[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20191107.154922.1123372183066604716.davem@davemloft.net>
Date: Thu, 07 Nov 2019 15:49:22 -0800 (PST)
From: David Miller <davem@...emloft.net>
To: Mark-MC.Lee@...iatek.com
Cc: sean.wang@...iatek.com, john@...ozen.org, matthias.bgg@...il.com,
andrew@...n.ch, robh+dt@...nel.org, mark.rutland@....com,
opensource@...rst.com, devicetree@...r.kernel.org,
netdev@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org,
jakub.kicinski@...ronome.com
Subject: Re: [PATCH net] net: ethernet: mediatek: rework GDM setup flow
From: MarkLee <Mark-MC.Lee@...iatek.com>
Date: Thu, 7 Nov 2019 18:51:35 +0800
> + for (i = 0; i < 2; i++) {
This is a regression, because in the existing code...
> - for (i = 0; i < MTK_MAC_COUNT; i++) {
the proper macro is used instead of a magic constant.
You're doing so many things in one change, it's hard to review
and audit.
If you're going to consolidate code, do that only in one change.
Then make other functional changes such as putting the chip into
GDMA_DROP_ALL mode during the stop operation etc.
Powered by blists - more mailing lists