[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Carlos Santos]

Hello,

configs/beaglebone_qt5_defconfig contain BR2_GCC_VERSION_5_X=y because
the Linux 4.1 kernel used on the board does not compile with gcc 6.x
and later.

At the same time it contains BR2_TOOLCHAIN_BUILDROOT_GLIBC=y, which
invalidated the CGG version selection because
since commit a7914499b80 glbc requires GCC 6.2 or later.

Switching to uClibc is not an option, since ti-sgx-um requires glibc.

-- 
Carlos Santos <unixmania@gmail.com>

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Yegor Yefremov]

Hello Carlos,

On Fri, Oct 11, 2019 at 4:03 AM Carlos Santos <unixmania@gmail.com> wrote:
>
> Hello,
>
> configs/beaglebone_qt5_defconfig contain BR2_GCC_VERSION_5_X=y because
> the Linux 4.1 kernel used on the board does not compile with gcc 6.x
> and later.
>
> At the same time it contains BR2_TOOLCHAIN_BUILDROOT_GLIBC=y, which
> invalidated the CGG version selection because
> since commit a7914499b80 glbc requires GCC 6.2 or later.
>
> Switching to uClibc is not an option, since ti-sgx-um requires glibc.

I've also seen the issue with the toolchain. I'm not familiar with 3D
stuff. My intention was to make both beaglebone defconfigs coherent in
regards to the bootloader

@Lothar can help here?

If you fix the defcongif, please also drop the /dev/ttyOx as a console.

Regards,
Yegor.

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Romain Naour]

Hi Carlos, All,

Le 11/10/2019 ? 04:03, Carlos Santos a ?crit?:
> Hello,
> 
> configs/beaglebone_qt5_defconfig contain BR2_GCC_VERSION_5_X=y because
> the Linux 4.1 kernel used on the board does not compile with gcc 6.x
> and later.
> 
> At the same time it contains BR2_TOOLCHAIN_BUILDROOT_GLIBC=y, which
> invalidated the CGG version selection because
> since commit a7914499b80 glbc requires GCC 6.2 or later.
> 
> Switching to uClibc is not an option, since ti-sgx-um requires glibc.
> 

We recently (June 2019) talked about this defconfig [1] but this was before
merging glibc 2.30 (August 2019).

There is an old patch [2] [3] in patchwork bumping the kernel (to 4.9) used in
this defconfig but it's marked "Changes Requested" due to issues.

If we really want to keep this defconfig (as is), we have to use an external
toolchain old enough to build the kernel and TI's modules.

[1] http://lists.busybox.net/pipermail/buildroot/2019-June/253504.html
[2] http://patchwork.ozlabs.org/patch/939705/
[3] http://lists.busybox.net/pipermail/buildroot/2018-June/224863.html

Best regards,
Romain

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Thomas Petazzoni]

On Sun, 13 Oct 2019 11:29:14 +0200
Romain Naour <romain.naour@gmail.com> wrote:

> There is an old patch [2] [3] in patchwork bumping the kernel (to 4.9) used in
> this defconfig but it's marked "Changes Requested" due to issues.
> 
> If we really want to keep this defconfig (as is), we have to use an external
> toolchain old enough to build the kernel and TI's modules.

The right way to fix this defconfig is certainly not to use an older
toolchain, but rather to use a more modern kernel. It's BeagleBone,
definitely not a legacy/ancient platform, it's still in widespread use,
there is quite certainly a kernel + SGX OpenGL stuff that is reasonably
recent ?

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Romain Naour]

Hello Thomas,

Le 13/10/2019 ? 15:42, Thomas Petazzoni a ?crit?:
> On Sun, 13 Oct 2019 11:29:14 +0200
> Romain Naour <romain.naour@gmail.com> wrote:
> 
>> There is an old patch [2] [3] in patchwork bumping the kernel (to 4.9) used in
>> this defconfig but it's marked "Changes Requested" due to issues.
>>
>> If we really want to keep this defconfig (as is), we have to use an external
>> toolchain old enough to build the kernel and TI's modules.
> 
> The right way to fix this defconfig is certainly not to use an older
> toolchain, but rather to use a more modern kernel. It's BeagleBone,
> definitely not a legacy/ancient platform, it's still in widespread use,
> there is quite certainly a kernel + SGX OpenGL stuff that is reasonably
> recent ?

I'm agree with you, but we need someone brave enough to work on this :)
I don't have a Beaglebone to do a runtime test.

Best regards,
Romain

> 
> Best regards,
> 
> Thomas
> 

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Carlos Santos]

I have a beaglebone black and can make some experiments this week.

