Thread (26 messages) 26 messages, 2 authors, 2023-08-24
STALE1067d REVIEWED: 1 (1M)
Revisions (25)
  1. v6 [diff vs current]
  2. v7 [diff vs current]
  3. v8 [diff vs current]
  4. v9 [diff vs current]
  5. v10 [diff vs current]
  6. v11 [diff vs current]
  7. v12 [diff vs current]
  8. v13 current
  9. v14 [diff vs current]
  10. v15 [diff vs current]
  11. v16 [diff vs current]
  12. v17 [diff vs current]
  13. v18 [diff vs current]
  14. v19 [diff vs current]
  15. v20 [diff vs current]
  16. v21 [diff vs current]
  17. v22 [diff vs current]
  18. v23 [diff vs current]
  19. v24 [diff vs current]
  20. v25 [diff vs current]
  21. v26 [diff vs current]
  22. v27 [diff vs current]
  23. v28 [diff vs current]
  24. v29 [diff vs current]
  25. v30 [diff vs current]

[PATCH v13 05/24] iov_iter: skip copy if src == dst for direct data placement

From: Aurelien Aptel <hidden>
Date: 2023-08-22 15:05:26
Also in: linux-nvme
Subsystem: library code, the rest, userspace copyin/copyout (uiovec) · Maintainers: Andrew Morton, Linus Torvalds, Alexander Viro

From: Ben Ben-Ishay <redacted>

When using direct data placement (DDP) the NIC could write the payload
directly into the destination buffer and constructs SKBs such that
they point to this data. To skip copies when SKB data already resides
in the destination buffer we check if (src == dst), and skip the copy
when it's true.

Signed-off-by: Ben Ben-Ishay <redacted>
Signed-off-by: Boris Pismenny <borisp@nvidia.com>
Signed-off-by: Or Gerlitz <redacted>
Signed-off-by: Yoray Zack <redacted>
Signed-off-by: Shai Malin <redacted>
Signed-off-by: Aurelien Aptel <redacted>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 lib/iov_iter.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index e4dc809d1075..f01ee0e28295 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -313,9 +313,15 @@ size_t _copy_to_iter(const void *addr, size_t bytes, struct iov_iter *i)
 		return 0;
 	if (user_backed_iter(i))
 		might_fault();
+	/*
+	 * When using direct data placement (DDP) the hardware writes
+	 * data directly to the destination buffer, and constructs
+	 * IOVs such that they point to this data.
+	 * Thus, when the src == dst we skip the memcpy.
+	 */
 	iterate_and_advance(i, bytes, base, len, off,
 		copyout(base, addr + off, len),
-		memcpy(base, addr + off, len)
+		(base != addr + off) && memcpy(base, addr + off, len)
 	)
 
 	return bytes;
-- 
2.34.1
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help