buffers: bring back cache occupancy improvement

The improvement was removed in 40129,
causing 5-40% regressions in AVF tests.

There is a memory-speed trade-off,
this change prefers speed over memory efficiency.

Ideally, the choice should be configurable,
but that is not easy to achieve, considering
how early is vlib_buffer_main_init called.

Type: fix
Fixes: 038dad7ef29b0b724071edb5f8cc7a9845584454

Change-Id: I4746f3634abe6d233c9d092a372de05b3d1ae4b6
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
(cherry picked from commit 04fd51c03c428859bae949a8294ee0f9c062a44b)
This commit is contained in:
Vratko Polak
2024-02-06 12:45:59 +01:00
parent 8cbf84dce0
commit ccfc24f745

View File

@ -473,6 +473,10 @@ vlib_buffer_alloc_size (uword ext_hdr_size, uword data_size)
uword alloc_size = ext_hdr_size + sizeof (vlib_buffer_t) + data_size; uword alloc_size = ext_hdr_size + sizeof (vlib_buffer_t) + data_size;
alloc_size = round_pow2 (alloc_size, VLIB_BUFFER_ALIGN); alloc_size = round_pow2 (alloc_size, VLIB_BUFFER_ALIGN);
/* in case when we have even number of 'cachelines', we add one more for
* better cache occupancy */
alloc_size |= VLIB_BUFFER_ALIGN;
return alloc_size; return alloc_size;
} }