@@ -573,12 +573,13 @@ NOTE: Za128rs is a new extension name capturing this feature. The
573
573
minimum reservation set size is effectively determined by the size of
574
574
atomic accesses in the A extension.
575
575
576
- - Misaligned loads and stores to main memory regions with both the
576
+ - *Zicclsm* Misaligned loads and stores to main memory regions with both the
577
577
cacheability and coherence PMAs must be supported.
578
578
579
- NOTE: Even when supported, misaligned loads and stores might execute
580
- extremely slowly. Standard software distributions should assume their
581
- existence only for correctness, not for performance.
579
+ NOTE: This introduces a new extension name for this feature. Even
580
+ though mandated, misaligned loads and stores might execute extremely
581
+ slowly. Standard software distributions should assume their existence
582
+ only for correctness, not for performance.
582
583
583
584
- Within main-memory regions, fetching a naturally aligned instruction of
584
585
up to four bytes must be atomic.
@@ -800,10 +801,11 @@ retroactively add to the 2020 and 2022 profiles as an exception.
800
801
801
802
- Main memory regions with both the cacheability and coherence PMAs must
802
803
support instruction fetch, AMOArithmetic, and RsrvEventual.
803
- - Misaligned loads and stores to main memory regions with both the
804
+
805
+ - *Zicclsm* Misaligned loads and stores to main memory regions with both the
804
806
cacheability and coherence PMAs must be supported.
805
807
806
- NOTE: Even when supported , misaligned loads and stores might execute
808
+ NOTE: Even though mandated , misaligned loads and stores might execute
807
809
extremely slowly. Standard software distributions should assume their
808
810
existence only for correctness, not for performance.
809
811
0 commit comments