[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

Hey all,

I've pushed v2015.01-rc1 out to the repository and tarballs should exist
soon.

The merge window is now closed.  Per both the min-summit and the follow
up emails, I'm going to try and get into the every 2 week RC tagging
groove.  And, I'm going to be relying on everyones best judgement about
what level of changes they're going to put into a pull request to me.

When I do a buildman of most things I see:
  blackfin: +   ip04
       arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
   powerpc: +   kmcoge4 taishan kmlion1 dlvision  

And I'd really like to see those sorted out, one way or another for
-rc2.

As always, if anything is broken please speak up.

Thanks all!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141110/36783cc5/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Masahiro Yamada]

On Mon, 10 Nov 2014 16:29:38 -0500
Tom Rini <trini@ti.com> wrote:

> Hey all,
> 
> I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> soon.


Half a day has passed since Tom's  "I've pushed v2015.rc-rc1" mail,
but I still don't see it.

I have been often wondering why we can see the release
much lator of Tom's announce mail.

When I push my u-boot-uniphier repo, I can see the update on the public repo.


Best Regards
Masahiro Yamada

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Andreas Bießmann]

Hi Tom,

On 11/10/2014 10:29 PM, Tom Rini wrote:
> Hey all,
> 
> I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> soon.
> 
> The merge window is now closed.  Per both the min-summit and the follow
> up emails, I'm going to try and get into the every 2 week RC tagging
> groove.  And, I'm going to be relying on everyones best judgement about
> what level of changes they're going to put into a pull request to me.
> 
> When I do a buildman of most things I see:
>   blackfin: +   ip04
>        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517

there must be something wrong:

---8<---
abiessmann at punisher % git describe
v2015.01-rc1-14-g63f7af9
abiessmann@punisher %
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make O=/tmp/build-tricorder CROSS_COMPILE=arm-arm926ejs-linux-gnueabi-
-s -j16 tricorder_defconfig
In file included from scripts/kconfig/zconf.tab.c:2537:0:
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c: In function
?get_symbol_str?:
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c:590:18: warning:
?jump? may be used uninitialized in this function [-Wmaybe-uninitialized]
     jump->offset = strlen(r->s);
                  ^
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c:551:19: note: ?jump?
was declared here
  struct jump_key *jump;
                   ^
#
# configuration written to .config
#
#
# configuration written to spl/.config
#
abiessmann at punisher %
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make O=/tmp/build-tricorder CROSS_COMPILE=arm-arm926ejs-linux-gnueabi-
-s -j16
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make    36.56s user 1.68s system 404% cpu 9.463 total
abiessmann at punisher % rm -rf /tmp/build-tricorder
abiessmann at punisher %
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make O=/tmp/build-tricorder CROSS_COMPILE=arm-arm926ejs-linux-gnueabi-
-s -j16 tricorder_flash_defconfig
In file included from scripts/kconfig/zconf.tab.c:2537:0:
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c: In function
?get_symbol_str?:
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c:590:18: warning:
?jump? may be used uninitialized in this function [-Wmaybe-uninitialized]
     jump->offset = strlen(r->s);
                  ^
/home/abiessmann/src/u-boot/scripts/kconfig/menu.c:551:19: note: ?jump?
was declared here
  struct jump_key *jump;
                   ^
#
# configuration written to .config
#
#
# configuration written to spl/.config
#
abiessmann at punisher %
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make O=/tmp/build-tricorder CROSS_COMPILE=arm-arm926ejs-linux-gnueabi-
-s -j16
PATH=/opt/CS.Toolchain-0.3.0-rc1/arm-arm926ejs-linux-gnueabi/bin:$PATH
make    35.62s user 1.58s system 526% cpu 7.063 total
abiessmann at punisher %
abiessmann at punisher % ./tools/buildman/buildman -o /tmp/testbuild -c 1
-b master tricorder
boards.cfg is up to date. Nothing to do.
Building 1 commit for 2 boards (2 threads, 4 jobs per thread)
Cloning repo for thread 0
Cloning repo for thread 1
    2    0    0 /2      0:00:10  : tricorder
./tools/buildman/buildman -o /tmp/testbuild -c 1 -b master tricorder
73.72s user 4.00s system 338% cpu 22.986 total
abiessmann at punisher %
--->8---

Best regards

Andreas Bie?mann

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

On Wed, Nov 12, 2014 at 10:12:30AM +0100, Andreas Bie?mann wrote:
> Hi Tom,
> 
> On 11/10/2014 10:29 PM, Tom Rini wrote:
> > Hey all,
> > 
> > I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> > soon.
> > 
> > The merge window is now closed.  Per both the min-summit and the follow
> > up emails, I'm going to try and get into the every 2 week RC tagging
> > groove.  And, I'm going to be relying on everyones best judgement about
> > what level of changes they're going to put into a pull request to me.
> > 
> > When I do a buildman of most things I see:
> >   blackfin: +   ip04
> >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
> 
> there must be something wrong:

Compiler fun!  With Linaro 2013.03 gcc 4.7.3 we're 260 bytes over the
limit on SPL.  With ELDK 5.4's 4.7.x we're 543 bytes under the limit.
What I think I had been thinking was that you could move
CONFIG_SPL_STACK into DDR and bump the max size a bit or based on
CONFIG_FLASHCARD or not build NAND or MMC into MLO but not both (and
that'd be a bigger size savings).

