[U-Boot] U-Boot, efi-x86_app and gcc-8.1?
Hey guys,
Looking at https://travis-ci.org/trini/u-boot/jobs/534037240 this is the only board that doesn't build (having switched ARM to gcc-8.3 from Arm/Linaro) and the error is pretty odd. Do you have any idea what's going on? Thanks!
Hi Tom,
On Sat, 18 May 2019 at 06:13, Tom Rini trini@konsulko.com wrote:
Hey guys,
Looking at https://travis-ci.org/trini/u-boot/jobs/534037240 this is the only board that doesn't build (having switched ARM to gcc-8.3 from Arm/Linaro) and the error is pretty odd. Do you have any idea what's going on? Thanks!
I wonder if the assembler output of image.c would give any clues? I'm assuming it isn't a toolchain bug.
Regards, Simon
Hi Tom,
On Sun, May 19, 2019 at 12:07 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 18 May 2019 at 06:13, Tom Rini trini@konsulko.com wrote:
Hey guys,
Looking at https://travis-ci.org/trini/u-boot/jobs/534037240 this is the only board that doesn't build (having switched ARM to gcc-8.3 from Arm/Linaro) and the error is pretty odd. Do you have any idea what's going on? Thanks!
I wonder if the assembler output of image.c would give any clues? I'm assuming it isn't a toolchain bug.
I've analysed this before, and it is a gcc regression. See my bug report @ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87793
The bug is fixed in gcc 8.3 and 9.0.
Regards, Bin
On Sun, May 19, 2019 at 08:55:53AM +0800, Bin Meng wrote:
Hi Tom,
On Sun, May 19, 2019 at 12:07 AM Simon Glass sjg@chromium.org wrote:
Hi Tom,
On Sat, 18 May 2019 at 06:13, Tom Rini trini@konsulko.com wrote:
Hey guys,
Looking at https://travis-ci.org/trini/u-boot/jobs/534037240 this is the only board that doesn't build (having switched ARM to gcc-8.3 from Arm/Linaro) and the error is pretty odd. Do you have any idea what's going on? Thanks!
I wonder if the assembler output of image.c would give any clues? I'm assuming it isn't a toolchain bug.
I've analysed this before, and it is a gcc regression. See my bug report @ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87793
The bug is fixed in gcc 8.3 and 9.0.
Ugh. And gcc-8.2 from Bootlin has the problem too (along with being not relocatable). Thanks!
participants (3)
-
Bin Meng -
Simon Glass -
Tom Rini