[PATCH net v2] net: ethernet: xscale: Check for PTP support properly

[PATCH net v2] net: ethernet: xscale: Check for PTP support properly

[Linus Walleij]

In ixp4xx_get_ts_info() ixp46x_ptp_find() is called
unconditionally despite this feature only existing on
ixp46x, leading to the following splat from tcpdump:

root@OpenWrt:~# tcpdump -vv -X -i eth0
(...)
Unable to handle kernel NULL pointer dereference at virtual address
  00000238 when read
(...)
Call trace:
 ptp_clock_index from ixp46x_ptp_find+0x1c/0x38
 ixp46x_ptp_find from ixp4xx_get_ts_info+0x4c/0x64
 ixp4xx_get_ts_info from __ethtool_get_ts_info+0x90/0x108
 __ethtool_get_ts_info from __dev_ethtool+0xa00/0x2648
 __dev_ethtool from dev_ethtool+0x160/0x234
 dev_ethtool from dev_ioctl+0x2cc/0x460
 dev_ioctl from sock_ioctl+0x1ec/0x524
 sock_ioctl from sys_ioctl+0x51c/0xa94
 sys_ioctl from ret_fast_syscall+0x0/0x44
 (...)
Segmentation fault

Check for ixp46x in ixp46x_ptp_find() before trying to set up
PTP to avoid this.

Fixes: 9055a2f59162 ("ixp4xx_eth: make ptp support a platform driver")
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
Changes in v2:
- Use the correct Fixes-tag.
- Move check to ixp46x_ptp_find() to simplify things.
- Link to v1: https://lore.kernel.org/r/20260116-ixp4xx-fix-ethernet-v1-1-9ab3b411c77e@kernel.org
---
 drivers/net/ethernet/xscale/ixp4xx_eth.c | 5 +----
 drivers/net/ethernet/xscale/ptp_ixp46x.c | 3 +++
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/xscale/ixp4xx_eth.c b/drivers/net/ethernet/xscale/ixp4xx_eth.c
index e1e7f65553e7..b0faa0f1780d 100644
--- a/drivers/net/ethernet/xscale/ixp4xx_eth.c
+++ b/drivers/net/ethernet/xscale/ixp4xx_eth.c
@@ -403,15 +403,12 @@ static int ixp4xx_hwtstamp_set(struct net_device *netdev,
 	int ret;
 	int ch;
 
-	if (!cpu_is_ixp46x())
-		return -EOPNOTSUPP;
-
 	if (!netif_running(netdev))
 		return -EINVAL;
 
 	ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
 	if (ret)
-		return ret;
+		return -EOPNOTSUPP;
 
 	ch = PORT2CHANNEL(port);
 	regs = port->timesync_regs;
diff --git a/drivers/net/ethernet/xscale/ptp_ixp46x.c b/drivers/net/ethernet/xscale/ptp_ixp46x.c
index 94203eb46e6b..93c64db22a69 100644
--- a/drivers/net/ethernet/xscale/ptp_ixp46x.c
+++ b/drivers/net/ethernet/xscale/ptp_ixp46x.c
@@ -232,6 +232,9 @@ static struct ixp_clock ixp_clock;
 
 int ixp46x_ptp_find(struct ixp46x_ts_regs *__iomem *regs, int *phc_index)
 {
+	if (!cpu_is_ixp46x())
+		return -ENODEV;
+
 	*regs = ixp_clock.regs;
 	*phc_index = ptp_clock_index(ixp_clock.ptp_clock);
 

---
base-commit: 05f7e89ab9731565d8a62e3b5d1ec206485eeb0b
change-id: 20260116-ixp4xx-fix-ethernet-4fa36d900ccc

Best regards,
-- 
Linus Walleij <linusw@kernel.org>

Re: [PATCH net v2] net: ethernet: xscale: Check for PTP support properly

[Andrew Lunn]

On Wed, Feb 11, 2026 at 01:58:21PM +0100, Linus Walleij wrote:
> In ixp4xx_get_ts_info() ixp46x_ptp_find() is called
> unconditionally despite this feature only existing on
> ixp46x, leading to the following splat from tcpdump:
> 
> root@OpenWrt:~# tcpdump -vv -X -i eth0
> (...)
> Unable to handle kernel NULL pointer dereference at virtual address
>   00000238 when read
> (...)
> Call trace:
>  ptp_clock_index from ixp46x_ptp_find+0x1c/0x38
>  ixp46x_ptp_find from ixp4xx_get_ts_info+0x4c/0x64
>  ixp4xx_get_ts_info from __ethtool_get_ts_info+0x90/0x108
>  __ethtool_get_ts_info from __dev_ethtool+0xa00/0x2648
>  __dev_ethtool from dev_ethtool+0x160/0x234
>  dev_ethtool from dev_ioctl+0x2cc/0x460
>  dev_ioctl from sock_ioctl+0x1ec/0x524
>  sock_ioctl from sys_ioctl+0x51c/0xa94
>  sys_ioctl from ret_fast_syscall+0x0/0x44
>  (...)
> Segmentation fault
> 
> Check for ixp46x in ixp46x_ptp_find() before trying to set up
> PTP to avoid this.

