diff mbox series

[v2,net,1/1] net: fec: using the standard return codes when xdp xmit errors

Message ID 20230510200523.1352951-1-shenwei.wang@nxp.com
State New
Headers
Series [v2,net,1/1] net: fec: using the standard return codes when xdp xmit errors |

Commit Message

Shenwei Wang May 10, 2023, 8:05 p.m. UTC 
  This patch standardizes the inconsistent return values for unsuccessful
XDP transmits by using standardized error codes (-EBUSY or -ENOMEM).

Fixes: 26312c685ae0 ("net: fec: correct the counting of XDP sent frames")
Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
---
 v2:
  - focusing on code clean up per Simon's feedback.

 drivers/net/ethernet/freescale/fec_main.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

--
2.34.1

Comments

Horatiu Vultur May 11, 2023, 7:24 a.m. UTC | #1 
The 05/10/2023 15:05, Shenwei Wang wrote:
> 
> This patch standardizes the inconsistent return values for unsuccessful
> XDP transmits by using standardized error codes (-EBUSY or -ENOMEM).

Shouldn't this patch target net-next instead of net? As Simon suggested
here [1], or maybe is just me who misunderstood that part.
Also it is nice to CC people who comment at your previous patches in all
the next versions.

Just a small thing, if there is only 1 patch in the series, you don't
need to add 1/1 in the subject.

[1] https://lore.kernel.org/netdev/20230509193845.1090040-1-shenwei.wang@nxp.com/T/#m4b6b21c75512391496294fc78db2fbdf687f1381

> 
> Fixes: 26312c685ae0 ("net: fec: correct the counting of XDP sent frames")
> Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
> ---
>  v2:
>   - focusing on code clean up per Simon's feedback.
> 
>  drivers/net/ethernet/freescale/fec_main.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
> index 42ec6ca3bf03..6a021fe24dfe 100644
> --- a/drivers/net/ethernet/freescale/fec_main.c
> +++ b/drivers/net/ethernet/freescale/fec_main.c
> @@ -3798,8 +3798,7 @@ static int fec_enet_txq_xmit_frame(struct fec_enet_private *fep,
>         entries_free = fec_enet_get_free_txdesc_num(txq);
>         if (entries_free < MAX_SKB_FRAGS + 1) {
>                 netdev_err(fep->netdev, "NOT enough BD for SG!\n");
> -               xdp_return_frame(frame);
> -               return NETDEV_TX_BUSY;
> +               return -EBUSY;
>         }
> 
>         /* Fill in a Tx ring entry */
> @@ -3813,7 +3812,7 @@ static int fec_enet_txq_xmit_frame(struct fec_enet_private *fep,
>         dma_addr = dma_map_single(&fep->pdev->dev, frame->data,
>                                   frame->len, DMA_TO_DEVICE);
>         if (dma_mapping_error(&fep->pdev->dev, dma_addr))
> -               return FEC_ENET_XDP_CONSUMED;
> +               return -ENOMEM;
> 
>         status |= (BD_ENET_TX_INTR | BD_ENET_TX_LAST);
>         if (fep->bufdesc_ex)
> @@ -3869,7 +3868,7 @@ static int fec_enet_xdp_xmit(struct net_device *dev,
>         __netif_tx_lock(nq, cpu);
> 
>         for (i = 0; i < num_frames; i++) {
> -               if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) != 0)
> +               if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) < 0)
>                         break;
>                 sent_frames++;
>         }
> --
> 2.34.1
> 
>
Simon Horman May 11, 2023, 1:02 p.m. UTC | #2 
On Thu, May 11, 2023 at 09:24:52AM +0200, Horatiu Vultur wrote:
> The 05/10/2023 15:05, Shenwei Wang wrote:
> > 
> > This patch standardizes the inconsistent return values for unsuccessful
> > XDP transmits by using standardized error codes (-EBUSY or -ENOMEM).
> 
> Shouldn't this patch target net-next instead of net? As Simon suggested
> here [1], or maybe is just me who misunderstood that part.
> Also it is nice to CC people who comment at your previous patches in all
> the next versions.
> 
> Just a small thing, if there is only 1 patch in the series, you don't
> need to add 1/1 in the subject.
> 
> [1] https://lore.kernel.org/netdev/20230509193845.1090040-1-shenwei.wang@nxp.com/T/#m4b6b21c75512391496294fc78db2fbdf687f1381
> 
> > 
> > Fixes: 26312c685ae0 ("net: fec: correct the counting of XDP sent frames")

Hi Shenwei,

I agree with Horatiu.

