tegra: correct cpu subtype

Tegra 2 processors have only 16 double-precision registers. The change
introduced by 8dcc1087602e ("toolchain: ARM: Fix toolchain compilation
for gcc 8.x") switched accidentally the toolchain for tegra target to cpu
type with 32 double-precision registers. This stems from gcc defaults
which assume "vfpv3-d32" if only "vfpv3" as mfpu is specified. That
change resulted in unusable image, in which kernel will kill userspace as
soon as it causing "Illegal instruction".

Ref: https://forum.openwrt.org/t/gcc-was-broken-on-mvebu-armada-370-device-after-commit-on-2019-03-25/43272
Fixes: 8dcc1087602e ("toolchain: ARM: Fix toolchain compilation for
gcc 8.x")
Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>

diff --git a/target/linux/tegra/Makefile b/target/linux/tegra/Makefile
index 0b48fc16baa24b636f51950461f897b0ea378e92..5dd4d439849e9f5b1523fcad9522e456d7fef0a2 100644 (file)
--- a/target/linux/tegra/Makefile
+++ b/target/linux/tegra/Makefile
@@ -11,7 +11,7 @@ BOARD := tegra
 BOARDNAME := NVIDIA Tegra
 FEATURES := audio boot-part display ext4 fpu gpio pci pcie rootfs-part rtc squashfs usb
 CPU_TYPE := cortex-a9
 BOARDNAME := NVIDIA Tegra
 FEATURES := audio boot-part display ext4 fpu gpio pci pcie rootfs-part rtc squashfs usb
 CPU_TYPE := cortex-a9

-CPU_SUBTYPE := vfpv3
+CPU_SUBTYPE := vfpv3-d16

 
 KERNEL_PATCHVER := 5.4
 KERNEL_TESTING_PATCHVER := 5.4
 
 KERNEL_PATCHVER := 5.4
 KERNEL_TESTING_PATCHVER := 5.4