pan/bi: Fix M1/M2 decoding in disassembler

C's definition of the % operator has a footgun around sign conversion.
Avoid it and just use bitwise arithemtic instead like the hardware
would, fixing the disassembly and making buggy assembly more obvious.

Fixes: 08a9e5e3e8 ("pan/bi: Decode M values in disasm")
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8723>
This commit is contained in:
Alyssa Rosenzweig 2021-01-08 21:38:09 -05:00 committed by Marge Bot
parent 6f28a4449a
commit a69c73988b
1 changed files with 9 additions and 7 deletions

View File

@ -611,15 +611,17 @@ static bool dump_clause(FILE *fp, uint32_t *words, unsigned *size, unsigned offs
consts.raw[const_idx + 1] = const1;
/* Calculate M values from A, B and 4-bit
* unsigned arithmetic */
* unsigned arithmetic. Mathematically it
* should be (A - B) % 16 but we use this
* alternate form to avoid sign issues */
signed A1 = bits(words[2], 0, 4);
signed B1 = bits(words[3], 28, 32);
signed A2 = bits(words[1], 0, 4);
signed B2 = bits(words[2], 28, 32);
unsigned A1 = bits(words[2], 0, 4);
unsigned B1 = bits(words[3], 28, 32);
unsigned A2 = bits(words[1], 0, 4);
unsigned B2 = bits(words[2], 28, 32);
unsigned M1 = (A1 - B1) % 16;
unsigned M2 = (A2 - B2) % 16;
unsigned M1 = (16 + A1 - B1) & 0xF;
unsigned M2 = (16 + A2 - B2) & 0xF;
decode_M(&consts.mods[const_idx], M1, M2, false);