[U-Boot] [PATCH] eth: dtsec: fix TBI ANA setting bug in dtsec_configure_serdes()

[York Sun]

Shaohui,

Please comment/confirm.

York

On 12/09/2015 10:25 PM, 李远正 wrote:
> The value TBIANA_SETTINGS is also work on the p2041rdb board, so it can work 
> on the both phy and phyless connections.
> 
> 
> 
> At 2015-11-26 12:17:42, "Shaohui Xie" <Shaohui.Xie at freescale.com> wrote:
>>> -----Original Message-----
>>> From: York Sun [mailto:yorksun at freescale.com]
>>> Sent: Thursday, November 26, 2015 1:03 AM
>>> To: Xie Shaohui-B21989; Ciubotariu Codrin Constantin-B43658
>>> Cc: Yuanzheng Li; u-boot at lists.denx.de; Liu Dave-R63238
>>> Subject: Re: [PATCH] eth: dtsec: fix TBI ANA setting bug in
>>> dtsec_configure_serdes()
>>> 
>>> + Experts
>>> 
>>> On 11/24/2015 11:04 PM, Yuanzheng Li wrote:
>>> > The TBI_ANA register is configurated with the wrong value 0x4001,
>>> > refer to QorIQ Data Path Acceleration Architecture (DPAA) Reference
>>> Manual.
>>> > It set the reserved areas, bit 1 and bit 11 to bit 15 in big endian,
>>> > which should be cleared. But the normal functions of the
>>> > auto-negotiation, e.g. Pause and Full Duplex, do not be set.
>>[S.H] The value 0x4001 is so special, as mentioned, it does not match the 
>>DPAA RM, So I dig some docs and found AN3869, in which the 0x4001 is the
>>desired one to write to TBI_ANA for SGMII.
>> 
>>> >
>>> > There is no problem in the p2041rdb board, because the ppc is
>>> > connected directly with the phy chip which support auto-negotiation by
>>> > default in SGMII interface. But the link problem will occur when it is
>>> > connected with a switch chip like BCM5389, the switch chip disable
>>> > auto-negotiation by default, and the ppc also disable
>>> > auto-negotiation, then there is no link between them.
>>[S.H] This seems a phyless connection. If BCM5389 disables AN, then PPC should
>>Also disable AN by setting TBI_CR.
>>
>>> >
>>> > So use the vlue TBIANA_SETTINGS to enable the ppc's auto-negotiation.
>>[S.H] The TBI_ANA is " AN Advertisement Register ", it does not enable/disable AN.
>>A proper way should be to distinguish the phy and phyless connections and configure
>>The TBI registers accordingly.
>>
>>Thanks,
>>Shaohui
>>
>>> >
>>> > Signed-off-by: Yuanzheng Li <liyuanzheng01 at 163.com>
>>> > Cc: York Sun <yorksun at freescale.com>
>>> > ---
>>> >  drivers/net/fm/eth.c |    2 +-
>>> >  1 files changed, 1 insertions(+), 1 deletions(-)
>>> >
>>> > diff --git a/drivers/net/fm/eth.c b/drivers/net/fm/eth.c index
>>> > eb8e936..78c0988 100644
>>> > --- a/drivers/net/fm/eth.c
>>> > +++ b/drivers/net/fm/eth.c
>>> > @@ -81,7 +81,7 @@ qsgmii_loop:
>>> >  	tsec_local_mdio_write(phyregs, in_be32(&regs->tbipa), 0, TBI_TBICON,
>>> >  			TBICON_CLK_SELECT);
>>> >  	tsec_local_mdio_write(phyregs, in_be32(&regs->tbipa), 0, TBI_ANA,
>>> > -			TBIANA_SGMII_ACK);
>>> > +			TBIANA_SETTINGS);
>>> >  	tsec_local_mdio_write(phyregs, in_be32(&regs->tbipa), 0,
>>> >  			TBI_CR, TBICR_SETTINGS);
>>> >  #endif
>>> >
>>> 
>>> Yuanzheng,
>>> 
>>> After your change, TBIANA_SGMII_ACK is not used and should be removed.
>>> 
>>> Shaohui and Codrin,
>>> 
>>> Please comment.
>>> 
>>> York
> 
> 
> 
>  
>