[PATCH v2 0/5] hw: aspeed: Init all UART's with serial devices

Peter Delevoryas posted 5 patches 1 year, 10 months ago
hw/arm/aspeed_ast10x0.c     | 32 +++++++++++++++++++++++++++-----
hw/arm/aspeed_ast2600.c     | 27 ++++++++++++++++++++++-----
hw/arm/aspeed_soc.c         | 33 +++++++++++++++++++++++++++++----
include/hw/arm/aspeed_soc.h | 10 ++++++++++
4 files changed, 88 insertions(+), 14 deletions(-)
[PATCH v2 0/5] hw: aspeed: Init all UART's with serial devices
Posted by Peter Delevoryas 1 year, 10 months ago
v2:
- Rebased on Cedric's irq proposal. [1]
- Added "Introduce common UART init function" patch
- Added "Add uarts_num SoC attribute" patch
- Rewrote last commit's message for clarity

I tried testing this by running acceptance tests, particularly the
boot_linux_console.py file, but I had to disable the raspi2_initrd case.
It's not related to my changes (A/B tested and it fails on upstream/master
too), but thought I would mention that.

I also manually tested several machines:

AST2400: https://github.com/facebook/openbmc/releases/download/v2021.49.0/wedge100.mtd
AST2500: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fby3.mtd
AST2600: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fuji.mtd
AST1030: https://github.com/peterdelevoryas/OpenBIC/releases/download/oby35-cl-2022.13.01/Y35BCL.elf

[1] https://lore.kernel.org/qemu-devel/20220516055620.2380197-1-clg@kaod.org/

Peter Delevoryas (5):
  hw: aspeed: Add missing UART's
  hw: aspeed: Add uarts_num SoC attribute
  hw: aspeed: Ensure AST1030 respects uart-default
  hw: aspeed: Introduce common UART init function
  hw: aspeed: Init all UART's with serial devices

 hw/arm/aspeed_ast10x0.c     | 32 +++++++++++++++++++++++++++-----
 hw/arm/aspeed_ast2600.c     | 27 ++++++++++++++++++++++-----
 hw/arm/aspeed_soc.c         | 33 +++++++++++++++++++++++++++++----
 include/hw/arm/aspeed_soc.h | 10 ++++++++++
 4 files changed, 88 insertions(+), 14 deletions(-)

-- 
2.30.2
Re: [PATCH v2 0/5] hw: aspeed: Init all UART's with serial devices
Posted by Cédric Le Goater 1 year, 10 months ago
On 5/16/22 08:23, Peter Delevoryas wrote:
> v2:
> - Rebased on Cedric's irq proposal. [1]
> - Added "Introduce common UART init function" patch
> - Added "Add uarts_num SoC attribute" patch
> - Rewrote last commit's message for clarity

Looks good to me.

> I tried testing this by running acceptance tests, particularly the
> boot_linux_console.py file, but I had to disable the raspi2_initrd case.
> It's not related to my changes (A/B tested and it fails on upstream/master
> too), but thought I would mention that.>
> I also manually tested several machines:
> 
> AST2400: https://github.com/facebook/openbmc/releases/download/v2021.49.0/wedge100.mtd
> AST2500: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fby3.mtd
> AST2600: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fuji.mtd
> AST1030: https://github.com/peterdelevoryas/OpenBIC/releases/download/oby35-cl-2022.13.01/Y35BCL.elf
> 
> [1] https://lore.kernel.org/qemu-devel/20220516055620.2380197-1-clg@kaod.org/
I have quite a few images which I run manually. OpenBMC is providing
images, Aspeed also. Joel did a small tool for the IBM rainier :

   https://github.com/shenki/qemu-boot-test

Having an automated framework for Aspeed machines pulling images
from different places would be nice but we cannot put all under
QEMU.


I like the buildroot CI using the QEMU boards. See a pipeline example:

   https://gitlab.com/legoater/buildroot/-/pipelines/539559209

but that's buildroot only and integrating a custom uboot might be
difficult.

For my ppc needs, I did a quick-and-dirty non-regression test suite :

   https://github.com/legoater/qemu-ppc-boot


Anyway, I haven't found a solution nor spent much time on it but
I am interested !


C.