>  	ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
>  	if (ret)
> -		return ret;
> +		return -EOPNOTSUPP;

Changing the error code is unusual. Maybe add to the commit message
why this is needed, or somebody might ask :-)

    Andrew

Re: [PATCH net v2] net: ethernet: xscale: Check for PTP support properly

[Linus Walleij]

On Wed, Feb 11, 2026 at 4:43 PM Andrew Lunn <andrew@lunn.ch> wrote:
> On Wed, Feb 11, 2026 at 01:58:21PM +0100, Linus Walleij wrote:
> > In ixp4xx_get_ts_info() ixp46x_ptp_find() is called
> > unconditionally despite this feature only existing on
> > ixp46x, leading to the following splat from tcpdump:
> >
> > root@OpenWrt:~# tcpdump -vv -X -i eth0
> > (...)
> > Unable to handle kernel NULL pointer dereference at virtual address
> >   00000238 when read
> > (...)
> > Call trace:
> >  ptp_clock_index from ixp46x_ptp_find+0x1c/0x38
> >  ixp46x_ptp_find from ixp4xx_get_ts_info+0x4c/0x64
> >  ixp4xx_get_ts_info from __ethtool_get_ts_info+0x90/0x108
> >  __ethtool_get_ts_info from __dev_ethtool+0xa00/0x2648
> >  __dev_ethtool from dev_ethtool+0x160/0x234
> >  dev_ethtool from dev_ioctl+0x2cc/0x460
> >  dev_ioctl from sock_ioctl+0x1ec/0x524
> >  sock_ioctl from sys_ioctl+0x51c/0xa94
> >  sys_ioctl from ret_fast_syscall+0x0/0x44
> >  (...)
> > Segmentation fault
> >
> > Check for ixp46x in ixp46x_ptp_find() before trying to set up
> > PTP to avoid this.
>
> >       ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
> >       if (ret)
> > -             return ret;
> > +             return -EOPNOTSUPP;
>
> Changing the error code is unusual. Maybe add to the commit message
> why this is needed, or somebody might ask :-)

Tricky one, but if you look close changing the error code is what I
try not to do! :D

Before this patch ixp4xx_hwtstamp_set() returned -EOPNOTSUPP:

-       if (!cpu_is_ixp46x())
-               return -EOPNOTSUPP;

After moving the check into ixp46x_ptp_find() as requested in the
review of v1 I have to return some error code from that new check,
but it is now called from more places, so it returns

int ixp46x_ptp_find(struct ixp46x_ts_regs *__iomem *regs, int *phc_index)
 {
+       if (!cpu_is_ixp46x())
+               return -ENODEV;

So this:

     ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
        if (ret)
-               return ret;
+               return -EOPNOTSUPP;

Is put into ixp4xx_hwtstamp_set() so it will return the same error
code as before if PTP was not found.

(I can copyedit this reasoning into the commit message if you want.)

Yours,
Linus Walleij

Re: [PATCH net v2] net: ethernet: xscale: Check for PTP support properly

[Jacob Keller]

On 2/11/2026 1:03 PM, Linus Walleij wrote:
> On Wed, Feb 11, 2026 at 4:43 PM Andrew Lunn <andrew@lunn.ch> wrote:
>> On Wed, Feb 11, 2026 at 01:58:21PM +0100, Linus Walleij wrote:
>>> In ixp4xx_get_ts_info() ixp46x_ptp_find() is called
>>> unconditionally despite this feature only existing on
>>> ixp46x, leading to the following splat from tcpdump:
>>>
>>> root@OpenWrt:~# tcpdump -vv -X -i eth0
>>> (...)
>>> Unable to handle kernel NULL pointer dereference at virtual address
>>>    00000238 when read
>>> (...)
>>> Call trace:
>>>   ptp_clock_index from ixp46x_ptp_find+0x1c/0x38
>>>   ixp46x_ptp_find from ixp4xx_get_ts_info+0x4c/0x64
>>>   ixp4xx_get_ts_info from __ethtool_get_ts_info+0x90/0x108
>>>   __ethtool_get_ts_info from __dev_ethtool+0xa00/0x2648
>>>   __dev_ethtool from dev_ethtool+0x160/0x234
>>>   dev_ethtool from dev_ioctl+0x2cc/0x460
>>>   dev_ioctl from sock_ioctl+0x1ec/0x524
>>>   sock_ioctl from sys_ioctl+0x51c/0xa94
>>>   sys_ioctl from ret_fast_syscall+0x0/0x44
>>>   (...)
>>> Segmentation fault
>>>
>>> Check for ixp46x in ixp46x_ptp_find() before trying to set up
>>> PTP to avoid this.
>>
>>>        ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
>>>        if (ret)
>>> -             return ret;
>>> +             return -EOPNOTSUPP;
>>
>> Changing the error code is unusual. Maybe add to the commit message
>> why this is needed, or somebody might ask :-)
> 
> Tricky one, but if you look close changing the error code is what I
> try not to do! :D
> 
> Before this patch ixp4xx_hwtstamp_set() returned -EOPNOTSUPP:
> 
> -       if (!cpu_is_ixp46x())
> -               return -EOPNOTSUPP;
> 
> After moving the check into ixp46x_ptp_find() as requested in the
> review of v1 I have to return some error code from that new check,
> but it is now called from more places, so it returns
> 
> int ixp46x_ptp_find(struct ixp46x_ts_regs *__iomem *regs, int *phc_index)
>   {
> +       if (!cpu_is_ixp46x())
> +               return -ENODEV;
> 
> So this:
> 
>       ret = ixp46x_ptp_find(&port->timesync_regs, &port->phc_index);
>          if (ret)
> -               return ret;
> +               return -EOPNOTSUPP;
> 
> Is put into ixp4xx_hwtstamp_set() so it will return the same error
> code as before if PTP was not found.
> 
> (I can copyedit this reasoning into the commit message if you want.)
> 
> Yours,
> Linus Walleij
> 

That would be helpful to have in the commit message, since it is not 
immediately obvious.