diff mbox series

[net] net: fec: don't save PTP state if PTP is unsupported

Message ID 20241008061153.1977930-1-wei.fang@nxp.com (mailing list archive)
State In Next, archived
Headers show
Series [net] net: fec: don't save PTP state if PTP is unsupported | expand

Commit Message

Wei Fang Oct. 8, 2024, 6:11 a.m. UTC
Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
these platforms fec_ptp_init() is not called and the related members
in fep are not initialized. However, fec_ptp_save_state() is called
unconditionally, which causes the kernel to panic. Therefore, add a
condition so that fec_ptp_save_state() is not called if PTP is not
supported.

Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
Reported-by: Guenter Roeck <linux@roeck-us.net>
Closes: https://lore.kernel.org/lkml/353e41fe-6bb4-4ee9-9980-2da2a9c1c508@roeck-us.net/
Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
 drivers/net/ethernet/freescale/fec_main.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Comments

Csókás Bence Oct. 8, 2024, 9:02 a.m. UTC | #1
On 2024. 10. 08. 8:11, Wei Fang wrote:
> Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> these platforms fec_ptp_init() is not called and the related members
> in fep are not initialized. However, fec_ptp_save_state() is called
> unconditionally, which causes the kernel to panic. Therefore, add a
> condition so that fec_ptp_save_state() is not called if PTP is not
> supported.
> 
> Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
> Reported-by: Guenter Roeck <linux@roeck-us.net>
> Closes: https://lore.kernel.org/lkml/353e41fe-6bb4-4ee9-9980-2da2a9c1c508@roeck-us.net/
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> ---
>   drivers/net/ethernet/freescale/fec_main.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
> index 60fb54231ead..1b55047c0237 100644
> --- a/drivers/net/ethernet/freescale/fec_main.c
> +++ b/drivers/net/ethernet/freescale/fec_main.c
> @@ -1077,7 +1077,8 @@ fec_restart(struct net_device *ndev)
>   	u32 rcntl = OPT_FRAME_SIZE | 0x04;
>   	u32 ecntl = FEC_ECR_ETHEREN;
>   
> -	fec_ptp_save_state(fep);
> +	if (fep->bufdesc_ex)
> +		fec_ptp_save_state(fep);
>   
>   	/* Whack a reset.  We should wait for this.
>   	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
> @@ -1340,7 +1341,8 @@ fec_stop(struct net_device *ndev)
>   			netdev_err(ndev, "Graceful transmit stop did not complete!\n");
>   	}
>   
> -	fec_ptp_save_state(fep);
> +	if (fep->bufdesc_ex)
> +		fec_ptp_save_state(fep);
>   
>   	/* Whack a reset.  We should wait for this.
>   	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC

Thanks, it seems fec_ptp_restore_state() was properly guarded, but 
save_state() was not. Sorry for that.

Reviewed-by: Csókás, Bence <csokas.bence@prolan.hu>
Simon Horman Oct. 9, 2024, 11:54 a.m. UTC | #2
On Tue, Oct 08, 2024 at 02:11:53PM +0800, Wei Fang wrote:
> Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> these platforms fec_ptp_init() is not called and the related members
> in fep are not initialized. However, fec_ptp_save_state() is called
> unconditionally, which causes the kernel to panic. Therefore, add a
> condition so that fec_ptp_save_state() is not called if PTP is not
> supported.
> 
> Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
> Reported-by: Guenter Roeck <linux@roeck-us.net>
> Closes: https://lore.kernel.org/lkml/353e41fe-6bb4-4ee9-9980-2da2a9c1c508@roeck-us.net/
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> ---
>  drivers/net/ethernet/freescale/fec_main.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
> index 60fb54231ead..1b55047c0237 100644
> --- a/drivers/net/ethernet/freescale/fec_main.c
> +++ b/drivers/net/ethernet/freescale/fec_main.c
> @@ -1077,7 +1077,8 @@ fec_restart(struct net_device *ndev)
>  	u32 rcntl = OPT_FRAME_SIZE | 0x04;
>  	u32 ecntl = FEC_ECR_ETHEREN;
>  
> -	fec_ptp_save_state(fep);
> +	if (fep->bufdesc_ex)
> +		fec_ptp_save_state(fep);

Hi,

I am wondering if you considered adding this check to (the top of)
fec_ptp_save_state. It seems like it would both lead to a smaller
change and be less error-prone to use.

