Use mmap(MAP_ANONYMOUS) to allocate queue buffers

Internal buffers for QPs, CQs, SRQs etc. are allocated with
mlx4_alloc_buf(), which rounds the buffer's size to the page size and
then allocates page aligned memory using posix_memalign().

However, this allocation is quite wasteful on architectures using 64K
pages (ia64 for example) because we then hit glibc's MMAP_THRESHOLD
malloc parameter and chunks are allocated using mmap.  Thus we end up
allocating:

  (requested size rounded to the page size) + (page size) + (malloc overhead)

rounded internally to the page size.

So for example, if we request a buffer of page_size bytes, we end up
consuming 3 pages.  In short, for each buffer we allocate, there is an
overhead of 2 pages.  This is quite visible on large clusters where
the number of QPs can reach several thousands.

This patch replaces the call to posix_memalign() in mlx4_alloc_buf()
with a direct call to mmap().

Signed-off-by: Sebastien Dugue <sebastien.dugue@bull.net>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

1 files changed, 7 insertions, 8 deletions

diff --git a/src/buf.c b/src/buf.c
index 0e5f9b6..bbaff12 100644
--- a/src/buf.c
+++ b/src/buf.c
@@ -61,16 +61,15 @@ int mlx4_alloc_buf(struct mlx4_buf *buf, size_t size, int page_size)
 {
 	int ret;
 
-	ret = posix_memalign(&buf->buf, page_size, align(size, page_size));
-	if (ret)
-		return ret;
+	buf->length = align(size, page_size);
+	buf->buf = mmap(NULL, buf->length, PROT_READ | PROT_WRITE,
+			MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+	if (buf->buf == MAP_FAILED)
+		return errno;
 
 	ret = ibv_dontfork_range(buf->buf, size);
 	if (ret)
-		free(buf->buf);
-
-	if (!ret)
-		buf->length = size;
+		munmap(buf->buf, buf->length);
 
 	return ret;
 }
@@ -78,5 +77,5 @@ int mlx4_alloc_buf(struct mlx4_buf *buf, size_t size, int page_size)
 void mlx4_free_buf(struct mlx4_buf *buf)
 {
 	ibv_dofork_range(buf->buf, buf->length);
-	free(buf->buf);
+	munmap(buf->buf, buf->length);
 }