This patchset is prompted by a Coverity issue CID 1659590, where it thinks that we use tainted data from the guest to index an array in bh_utf8_decode() without bounds-checking. In fact I think this is a false positive because the array is 256 bytes and the data from the guest is always a byte, but both Coverity and I get confused because the chain of functions operating on this data switch between "uint8_t", "int" and "uint32_t" types to hold it as it passes down.
The patchset standardizes this all to uint8_t. Peter Maydell (2): ui/vt100: Standardize on uint8_t for "ch" byte variables ui/vt100: Take byte as uint8_t in bh_utf8_decode() ui/vt100.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) -- 2.43.0
