arch/arm64: Enable FPU for qemu-armv8a:netnsh_smp and netnsh_smp_hv
When reading https://github.com/apache/nuttx-apps/pull/1793, I found a previous problem https://github.com/apache/nuttx-apps/pull/1454#discussion_r1045718067 now still exists on current master (inf in iperf bandwidth). By step-by-step debugging, I found a value of 1000000.0 stored in register d8 is overwritten as 0 after context switch, then fdiv results in inf because of divided by 0. Finally I found that ARCH_FPU is not enabled while compiler is using FPU, and may lose values in floating-point registers. Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
This commit is contained in:
parent
ad6361f0cc
commit
3d5eae2497
2 changed files with 0 additions and 2 deletions
|
@ -5,7 +5,6 @@
|
|||
# You can then do "make savedefconfig" to generate a new defconfig file that includes your
|
||||
# modifications.
|
||||
#
|
||||
# CONFIG_ARCH_FPU is not set
|
||||
CONFIG_ALLOW_BSD_COMPONENTS=y
|
||||
CONFIG_ARCH="arm64"
|
||||
CONFIG_ARCH_ARM64=y
|
||||
|
|
|
@ -5,7 +5,6 @@
|
|||
# You can then do "make savedefconfig" to generate a new defconfig file that includes your
|
||||
# modifications.
|
||||
#
|
||||
# CONFIG_ARCH_FPU is not set
|
||||
CONFIG_ALLOW_BSD_COMPONENTS=y
|
||||
CONFIG_ARCH="arm64"
|
||||
CONFIG_ARCH_ARM64=y
|
||||
|
|
Loading…
Reference in a new issue