linux-poplar fails in thud branch

[mirzak]

Hi,

Today I tested running a poplar board using the thud branch of the oe-rpb-manifest but I am getting a lot of errors during the linux-poplar build due to usage of gcc8,

example of error:

drivers/msp/pvr/drv_pvr_intf.c:255:56: error: argument to 'sizeof' in 'strncpy' call is the same expression as the source; did you mean to use the size of the destination? [-Werror=sizeof-pointer-memaccess]

I tried fixing a couple of them but new errors kept appearing.

Is this something someone else has experienced and know if there are any plans to update the linux-poplar fork. I suspect not as it has been pretty dead for 2 years.

I am trying to re-run the build with:

GCCVERSION = "7.%"

[fboudra]

@petegriffin any infos?

[mirzak]

By the way, downgrading the gcc version to 7 "worked".

[petegriffin]

@mirzak @fboudra I think as a temporary workaround we should force this recipe to build with gcc 7. I have also pinged a mail to the person in HiSilicon and Linaro who was working on this to see if there are any plans to update things.

[shawnguo2]

@mirzak Unfortunately, there is no plan to update linux-poplar, so you will have to live with GCC7.x or migrate to GCC8.x on your own.

[mirzak]

Thanks for the quick feedback.

That raises the question if the poplar support should be dropped in meta-96boards/thud and newer?

Or as suggested force gcc7 build of linux-poplar, I would create a PR but it is not clear to me if it is possible to specify gcc version per package in Yocto.

During my tests I set it globally.

[petegriffin]

For Chromium (which officially only supports clang, and frequently breaks with gcc), I use the following patch to force just that recipe to compiled with clang petegriffin/meta-browser@8748a31. Hopefully something similar can be done for gcc7

[mirzak]

Thanks, I will give it a go and report back.

[daniel-thompson]

On Wed, May 15, 2019 at 12:48:24AM -0700, Mirza Krak wrote: Thanks for the quick feedback. That raises the question if the `poplar` support should be dropped in meta-96boards/thud and newer?

AFAICT there's already a PR to fix this in the 96boards-poplar/linux project: https://github.com/96boards-poplar/linux/pull/3/commits It's fine to no longer develop the kernel but it is a shame we cannot find a way to park the poplar kernel in a "patches welcome" state.

…

-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #310 (comment)

[shawnguo2]

On Wed, May 15, 2019 at 12:48:24AM -0700, Mirza Krak wrote: Thanks for the quick feedback. That raises the question if the poplar support should be dropped in meta-96boards/thud and newer?
AFAICT there's already a PR to fix this in the 96boards-poplar/linux project: https://github.com/96boards-poplar/linux/pull/3/commits It's fine to no longer develop the kernel but it is a shame we cannot find a way to park the poplar kernel in a "patches welcome" state.

Agreed. Not sure how I missed the PR. I will handle it tomorrow. Thanks for pointing me to the PR, @daniel-thompson

[shawnguo2]

@mirzak I tested and merged 96boards-poplar/linux#3

[mirzak]

Thanks @shawnguo2, I have also tested it and it worked well for me as well.

I will create a PR to bump the linux-poplar recipe.

[mirzak]

#312 created. Though notice that there seems not have been a revision bump in a while so there are more changes then discussed here.

I have boot tested the change