> 
> Peter Delevoryas (5):
>    hw: aspeed: Add missing UART's
>    hw: aspeed: Add uarts_num SoC attribute
>    hw: aspeed: Ensure AST1030 respects uart-default
>    hw: aspeed: Introduce common UART init function
>    hw: aspeed: Init all UART's with serial devices
> 
>   hw/arm/aspeed_ast10x0.c     | 32 +++++++++++++++++++++++++++-----
>   hw/arm/aspeed_ast2600.c     | 27 ++++++++++++++++++++++-----
>   hw/arm/aspeed_soc.c         | 33 +++++++++++++++++++++++++++++----
>   include/hw/arm/aspeed_soc.h | 10 ++++++++++
>   4 files changed, 88 insertions(+), 14 deletions(-)
>
Re: [PATCH v2 0/5] hw: aspeed: Init all UART's with serial devices
Posted by Peter Delevoryas 1 year, 10 months ago

> On May 16, 2022, at 12:18 AM, Cédric Le Goater <clg@kaod.org> wrote:
> 
> On 5/16/22 08:23, Peter Delevoryas wrote:
>> v2:
>> - Rebased on Cedric's irq proposal. [1]
>> - Added "Introduce common UART init function" patch
>> - Added "Add uarts_num SoC attribute" patch
>> - Rewrote last commit's message for clarity
> 
> Looks good to me.
> 
>> I tried testing this by running acceptance tests, particularly the
>> boot_linux_console.py file, but I had to disable the raspi2_initrd case.
>> It's not related to my changes (A/B tested and it fails on upstream/master
>> too), but thought I would mention that.>
>> I also manually tested several machines:
>> AST2400: https://github.com/facebook/openbmc/releases/download/v2021.49.0/wedge100.mtd
>> AST2500: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fby3.mtd
>> AST2600: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fuji.mtd
>> AST1030: https://github.com/peterdelevoryas/OpenBIC/releases/download/oby35-cl-2022.13.01/Y35BCL.elf
>> [1] https://lore.kernel.org/qemu-devel/20220516055620.2380197-1-clg@kaod.org/
> I have quite a few images which I run manually. OpenBMC is providing
> images, Aspeed also. Joel did a small tool for the IBM rainier :
> 
>  https://github.com/shenki/qemu-boot-test
> 
> Having an automated framework for Aspeed machines pulling images
> from different places would be nice but we cannot put all under
> QEMU.
> 
> 
> I like the buildroot CI using the QEMU boards. See a pipeline example:
> 
>  https://gitlab.com/legoater/buildroot/-/pipelines/539559209 
> but that's buildroot only and integrating a custom uboot might be
> difficult.
> 
> For my ppc needs, I did a quick-and-dirty non-regression test suite :
> 
>  https://github.com/legoater/qemu-ppc-boot
> 
> 
> Anyway, I haven't found a solution nor spent much time on it but
> I am interested !

Oh thanks for the links, I’m somewhat interested in looking into this further.

> 
> 
> C.
> 
> 
> 
> 
>> Peter Delevoryas (5):
>>   hw: aspeed: Add missing UART's
>>   hw: aspeed: Add uarts_num SoC attribute
>>   hw: aspeed: Ensure AST1030 respects uart-default
>>   hw: aspeed: Introduce common UART init function
>>   hw: aspeed: Init all UART's with serial devices
>>  hw/arm/aspeed_ast10x0.c     | 32 +++++++++++++++++++++++++++-----
>>  hw/arm/aspeed_ast2600.c     | 27 ++++++++++++++++++++++-----
>>  hw/arm/aspeed_soc.c         | 33 +++++++++++++++++++++++++++++----
>>  include/hw/arm/aspeed_soc.h | 10 ++++++++++
>>  4 files changed, 88 insertions(+), 14 deletions(-)
> 

Re: [PATCH v2 0/5] hw: aspeed: Init all UART's with serial devices
Posted by Andrew Jeffery 1 year, 10 months ago

On Mon, 16 May 2022, at 16:48, Cédric Le Goater wrote:
> On 5/16/22 08:23, Peter Delevoryas wrote:
>> v2:
>> - Rebased on Cedric's irq proposal. [1]
>> - Added "Introduce common UART init function" patch
>> - Added "Add uarts_num SoC attribute" patch
>> - Rewrote last commit's message for clarity
>
> Looks good to me.
>
>> I tried testing this by running acceptance tests, particularly the
>> boot_linux_console.py file, but I had to disable the raspi2_initrd case.
>> It's not related to my changes (A/B tested and it fails on upstream/master
>> too), but thought I would mention that.>
>> I also manually tested several machines:
>> 
>> AST2400: https://github.com/facebook/openbmc/releases/download/v2021.49.0/wedge100.mtd
>> AST2500: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fby3.mtd
>> AST2600: https://github.com/facebook/openbmc/releases/download/v2021.49.0/fuji.mtd
>> AST1030: https://github.com/peterdelevoryas/OpenBIC/releases/download/oby35-cl-2022.13.01/Y35BCL.elf
>> 
>> [1] https://lore.kernel.org/qemu-devel/20220516055620.2380197-1-clg@kaod.org/
> I have quite a few images which I run manually. OpenBMC is providing
> images, Aspeed also. Joel did a small tool for the IBM rainier :
>
>    https://github.com/shenki/qemu-boot-test
>
> Having an automated framework for Aspeed machines pulling images
> from different places would be nice but we cannot put all under
> QEMU.

For what it's worth I run this as a smoke test before pushing updates to openbmc/qemu:

https://github.com/openbmc/openbmc-build-scripts/blob/master/scripts/test-qemu

Andrew