Merge of 3.14.52 GA into master and jethro

Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

Hello folks,

I finally got the time slot to finish the 3.14.52 review and merge. I
have merged them, complete, on master-next and jethro-next.

Please give the jethro-next a good test so we avoid bad surprises. I
intend to merge jethro-next on Feb 26th.

The changes under jethro-next for test are:

88ca952 linux-fslc-imx: Upgrade to the 3.14.52-1.1.0 GA based fork
8383610 EULA: Update to LA_OPT_BASE_LICENSE v10 November 2015
d2f2a2e firmware-imx: Upgrade to 5.3 version
473ef3d formfacor: Add generic machconfig formfactors for i.MX
6UltraLite and 7Dual
cc5c889 imx-uuc: Upgrade to 0.5.1 version
a25d68b imx-vpu: Upgrade to 5.4.32
58cb668 imx-test.inc: Add depends to support x11 builds
2cc46cd imx-test: Upgrade to 5.3 version
8d2a0c1 imx-kobs: Upgrade to 5.3 version
5e39560 u-boot-imx: Upgrade to 3.14.52-1.1.0_ga release
19df7f9 imx6qpsabresd: Create i.MX 6QuadPlus Sabre SD machine configuration
875f583 imx7dsabresd: Add epdc U-Boot configuration
df938d8 imx6ulevk: Add Murata BT/WiFi module support
7e3590c imx6sxsabresd: Add mqs sound support for the expansion board
eaae01d imx6sxsabresd: Add Murata BT/WiFi module support
03b5134 imx6slevk: Add Murata BT/WiFi module support
ce6798f imx6slevk: Add epdc uboot configuration and epdc firmware integration
d6329e0 imx6qsabred: Add Murata BT/WiFi module support
ea16b3f imx6qsabresd: Remove imx6q-sabresd-uart device tree
ef25822 imx6dlsabresd.conf: Add epdc boot configurations and epdc firmware
2cd8d6a imx6dlsabred: Add Murata BT/WiFi module support
e118b12 linux-imx: Upgrade to 3.14.52-1.1.0_ga Consolidatd GA release
84d6468 xf86-video-imxfb-vivante: Upgrade to 5.0.11.p7.4 version
2fa1970 weston: Add patches to fix bugs
feb4b08 kernel-module-imx-gpu-viv: Upgrade to 5.0.11.p7.4 version
c03c371 imx-gpu-viv: Add dependency on libpng
fb87643 imx-gpu-viv: Update to 5.0.11-p7.4 version
0fb8b89 libfslvpuwrap: Upgrade to 1.0.62 version
8401db5 libfslparser: Upgrade to 4.0.8 version
707f0cc libfslcodec: Upgrade to 4.0.8 version
02a3d11 gst1.0-fsl-plugins: Upgrade to 4.0.8 version
552d928 imx-base.inc: Move IMAGE_FSTYPES override to a SoC variable
2139719 imx-base: Set UBOOT_ENTRYPOINT for i.MX6UL
8997374 imx-gpu-viv: Remove wrong imx-gpu-viv-tools-dbg packaging
291aec5 fsl-vivante-kernel-driver-handler.bbclass: Postpone defconfig changes
e1bd3fe xf86-video-imxfb-vivante: Stop using swaprep.h header
6e2c0ed xf86-video-imxfb-vivante: Stop using Git to write local version
9cbf0f9 xf86-video-imxfb-vivante: Move patches attribution to the recipe
4d2af88 xf86-video-imxfb-vivante: Stop running 'make' in debug mode

If you are against any of those patches please say so.

Best Regards,

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Sat, Feb 20, 2016 at 5:38 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:
> Hello folks,
>
> I finally got the time slot to finish the 3.14.52 review and merge. I
> have merged them, complete, on master-next and jethro-next.
>
> Please give the jethro-next a good test so we avoid bad surprises. I
> intend to merge jethro-next on Feb 26th.

Otavio,

I ran into an issue building the 3.14-1.1.x-imx kernel using my custom
kernel configuration - duplicate symbols linking the kernel image.

The source of the conflict is having both of these patches in place:

3241216e250430a27b30eaf7f7df3e5e006acfb5 MGS-577 ARM: Export the cache
management symbols when !MULTI_CACHE
6325d57e9677b417552ac76ada0ee91fa4c98566 arm: Export cache flush
management symbols when !MULTI_CACHE

and building for just ARMv7 CPUs.  Reverting one or the other of them
fixes the issue.  I think that 3241216e is probably the better patch.