Carlos Santos <unixmania@gmail.com>

Em 13 de out de 2019, ?(s) 10:57, Romain Naour <romain.naour@gmail.com> escreveu:

> Hello Thomas,
> 
>> Le 13/10/2019 ? 15:42, Thomas Petazzoni a ?crit :
>> On Sun, 13 Oct 2019 11:29:14 +0200
>> Romain Naour <romain.naour@gmail.com> wrote:
>> 
>>> There is an old patch [2] [3] in patchwork bumping the kernel (to 4.9) used in
>>> this defconfig but it's marked "Changes Requested" due to issues.
>>> 
>>> If we really want to keep this defconfig (as is), we have to use an external
>>> toolchain old enough to build the kernel and TI's modules.
>> 
>> The right way to fix this defconfig is certainly not to use an older
>> toolchain, but rather to use a more modern kernel. It's BeagleBone,
>> definitely not a legacy/ancient platform, it's still in widespread use,
>> there is quite certainly a kernel + SGX OpenGL stuff that is reasonably
>> recent ?
> 
> I'm agree with you, but we need someone brave enough to work on this :)
> I don't have a Beaglebone to do a runtime test.
> 
> Best regards,
> Romain
> 
>> 
>> Best regards,
>> 
>> Thomas
>> 
> 

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Lothar Felten]

On Sun, 2019-10-13 at 11:57 -0300, Carlos Santos wrote:
> I have a beaglebone black and can make some experiments this week.

Hello Carlos,

I have four patches, they allow you to build a new kernel, kernel
module and userspace tools including libgbm for the beaglebone with the
current toolchain.
You can run the GLES(1) demos, I had no success with the GLES2 demos.
The last patch is for qt5base, but it is incomplete, it won't compile
the qpa backend yet.
I'll post the four patches with RFC prefix to the list.

Best regards,
Lothar
> 
> Carlos Santos <unixmania@gmail.com>
> 
> Em 13 de out de 2019, ?(s) 10:57, Romain Naour <
> romain.naour at gmail.com> escreveu:
> 
> > Hello Thomas,
> > 
> > > Le 13/10/2019 ? 15:42, Thomas Petazzoni a ?crit :
> > > On Sun, 13 Oct 2019 11:29:14 +0200
> > > Romain Naour <romain.naour@gmail.com> wrote:
> > > 
> > > > There is an old patch [2] [3] in patchwork bumping the kernel
> > > > (to 4.9) used in
> > > > this defconfig but it's marked "Changes Requested" due to
> > > > issues.
> > > > 
> > > > If we really want to keep this defconfig (as is), we have to
> > > > use an external
> > > > toolchain old enough to build the kernel and TI's modules.
> > > 
> > > The right way to fix this defconfig is certainly not to use an
> > > older
> > > toolchain, but rather to use a more modern kernel. It's
> > > BeagleBone,
> > > definitely not a legacy/ancient platform, it's still in
> > > widespread use,
> > > there is quite certainly a kernel + SGX OpenGL stuff that is
> > > reasonably
> > > recent ?
> > 
> > I'm agree with you, but we need someone brave enough to work on
> > this :)
> > I don't have a Beaglebone to do a runtime test.
> > 
> > Best regards,
> > Romain
> > 
> > > 
> > > Best regards,
> > > 
> > > Thomas
> > > 

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Carlos Santos]

On Tue, Oct 15, 2019 at 6:31 PM Lothar Felten <lothar.felten@gmail.com> wrote:
>
> On Sun, 2019-10-13 at 11:57 -0300, Carlos Santos wrote:
> > I have a beaglebone black and can make some experiments this week.
>
> Hello Carlos,
>
> I have four patches, they allow you to build a new kernel, kernel
> module and userspace tools including libgbm for the beaglebone with the
> current toolchain.
> You can run the GLES(1) demos, I had no success with the GLES2 demos.
> The last patch is for qt5base, but it is incomplete, it won't compile
> the qpa backend yet.
> I'll post the four patches with RFC prefix to the list.
>
> Best regards,
> Lothar

Thanks. I will test them as time permits.

-- 
Carlos Santos <unixmania@gmail.com>

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Carlos Santos]

On Tue, Oct 15, 2019 at 10:07 PM Carlos Santos <unixmania@gmail.com> wrote:
>
> On Tue, Oct 15, 2019 at 6:31 PM Lothar Felten <lothar.felten@gmail.com> wrote:
> >
> > On Sun, 2019-10-13 at 11:57 -0300, Carlos Santos wrote:
> > > I have a beaglebone black and can make some experiments this week.
> >
> > Hello Carlos,
> >
> > I have four patches, they allow you to build a new kernel, kernel
> > module and userspace tools including libgbm for the beaglebone with the
> > current toolchain.
> > You can run the GLES(1) demos, I had no success with the GLES2 demos.
> > The last patch is for qt5base, but it is incomplete, it won't compile
> > the qpa backend yet.
> > I'll post the four patches with RFC prefix to the list.
> >
> > Best regards,
> > Lothar
>
> Thanks. I will test them as time permits.