Also, this is not a fix. So it should not have a Fixes tag.

After waiting for further review please send a v3 with these updates.

pw-bot: cr
Shenwei Wang May 11, 2023, 1:54 p.m. UTC | #3 
> -----Original Message-----
> From: Horatiu Vultur <horatiu.vultur@microchip.com>
> Sent: Thursday, May 11, 2023 2:25 AM
> To: Shenwei Wang <shenwei.wang@nxp.com>
> Cc: Wei Fang <wei.fang@nxp.com>; David S. Miller <davem@davemloft.net>;
> Eric Dumazet <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>;
> Paolo Abeni <pabeni@redhat.com>; Clark Wang <xiaoning.wang@nxp.com>; dl-
> linux-imx <linux-imx@nxp.com>; Alexei Starovoitov <ast@kernel.org>; Daniel
> Borkmann <daniel@iogearbox.net>; Jesper Dangaard Brouer
> <hawk@kernel.org>; John Fastabend <john.fastabend@gmail.com>; Alexander
> Lobakin <alexandr.lobakin@intel.com>; netdev@vger.kernel.org; linux-
> kernel@vger.kernel.org; imx@lists.linux.dev
> Subject: [EXT] Re: [PATCH v2 net 1/1] net: fec: using the standard return codes
> when xdp xmit errors
>
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report this
> email' button
>
>
> The 05/10/2023 15:05, Shenwei Wang wrote:
> >
> > This patch standardizes the inconsistent return values for
> > unsuccessful XDP transmits by using standardized error codes (-EBUSY or -
> ENOMEM).
>
> Shouldn't this patch target net-next instead of net? As Simon suggested here [1],
> or maybe is just me who misunderstood that part.

-               xdp_return_frame(frame);
-               return NETDEV_TX_BUSY;
+               return -EBUSY;
As the "xdp_return_frame(frame)" should not be called when error branch, the patch
fix it too.

> Also it is nice to CC people who comment at your previous patches in all the next
> versions.
>
> Just a small thing, if there is only 1 patch in the series, you don't need to add 1/1
> in the subject.
>

The patch was generated by using "git format-patch -n1". The strange thing is that
it appends "1/1" on some machines but not on others. The output seems inconsistent.
However, using "git format-patch -N1" can fix this.

