Thread (17 messages) 17 messages, 4 authors, 2016-02-17

Re: [PATCH V3 1/6] stm class: Add ioctl get_options interface

From: Mathieu Poirier <mathieu.poirier@linaro.org>
Date: 2016-02-12 18:33:21
Also in: linux-arm-kernel, lkml

On 12 February 2016 at 08:18, Alexander Shishkin
[off-list ref] wrote:
Chunyan Zhang [off-list ref] writes:
quoted
There is already an interface of set_options, but no get_options yet.
Before setting any options, one would may want to see the current
status of that option by means of get_options interface. This
interface has been used in CoreSight STM driver.
I'm not sure I understand the reasoning behind this. If a userspace
program opens a communication channel and wants to configure certain
features on it, why does its choice depend on what has been configured
for this channel previously? It can be anything at all. Most likely,
it's either unconfigured or configured to its default values, but why
does this matter for a new writer?
A client may wish to change the settings (invariant/guaranteed) it has
on a specific channel - it may even want to so do multiple times.  The
idea behind introducing a get_options() was to probe the specific
settings of a channel before going a head with a new configuration.
In hindsight it may not be needed as a client should simply go ahead
and push down the configuration it wants.
Regards,
--
Alex
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help