[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <AF233D1473C1364ABD51D28909A1B1B75C0B0E38@pgsmsx114.gar.corp.intel.com>
Date: Thu, 25 Apr 2019 06:51:47 +0000
From: "Ong, Boon Leong" <boon.leong.ong@...el.com>
To: Andrew Lunn <andrew@...n.ch>
CC: "David S. Miller" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Kweh, Hock Leong" <hock.leong.kweh@...el.com>,
"Voon, Weifeng" <weifeng.voon@...el.com>
Subject: RE: [PATCH 4/7] net: stmmac: introducing support for DWC xPCS logics
>-----Original Message-----
>From: Andrew Lunn [mailto:andrew@...n.ch]
>Sent: Thursday, April 25, 2019 11:21 AM
>To: Ong, Boon Leong <boon.leong.ong@...el.com>
>Cc: David S. Miller <davem@...emloft.net>; netdev@...r.kernel.org; linux-
>kernel@...r.kernel.org; Kweh, Hock Leong <hock.leong.kweh@...el.com>;
>Voon, Weifeng <weifeng.voon@...el.com>
>Subject: Re: [PATCH 4/7] net: stmmac: introducing support for DWC xPCS
>logics
>
>> >> +/* MII MMD AN Status defines for C37 AN SGMII Status */
>> >> +#define AN_STAT_C37_AN_CMPLT BIT(0) /* AN
>Complete Intr */
>> >> +#define AN_STAT_C37_AN_FD BIT(1) /* Full Duplex */
>> >> +#define AN_STAT_C37_AN_SPEED_SHIFT 2 /* AN Speed shift */
>> >> +#define AN_STAT_C37_AN_SPEED GENMASK(3, 2) /* AN
>Speed */
>> >> +#define AN_STAT_C37_AN_10MBPS 0x0 /* 10 Mbps */
>> >> +#define AN_STAT_C37_AN_100MBPS 0x1 /* 100 Mbps
>> >*/
>> >> +#define AN_STAT_C37_AN_1000MBPS 0x2 /* 1000 Mbps
>> >*/
>> >> +#define AN_STAT_C37_AN_LNKSTS BIT(4) /* Link Status
>*/
>> >
>> >Is these are standardized, not proprietary, consider adding them to
>> >include/uapi/linux/mii.h so similar.
>>
>> Yeah, it does look very standardized. However, per DW spec, they are
>> vendor-specific register set which uses MDIO_MMD_VEND2 to access.
>
>Maybe you can go get the 802.3 standard and check section 37. If these
>registers are the same, they could still be placed in mii.h. What MMD
>you use is a different matter.
>
I cross-checked on 802.3 section 37.2.5.1 Management registers, the spec only
talks about BMSR and extended status. I also checked on 802.3 section 45.2
MDIO interface registers, section 45.2.7 AN registers, section 45.2.9 &
45.2.10 Vendor specific MMD1 & MMD2, there are no field format that
matches the above bit fields. So, it should be reasonable to park these #defines
in this local file for now.
Thanks.
Boon Leong
Powered by blists - more mailing lists