Thanks,
Shenwei

> [1]
> https://lore.kern/
> el.org%2Fnetdev%2F20230509193845.1090040-1-
> shenwei.wang%40nxp.com%2FT%2F%23m4b6b21c75512391496294fc78db2fbd
> f687f1381&data=05%7C01%7Cshenwei.wang%40nxp.com%7Ca92a1dbe84f348
> 42e79e08db51f0d24f%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C
> 638193866998878533%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMD
> AiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&s
> data=MO7RbDtCUU82clB2CyltzpvEdc7%2FjwLWvcCzgGN8mDc%3D&reserved=0
>
> >
> > Fixes: 26312c685ae0 ("net: fec: correct the counting of XDP sent
> > frames")
> > Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
> > ---
> >  v2:
> >   - focusing on code clean up per Simon's feedback.
> >
> >  drivers/net/ethernet/freescale/fec_main.c | 7 +++----
> >  1 file changed, 3 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/freescale/fec_main.c
> > b/drivers/net/ethernet/freescale/fec_main.c
> > index 42ec6ca3bf03..6a021fe24dfe 100644
> > --- a/drivers/net/ethernet/freescale/fec_main.c
> > +++ b/drivers/net/ethernet/freescale/fec_main.c
> > @@ -3798,8 +3798,7 @@ static int fec_enet_txq_xmit_frame(struct
> fec_enet_private *fep,
> >         entries_free = fec_enet_get_free_txdesc_num(txq);
> >         if (entries_free < MAX_SKB_FRAGS + 1) {
> >                 netdev_err(fep->netdev, "NOT enough BD for SG!\n");
> > -               xdp_return_frame(frame);
> > -               return NETDEV_TX_BUSY;
> > +               return -EBUSY;
> >         }
> >
> >         /* Fill in a Tx ring entry */
> > @@ -3813,7 +3812,7 @@ static int fec_enet_txq_xmit_frame(struct
> fec_enet_private *fep,
> >         dma_addr = dma_map_single(&fep->pdev->dev, frame->data,
> >                                   frame->len, DMA_TO_DEVICE);
> >         if (dma_mapping_error(&fep->pdev->dev, dma_addr))
> > -               return FEC_ENET_XDP_CONSUMED;
> > +               return -ENOMEM;
> >
> >         status |= (BD_ENET_TX_INTR | BD_ENET_TX_LAST);
> >         if (fep->bufdesc_ex)
> > @@ -3869,7 +3868,7 @@ static int fec_enet_xdp_xmit(struct net_device *dev,
> >         __netif_tx_lock(nq, cpu);
> >
> >         for (i = 0; i < num_frames; i++) {
> > -               if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) != 0)
> > +               if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) < 0)
> >                         break;
> >                 sent_frames++;
> >         }
> > --
> > 2.34.1
> >
> >
>
> --
> /Horatiu
Jakub Kicinski May 12, 2023, 12:55 a.m. UTC | #4 
On Wed, 10 May 2023 15:05:23 -0500 Shenwei Wang wrote:
> -		xdp_return_frame(frame);

This line is a bug fix (double free).

I'm going to apply v2, it's good enough.
Jakub Kicinski May 12, 2023, 12:58 a.m. UTC | #5 
On Thu, 11 May 2023 17:55:42 -0700 Jakub Kicinski wrote:
> On Wed, 10 May 2023 15:05:23 -0500 Shenwei Wang wrote:
> > -		xdp_return_frame(frame);  
> 
> This line is a bug fix (double free).
> 
> I'm going to apply v2, it's good enough.

Let me take that back, I'll reply to v2.
diff mbox series

Patch

diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 42ec6ca3bf03..6a021fe24dfe 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -3798,8 +3798,7 @@  static int fec_enet_txq_xmit_frame(struct fec_enet_private *fep,
 	entries_free = fec_enet_get_free_txdesc_num(txq);
 	if (entries_free < MAX_SKB_FRAGS + 1) {
 		netdev_err(fep->netdev, "NOT enough BD for SG!\n");
-		xdp_return_frame(frame);
-		return NETDEV_TX_BUSY;
+		return -EBUSY;
 	}

 	/* Fill in a Tx ring entry */
@@ -3813,7 +3812,7 @@  static int fec_enet_txq_xmit_frame(struct fec_enet_private *fep,
 	dma_addr = dma_map_single(&fep->pdev->dev, frame->data,
 				  frame->len, DMA_TO_DEVICE);
 	if (dma_mapping_error(&fep->pdev->dev, dma_addr))
-		return FEC_ENET_XDP_CONSUMED;
+		return -ENOMEM;

 	status |= (BD_ENET_TX_INTR | BD_ENET_TX_LAST);
 	if (fep->bufdesc_ex)
@@ -3869,7 +3868,7 @@  static int fec_enet_xdp_xmit(struct net_device *dev,
 	__netif_tx_lock(nq, cpu);

 	for (i = 0; i < num_frames; i++) {
-		if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) != 0)
+		if (fec_enet_txq_xmit_frame(fep, txq, frames[i]) < 0)
 			break;
 		sent_frames++;
 	}