[PATCH net 1/2] bnxt_en: Need memory barrier when processing the completion ring.
From: Michael Chan <michael.chan@broadcom.com>
Date: 2016-05-04 20:56:45
Subsystem:
broadcom bnxt_en 50 gigabit ethernet driver, networking drivers, the rest · Maintainers:
Michael Chan, Pavan Chebbi, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
The code determines if the next ring entry is valid before proceeding further to read the rest of the entry. The CPU can re-order and read the rest of the entry first, possibly reading a stale entry, if DMA of a new entry happens right after reading it. This issue can be readily seen on a ppc64 system, causing it to crash. Signed-off-by: Michael Chan <michael.chan@broadcom.com> --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 72eb29e..f33ff20 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c@@ -1388,6 +1388,10 @@ static int bnxt_poll_work(struct bnxt *bp, struct bnxt_napi *bnapi, int budget) if (!TX_CMP_VALID(txcmp, raw_cons)) break; + /* The valid test of the entry must be done first before + * reading any further. + */ + rmb(); if (TX_CMP_TYPE(txcmp) == CMP_TYPE_TX_L2_CMP) { tx_pkts++; /* return full budget so NAPI will complete. */
--
1.8.3.1