v1-0003-WIP-Attempt-alignment-preamble-better-suited-to-W.patch
application/x-patch
Filename: v1-0003-WIP-Attempt-alignment-preamble-better-suited-to-W.patch
Type: application/x-patch
Part: 2
Message:
vectorized CRC on ARM64
Patch
Same data as JSON:
GET /api/v1/attachments/:id/patch
the parsed metadata as JSON — format, series position, per-file stats; never the diff bytes.
API reference →
Format: format-patch
Series: patch v1-0003
Subject: WIP: Attempt alignment preamble better suited to WAL
| File | + | − |
|---|---|---|
| src/port/pg_crc32c_armv8.c | 19 | 0 |
From a0aad5bf044170e6a8ceaf368fcfff68177cb9fb Mon Sep 17 00:00:00 2001
From: John Naylor <john.naylor@postgresql.org>
Date: Wed, 14 May 2025 01:59:41 +0700
Subject: [PATCH v1 3/3] WIP: Attempt alignment preamble better suited to WAL
---
src/port/pg_crc32c_armv8.c | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)
diff --git a/src/port/pg_crc32c_armv8.c b/src/port/pg_crc32c_armv8.c
index f67de2016b4..da6116efa21 100644
--- a/src/port/pg_crc32c_armv8.c
+++ b/src/port/pg_crc32c_armv8.c
@@ -119,6 +119,24 @@ pg_comp_crc32c_pmull(pg_crc32c crc, const void *data, size_t len)
const char *buf = data;
/* Align to 16 bytes to prevent straddling cacheline boundaries. */
+#if 1
+
+ /*
+ * WIP: WAL is 4-byte aligned, so in that case the the first loop will be skipped. Is this
+ * better?
+ */
+ for (; len && ((uintptr_t) buf & 3); --len)
+ {
+ crc0 = __crc32cb(crc0, *buf++);
+ }
+ if (((uintptr_t) buf & 12) && len >= 4)
+ {
+ crc0 = __crc32cw(crc0, *(const uint64_t *) buf);
+ buf += 4;
+ len -= 4;
+ }
+#else
+ /* original */
for (; len && ((uintptr_t) buf & 7); --len)
{
crc0 = __crc32cb(crc0, *buf++);
@@ -129,6 +147,7 @@ pg_comp_crc32c_pmull(pg_crc32c crc, const void *data, size_t len)
buf += 8;
len -= 8;
}
+#endif
if (len >= 64)
{
--
2.49.0