Commit f2e42d9
authored
[mlir][IntRangeInference] Handle ceildivsi(INT_MIN, x > 1) as expected (llvm#116284)
Fixes llvm#115293
While the definition of ceildivsi is integer division, rounding up, most
implementations will use `-(-a / b)` for dividing `a ceildiv b` with `a`
negative and `b` positive.
Mathematically, and for most integers, these two definitions are
equivalent. However, with `a == INT_MIN`, the initial negation is a
noop, which means that, while divinding and rounding up would give a
negative result, `-((- INT_MIN) / b)` is `-(INT_MIN / b)`, which is
positive.
This commit adds a special case to ceilDivSI inference to handle this
case and bring it in line with the operational instead of the
mathematical semantics of ceiling division.1 parent d82422f commit f2e42d9
File tree
2 files changed
+22
-0
lines changed- mlir
- lib/Interfaces/Utils
- test/Dialect/Arith
2 files changed
+22
-0
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
375 | 375 | | |
376 | 376 | | |
377 | 377 | | |
| 378 | + | |
| 379 | + | |
| 380 | + | |
| 381 | + | |
| 382 | + | |
| 383 | + | |
| 384 | + | |
| 385 | + | |
| 386 | + | |
378 | 387 | | |
379 | 388 | | |
380 | 389 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
249 | 249 | | |
250 | 250 | | |
251 | 251 | | |
| 252 | + | |
| 253 | + | |
| 254 | + | |
| 255 | + | |
| 256 | + | |
| 257 | + | |
| 258 | + | |
| 259 | + | |
| 260 | + | |
| 261 | + | |
| 262 | + | |
| 263 | + | |
| 264 | + | |
252 | 265 | | |
253 | 266 | | |
254 | 267 | | |
| |||
0 commit comments