Or maybe the answer is that I need to stop using that Linaro toolchain
for 4.7.x testing.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141112/5dad18d7/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

On Wed, Nov 12, 2014 at 12:04:39PM -0500, Tom Rini wrote:
> On Wed, Nov 12, 2014 at 10:12:30AM +0100, Andreas Bie?mann wrote:
> > Hi Tom,
> > 
> > On 11/10/2014 10:29 PM, Tom Rini wrote:
> > > Hey all,
> > > 
> > > I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> > > soon.
> > > 
> > > The merge window is now closed.  Per both the min-summit and the follow
> > > up emails, I'm going to try and get into the every 2 week RC tagging
> > > groove.  And, I'm going to be relying on everyones best judgement about
> > > what level of changes they're going to put into a pull request to me.
> > > 
> > > When I do a buildman of most things I see:
> > >   blackfin: +   ip04
> > >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
> > 
> > there must be something wrong:
> 
> Compiler fun!  With Linaro 2013.03 gcc 4.7.3 we're 260 bytes over the
> limit on SPL.  With ELDK 5.4's 4.7.x we're 543 bytes under the limit.
> What I think I had been thinking was that you could move
> CONFIG_SPL_STACK into DDR and bump the max size a bit or based on
> CONFIG_FLASHCARD or not build NAND or MMC into MLO but not both (and
> that'd be a bigger size savings).
> 
> Or maybe the answer is that I need to stop using that Linaro toolchain
> for 4.7.x testing.

... ELDK 5.5.x also has a similar slight size overflow and is 4.8.x.  I
guess for now I'll default to use ELDK 5.4 for ARM but people may need
to consider a something-something for newer toolchains sooner rather
than later.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141112/2ff1d68c/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

On Wed, Nov 12, 2014 at 04:10:06PM -0500, Tom Rini wrote:
> On Wed, Nov 12, 2014 at 12:04:39PM -0500, Tom Rini wrote:
> > On Wed, Nov 12, 2014 at 10:12:30AM +0100, Andreas Bie?mann wrote:
> > > Hi Tom,
> > > 
> > > On 11/10/2014 10:29 PM, Tom Rini wrote:
> > > > Hey all,
> > > > 
> > > > I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> > > > soon.
> > > > 
> > > > The merge window is now closed.  Per both the min-summit and the follow
> > > > up emails, I'm going to try and get into the every 2 week RC tagging
> > > > groove.  And, I'm going to be relying on everyones best judgement about
> > > > what level of changes they're going to put into a pull request to me.
> > > > 
> > > > When I do a buildman of most things I see:
> > > >   blackfin: +   ip04
> > > >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
> > > 
> > > there must be something wrong:
> > 
> > Compiler fun!  With Linaro 2013.03 gcc 4.7.3 we're 260 bytes over the
> > limit on SPL.  With ELDK 5.4's 4.7.x we're 543 bytes under the limit.
> > What I think I had been thinking was that you could move
> > CONFIG_SPL_STACK into DDR and bump the max size a bit or based on
> > CONFIG_FLASHCARD or not build NAND or MMC into MLO but not both (and
> > that'd be a bigger size savings).
> > 
> > Or maybe the answer is that I need to stop using that Linaro toolchain
> > for 4.7.x testing.
> 
> ... ELDK 5.5.x also has a similar slight size overflow and is 4.8.x.  I
> guess for now I'll default to use ELDK 5.4 for ARM but people may need
> to consider a something-something for newer toolchains sooner rather
> than later.

