2018-04-10 13:22:38

by Jia-Ju Bai

Subject: [PATCH 2/2] ntb: ntb_transport: Replace GFP_ATOMIC with GFP_KERNEL in ntb_transport_create_queue

ntb_transport_create_queue() is never called in atomic context.

ntb_transport_create_queue() is only called by ntb_netdev_probe(),
which is set as ".probe" in struct ntb_transport_client.

Despite never getting called from atomic context,
ntb_transport_create_queue() calls kzalloc_node() with GFP_ATOMIC,
which does not sleep for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL,
which can sleep and improve the possibility of sucessful allocation.

This is found by a static analysis tool named DCNS written by myself.
And I also manually check it

Signed-off-by: Jia-Ju Bai <[email protected]>
---
drivers/ntb/ntb_transport.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/ntb/ntb_transport.c b/drivers/ntb/ntb_transport.c
index f58d8e3..2c0c8bc 100644
--- a/drivers/ntb/ntb_transport.c
+++ b/drivers/ntb/ntb_transport.c
@@ -1825,7 +1825,7 @@ struct ntb_transport_qp *
qp->rx_dma_chan ? "DMA" : "CPU");

for (i = 0; i < NTB_QP_DEF_NUM_ENTRIES; i++) {
- entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
+ entry = kzalloc_node(sizeof(*entry), GFP_KERNEL, node);
if (!entry)
goto err1;

@@ -1836,7 +1836,7 @@ struct ntb_transport_qp *
qp->rx_alloc_entry = NTB_QP_DEF_NUM_ENTRIES;

for (i = 0; i < qp->tx_max_entry; i++) {
- entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
+ entry = kzalloc_node(sizeof(*entry), GFP_KERNEL, node);
if (!entry)
goto err2;

--
1.9.1

2018-05-22 00:35:31

by Jon Mason

Subject: Re: [PATCH 2/2] ntb: ntb_transport: Replace GFP_ATOMIC with GFP_KERNEL in ntb_transport_create_queue

On Tue, Apr 10, 2018 at 09:17:54PM +0800, Jia-Ju Bai wrote:
> ntb_transport_create_queue() is never called in atomic context.
>
> ntb_transport_create_queue() is only called by ntb_netdev_probe(),
> which is set as ".probe" in struct ntb_transport_client.
>
> Despite never getting called from atomic context,
> ntb_transport_create_queue() calls kzalloc_node() with GFP_ATOMIC,
> which does not sleep for allocation.
> GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL,
> which can sleep and improve the possibility of sucessful allocation.
>
> This is found by a static analysis tool named DCNS written by myself.
> And I also manually check it
>
> Signed-off-by: Jia-Ju Bai <[email protected]>

Both patches have been applied to my ntb-next branch.

Thanks,
Jon

> ---
> drivers/ntb/ntb_transport.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/ntb/ntb_transport.c b/drivers/ntb/ntb_transport.c
> index f58d8e3..2c0c8bc 100644
> --- a/drivers/ntb/ntb_transport.c
> +++ b/drivers/ntb/ntb_transport.c
> @@ -1825,7 +1825,7 @@ struct ntb_transport_qp *
> qp->rx_dma_chan ? "DMA" : "CPU");
>
> for (i = 0; i < NTB_QP_DEF_NUM_ENTRIES; i++) {
> - entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
> + entry = kzalloc_node(sizeof(*entry), GFP_KERNEL, node);
> if (!entry)
> goto err1;
>
> @@ -1836,7 +1836,7 @@ struct ntb_transport_qp *
> qp->rx_alloc_entry = NTB_QP_DEF_NUM_ENTRIES;
>
> for (i = 0; i < qp->tx_max_entry; i++) {
> - entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
> + entry = kzalloc_node(sizeof(*entry), GFP_KERNEL, node);
> if (!entry)
> goto err2;
>
> --
> 1.9.1
>