-Matt

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

Hello Matt,

On Tue, Feb 23, 2016 at 12:23 PM, Matt Madison <matt@madison.systems> wrote:
> On Sat, Feb 20, 2016 at 5:38 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>> Hello folks,
>>
>> I finally got the time slot to finish the 3.14.52 review and merge. I
>> have merged them, complete, on master-next and jethro-next.
>>
>> Please give the jethro-next a good test so we avoid bad surprises. I
>> intend to merge jethro-next on Feb 26th.
>
> Otavio,
>
> I ran into an issue building the 3.14-1.1.x-imx kernel using my custom
> kernel configuration - duplicate symbols linking the kernel image.
>
> The source of the conflict is having both of these patches in place:
>
> 3241216e250430a27b30eaf7f7df3e5e006acfb5 MGS-577 ARM: Export the cache
> management symbols when !MULTI_CACHE
> 6325d57e9677b417552ac76ada0ee91fa4c98566 arm: Export cache flush
> management symbols when !MULTI_CACHE
>
> and building for just ARMv7 CPUs.  Reverting one or the other of them
> fixes the issue.  I think that 3241216e is probably the better patch.

Thanks for your report and detailed analysis. I fixed jethro-next and
master-next, please check.

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Tue, Feb 23, 2016 at 5:01 PM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:
> Hello Matt,
[...]
> Thanks for your report and detailed analysis. I fixed jethro-next and
> master-next, please check.

Thanks, Otavio, that's better now.

-Matt


>
> --
> 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

Re: Merge of 3.14.52 GA into master and jethro

[Javier Viguera]

On 20/02/16 14:38, Otavio Salvador wrote:
> Hello folks,
>
> I finally got the time slot to finish the 3.14.52 review and merge. I
> have merged them, complete, on master-next and jethro-next.
>
> Please give the jethro-next a good test so we avoid bad surprises. I
> intend to merge jethro-next on Feb 26th.

> 02a3d11 gst1.0-fsl-plugins: Upgrade to 4.0.8 version

Hi Otavio.

gst1.0-fsl-plugins fails to build:

ERROR: Command Error: exit status: 1  Output:
Applying patch 0001-gst1.0-fsl-plugins-fix-appending-data-to-variable.patch
patching file configure.ac
Hunk #1 FAILED at 203.

The failing patch is one I sent for version 4.0.7. Seems that it has 
been included in version 4.0.8, so the patch is not needed anymore in 
the recipe. It needs to be removed.

-- 
Regards,

Javier Viguera

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

On Thu, Feb 25, 2016 at 6:43 AM, Javier Viguera <javier.viguera@digi.com> wrote:
...
> gst1.0-fsl-plugins fails to build:
>
> ERROR: Command Error: exit status: 1  Output:
> Applying patch 0001-gst1.0-fsl-plugins-fix-appending-data-to-variable.patch
> patching file configure.ac
> Hunk #1 FAILED at 203.
>
> The failing patch is one I sent for version 4.0.7. Seems that it has been
> included in version 4.0.8, so the patch is not needed anymore in the recipe.
> It needs to be removed.

I applied the patches you sent; thanks for taking the time to cook the patches.

Please give it a new try.

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Javier Viguera]

Hi Otavio,

On 25/02/16 19:45, Otavio Salvador wrote:
>
> I applied the patches you sent; thanks for taking the time to cook the patches.
>
> Please give it a new try.

I tested 'jethro-next' branch and worked fine. I guess 'master-next' 
should work as well.

Thanks.

Javier Viguera

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Sat, Feb 20, 2016 at 5:38 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:
> Hello folks,
>
> I finally got the time slot to finish the 3.14.52 review and merge. I
> have merged them, complete, on master-next and jethro-next.
>
> Please give the jethro-next a good test so we avoid bad surprises. I
> intend to merge jethro-next on Feb 26th.

I've done some basic testing of both jethro-next and master-next on my
wandboard-quad and a custom 6q-based system, and they're both working
OK, so far.  I'll try images on my imx6ulevk today.

I did have some problems with ext4 filesystem support in U-Boot, both
2015.10 and 2016.01, but I haven't had a chance to dig into that yet.