Lothar,

I'm attempting to test your last patches on my Beaglebone Black but
had no success. I event attempted the plain beaglebone_defconfig to
simplify the build but it still hangs on boot:

U-Boot 2013.10-dirty (Nov 14 2013 - 12:59:11)

I2C:   ready
DRAM:  512 MiB
WARNING: Caches not enabled
NAND:  0 MiB
MMC:   OMAP SD/MMC: 0, OMAP SD/MMC: 1
*** Warning - readenv() failed, using default environment

Net:   <ethaddr> not set. Validating first E-fuse MAC
cpsw, usb_ether
Hit any key to stop autoboot:  0
mmc0 is current device
SD/MMC found on device 0
reading uEnv.txt
376 bytes read in 3 ms (122.1 KiB/s)
Loaded environment from uEnv.txt
Importing environment from mmc ...
Running uenvcmd ...
reading /zImage
6076272 bytes read in 568 ms (10.2 MiB/s)
reading /am335x-boneblack.dtb
36717 bytes read in 8 ms (4.4 MiB/s)
bootargs=console=ttyO0,115200n8 root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait
Kernel image @ 0x80200000 [ 0x000000 - 0x5cb770 ]
## Flattened Device Tree blob at 80f80000
   Booting using the fdt blob at 0x80f80000
   Using Device Tree in place at 80f80000, end 80f8bf6c

Starting kernel ...

[hangs here]

This is with the master branch as of commit
3db8622f311f46cea9dcc2f9dbed48ca543b2e24 plus the reversion of commit
1e55b66e975d234f5bff992785762e8f7ee5f5b9.

-- 
Carlos Santos <unixmania@gmail.com>

[Buildroot] beaglebone_qt5 does not build since commit a7914499b80

[Peter Korsgaard]

>>>>> "Carlos" == Carlos Santos <unixmania@gmail.com> writes:

Hi,

 > Lothar,

 > I'm attempting to test your last patches on my Beaglebone Black but
 > had no success. I event attempted the plain beaglebone_defconfig to
 > simplify the build but it still hangs on boot:

 > U-Boot 2013.10-dirty (Nov 14 2013 - 12:59:11)

 > I2C:   ready
 > DRAM:  512 MiB
 > WARNING: Caches not enabled
 > NAND:  0 MiB
 > MMC:   OMAP SD/MMC: 0, OMAP SD/MMC: 1
 > *** Warning - readenv() failed, using default environment

 > Net:   <ethaddr> not set. Validating first E-fuse MAC
 > cpsw, usb_ether
 > Hit any key to stop autoboot:  0
 > mmc0 is current device
 > SD/MMC found on device 0
 > reading uEnv.txt
 > 376 bytes read in 3 ms (122.1 KiB/s)
 > Loaded environment from uEnv.txt
 > Importing environment from mmc ...
 > Running uenvcmd ...
 > reading /zImage
 > 6076272 bytes read in 568 ms (10.2 MiB/s)
 > reading /am335x-boneblack.dtb
 > 36717 bytes read in 8 ms (4.4 MiB/s)
 > bootargs=console=ttyO0,115200n8 root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait
 > Kernel image @ 0x80200000 [ 0x000000 - 0x5cb770 ]
 > ## Flattened Device Tree blob at 80f80000
 >    Booting using the fdt blob at 0x80f80000
 >    Using Device Tree in place at 80f80000, end 80f8bf6c

 > Starting kernel ...

 > [hangs here]

Did you try console=ttyS0,115200n8? It has been a few years since I last
booted a beaglebone, but considering the recent change to the defconfig
it looks worth a try:

commit 68b5b79b2fb316fd7c6c40c12a7e9ddaeada837d
Author: Yegor Yefremov <yegorslists@googlemail.com>
Date:   Mon Sep 30 08:04:47 2019 +0200

    configs/beaglebone_defconfig: use default console device

    OMAP kernels use 8250 driver by default. Hence the name of
    the console device is not /dev/ttyO0 but /dev/ttyS0.
    Use /dev/console in order to handle the console independently
    of the selected driver.

    Tested in BeagleBone Black board.

    Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
    Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

-- 
Bye, Peter Korsgaard