This PR fixes a SIGFPE crash on x86_64 when evaluating `INT_MIN / -1` or `INT_MIN % -1` for signed integer types. On x86_64, the `idiv` instruction traps when the quotient overflows the destination register. For signed integers, `INT_MIN / -1` produces a result that overflows (e.g., `-2147483648 / -1 = 2147483648` which doesn't fit in Int32). ARM64's `sdiv` instruction wraps instead of trapping. The fix: - For Int8/Int16/Int32: widen to the next larger type before dividing/modding, then truncate back - For Int64: explicitly check for the overflow case and return the wrapped result Fixes #11612 🤖 Prepared with Claude Code
36 lines
1.2 KiB
Text
36 lines
1.2 KiB
Text
/-!
|
|
# Signed integer division overflow test
|
|
|
|
On x86_64, INT_MIN / -1 and INT_MIN % -1 cause a SIGFPE (floating point exception)
|
|
because the result overflows the destination register. This test verifies the fix.
|
|
|
|
The expected behavior (per BitVec.sdiv semantics): return minValue (wrapped result) for div,
|
|
and 0 for mod.
|
|
-/
|
|
|
|
-- Test that signed division handles overflow correctly
|
|
-- INT_MIN / -1 should return INT_MIN (wrapped)
|
|
|
|
#guard (Int8.minValue / -1 : Int8) == Int8.minValue
|
|
#guard (Int16.minValue / -1 : Int16) == Int16.minValue
|
|
#guard (Int32.minValue / -1 : Int32) == Int32.minValue
|
|
#guard (Int64.minValue / -1 : Int64) == Int64.minValue
|
|
|
|
-- Test that signed mod handles overflow correctly
|
|
-- INT_MIN % -1 should return 0
|
|
|
|
#guard (Int8.minValue % -1 : Int8) == 0
|
|
#guard (Int16.minValue % -1 : Int16) == 0
|
|
#guard (Int32.minValue % -1 : Int32) == 0
|
|
#guard (Int64.minValue % -1 : Int64) == 0
|
|
|
|
-- Also test via #eval to exercise the C runtime
|
|
#eval (Int8.minValue / -1 : Int8)
|
|
#eval (Int16.minValue / -1 : Int16)
|
|
#eval (Int32.minValue / -1 : Int32)
|
|
#eval (Int64.minValue / -1 : Int64)
|
|
|
|
#eval (Int8.minValue % -1 : Int8)
|
|
#eval (Int16.minValue % -1 : Int16)
|
|
#eval (Int32.minValue % -1 : Int32)
|
|
#eval (Int64.minValue % -1 : Int64)
|