-Matt

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

On Fri, Feb 26, 2016 at 9:57 AM, Matt Madison <matt@madison.systems> wrote:
> On Sat, Feb 20, 2016 at 5:38 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>> Hello folks,
>>
>> I finally got the time slot to finish the 3.14.52 review and merge. I
>> have merged them, complete, on master-next and jethro-next.
>>
>> Please give the jethro-next a good test so we avoid bad surprises. I
>> intend to merge jethro-next on Feb 26th.
>
> I've done some basic testing of both jethro-next and master-next on my
> wandboard-quad and a custom 6q-based system, and they're both working
> OK, so far.  I'll try images on my imx6ulevk today.

Great!

> I did have some problems with ext4 filesystem support in U-Boot, both
> 2015.10 and 2016.01, but I haven't had a chance to dig into that yet.

What kind of problems did you have?

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Fri, Feb 26, 2016 at 5:03 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:
> On Fri, Feb 26, 2016 at 9:57 AM, Matt Madison <matt@madison.systems> wrote:
[...]
>> I did have some problems with ext4 filesystem support in U-Boot, both
>> 2015.10 and 2016.01, but I haven't had a chance to dig into that yet.
>
> What kind of problems did you have?

ext4ls either returns garbage filenames (seen on my Wandboard with
U-Boot 2016.01) or empty listings of directories that contain files
(seen on my custom 6q board with 2015.10).  The latter was more
problematic at the time, as I was trying to boot off an ext4 partition
and U-Boot couldn't find the kernel.

-Matt

>
> --
> 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

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

Hello Matt,

(Added Fabio on Cc)

On Fri, Feb 26, 2016 at 10:16 AM, Matt Madison <matt@madison.systems> wrote:
> On Fri, Feb 26, 2016 at 5:03 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>> On Fri, Feb 26, 2016 at 9:57 AM, Matt Madison <matt@madison.systems> wrote:
> [...]
>>> I did have some problems with ext4 filesystem support in U-Boot, both
>>> 2015.10 and 2016.01, but I haven't had a chance to dig into that yet.
>>
>> What kind of problems did you have?
>
> ext4ls either returns garbage filenames (seen on my Wandboard with
> U-Boot 2016.01) or empty listings of directories that contain files
> (seen on my custom 6q board with 2015.10).  The latter was more
> problematic at the time, as I was trying to boot off an ext4 partition
> and U-Boot couldn't find the kernel.

This is interesting. Fabio did you see something like that?

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

On Fri, Feb 26, 2016 at 10:21 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:

> This is interesting. Fabio did you see something like that?

Is there a pre-built image I can use to reproduce the problem?

I tried to use the 'master-next' images for mx6qsabresd, but the URLs
are all broken:
http://ci.ossystems.com.br/public/fsl-community-bsp/master-next/imx6qsabresd

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

On Fri, Feb 26, 2016 at 7:33 PM, Fabio Estevam <festevam@gmail.com> wrote:
> On Fri, Feb 26, 2016 at 10:21 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>
>> This is interesting. Fabio did you see something like that?
>
> Is there a pre-built image I can use to reproduce the problem?
>
> I tried to use the 'master-next' images for mx6qsabresd, but the URLs
> are all broken:
> http://ci.ossystems.com.br/public/fsl-community-bsp/master-next/imx6qsabresd

http://ci.ossystems.com.br/public/fsl-community-bsp/master-next/imx6qsabresd/framebuffer/1040/images

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

Hi Otavio,

On Sat, Feb 27, 2016 at 8:13 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:

> http://ci.ossystems.com.br/public/fsl-community-bsp/master-next/imx6qsabresd/framebuffer/1040/images

Thanks for providing the pre-built images. They are very helpful.

I tried it and booted fine for the first time on my mx6qsabresd. I
could reach the prompt successfully.

