On Sun, Jul 28, 2019 at 5:08 PM Heinrich Schuchardt xypron.glpk@gmx.de wrote:
On 7/28/19 10:56 AM, Bin Meng wrote:
On Sun, Jul 28, 2019 at 4:33 PM Heinrich Schuchardt xypron.glpk@gmx.de wrote:
On 7/28/19 9:41 AM, Bin Meng wrote:
+Stefan
Hi Heinrich,
On Sun, Jul 28, 2019 at 2:32 PM Heinrich Schuchardt xypron.glpk@gmx.de wrote:
Hello Tom,
origin/master commit bfe28bc70fc8b192878d6e8b7f7db344091643cf fails to build conga-qeval20-qa3-e3845 on Debian Buster:
I cannot reproduce it. Did you set CONFIG_BUILD_ROM = y?
My expectation is that if I call 'make conga-qeval20-qa3-e3845_defconfig' it sets all necessary configuration options to build successfully. This defconfig results in
# CONFIG_BUILD_ROM is not set
Sorry I was not clear. CONFIG_BUILD_ROM should not be set. When it is unset, it does not touch descriptor.bin. I thought you previously had it enabled hence binman tried to find descriptor.bin.
Setting CONFIG_BUILD_ROM=y does not solve the problem.
Did you delete all *.bin files before testing? 'make mrproper' does not delete *.bin from the build path. Shouldn't it remove generated *.bin?
*.bin in the board directory is kept there because it's not generated during U-Boot build.
I still cannot reproduce this. Not sure what's wrong.
In my .bashrc I had export BUILD_ROM=y. In earlier releases qemu-x86_defconfig needed this. If I unset this variable the build works on my local system. Thanks for the hint.
Good to hear.
This still leaves me with the question what is broken on Gitlab (Travis CI ran fine).
I tried to look at the Gitlab log but when I opened https://gitlab.denx.de/u-boot/custodians/u-boot-efi/-/jobs/2155, it gave me 404.
Regards, Bin