[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <s5hefzfivfh.wl-tiwai@suse.de>
Date: Fri, 03 Feb 2017 08:29:38 +0100
From: Takashi Iwai <tiwai@...e.de>
To: "Satendra Singh Thakur" <satendra.t@...sung.com>
Cc: <alsa-devel@...a-project.org>, <perex@...ex.cz>,
<o-takashi@...amocchi.jp>, <hemanshu.s@...sung.com>,
<nishant.s4@...sung.com>, <linux-kernel@...r.kernel.org>
Subject: Re: [RFC v2] [ALSA][CONTROL] 1. Added 2 ioctls for reading/writing values of multiple controls in one go (at once) 2. Restructured the code for read/write of a single control
On Fri, 03 Feb 2017 06:28:08 +0100,
Satendra Singh Thakur wrote:
>
> From: satendra singh thakur <satendra.t@...sung.com>
>
> 1.Added 2 ioctls in alsa driver's control interface
> -Added an ioctl to read values of multiple elements at once
> -Added an ioctl to write values of multiple elements at once
> -In the absence of above ioctls user needs to call N ioctls to
> read/write value of N elements which requires N context switches
> -Proposed ioctls will allow accessing N elements' values in a single
> context switch
> -Above mentioned ioctl will be useful for alsa utils such as amixer
> which reads all controls of given sound card
> 2. Restructured/Combined two functions snd_ctl_elem_read_user and
> snd_ctl_elem_write_user into a single function snd_ctl_elem_rw_user
> -These functions were having most of the code which was similar
> to each other
> -Thus, there was redundant/duplicate code which was removed
> and a single function was formed/defined.
> -Removed functions snd_ctl_elem_read_user and snd_ctl_elem_write_user
> having redundant/duplicate code
> 3. This is version 2 of RFC, here we combine previous 2 patches
> submitted yesterday (02-Feb) with prefix [RFC] [ALSA][CONTROL]
> -Fixed doxygen comments related warnings
> -Fixed stack frame related warning
>
> Signed-off-by: Satendra Singh Thakur <satendra.t@...sung.com>
Thanks for the patch.
But, could you answer to the question Clemens posted for v1 patch?
Namely, do N-times context switching really matter? Does it give the
severe performance issue?
If we have a measured number, this thing is worth to consider.
OTOH, without the actual measurement but "just because it must be
better", it's no good reason for adding a new API/ABI.
thanks,
Takashi
Powered by blists - more mailing lists