However in the second boot it failed to found the rootfs, which seems
to match what Matt reported:

ALSA device list:
  #0: wm8962-audio
  #1: imx-hdmi-soc
EXT3-fs (mmcblk2p2): error: couldn't mount because of unsupported
optional features (2c0)
EXT2-fs (mmcblk2p2): error: couldn't mount because of unsupported
optional features (2c4)
JBD2: recovery failed
EXT4-fs (mmcblk2p2): error loading journal
List of all partitions:
1f00            4096 mtdblock0  (driver?)
b300         3872256 mmcblk2  driver: mmcblk
  b301            8192 mmcblk2p1 04277521-01
  b302          761856 mmcblk2p2 04277521-02
b308         7757824 mmcblk3  driver: mmcblk
b320             128 mmcblk3rpmb  (driver?)
b318            1024 mmcblk3boot1  (driver?)
b310            1024 mmcblk3boot0  (driver?)
No filesystem could mount root, tried:  ext3 ext2 ext4 vfat fuseblk
Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(179,2)
CPU2: stopping
CPU: 2 PID: 0 Comm: swapper/2 Not tainted 3.14.61-fslc+g327d5c9 #1
[<80014f44>] (unwind_backtrace) from [<80011714>] (show_stack+0x10/0x14)
[<80011714>] (show_stack) from [<806fe774>] (dump_stack+0x7c/0xbc)
[<806fe774>] (dump_stack) from [<80013a44>] (handle_IPI+0x144/0x158)
[<80013a44>] (handle_IPI) from [<80008598>] (gic_handle_irq+0x58/0x5c)
[<80008598>] (gic_handle_irq) from [<80012280>] (__irq_svc+0x40/0x70)
Exception stack(0xa80a9fa0 to 0xa80a9fe8)
9fa0: ffffffed 2ad0f000 80a31790 8001d280 a80a8000 80a30574 8070774c a80a8030
9fc0: 80a99d7d 80a99d7d a80a8000 00000000 00000000 a80a9fe8 8000f058 8000f044
9fe0: 60000113 ffffffff
[<80012280>] (__irq_svc) from [<8000f044>] (arch_cpu_idle+0x2c/0x44)
[<8000f044>] (arch_cpu_idle) from [<8006dc54>] (cpu_startup_entry+0x100/0x14c)
[<8006dc54>] (cpu_startup_entry) from [<10008624>] (0x10008624)
CPU3: stopping
CPU: 3 PID: 0 Comm: swapper/3 Not tainted 3.14.61-fslc+g327d5c9 #1
[<80014f44>] (unwind_backtrace) from [<80011714>] (show_stack+0x10/0x14)
[<80011714>] (show_stack) from [<806fe774>] (dump_stack+0x7c/0xbc)
[<806fe774>] (dump_stack) from [<80013a44>] (handle_IPI+0x144/0x158)
[<80013a44>] (handle_IPI) from [<80008598>] (gic_handle_irq+0x58/0x5c)
[<80008598>] (gic_handle_irq) from [<80012280>] (__irq_svc+0x40/0x70)
Exception stack(0xa80abfa0 to 0xa80abfe8)
bfa0: ffffffed 2ad17000 80a31790 8001d280 a80aa000 80a30574 8070774c a80aa030
bfc0: 80a99d7d 80a99d7d a80aa000 00000000 00000000 a80abfe8 8000f058 8000f044
bfe0: 60000113 ffffffff
[<80012280>] (__irq_svc) from [<8000f044>] (arch_cpu_idle+0x2c/0x44)
[<8000f044>] (arch_cpu_idle) from [<8006dc54>] (cpu_startup_entry+0x100/0x14c)
[<8006dc54>] (cpu_startup_entry) from [<10008624>] (0x10008624)
CPU0: stopping
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.14.61-fslc+g327d5c9 #1
[<80014f44>] (unwind_backtrace) from [<80011714>] (show_stack+0x10/0x14)
[<80011714>] (show_stack) from [<806fe774>] (dump_stack+0x7c/0xbc)
[<806fe774>] (dump_stack) from [<80013a44>] (handle_IPI+0x144/0x158)
[<80013a44>] (handle_IPI) from [<80008598>] (gic_handle_irq+0x58/0x5c)
[<80008598>] (gic_handle_irq) from [<80012280>] (__irq_svc+0x40/0x70)
Exception stack(0x80a29f68 to 0x80a29fb0)
9f60:                   ffffffed 2acff000 80a31790 8001d280 80a28000 80a30574
9f80: 8070774c 80a28038 80a99d7d 80a99d7d 80a28000 00000000 00000000 80a29fb0
9fa0: 8000f058 8000f044 60000113 ffffffff
[<80012280>] (__irq_svc) from [<8000f044>] (arch_cpu_idle+0x2c/0x44)
[<8000f044>] (arch_cpu_idle) from [<8006dc54>] (cpu_startup_entry+0x100/0x14c)
[<8006dc54>] (cpu_startup_entry) from [<809c4b2c>] (start_kernel+0x350/0x35c)
[<809c4b2c>] (start_kernel) from [<10008074>] (0x10008074)

If I insert the SD card on my host PC, I can only mount the zImage/dtb
partition and the rootfs does not mount in the PC too.

