ARM: imx_v6_v7_defconfig: Select SMSC_PHY

Hi Leonard,

On Wed, Mar 22, 2017 at 04:27:37PM +0200, Leonard Crestez wrote:
> The imx6sl-evk board has a LAN8720A ethernet phy supported by SMSC_PHY.
> Add this driver to the default imx config since the device is present on
> one of the evaluation boards.

Upstream currently no evaluation board support.

This may leave to Shawn to decide whether apply it.

> 
> This used to work mostly fine with the generic phy driver but since
> commit 0878fff1f42c18e448ab5b8b4f6a3eb32365b5b6 that driver no longer

I guess you should get a checkpatch error here.

> performs a soft reset on startup. This causes netboot to sometimes
> timeout on DHCP because RX doesn't work right. DHCP is eventually retried and

Probably another checkpatch warning.

> it works the second time but it takes 90+ seconds to get a login prompt.
> 
> This was generated with "make savedefconfig" and it includes a few
> additional minor cleanups.
> 

It is a bit strange, after apply your patch, i still get a lot difference
as follows when savedefconfig:

I don't know what's wrong. Toolchains difference?
Quite strange...

Regards
Dong Aisheng

> Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
> ---
> 
> I also tried to do some debugging in the fec driver and it apparently receives
> corrupted packets when this happens. If I hack it to go into promiscuous mode
> unconditionally it gets a whole bunch of rx errors (crc errors, length errors
> and so on). So the phy config is probably wrong and is confusing the mac?
> 
> In theory it might be possible to make that driver "just work" with phy
> settings from uboot but it's not clear it's worthwhile.
> 
>  arch/arm/configs/imx_v6_v7_defconfig | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/arch/arm/configs/imx_v6_v7_defconfig b/arch/arm/configs/imx_v6_v7_defconfig
> index eaba3b1..e605389 100644
> --- a/arch/arm/configs/imx_v6_v7_defconfig
> +++ b/arch/arm/configs/imx_v6_v7_defconfig
> @@ -143,6 +143,7 @@ CONFIG_SMSC911X=y
>  # CONFIG_NET_VENDOR_STMICRO is not set
>  CONFIG_AT803X_PHY=y
>  CONFIG_MICREL_PHY=y
> +CONFIG_SMSC_PHY=y
>  CONFIG_USB_PEGASUS=m
>  CONFIG_USB_RTL8150=m
>  CONFIG_USB_RTL8152=m
> @@ -152,7 +153,6 @@ CONFIG_BRCMFMAC=m
>  CONFIG_WL12XX=m
>  CONFIG_WLCORE_SDIO=m
>  # CONFIG_WILINK_PLATFORM_DATA is not set
> -# CONFIG_INPUT_MOUSEDEV_PSAUX is not set
>  CONFIG_INPUT_EVDEV=y
>  CONFIG_INPUT_EVBUG=m
>  CONFIG_KEYBOARD_GPIO=y
> @@ -376,7 +376,6 @@ CONFIG_NLS_ISO8859_1=y
>  CONFIG_NLS_ISO8859_15=m
>  CONFIG_NLS_UTF8=y
>  CONFIG_PRINTK_TIME=y
> -CONFIG_DEBUG_FS=y
>  CONFIG_MAGIC_SYSRQ=y
>  # CONFIG_SCHED_DEBUG is not set
>  CONFIG_PROVE_LOCKING=y
> -- 
> 2.7.4
>

On Fri, 2017-03-24 at 14:43 +0800, Dong Aisheng wrote:
> Hi Leonard,
> 
> On Wed, Mar 22, 2017 at 04:27:37PM +0200, Leonard Crestez wrote:
> > 
> > The imx6sl-evk board has a LAN8720A ethernet phy supported by SMSC_PHY.
> > Add this driver to the default imx config since the device is present on
> > one of the evaluation boards.
> Upstream currently no evaluation board support.
> 
> This may leave to Shawn to decide whether apply it.

Really? But the dts for this eval board is included in upstream. Shouldn't the
default config try to work with the default imx*.dts files included?

