| Message ID | 20191208173026.25745-1-netanel@amazon.com |
|---|---|
| State | Not Applicable |
| Delegated to: | David Miller |
| Headers | show |
| Series | [V1,net] net: ena: fix napi handler misbehavior when the napi budget is zero | expand |
From: Netanel Belgazal <netanel@amazon.com> Date: Sun, 8 Dec 2019 17:30:26 +0000 > In netpoll the napi handler could be called with budget equal to zero. > Current ENA napi handler doesn't take that into consideration. > > The napi handler handles Rx packets in a do-while loop. > Currently, the budget check happens only after decrementing the > budget, therefore the napi handler, in rare cases, could run over > MAX_INT packets. > > In addition to that, this moves all budget related variables to int > calculation and stop mixing u32 to avoid ambiguity > > Signed-off-by: Netanel Belgazal <netanel@amazon.com> Bug fixes need to have an appropriate Fixes: tag. Please repost with that added, thank you.
I just did it. Regards, Netanel On 12/9/19, 8:35 PM, "David Miller" <davem@davemloft.net> wrote: From: Netanel Belgazal <netanel@amazon.com> Date: Sun, 8 Dec 2019 17:30:26 +0000 > In netpoll the napi handler could be called with budget equal to zero. > Current ENA napi handler doesn't take that into consideration. > > The napi handler handles Rx packets in a do-while loop. > Currently, the budget check happens only after decrementing the > budget, therefore the napi handler, in rare cases, could run over > MAX_INT packets. > > In addition to that, this moves all budget related variables to int > calculation and stop mixing u32 to avoid ambiguity > > Signed-off-by: Netanel Belgazal <netanel@amazon.com> Bug fixes need to have an appropriate Fixes: tag. Please repost with that added, thank you.
diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c index c487d2a7d6dd..b4a145220aba 100644 --- a/drivers/net/ethernet/amazon/ena/ena_netdev.c +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c @@ -1238,8 +1238,8 @@ static int ena_io_poll(struct napi_struct *napi, int budget) struct ena_napi *ena_napi = container_of(napi, struct ena_napi, napi); struct ena_ring *tx_ring, *rx_ring; - u32 tx_work_done; - u32 rx_work_done; + int tx_work_done; + int rx_work_done = 0; int tx_budget; int napi_comp_call = 0; int ret; @@ -1256,7 +1256,11 @@ static int ena_io_poll(struct napi_struct *napi, int budget) } tx_work_done = ena_clean_tx_irq(tx_ring, tx_budget); - rx_work_done = ena_clean_rx_irq(rx_ring, napi, budget); + /* On netpoll the budget is zero and the handler should only clean the + * tx completions. + */ + if (likely(budget)) + rx_work_done = ena_clean_rx_irq(rx_ring, napi, budget); /* If the device is about to reset or down, avoid unmask * the interrupt and return 0 so NAPI won't reschedule
In netpoll the napi handler could be called with budget equal to zero. Current ENA napi handler doesn't take that into consideration. The napi handler handles Rx packets in a do-while loop. Currently, the budget check happens only after decrementing the budget, therefore the napi handler, in rare cases, could run over MAX_INT packets. In addition to that, this moves all budget related variables to int calculation and stop mixing u32 to avoid ambiguity Signed-off-by: Netanel Belgazal <netanel@amazon.com> --- drivers/net/ethernet/amazon/ena/ena_netdev.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)