This is the error I got on my PC:

"Unable to mount 780 MB Volume
Error mounting /dev/mmcblk0p2 at
/media/fabio/167af0df-63c6-426f-8461-763d24a0b678: Command-line `mount
-t "ext4" -o "uhelper=udisks2,nodev,nosuid" "/dev/mmcblk0p2"
"/media/fabio/167af0df-63c6-426f-8461-763d24a0b678"' exited with
non-zero exit status 32: mount: wrong fs type, bad option, bad
superblock on /dev/mmcblk0p2,
       missing codepage or helper program, or other error
       In some cases useful info is found in syslog - try
       dmesg | tail  or so"

All subsequent boots failed as well.

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

On Sat, Feb 27, 2016 at 9:03 AM, Fabio Estevam <festevam@gmail.com> wrote:
> On Sat, Feb 27, 2016 at 8:13 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>
>> http://ci.ossystems.com.br/public/fsl-community-bsp/master-next/imx6qsabresd/framebuffer/1040/images
>
> Thanks for providing the pre-built images. They are very helpful.

You are welcome. We are doing some massive changes on our autobuilder
infrastructure and we had some issues last days due those changes.

> I tried it and booted fine for the first time on my mx6qsabresd. I
> could reach the prompt successfully.
>
> However in the second boot it failed to found the rootfs, which seems
> to match what Matt reported:
>
> ALSA device list:
>   #0: wm8962-audio
>   #1: imx-hdmi-soc
> EXT3-fs (mmcblk2p2): error: couldn't mount because of unsupported
> optional features (2c0)
> EXT2-fs (mmcblk2p2): error: couldn't mount because of unsupported
> optional features (2c4)
> JBD2: recovery failed
> EXT4-fs (mmcblk2p2): error loading journal
...
> "Unable to mount 780 MB Volume
> Error mounting /dev/mmcblk0p2 at
> /media/fabio/167af0df-63c6-426f-8461-763d24a0b678: Command-line `mount
> -t "ext4" -o "uhelper=udisks2,nodev,nosuid" "/dev/mmcblk0p2"
> "/media/fabio/167af0df-63c6-426f-8461-763d24a0b678"' exited with
> non-zero exit status 32: mount: wrong fs type, bad option, bad
> superblock on /dev/mmcblk0p2,
>        missing codepage or helper program, or other error
>        In some cases useful info is found in syslog - try
>        dmesg | tail  or so"
>
> All subsequent boots failed as well.

I think I know what is going on; instead of plugging out the power
cable, issue a reboot in the first time. It will work afterwards it
seems.

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Sat, Feb 27, 2016 at 4:12 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:
> On Sat, Feb 27, 2016 at 9:03 AM, Fabio Estevam <festevam@gmail.com> wrote:
[...]
>> However in the second boot it failed to found the rootfs, which seems
>> to match what Matt reported:
[...]
>> All subsequent boots failed as well.
>
> I think I know what is going on; instead of plugging out the power
> cable, issue a reboot in the first time. It will work afterwards it
> seems.

The problem I was reporting was U-Boot being unable to read the ext4
rootfs properly (using ext4ls, for instance).  Linux had no problem
with it.

And it looks like the problem is that U-Boot's ext4 filesystem support
doesn't understand some of the more recent features that have been
added.  In particular, I found that by adding

EXTRA_IMAGECMD_append_ext4 = " -O ^flex_bg"

to my image recipe, to disable that particular ext4 feature, clears up
the problem I was seeing on my Wandboard, at least with jethro builds.
That seems to be enough, at least - I couldn't really tell from
reading the U-Boot ext4 code which features it really does support.

This doesn't affect the default SDcard images that much, since the
kernel and device tree get put into a FAT partition anyway, so U-Boot
doesn't have to read the rootfs partition.  It was more of a problem
for my custom board, because there I don't use the default layout, and
instead bundle the kernel and device tree with the rootfs.

