[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[luo penghao]

This statement is redundant in the context, because there will be
an identical statement next. otherwise, the variable initialization
is actually unnecessary.

The clang_analyzer complains as follows:

drivers/net/ethernet/intel/e1000/e1000_ethtool.c:1218:2 warning:

Value stored to 'ctrl_reg' is never read.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: luo penghao <luo.penghao@zte.com.cn>
---
 drivers/net/ethernet/intel/e1000/e1000_ethtool.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
index 0a57172..8951f2a 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
@@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
 		e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
 	}
 
-	ctrl_reg = er32(CTRL);
-
 	/* force 1000, set loopback */
 	e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);
 
-- 
2.15.2

[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[jesse.brandeburg]

On 10/18/21 1:53 AM, luo penghao <cgel.zte@gmail.com> wrote:
> This statement is redundant in the context, because there will be
> an identical statement next. otherwise, the variable initialization
> is actually unnecessary.
> 
> The clang_analyzer complains as follows:
> 
> drivers/net/ethernet/intel/e1000/e1000_ethtool.c:1218:2 warning:
> 
> Value stored to 'ctrl_reg' is never read.
> 
> Reported-by: Zeal Robot <zealci@zte.com.cn>
> Signed-off-by: luo penghao <luo.penghao@zte.com.cn>
> ---
>   drivers/net/ethernet/intel/e1000/e1000_ethtool.c | 2 --
>   1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> index 0a57172..8951f2a 100644
> --- a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> +++ b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> @@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
>   		e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
>   	}
>   
> -	ctrl_reg = er32(CTRL);

Thanks for your patch, but this change is not safe. you're removing a read that could do two things. The first is that the read "flushes" the write just above to PCI (it's a PCI barrier), and the second is that the read can have some side effects.

If this change must be done, the code should be to remove the assignment to ctrl_reg, but leave the read, so the line would just look like:
        er32(CTRL);

This will get rid of the warning and not change the flow from the hardware perspective.

> -
>   	/* force 1000, set loopback */
>   	e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);
>   
> 
> 

[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[Jesse Brandeburg]

On 10/18/2021 9:10 AM, jesse.brandeburg at intel.com wrote:
> On 10/18/21 1:53 AM, luo penghao <cgel.zte@gmail.com> wrote:
>> This statement is redundant in the context, because there will be
>> an identical statement next. otherwise, the variable initialization
>> is actually unnecessary.
>>
>> The clang_analyzer complains as follows:
>>
>> drivers/net/ethernet/intel/e1000/e1000_ethtool.c:1218:2 warning:
>>
>> Value stored to 'ctrl_reg' is never read.
>>
>> Reported-by: Zeal Robot <zealci@zte.com.cn>
>> Signed-off-by: luo penghao <luo.penghao@zte.com.cn>
>> ---
>> ? drivers/net/ethernet/intel/e1000/e1000_ethtool.c | 2 --
>> ? 1 file changed, 2 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c 
>> b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
>> index 0a57172..8951f2a 100644
>> --- a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
>> +++ b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
>> @@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct 
>> e1000_adapter *adapter)
>> ????????? e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
>> ????? }
>> -??? ctrl_reg = er32(CTRL);
> 

Please see my reply below.

> Thanks for your patch, but this change is not safe. you're removing a 
> read that could do two things. The first is that the read "flushes" the 
> write just above to PCI (it's a PCI barrier), and the second is that the 
> read can have some side effects.
> 
> If this change must be done, the code should be to remove the assignment 
> to ctrl_reg, but leave the read, so the line would just look like:
>  ?????? er32(CTRL);
> 
> This will get rid of the warning and not change the flow from the 
> hardware perspective.
> 
>> -
>> ????? /* force 1000, set loopback */
>> ????? e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);

my previous mail seems to have not made it to netdev/patchwork, so 
re-sending.

NAK. Please respin if you must make this change.

[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[Simon Horman]

On Mon, Oct 18, 2021 at 08:53:05AM +0000, luo penghao wrote:

nit: s/linux-next/net-next/ in subject

> This statement is redundant in the context, because there will be
> an identical statement next. otherwise, the variable initialization
> is actually unnecessary.
> 
> The clang_analyzer complains as follows:
> 
> drivers/net/ethernet/intel/e1000/e1000_ethtool.c:1218:2 warning:
> 
> Value stored to 'ctrl_reg' is never read.

I agree this does seem to be the case.

> Reported-by: Zeal Robot <zealci@zte.com.cn>
> Signed-off-by: luo penghao <luo.penghao@zte.com.cn>

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

> ---
>  drivers/net/ethernet/intel/e1000/e1000_ethtool.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> index 0a57172..8951f2a 100644
> --- a/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> +++ b/drivers/net/ethernet/intel/e1000/e1000_ethtool.c
> @@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
>  		e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
>  	}
>  
> -	ctrl_reg = er32(CTRL);
> -
>  	/* force 1000, set loopback */
>  	e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);
>  
> -- 
> 2.15.2
> 
> 

[Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

[Jesse Brandeburg]

Apologies for the duplicates, mail from my intel account going out
through outlook.com is not being delivered.

On Wed, Oct 20, 2021 at 7:00 AM Simon Horman <horms@kernel.org> wrote:

> > Value stored to 'ctrl_reg' is never read.
>
> I agree this does seem to be the case.
>
> > Reported-by: Zeal Robot <zealci@zte.com.cn>
> > Signed-off-by: luo penghao <luo.penghao@zte.com.cn>
>
> Reviewed-by: Simon Horman <horms@kernel.org>

Thanks for the review, but (davem/kuba) please do not apply.

> > @@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
> >               e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
> >       }
> >
> > -     ctrl_reg = er32(CTRL);

Thanks for your patch, but this change is not safe. you're removing a
read that could do two things. The first is that the read "flushes"
the write just above to PCI (it's a PCI barrier), and the second is
that the read can have some side effects.

If this change must be done, the code should be to remove the
assignment to ctrl_reg, but leave the read, so the line would just
look like:
        er32(CTRL);

This will get rid of the warning and not change the flow from the
hardware perspective.

> > -
> >       /* force 1000, set loopback */
> >       e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);
> >

Please do not apply this.