Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 26d92f92 authored by Sathya Perla's avatar Sathya Perla Committed by David S. Miller
Browse files

be2net: fix bug in rx page posting



Pages are posted to the rxq in such a way that more than one frag
can share the page. The last frag that uses the page unmaps the
page.  In the case when a page is not fully used (due to lack of space in rxq)
the last frag that uses the page is not being set as a "last_page_user";
instead, the next frag in the rxq is incorrectly being set.

The fix has also been tested on ppc64 with 64k pages...

Signed-off-by: default avatarSathya Perla <sathyap@serverengines.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent b94b5028
Loading
Loading
Loading
Loading
+5 −3
Original line number Diff line number Diff line
@@ -910,7 +910,7 @@ static inline struct page *be_alloc_pages(u32 size)
static void be_post_rx_frags(struct be_adapter *adapter)
{
	struct be_rx_page_info *page_info_tbl = adapter->rx_obj.page_info_tbl;
	struct be_rx_page_info *page_info = NULL;
	struct be_rx_page_info *page_info = NULL, *prev_page_info = NULL;
	struct be_queue_info *rxq = &adapter->rx_obj.q;
	struct page *pagep = NULL;
	struct be_eth_rx_d *rxd;
@@ -941,7 +941,6 @@ static void be_post_rx_frags(struct be_adapter *adapter)
		rxd = queue_head_node(rxq);
		rxd->fragpa_lo = cpu_to_le32(frag_dmaaddr & 0xFFFFFFFF);
		rxd->fragpa_hi = cpu_to_le32(upper_32_bits(frag_dmaaddr));
		queue_head_inc(rxq);

		/* Any space left in the current big page for another frag? */
		if ((page_offset + rx_frag_size + rx_frag_size) >
@@ -949,10 +948,13 @@ static void be_post_rx_frags(struct be_adapter *adapter)
			pagep = NULL;
			page_info->last_page_user = true;
		}

		prev_page_info = page_info;
		queue_head_inc(rxq);
		page_info = &page_info_tbl[rxq->head];
	}
	if (pagep)
		page_info->last_page_user = true;
		prev_page_info->last_page_user = true;

	if (posted) {
		atomic_add(posted, &rxq->used);