linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Mauro Carvalho Chehab <mchehab@s-opensource.com>
To: Sean Young <sean@mess.org>
Cc: Linux Media Mailing List <linux-media@vger.kernel.org>,
	Mauro Carvalho Chehab <mchehab@infradead.org>
Subject: Re: [PATCH 03/20] [media] lirc.h: remove several unused ioctls
Date: Tue, 12 Jul 2016 11:05:24 -0300	[thread overview]
Message-ID: <20160712110524.57e2b674@recife.lan> (raw)
In-Reply-To: <20160712135438.GA11183@gofer.mess.org>

Em Tue, 12 Jul 2016 14:54:38 +0100
Sean Young <sean@mess.org> escreveu:

> On Tue, Jul 12, 2016 at 02:35:56PM +0100, Sean Young wrote:
> > On Tue, Jul 12, 2016 at 10:23:00AM -0300, Mauro Carvalho Chehab wrote:  
> > > Em Tue, 12 Jul 2016 14:14:06 +0100
> > > Sean Young <sean@mess.org> escreveu:  
> > > > On Tue, Jul 12, 2016 at 09:41:57AM -0300, Mauro Carvalho Chehab wrote:  
> > -snip-  
> > > > > -#define LIRC_SET_REC_DUTY_CYCLE        _IOW('i', 0x00000016, __u32)    
> > > > 
> > > > Also remove LIRC_CAN_SET_REC_DUTY_CYCLE and 
> > > > LIRC_CAN_SET_REC_DUTY_CYCLE_RANGE.  
> > > 
> > > Removing the "LIRC_CAN" macros can break userspace, as some app could
> > > be using it to print the LIRC features. That's why I opted to keep
> > > them, but to document that those features are unused - this is at
> > > the next patch (04/20).  
> > 
> > How is that different from removing the ioctls? Might as well go the whole
> > hog.  

If someone implemented LIRC_GET_FEATURES and handled all flags, such
program would break, as the API change.

> 
> Ah you meant that if someone later adds a new feature then we might reuse
> an existing bit. Oops, sorry.

Yes, this is another possibility. In such case, the ABI will also 
break, with is more severe than a pure API change.

Removing the ioctl declarations will still work for compiled programs,
as they'll still receive an error code when the ioctl is issued.

> 
> > Also note that LIRC_CAN_SET_REC_DUTY_CYCLE has the same value as
> > LIRC_CAN_MEASURE_CARRIER, so if some userspace program uses this it might
> > end up in the mistaken belief its supports LIRC_CAN_SET_REC_DUTY_CYCLE.  
> 
> So there is an argument for removing LIRC_CAN_SET_REC_DUTY_CYCLE, but
> that should be a separate patch.

Yes. Yet, IMHO, the best would be to put those unused LIRC_CAN into a:

#ifndef __KERNEL

macro block, to:

1) avoid the risk of breaking userspace;
2) be clear that those are deprecated stuff and should not be used on
   newer programs;
3) Reserve the bits for not be used, to avoid possible conflicts.

Regards,
Mauro

  reply	other threads:[~2016-07-12 14:05 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-07-12 12:41 [PATCH 00/20] Improve LIRC documentation Mauro Carvalho Chehab
2016-07-12 12:41 ` [PATCH 01/20] [media] doc-rst: Document ioctl LIRC_GET_FEATURES Mauro Carvalho Chehab
2016-07-12 12:41 ` [PATCH 02/20] [media] doc-rst: add media/uapi/rc/lirc-header.rst Mauro Carvalho Chehab
2016-07-12 12:41 ` [PATCH 03/20] [media] lirc.h: remove several unused ioctls Mauro Carvalho Chehab
2016-07-12 13:14   ` Sean Young
2016-07-12 13:23     ` Mauro Carvalho Chehab
2016-07-12 13:35       ` Sean Young
2016-07-12 13:54         ` Sean Young
2016-07-12 14:05           ` Mauro Carvalho Chehab [this message]
2016-07-12 12:41 ` [PATCH 04/20] [media] doc-rst: remove not used ioctls from documentation Mauro Carvalho Chehab
2016-07-12 12:41 ` [PATCH 05/20] [media] doc-rst: Fix LIRC_GET_FEATURES references Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 06/20] [media] doc-rst: document ioctl LIRC_GET_SEND_MODE Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 07/20] [media] doc-rst: fix some lirc cross-references Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 08/20] [media] doc-rst: document ioctl LIRC_GET_REC_MODE Mauro Carvalho Chehab
2016-07-12 13:01   ` Sean Young
2016-07-12 13:27     ` Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 09/20] [media] doc-rst: document LIRC_GET_REC_RESOLUTION Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 10/20] [media] doc-rst: document LIRC_SET_SEND_DUTY_CYCLE Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 11/20] [media] doc-rst: document LIRC_GET_*_TIMEOUT ioctls Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 12/20] [media] doc-rst: Document LIRC_GET_LENGTH ioctl Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 13/20] [media] doc-rst: document LIRC set carrier ioctls Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 14/20] [media] doc-rst: document LIRC_SET_TRANSMITTER_MASK Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 15/20] [media] doc-rst: document LIRC_SET_REC_TIMEOUT Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 16/20] [media] doc-rst: document LIRC_SET_REC_TIMEOUT_REPORTS Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 17/20] [media] doc-rst: add documentation for LIRC_SET_MEASURE_CARRIER_MODE Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 18/20] [media] doc-rst: document LIRC_SET_WIDEBAND_RECEIVER Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 19/20] [media] doc-rst: Document LIRC set mode ioctls Mauro Carvalho Chehab
2016-07-12 12:42 ` [PATCH 20/20] [media] doc-rst: reorganize LIRC ReST files Mauro Carvalho Chehab

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160712110524.57e2b674@recife.lan \
    --to=mchehab@s-opensource.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@infradead.org \
    --cc=sean@mess.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).