock and bug hang fixes.
>>     - crypto - Update testmgr DMA
>>     - ipu MXC IPUv3 fb: Fix MXCFB_CSC_UPDATE ioctl
>>     - ARM - Drop PL310 errata
>>
>>     Signed-off-by: Lauren Post <lauren.post at freescale.com>
>>     Signed-off-by: Otavio Salvador <otavio at ossystems.com.br>
>>
> Just let me understand. I see in the patch
>
> +# back to -1.0.0_ga branch for all patch release
> +SRCBRANCH = "imx_3.10.17_1.0.0_ga"
> +SRCREV = "33597e348b2d60dd5c71890ef7b7d3d3fd6e4e97"
> +LOCALVERSION = "-1.0.2_ga"
>
> What make the trick for 1.02 (fetching the kernel sources I don't see
> 1.02). To me it seems that the patch removes all patches and switches
> back to 1.00.

Freescale is using the 1.0.0 branch to host all patches done on top of
the initial 3.10.17-1.0.0 GA release.

As you can see on the log below, the top commit (the one referenced on
the SRCREV) is also the one used for the tag
(rel_imx_3.10.17_1.0.2_ga).

commit 33597e348b2d60dd5c71890ef7b7d3d3fd6e4e97 (tag:
rel_imx_3.10.17_1.0.2_ga, freescale/imx_3.10.17_1.0.0_ga)
Author: Horia Geanta <horia.geanta at freescale.com>
Date:   Wed Jul 23 11:59:38 2014 +0300

    crypto: testmgr - avoid DMA mapping from text, rodata, stack

    With DMA_API_DEBUG set, following warnings are emitted
    (tested on CAAM accelerator):
    DMA-API: device driver maps memory from kernel text or rodata
    DMA-API: device driver maps memory from stack
    and the culprits are:
    -key in __test_aead and __test_hash
    -result in __test_hash

    MAX_KEYLEN is changed to accommodate maximum key length from
    existing test vectors in crypto/testmgr.h (131 bytes) and rounded.

    Signed-off-by: Horia Geanta <horia.geanta at freescale.com>
    Acked-by: Kim Phillips <kim.phillips at freescale.com>
    Signed-off-by: Herbert Xu <herbert at gondor.apana.org.au>

So what Lauren did was to adjust the branch name (for BitBake's
fetcher to be able to validate the hash existence) and bump the
SRCREV.

In future, if a 3.10.17-1.0.3 is released, the only change will be the
SRCREV bump.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750