[U-Boot] [PULL] Please pull u-boot-imx: u-boot-imx-20181217

[Tom Rini]

On Mon, Dec 17, 2018 at 04:57:26PM +0100, Stefano Babic wrote:

> Hi Tom,
> 
> please pull from u-boot-imx, u-boot-imx-20181217, thanks !
> 
> - fix missing MAINTANERS for embestmx6boards
> - drop Wolfgang's patch
> - apply V4 of "pico-imx7d: Increase the CONFIG_ENV_OFFSET size " that
> use fix values for size.
> 
> The following changes since commit d117d8f19b0625f88309e47a8a32c2faa384dddc:
> 
>   Merge branch 'master' of git://git.denx.de/u-boot-i2c (2018-12-13
> 09:36:55 -0500)
> 
> are available in the Git repository at:
> 
>   git://www.denx.de/git/u-boot-imx.git tags/u-boot-imx-20181217
> 
> for you to fetch changes up to 22f219feb807df91d958c3cc5a8634b6218b2565:
> 
>   pico-imx7d: Increase the CONFIG_ENV_OFFSET size (2018-12-17 16:51:02
> +0100)
> 
> ----------------------------------------------------------------
> imx for 2019.01
> 
> - introduce support for i.MX8M
> - fix size limit for Vhybrid / pico boards
> - several board fixes
> - w1 driver for MX2x / MX5x
> 
> ----------------------------------------------------------------

So, this breaks travis.  I know I've mentioned some of this off-list,
but for the record and to aid in future debugging of issues like this,
if we look at https://travis-ci.org/trini/u-boot/builds/472348357 we can
see just a single job fails.  So I rebased and put in as the first
commit a hack to .travis.yml where I remove every other job in the
matrix so it runs (relatively) quick.  Then I did a git bisect where for
the test I did:
$ git push -f github HEAD:TRAVIS-imx-PR-bisect
And then bisect good/bad depending on what happens.  This narrowed down
things to:
Author: Denis Zalevskiy <denis.zalevskiy at ge.com>
Date:   Wed Oct 17 10:33:28 2018 +0200

    bootcount: i2c: Add bus switching to the I2C bootcount driver

Being the culprit.  My _guess_ is that after looking at the commit we
have some boards that are prompting for
SYS_BOOTCOUNT_I2C_BUS/SYS_BOOTCOUNT_I2C_ADDR now and not having a value
set to they don't get built, but don't quite show up as an obvious
failure either (aside, Simon, any idea why?).  So, if we do:
$ ./tools/buildman/buildman -v --dry-run powerpc -x freescale
boards.cfg is up to date. Nothing to do.
Dry run, so not doing much. But I would do this:

Building current source for 47 boards (16 threads, 1 job per thread)
Build directory: ..

powerpc : 47 boards
   mpc8308_p1m sbc8349 sbc8349_PCI_33 sbc8349_PCI_66 ve8313 caddy2
   vme8349 hrcon hrcon_dh strider_con strider_con_dp strider_cpu
   strider_cpu_dp ids8313 kmcoge5ne kmeter1 kmopti2 kmsupx5 kmtegr1
   kmtepr2 kmvect1 suvd3 tuge1 tuxx1 TQM834x sbc8548 sbc8548_PCI_33
   sbc8548_PCI_33_PCIE sbc8548_PCI_66 sbc8548_PCI_66_PCIE socrates
   UCP1020 UCP1020_SPIFLASH controlcenterd_36BIT_SDCARD
   controlcenterd_36BIT_SDCARD_DEVELOP controlcenterd_TRAILBLAZER
   controlcenterd_TRAILBLAZER_DEVELOP kmcoge4 kmlion1 Cyrus_P5020
   Cyrus_P5040 xpedite520x xpedite537x xpedite550x sbc8641d xpedite517x
   MCR3000
Total boards to build for each commit: 47

There's the list of boards that it might be.  Doing a for B in ...;do
make O=/tmp/$B ${B}_defconfig oldconfig and I see that yup, ids8313
trips on the new symbols without having a value.  So those two bootcount
patches from Denis need a slight rework (the second depends on the
first) or to be dropped for now.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20181226/e133e9b2/attachment.sig>