With builds off master, the newer e2fsprogs from OE-Core master also
defaults to enabling the metadata_csum feature, and mke2fs issues a
warning about moving to a 3.18 or later kernel during the filesystem
creation (when run directly on my board with the 3.14-based kernel),
due to instability of that feature in older kernels.  So we could see
some other issues crop up going forward, too.

Thanks,
-Matt


>
> --
> 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
> --
> _______________________________________________
> meta-freescale mailing list
> meta-freescale@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-freescale

Re: Merge of 3.14.52 GA into master and jethro

[Otavio Salvador]

On Sat, Feb 27, 2016 at 10:23 AM, Matt Madison <matt@madison.systems> wrote:
> On Sat, Feb 27, 2016 at 4:12 AM, Otavio Salvador
> <otavio.salvador@ossystems.com.br> wrote:
>> On Sat, Feb 27, 2016 at 9:03 AM, Fabio Estevam <festevam@gmail.com> wrote:
> [...]
>>> However in the second boot it failed to found the rootfs, which seems
>>> to match what Matt reported:
> [...]
>>> All subsequent boots failed as well.
>>
>> I think I know what is going on; instead of plugging out the power
>> cable, issue a reboot in the first time. It will work afterwards it
>> seems.
>
> The problem I was reporting was U-Boot being unable to read the ext4
> rootfs properly (using ext4ls, for instance).  Linux had no problem
> with it.
>
> And it looks like the problem is that U-Boot's ext4 filesystem support
> doesn't understand some of the more recent features that have been
> added.  In particular, I found that by adding
>
> EXTRA_IMAGECMD_append_ext4 = " -O ^flex_bg"
>
> to my image recipe, to disable that particular ext4 feature, clears up
> the problem I was seeing on my Wandboard, at least with jethro builds.
> That seems to be enough, at least - I couldn't really tell from
> reading the U-Boot ext4 code which features it really does support.
>
> This doesn't affect the default SDcard images that much, since the
> kernel and device tree get put into a FAT partition anyway, so U-Boot
> doesn't have to read the rootfs partition.  It was more of a problem
> for my custom board, because there I don't use the default layout, and
> instead bundle the kernel and device tree with the rootfs.
>
> With builds off master, the newer e2fsprogs from OE-Core master also
> defaults to enabling the metadata_csum feature, and mke2fs issues a
> warning about moving to a 3.18 or later kernel during the filesystem
> creation (when run directly on my board with the 3.14-based kernel),
> due to instability of that feature in older kernels.  So we could see
> some other issues crop up going forward, too.

Great!

EXTRA_IMAGECMD_append_ext4 = " -O ^flex_bg -O ^metadata_csum"

Is worth test Fabio, it might solve the error you were seeing in the
first boot when plugging the power cable or doing a dirty reset.

-- 
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

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

On Sat, Feb 27, 2016 at 11:41 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:

> Great!
>
> EXTRA_IMAGECMD_append_ext4 = " -O ^flex_bg -O ^metadata_csum"
>
> Is worth test Fabio, it might solve the error you were seeing in the
> first boot when plugging the power cable or doing a dirty reset.

Ok, managed to get the ext4 corruption again.

If someone can generate me an image with 'EXTRA_IMAGECMD_append_ext4 =
" -O ^flex_bg -O ^metadata_csum"' I will be glad to test it.

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

On Sat, Feb 27, 2016 at 1:29 PM, Fabio Estevam <festevam@gmail.com> wrote:

> Ok, managed to get the ext4 corruption again.

Sorry, please discard this. I was using other board. Too many boards
on my table :-)

So far I only saw the ext4 corruption once. After reflashing the SD
card I no longer see it.

Re: Merge of 3.14.52 GA into master and jethro

[Julio Cruz Barroso]

Hi Fabio,

I cannot reproduce the issue.

U-Boot 2015.10+fslc+g1b6aee7 (Feb 26 2016 - 20:05:37 +0800)
....
=> ext4ls mmc 0:2
<DIR>       4096 .
<DIR>       4096 ..
<DIR>      16384 lost+found
<DIR>       4096 unit_tests
<DIR>       4096 media
<DIR>       4096 dev
<DIR>       4096 bin
<DIR>       4096 sbin
<DIR>       4096 usr
<DIR>       4096 run
<DIR>       4096 etc
<DIR>       4096 boot
<DIR>       4096 home
<DIR>       4096 mnt
<DIR>       4096 opt
<DIR>       4096 sys
<DIR>       4096 proc
<DIR>       4096 lib
<DIR>       1342 tmp
<DIR>       1159 var
=>
...

