diff options
author | Andrew Jones <ajones@ventanamicro.com> | 2022-07-18 19:20:28 +0200 |
---|---|---|
committer | Anup Patel <anup@brainfault.org> | 2022-07-30 11:39:42 +0530 |
commit | f27203525aa2fc20d0074feb892a6f8433f84e3a (patch) | |
tree | db3f886de35c8bd270564e97278ae9ab08833d95 /scripts/Kconfiglib/kconfiglib.py | |
parent | 7198e1d06f486173e6565ff6c718f30bf8c2ff30 (diff) | |
download | opensbi-f27203525aa2fc20d0074feb892a6f8433f84e3a.zip opensbi-f27203525aa2fc20d0074feb892a6f8433f84e3a.tar.gz opensbi-f27203525aa2fc20d0074feb892a6f8433f84e3a.tar.bz2 |
lib: utils/serial: Ensure baudrate is non-zero before using
RISC-V doesn't generate exceptions on divide-by-zero, but the result,
all bits set, is not likely what people expect either. In all cases
where we divide by baudrate there's a chance it's zero (when the DT
it came from is "bad"). To avoid difficult to debug situations, leave
baudrate dependent registers alone when baudrate is zero, as, also in
all cases, it appears we can skip initialization of those registers
and still [hopefully] have a functioning UART.
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Diffstat (limited to 'scripts/Kconfiglib/kconfiglib.py')
0 files changed, 0 insertions, 0 deletions