[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANh8QzxuHAu+L0swPC5V4Oca21Z5zpiULTm22VPShX_T-JVznQ@mail.gmail.com>
Date: Sat, 23 May 2020 11:55:30 +0200
From: "Fuzzey, Martin" <martin.fuzzey@...wbird.group>
To: Andy Duan <fugang.duan@....com>
Cc: "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
Rob Herring <robh+dt@...nel.org>,
Shawn Guo <shawnguo@...nel.org>, devicetree@...r.kernel.org
Subject: Re: [PATCH net 1/4] net: ethernet: fec: move GPR register offset and
bit into DT
Hi Andy,
> Fixes: da722186f654(net: fec: set GPR bit on suspend by DT configuration)
Just a nitpick maybe but I don't really think this need as Fixes: tag.
That commit didn't actually *break* anything AFAIK.
It added WoL support for *some* SoCs that didn't have any in mainline
and didn't hurt the others.
Of course it turned out to be insufficient for the multiple FEC case
so this patch series is a welcome improvement.
> struct fec_devinfo {
> u32 quirks;
> - u8 stop_gpr_reg;
> - u8 stop_gpr_bit;
> };
This structure has become redundant now that it only contains a single
u32 quirks field.
So we *could* go back to storing the quirks bitmask directly in
.driver_data as was done before.
It's a slight wastage to keep the, now unnecessary, indirection,
though the size impact is small
and it's only used at probe() time not on a hot path.
But switching back could be seen as code churn too...
I don't have a strong opinion on this, so just noting it to see what
others think.
Martin
Powered by blists - more mailing lists