Regards

Julio


-----Original Message-----
From: meta-freescale-bounces@yoctoproject.org [mailto:meta-freescale-bounces@yoctoproject.org] On Behalf Of Fabio Estevam
Sent: Sunday, February 28, 2016 12:59 AM
To: Otavio Salvador
Cc: Fabio Estevam; meta-freescale@yoctoproject.org
Subject: Re: [meta-freescale] Merge of 3.14.52 GA into master and jethro

On Sat, Feb 27, 2016 at 1:29 PM, Fabio Estevam <festevam@gmail.com> wrote:

> Ok, managed to get the ext4 corruption again.

Sorry, please discard this. I was using other board. Too many boards on my table :-)

So far I only saw the ext4 corruption once. After reflashing the SD card I no longer see it.
--
_______________________________________________
meta-freescale mailing list
meta-freescale@yoctoproject.org
https://lists.yoctoproject.org/listinfo/meta-freescale

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

Hi Julio,

On Mon, Feb 29, 2016 at 2:46 AM, Julio Cruz Barroso
<julio.cruz@smartmatic.com> wrote:
> Hi Fabio,
>
> I cannot reproduce the issue.

Same here with 2016.03-rc1:

=> ext4ls mmc 1:2
<DIR>       4096 .
<DIR>       4096 ..
<DIR>      16384 lost+found
<DIR>       4096 sbin
<DIR>       4096 dev
<DIR>       4096 run
<DIR>       4096 unit_tests
<SYM>          8 tmp
<DIR>       4096 bin
<DIR>       4096 media
<DIR>       4096 sys
<DIR>       4096 mnt
<DIR>       4096 home
<DIR>       4096 boot
<DIR>       4096 proc
<DIR>       4096 usr
<DIR>       4096 lib
<DIR>       4096 etc
<DIR>       4096 opt
<DIR>       4096 var

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Mon, Feb 29, 2016 at 3:33 AM, Fabio Estevam <festevam@gmail.com> wrote:
> Hi Julio,
>
> On Mon, Feb 29, 2016 at 2:46 AM, Julio Cruz Barroso
> <julio.cruz@smartmatic.com> wrote:
>> Hi Fabio,
>>
>> I cannot reproduce the issue.
[...]

>
> Same here with 2016.03-rc1:
[...]


Root-level directory listing generally works fine.  Try listing the
contents of some of the directories below.

Thanks,
-Matt

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

On Mon, Feb 29, 2016 at 8:48 AM, Matt Madison <matt@madison.systems> wrote:

> Root-level directory listing generally works fine.  Try listing the
> contents of some of the directories below.

Seems to work fine here too:

=> ext4ls mmc 1:2 unit_tests
<DIR>       4096 .
<DIR>       4096 ..
           24712 mxc_spi_test1.out
            1138 gpuinfo.sh
            1159 autorun-ata.sh
           20880 rtctest.out
             183 autorun-keypad.sh
         2363276 mxc_epdc_fb_test.out
             198 mxc_tuner_test.sh
             362 autorun-hantro.sh
         1843200 pansy-1280x720-565.rgb
           51652 mxc_l2switch_test.out
           24676 mxc_uart_xmit_test.out
         2913440 memtool
           14916 mxc_cec_test.out
            1108 autorun-asrc.sh
          576000 stefan_interlaced_320x240_5frames.yv12
           19292 check_audio.out
           13052 mxc_sdma_test.out
          382424 autorun-pxp.sh
             563 autorun-mmc-fdisk.sh
             448 gpu.sh
             624 autorun-mmc.sh
            5166 test-utils.sh
            1778 ipudev_config_file
            1754 autorun-fb.sh
             644 mxc_tve_test.sh
         2621440 plumbago-1280x1024-565.rgb
             486 autorun-sound.sh
          614400 daisy-640x480-565.rgb
            1068 config_enc
            1295 mxc_irlan_test.sh
            1342 autorun-rtc.sh
           31400 check_ioctls.out
             724 all-suite.txt
           19796 check_recording.out
            4957 autorun-ipu.sh
           24904 mxc_fb_test.out
             345 suspend_random_auto.sh
           15016 wdt_driver_test.out
          133940 mxc_vpu_test.out
           49064 mxc_v4l2_overlay.out
           19280 mxc_tuner_test.out
            1309 config_dec
            1463 dump-clocks-dot.sh
            1627 autorun-usb-gadget.sh
             760 autorun-oprofile.sh
           11876 mmdc2
           37696 etm
             579 autorun-mmc-blockrw.sh
             474 autorun-firi.sh
             936 autorun-mmc-fs.sh
           17508 mxc_i2c_slave_test.out
            2088 config_encdec
            3383 autorun.sh
            8381 autorun-v4l2.sh
            4063 autorun-owire.sh
           17312 mxc_fb_vsync_test.out
             473 dbg_monitor.sh
           30860 mxc_uart_stress_test.out
           15332 mxc_uart_test.out
           29644 mxc_v4l2_still.out
            2825 config_net
            1467 dryice.sh
              65 mxc_keyb_test.sh
          382424 pxp_test.out
           13616 check_mixer.out
           32936 mxc_dcic_test.out
           17228 check_2w1r.out
           24564 mxc_mlb_test.out
             408 autorun-mmc-mkfs.sh
           59408 mxc_ipudev_test.out
           38044 mxc_v4l2_tvin.out
             410 autorun-mxc_uart.sh
             388 autorun-vpu.sh
           14484 check_volume2.out
        27201512 mxc_epdc_v2_fb_test.out
           32900 mxc_asrc_test.out
           36980 mxc_v4l2_capture.out
           24320 mxc_isl29023.out
           47652 mxc_v4l2_output.out
          960000 rose-800x600-565.rgb
           24812 rtcwakeup.out
           15396 check_minor_open.out
           86400 desk240x180-565.rgb
          161684 audio8k16S.wav
         3304928 mxc_spdc_fb_test.out
           14184 check_synchronization.out
             272 suspend_quick_auto.sh
             950 dump-clocks.sh
           68010 akiyo.mp4
             164 autorun-wdog.sh
            1263 autorun-usb-host.sh
         1572864 wall-1024x768-565.rgb
           77076 cam2mpeg4.out
           15664 mxc_iim_test.out

Re: Merge of 3.14.52 GA into master and jethro

[Matt Madison]

On Mon, Feb 29, 2016 at 3:53 AM, Fabio Estevam <festevam@gmail.com> wrote:
> On Mon, Feb 29, 2016 at 8:48 AM, Matt Madison <matt@madison.systems> wrote:
>
>> Root-level directory listing generally works fine.  Try listing the
>> contents of some of the directories below.
>
> Seems to work fine here too:

Well, that figures.  :(  When I get a chance, I'll work on trying to
characterize the problem better, or at least get a reliably
reproducible test case together.

Thanks,
-Matt

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

Hi Matt,

On Sat, Feb 27, 2016 at 10:23 AM, Matt Madison <matt@madison.systems> wrote:

> The problem I was reporting was U-Boot being unable to read the ext4
> rootfs properly (using ext4ls, for instance).  Linux had no problem
> with it.
>
> And it looks like the problem is that U-Boot's ext4 filesystem support
> doesn't understand some of the more recent features that have been
> added.  In particular, I found that by adding
>
> EXTRA_IMAGECMD_append_ext4 = " -O ^flex_bg"
>
> to my image recipe, to disable that particular ext4 feature, clears up
> the problem I was seeing on my Wandboard, at least with jethro builds.
> That seems to be enough, at least - I couldn't really tell from
> reading the U-Boot ext4 code which features it really does support.

Care to report this problem to the U-boot mailing list? I think it is
worth to be reported and investigated.

Thanks

Re: Merge of 3.14.52 GA into master and jethro

[Fabio Estevam]

On Sat, Feb 27, 2016 at 9:12 AM, Otavio Salvador
<otavio.salvador@ossystems.com.br> wrote:

> I think I know what is going on; instead of plugging out the power
> cable, issue a reboot in the first time. It will work afterwards it
> seems.

Ok, so I reflashed the same image into the same SD card, and now I can
not see the rootfs corruption anymore.

I booted via 'reboot' command in Linux and also by unplugging the power.

Tested several boots and the rootfs mounted fine every time, so I
could not reproduce the original issue I reported earlier.