> > This used to work mostly fine with the generic phy driver but since
> > commit 0878fff1f42c18e448ab5b8b4f6a3eb32365b5b6 that driver no longer
> I guess you should get a checkpatch error here.

Because I didn't reference that other commit right? Sorry about this.

Should I submit again with an improved commit message?

> > performs a soft reset on startup. This causes netboot to sometimes
> > timeout on DHCP because RX doesn't work right. DHCP is eventually retried
> > and
> Probably another checkpatch warning.

Why, line too long?

> > This was generated with "make savedefconfig" and it includes a few
> > additional minor cleanups.
> > 
> It is a bit strange, after apply your patch, i still get a lot difference
> as follows when savedefconfig:
> 
> I don't know what's wrong. Toolchains difference?

Maybe, kconfig acts strangely sometimes. It's not clear option ordering is
guaranteed to be stable. I don't think it matters though.

--
Regards,
Leonard

On Tue, Mar 28, 2017 at 02:16:50PM +0300, Leonard Crestez wrote:
> On Fri, 2017-03-24 at 14:43 +0800, Dong Aisheng wrote:
> > Hi Leonard,
> > 
> > On Wed, Mar 22, 2017 at 04:27:37PM +0200, Leonard Crestez wrote:
> > > 
> > > The imx6sl-evk board has a LAN8720A ethernet phy supported by SMSC_PHY.
> > > Add this driver to the default imx config since the device is present on
> > > one of the evaluation boards.
> > Upstream currently no evaluation board support.
> > 
> > This may leave to Shawn to decide whether apply it.
> 
> Really? But the dts for this eval board is included in upstream. Shouldn't the
> default config try to work with the default imx*.dts files included?
> 
> > > This used to work mostly fine with the generic phy driver but since
> > > commit 0878fff1f42c18e448ab5b8b4f6a3eb32365b5b6 that driver no longer
> > I guess you should get a checkpatch error here.
> 
> Because I didn't reference that other commit right? Sorry about this.
> 
> Should I submit again with an improved commit message?

Yes, please, and make sure that checkpatch.pl is happy about the patch.

Shawn

On Tue, Mar 28, 2017 at 02:16:50PM +0300, Leonard Crestez wrote:
> On Fri, 2017-03-24 at 14:43 +0800, Dong Aisheng wrote:
> > Hi Leonard,
> > 
> > On Wed, Mar 22, 2017 at 04:27:37PM +0200, Leonard Crestez wrote:
> > > 
> > > The imx6sl-evk board has a LAN8720A ethernet phy supported by SMSC_PHY.
> > > Add this driver to the default imx config since the device is present on
> > > one of the evaluation boards.
> > Upstream currently no evaluation board support.
> > 
> > This may leave to Shawn to decide whether apply it.
> 
> Really? But the dts for this eval board is included in upstream. Shouldn't the
> default config try to work with the default imx*.dts files included?
> 

Oh.. i saw the board wrong, sorry for the noise.

Regards
Dong Aisheng

> > > This used to work mostly fine with the generic phy driver but since
> > > commit 0878fff1f42c18e448ab5b8b4f6a3eb32365b5b6 that driver no longer
> > I guess you should get a checkpatch error here.
> 
> Because I didn't reference that other commit right? Sorry about this.
> 
> Should I submit again with an improved commit message?
> 
> > > performs a soft reset on startup. This causes netboot to sometimes
> > > timeout on DHCP because RX doesn't work right. DHCP is eventually retried
> > > and
> > Probably another checkpatch warning.
> 
> Why, line too long?
> 
> > > This was generated with "make savedefconfig" and it includes a few
> > > additional minor cleanups.
> > > 
> > It is a bit strange, after apply your patch, i still get a lot difference
> > as follows when savedefconfig:
> > 
> > I don't know what's wrong. Toolchains difference?
> 
> Maybe, kconfig acts strangely sometimes. It's not clear option ordering is
> guaranteed to be stable. I don't think it matters though.
> 
> --
> Regards,
> Leonard