The TX path usually maps the full content of a page into a buffer
element. But there's specific skb layouts (ie. linearized TSO skbs)
where the HW header (1) requires a separate buffer element, and (2) is
page-contiguous with the packet data that's mapped into the next buffer
element.
Flag such buffer elements accordingly, so that HW can optimize its data
access for them.
Signed-off-by: Julian Wiedmann <j...@linux.ibm.com>
---
drivers/s390/net/qeth_core_main.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/s390/net/qeth_core_main.c
b/drivers/s390/net/qeth_core_main.c
index 9c9a6edb5384..4d1d053eebb7 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -4110,8 +4110,16 @@ static unsigned int qeth_fill_buffer(struct
qeth_qdio_out_buffer *buf,
buffer->element[element].addr = virt_to_phys(hdr);
buffer->element[element].length = hd_len;
buffer->element[element].eflags = SBAL_EFLAGS_FIRST_FRAG;
- /* remember to free cache-allocated HW header: */
- buf->is_header[element] = ((void *)hdr != skb->data);
+
+ /* HW header is allocated from cache: */
+ if ((void *)hdr != skb->data)
+ buf->is_header[element] = 1;
+ /* HW header was pushed and is contiguous with linear part: */
+ else if (length > 0 && !PAGE_ALIGNED(data) &&
+ (data == (char *)hdr + hd_len))
+ buffer->element[element].eflags |=
+ SBAL_EFLAGS_CONTIGUOUS;
+
element++;
}
--
2.17.1