>  
>  	/* Whack a reset.  We should wait for this.
>  	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
> @@ -1340,7 +1341,8 @@ fec_stop(struct net_device *ndev)
>  			netdev_err(ndev, "Graceful transmit stop did not complete!\n");
>  	}
>  
> -	fec_ptp_save_state(fep);
> +	if (fep->bufdesc_ex)
> +		fec_ptp_save_state(fep);
>  
>  	/* Whack a reset.  We should wait for this.
>  	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
> -- 
> 2.34.1
> 
>
Wei Fang Oct. 9, 2024, 12:43 p.m. UTC | #3
> -----Original Message-----
> From: Simon Horman <horms@kernel.org>
> Sent: 2024年10月9日 19:55
> To: Wei Fang <wei.fang@nxp.com>
> Cc: davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com; richardcochran@gmail.com; csokas.bence@prolan.hu;
> Shenwei Wang <shenwei.wang@nxp.com>; Clark Wang
> <xiaoning.wang@nxp.com>; linux@roeck-us.net; imx@lists.linux.dev;
> netdev@vger.kernel.org; linux-kernel@vger.kernel.org
> Subject: Re: [PATCH net] net: fec: don't save PTP state if PTP is unsupported
>
> On Tue, Oct 08, 2024 at 02:11:53PM +0800, Wei Fang wrote:
> > Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> > these platforms fec_ptp_init() is not called and the related members
> > in fep are not initialized. However, fec_ptp_save_state() is called
> > unconditionally, which causes the kernel to panic. Therefore, add a
> > condition so that fec_ptp_save_state() is not called if PTP is not
> > supported.
> >
> > Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
> > Reported-by: Guenter Roeck <linux@roeck-us.net>
> > Closes:
> https://lore.ker/
> nel.org%2Flkml%2F353e41fe-6bb4-4ee9-9980-2da2a9c1c508%40roeck-us.net
> %2F&data=05%7C02%7Cwei.fang%40nxp.com%7Cb10cac9ed8cd43284aae08
> dce85930cd%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C63864
> 0716999752935%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLC
> JQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=1
> gxwnxNjk91xX7I%2Foco%2F4OhBbxNCryhDMo72O9Jkr2w%3D&reserved=0
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > ---
> >  drivers/net/ethernet/freescale/fec_main.c | 6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/freescale/fec_main.c
> b/drivers/net/ethernet/freescale/fec_main.c
> > index 60fb54231ead..1b55047c0237 100644
> > --- a/drivers/net/ethernet/freescale/fec_main.c
> > +++ b/drivers/net/ethernet/freescale/fec_main.c
> > @@ -1077,7 +1077,8 @@ fec_restart(struct net_device *ndev)
> >     u32 rcntl = OPT_FRAME_SIZE | 0x04;
> >     u32 ecntl = FEC_ECR_ETHEREN;
> >
> > -   fec_ptp_save_state(fep);
> > +   if (fep->bufdesc_ex)
> > +           fec_ptp_save_state(fep);
>
> Hi,
>
> I am wondering if you considered adding this check to (the top of)
> fec_ptp_save_state. It seems like it would both lead to a smaller
> change and be less error-prone to use.
>

Yes, I considered this solution, but when I thought about it,
fec_ptp_save_state() and fec_ptp_restore_state() are a pair. If
the check is added to fec_ptp_save_state(), it is better to add
it to fec_ptp_restore_state(). However, considering that this is
not related to the current problem, and there are relatively few
calls to fec_ptp_restore_state(), I did not do this. If there are more
calls to fec_ptp_restore_state()/fec_ptp_restore_state() in the
future, I will consider it.

Thanks.

