[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+sbYW2SMbExwyw20TZtn3U13fkJvKCs5P8w5LdovN4Q3u0sjg@mail.gmail.com>
Date: Mon, 16 Jan 2017 13:38:40 +0530
From: Selvin Xavier <selvin.xavier@...adcom.com>
To: Leon Romanovsky <leon@...nel.org>
Cc: Doug Ledford <dledford@...hat.com>, linux-rdma@...r.kernel.org,
Linux Netdev List <netdev@...r.kernel.org>,
Michael Chan <michael.chan@...adcom.com>,
Eddie Wai <eddie.wai@...adcom.com>,
Devesh Sharma <devesh.sharma@...adcom.com>,
Somnath Kotur <somnath.kotur@...adcom.com>,
Sriharsha Basavapatna <sriharsha.basavapatna@...adcom.com>
Subject: Re: [PATCH for bnxt_re V3 03/21] bnxt_re: register with the NIC driver
On Mon, Jan 16, 2017 at 1:11 AM, Leon Romanovsky <leon@...nel.org> wrote:
> Recently, in one of our submission to netdev and rdma, we got
> a reminder that inline functions shouldn't be in *.c. Let for
> the compiler to decide.
>
> IMHO, it should be open-coded without wrappers and honestly I failed to
> understand why do you need so many wrappers for one line standard kernel
> functions.
Agreed. During the initial days of development, we had some debug code also
inside these wrapper functions. Since the debug code was removed before upstream
submission, this function ended up as a wrapper for standard kernel function.
It makes perfect sense to knock off these wrapper functions. I will
queue it for next
clean-up/bug fix patch series.
Thanks
Powered by blists - more mailing lists