... I'm actually running into that issue where buildman grabs things
from $PATH sometimes rather than the toolchain it found, re-running some
tests.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141112/9a70677d/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

On Wed, Nov 12, 2014 at 05:16:42PM -0500, Tom Rini wrote:
> On Wed, Nov 12, 2014 at 04:10:06PM -0500, Tom Rini wrote:
> > On Wed, Nov 12, 2014 at 12:04:39PM -0500, Tom Rini wrote:
> > > On Wed, Nov 12, 2014 at 10:12:30AM +0100, Andreas Bie?mann wrote:
> > > > Hi Tom,
> > > > 
> > > > On 11/10/2014 10:29 PM, Tom Rini wrote:
> > > > > Hey all,
> > > > > 
> > > > > I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> > > > > soon.
> > > > > 
> > > > > The merge window is now closed.  Per both the min-summit and the follow
> > > > > up emails, I'm going to try and get into the every 2 week RC tagging
> > > > > groove.  And, I'm going to be relying on everyones best judgement about
> > > > > what level of changes they're going to put into a pull request to me.
> > > > > 
> > > > > When I do a buildman of most things I see:
> > > > >   blackfin: +   ip04
> > > > >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
> > > > 
> > > > there must be something wrong:
> > > 
> > > Compiler fun!  With Linaro 2013.03 gcc 4.7.3 we're 260 bytes over the
> > > limit on SPL.  With ELDK 5.4's 4.7.x we're 543 bytes under the limit.
> > > What I think I had been thinking was that you could move
> > > CONFIG_SPL_STACK into DDR and bump the max size a bit or based on
> > > CONFIG_FLASHCARD or not build NAND or MMC into MLO but not both (and
> > > that'd be a bigger size savings).
> > > 
> > > Or maybe the answer is that I need to stop using that Linaro toolchain
> > > for 4.7.x testing.
> > 
> > ... ELDK 5.5.x also has a similar slight size overflow and is 4.8.x.  I
> > guess for now I'll default to use ELDK 5.4 for ARM but people may need
> > to consider a something-something for newer toolchains sooner rather
> > than later.
> 
> ... I'm actually running into that issue where buildman grabs things
> from $PATH sometimes rather than the toolchain it found, re-running some
> tests.

OK, purging the Ubuntu toolchain packages so that buildman really uses
what I tell it, I'm getting the best (build time) results with ELDK
5.5.2 for ARM.  After some patches that I'm about to push:
Summary of current source for 1166 boards (6 threads, 1 job per thread)
       arm: +   maxbcm odroid db-mv784mp-gp
+(maxbcm)   u16 reg;
+(maxbcm,maxbcm)       ^
+(maxbcm)   u16 devadr = CONFIG_PHY_BASE_ADDR;
+(maxbcm,maxbcm) Didn't find the file 'board/maxbcm/binary.0' in 'arm/current/.bm-work/
03/build' which is mandatory to generate the image
+(maxbcm,maxbcm,db-mv784mp-gp,db-mv784mp-gp) This file generally contains the DDR3 training code, and should be extracted from an existing bootable
+(maxbcm,maxbcm,db-mv784mp-gp,db-mv784mp-gp) image for your board. See 'kwbimage -x' to extract it from an existing image.
+(maxbcm,db-mv784mp-gp) Could not create image
+(maxbcm,db-mv784mp-gp) make[1]: *** [u-boot.kwb] Error 1
+(maxbcm,db-mv784mp-gp) make: *** [sub-make] Error 2
+(odroid,odroid)    addr = CONFIG_SYS_SDRAM_BASE + (i * SDRAM_BANK_SIZE);
+(odroid,odroid)                                 ^
+(odroid) board/samsung/common/board.c:94:2: note: containing loop
+(odroid,odroid)   for (i = 0; i < CONFIG_NR_DRAM_BANKS; i++) {
+(odroid,odroid)   ^
+(odroid) board/samsung/common/board.c:106:2: note: containing loop
+(db-mv784mp-gp,db-mv784mp-gp) Didn't find the file 'board/Marvell/db-mv784mp-gp/binary.0' in 'arm/current/.bm-work/03/build' which is mandatory to generate the image
w+(maxbcm) board/maxbcm/maxbcm.c: In function 'reset_phy':
w+(maxbcm) board/maxbcm/maxbcm.c:68:6: warning: unused variable 'reg' [-Wunused-variable]
w+(maxbcm) board/maxbcm/maxbcm.c:66:6: warning: unused variable 'devadr' [-Wunused-variable]
w+(odroid) board/samsung/common/board.c: In function 'dram_init':
w+(odroid) board/samsung/common/board.c:95:32: warning: iteration 4u invokes undefined behavior [-Waggressive-loop-optimizations]
w+(odroid) board/samsung/common/board.c: In function 'dram_init_banksize':
w+(odroid) board/samsung/common/board.c:107:32: warning: iteration 4u invokes undefined behavior [-Waggressive-loop-optimizations]

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141113/d1145883/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Nikita Kiryanov]

Hi Tom,

On 11/10/2014 11:29 PM, Tom Rini wrote:
> Hey all,
>
> I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> soon.
>
> The merge window is now closed.  Per both the min-summit and the follow
> up emails, I'm going to try and get into the every 2 week RC tagging
> groove.  And, I'm going to be relying on everyones best judgement about
> what level of changes they're going to put into a pull request to me.
>
> When I do a buildman of most things I see:
>    blackfin: +   ip04
>         arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517

I am able to build cm_t3517 without any problems.