> >
> >     /* Whack a reset.  We should wait for this.
> >      * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
> > @@ -1340,7 +1341,8 @@ fec_stop(struct net_device *ndev)
> >                     netdev_err(ndev, "Graceful transmit stop did not complete!\n");
> >     }
> >
> > -   fec_ptp_save_state(fep);
> > +   if (fep->bufdesc_ex)
> > +           fec_ptp_save_state(fep);
> >
> >     /* Whack a reset.  We should wait for this.
> >      * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
> > --
> > 2.34.1
> >
> >
Simon Horman Oct. 9, 2024, 12:52 p.m. UTC | #4
On Wed, Oct 09, 2024 at 12:43:50PM +0000, Wei Fang wrote:
> > -----Original Message-----
> > From: Simon Horman <horms@kernel.org>
> > Sent: 2024年10月9日 19:55
> > To: Wei Fang <wei.fang@nxp.com>
> > Cc: davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> > pabeni@redhat.com; richardcochran@gmail.com; csokas.bence@prolan.hu;
> > Shenwei Wang <shenwei.wang@nxp.com>; Clark Wang
> > <xiaoning.wang@nxp.com>; linux@roeck-us.net; imx@lists.linux.dev;
> > netdev@vger.kernel.org; linux-kernel@vger.kernel.org
> > Subject: Re: [PATCH net] net: fec: don't save PTP state if PTP is unsupported
> >
> > On Tue, Oct 08, 2024 at 02:11:53PM +0800, Wei Fang wrote:
> > > Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> > > these platforms fec_ptp_init() is not called and the related members
> > > in fep are not initialized. However, fec_ptp_save_state() is called
> > > unconditionally, which causes the kernel to panic. Therefore, add a
> > > condition so that fec_ptp_save_state() is not called if PTP is not
> > > supported.
> > >
> > > Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
> > > Reported-by: Guenter Roeck <linux@roeck-us.net>
> > > Closes:
> > https://lore.ker/
> > nel.org%2Flkml%2F353e41fe-6bb4-4ee9-9980-2da2a9c1c508%40roeck-us.net
> > %2F&data=05%7C02%7Cwei.fang%40nxp.com%7Cb10cac9ed8cd43284aae08
> > dce85930cd%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C63864
> > 0716999752935%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLC
> > JQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=1
> > gxwnxNjk91xX7I%2Foco%2F4OhBbxNCryhDMo72O9Jkr2w%3D&reserved=0
> > > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > > ---
> > >  drivers/net/ethernet/freescale/fec_main.c | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/net/ethernet/freescale/fec_main.c
> > b/drivers/net/ethernet/freescale/fec_main.c
> > > index 60fb54231ead..1b55047c0237 100644
> > > --- a/drivers/net/ethernet/freescale/fec_main.c
> > > +++ b/drivers/net/ethernet/freescale/fec_main.c
> > > @@ -1077,7 +1077,8 @@ fec_restart(struct net_device *ndev)
> > >     u32 rcntl = OPT_FRAME_SIZE | 0x04;
> > >     u32 ecntl = FEC_ECR_ETHEREN;
> > >
> > > -   fec_ptp_save_state(fep);
> > > +   if (fep->bufdesc_ex)
> > > +           fec_ptp_save_state(fep);
> >
> > Hi,
> >
> > I am wondering if you considered adding this check to (the top of)
> > fec_ptp_save_state. It seems like it would both lead to a smaller
> > change and be less error-prone to use.
> >
> 
> Yes, I considered this solution, but when I thought about it,
> fec_ptp_save_state() and fec_ptp_restore_state() are a pair. If
> the check is added to fec_ptp_save_state(), it is better to add
> it to fec_ptp_restore_state(). However, considering that this is
> not related to the current problem, and there are relatively few
> calls to fec_ptp_restore_state(), I did not do this. If there are more
> calls to fec_ptp_restore_state()/fec_ptp_restore_state() in the
> future, I will consider it.

Sure, I agree on your point regarding symmetry, which I had not considered
when I wrote my previous email.  And I agree that the patch is suitable for
net in it's current form.

Reviewed-by: Simon Horman <horms@kernel.org>

...
Guenter Roeck Oct. 9, 2024, 5:19 p.m. UTC | #5
On 10/7/24 23:11, Wei Fang wrote:
> Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> these platforms fec_ptp_init() is not called and the related members
> in fep are not initialized. However, fec_ptp_save_state() is called
> unconditionally, which causes the kernel to panic. Therefore, add a
> condition so that fec_ptp_save_state() is not called if PTP is not
> supported.
> 
> Fixes: a1477dc87dc4 ("net: fec: Restart PPS after link state change")
> Reported-by: Guenter Roeck <linux@roeck-us.net>
> Closes: https://lore.kernel.org/lkml/353e41fe-6bb4-4ee9-9980-2da2a9c1c508@roeck-us.net/
> Signed-off-by: Wei Fang <wei.fang@nxp.com>

Tested-by: Guenter Roeck <linux@roeck-us.net>
patchwork-bot+netdevbpf@kernel.org Oct. 10, 2024, 2:50 a.m. UTC | #6
Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Tue,  8 Oct 2024 14:11:53 +0800 you wrote:
> Some platforms (such as i.MX25 and i.MX27) do not support PTP, so on
> these platforms fec_ptp_init() is not called and the related members
> in fep are not initialized. However, fec_ptp_save_state() is called
> unconditionally, which causes the kernel to panic. Therefore, add a
> condition so that fec_ptp_save_state() is not called if PTP is not
> supported.
> 
> [...]

Here is the summary with links:
  - [net] net: fec: don't save PTP state if PTP is unsupported
    https://git.kernel.org/netdev/net/c/6be063071a45

You are awesome, thank you!
diff mbox series

Patch

diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 60fb54231ead..1b55047c0237 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -1077,7 +1077,8 @@  fec_restart(struct net_device *ndev)
 	u32 rcntl = OPT_FRAME_SIZE | 0x04;
 	u32 ecntl = FEC_ECR_ETHEREN;
 
-	fec_ptp_save_state(fep);
+	if (fep->bufdesc_ex)
+		fec_ptp_save_state(fep);
 
 	/* Whack a reset.  We should wait for this.
 	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC
@@ -1340,7 +1341,8 @@  fec_stop(struct net_device *ndev)
 			netdev_err(ndev, "Graceful transmit stop did not complete!\n");
 	}
 
-	fec_ptp_save_state(fep);
+	if (fep->bufdesc_ex)
+		fec_ptp_save_state(fep);
 
 	/* Whack a reset.  We should wait for this.
 	 * For i.MX6SX SOC, enet use AXI bus, we use disable MAC