-- 
Regards,
Nikita Kiryanov

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Tom Rini]

On Wed, Nov 12, 2014 at 08:24:47PM +0200, Nikita Kiryanov wrote:
> Hi Tom,
> 
> On 11/10/2014 11:29 PM, Tom Rini wrote:
> >Hey all,
> >
> >I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> >soon.
> >
> >The merge window is now closed.  Per both the min-summit and the follow
> >up emails, I'm going to try and get into the every 2 week RC tagging
> >groove.  And, I'm going to be relying on everyones best judgement about
> >what level of changes they're going to put into a pull request to me.
> >
> >When I do a buildman of most things I see:
> >   blackfin: +   ip04
> >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
> 
> I am able to build cm_t3517 without any problems.

OK, that failure was a race condition, sorry for the noise!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20141112/8413bdc2/attachment.pgp>

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Stefano Babic]

Hi Marek,

On 10/11/2014 22:29, Tom Rini wrote:
> Hey all,
> 
> I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> soon.
> 
> The merge window is now closed.  Per both the min-summit and the follow
> up emails, I'm going to try and get into the every 2 week RC tagging
> groove.  And, I'm going to be relying on everyones best judgement about
> what level of changes they're going to put into a pull request to me.
> 
> When I do a buildman of most things I see:
>   blackfin: +   ip04
>        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517

I confirm the issue with novena after pulling mainline into u-boot-imx.
It looks like that CONFIG_SPL_LDSCRIPT is somewhere unset (I do not why,
gw_ventana is built fine).

Marek, can you take a look at it ?

Thanks !

Stefano

-- 
=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sbabic at denx.de
=====================================================================

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Marek Vasut]

On Thursday, November 13, 2014 at 10:46:16 AM, Stefano Babic wrote:
> Hi Marek,
> 
> On 10/11/2014 22:29, Tom Rini wrote:
> > Hey all,
> > 
> > I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> > soon.
> > 
> > The merge window is now closed.  Per both the min-summit and the follow
> > up emails, I'm going to try and get into the every 2 week RC tagging
> > groove.  And, I'm going to be relying on everyones best judgement about
> > what level of changes they're going to put into a pull request to me.
> > 
> > When I do a buildman of most things I see:
> >   blackfin: +   ip04
> >   
> >        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate
> >        openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt
> >        tricorder_flash cm_t3517
> 
> I confirm the issue with novena after pulling mainline into u-boot-imx.
> It looks like that CONFIG_SPL_LDSCRIPT is somewhere unset (I do not why,
> gw_ventana is built fine).
> 
> Marek, can you take a look at it ?

Yes, you have an email fixing this.

Best regards,
Marek Vasut

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Dirk Eibach]

Hey Tom,

2014-11-10 22:29 GMT+01:00 Tom Rini <trini@ti.com>:
> ...
> When I do a buildman of most things I see:
>   blackfin: +   ip04
>        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
>    powerpc: +   kmcoge4 taishan kmlion1 dlvision
>
> And I'd really like to see those sorted out, one way or another for
> -rc2.

dlvision should be fixed with the "Update gdsys board support" series
that is already under review by Stefan Roese and Kim Phillips.

Cheers
Dirk

[U-Boot] [ANN] U-Boot v2015.01-rc1 released

[Valentin Longchamp]

Hi Tom,

On 11/10/2014 10:29 PM, Tom Rini wrote:
> Hey all,
> 
> I've pushed v2015.01-rc1 out to the repository and tarballs should exist
> soon.
> 
> The merge window is now closed.  Per both the min-summit and the follow
> up emails, I'm going to try and get into the every 2 week RC tagging
> groove.  And, I'm going to be relying on everyones best judgement about
> what level of changes they're going to put into a pull request to me.
> 
> When I do a buildman of most things I see:
>   blackfin: +   ip04
>        arm: +   Ippo_q8h_v5 novena openrd_base maxbcm openrd_ultimate openrd_client tricorder db-mv784mp-gp Linksprite_pcDuino3_fdt tricorder_flash cm_t3517
>    powerpc: +   kmcoge4 taishan kmlion1 dlvision  
> 
> And I'd really like to see those sorted out, one way or another for
> -rc2.
> 

For both kmcoge4 and kmlion1, the error is fixed by [1].

Hopefully York will be able to review this, also with the 2 patches for  the
migration of the keymile powerpc board to generic board, for -rc2.

Best Regards

Valentin

[1] http://patchwork.ozlabs.org/patch/403232/