[PATCH] libata-sff: trivial corrections to Kconfig help text

[PATCH] libata-sff: trivial corrections to Kconfig help text

[Stefan Richter]

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
---
 drivers/ata/Kconfig |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/ata/Kconfig b/drivers/ata/Kconfig
index 73f8833..aa85a98 100644
--- a/drivers/ata/Kconfig
+++ b/drivers/ata/Kconfig
@@ -168,10 +168,10 @@ config ATA_BMDMA
 	default y
 	help
 	  This option adds support for SFF ATA controllers with BMDMA
-	  capability.  BMDMA stands for bus-master DMA and the
-	  de-facto DMA interface for SFF controllers.
+	  capability.  BMDMA stands for bus-master DMA and is the
+	  de facto DMA interface for SFF controllers.
 
-	  If unuser, say Y.
+	  If unsure, say Y.
 
 if ATA_BMDMA
 

By the way, is there something that could be said in a few words in the
help text when to enable that option and when not?

Or could that option even be hidden and 'select'ed by bus master DMA
capable ATA SFF drivers?
-- 
Stefan Richter
-=====-==-=- -=-= ===-=
http://arcgraph.de/sr/

Re: [PATCH] libata-sff: trivial corrections to Kconfig help text

[Tejun Heo]

On 05/29/2010 12:47 PM, Stefan Richter wrote:
> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

Acked-by: Tejun Heo <tj@kernel.org>

> ---
>  drivers/ata/Kconfig |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/ata/Kconfig b/drivers/ata/Kconfig
> index 73f8833..aa85a98 100644
> --- a/drivers/ata/Kconfig
> +++ b/drivers/ata/Kconfig
> @@ -168,10 +168,10 @@ config ATA_BMDMA
>  	default y
>  	help
>  	  This option adds support for SFF ATA controllers with BMDMA
> -	  capability.  BMDMA stands for bus-master DMA and the
> -	  de-facto DMA interface for SFF controllers.
> +	  capability.  BMDMA stands for bus-master DMA and is the
> +	  de facto DMA interface for SFF controllers.
>  
> -	  If unuser, say Y.
> +	  If unsure, say Y.
>  
>  if ATA_BMDMA
>  
> 
> By the way, is there something that could be said in a few words in the
> help text when to enable that option and when not?

Hmmmm... maybe we can emphasize the 'if unsure, say Y' part?

> Or could that option even be hidden and 'select'ed by bus master DMA
> capable ATA SFF drivers?

I wanted to make the distinction clear so made the option explicit.
Opinions on whether that is actually a good idea would differ but at
the same time, I don't think this would cause any problem, would it?

Thanks.

-- 
tejun

Re: [PATCH] libata-sff: trivial corrections to Kconfig help text

[Stefan Richter]

Tejun Heo wrote:
> On 05/29/2010 12:47 PM, Stefan Richter wrote:
>> By the way, is there something that could be said in a few words in the
>> help text when to enable that option and when not?
> 
> Hmmmm... maybe we can emphasize the 'if unsure, say Y' part?

I did follow that recommendation but was still left curious. :-)

>> Or could that option even be hidden and 'select'ed by bus master DMA
>> capable ATA SFF drivers?
> 
> I wanted to make the distinction clear so made the option explicit.
> Opinions on whether that is actually a good idea would differ but at
> the same time, I don't think this would cause any problem, would it?

OK, I first came across it by way of make oldconfig.  Then I read the
changelog and still wondered.  Now I had a look at the menu layout in
gconfig and it makes sense to me now.

Just from the looks of the menu, would calling the prompt

	bool "SFF controllers with bus master DMA"

be sensible?  After all, this option does not just control whether a
certain capability is built into libata, switching it on pulls up the
Kconfig prompts for a whole range of controller drivers.  (Ok, the help
text is saying exactly this, but then the prompt could do so already...)

The 'hidden variable + select' way would be more friendly to "make
oldconfig", but I see now that it would take quite a few select
statements to be added.
-- 
Stefan Richter
-=====-==-=- -=-= ===-=
http://arcgraph.de/sr/

Re: [PATCH] libata-sff: trivial corrections to Kconfig help text

[Tejun Heo]

Hello,

On 05/29/2010 05:41 PM, Stefan Richter wrote:
>> Hmmmm... maybe we can emphasize the 'if unsure, say Y' part?
> 
> I did follow that recommendation but was still left curious. :-)

Maybe it should be changed to "You're supposed to be unsure, say Y,
dammit!" :-)

>>> Or could that option even be hidden and 'select'ed by bus master DMA
>>> capable ATA SFF drivers?
>>
>> I wanted to make the distinction clear so made the option explicit.
>> Opinions on whether that is actually a good idea would differ but at
>> the same time, I don't think this would cause any problem, would it?
> 
> OK, I first came across it by way of make oldconfig.  Then I read the
> changelog and still wondered.  Now I had a look at the menu layout in
> gconfig and it makes sense to me now.
> 
> Just from the looks of the menu, would calling the prompt
> 
> 	bool "SFF controllers with bus master DMA"
> 
> be sensible?  After all, this option does not just control whether a
> certain capability is built into libata, switching it on pulls up the
> Kconfig prompts for a whole range of controller drivers.  (Ok, the help
> text is saying exactly this, but then the prompt could do so already...)

Yes, sure.  Can you please combine this with the original patch?

> The 'hidden variable + select' way would be more friendly to "make
> oldconfig", but I see now that it would take quite a few select
> statements to be added.

Yeah, that and I *really* wanted to make the distinction explicit
because ambiguous dependeny has been a problem for some time.  Now a
driver will end up in a different config section which is always
visible depending on its dependencies, so I like this much better.

Thanks.

-- 
tejun

[PATCH] libata-sff: clarification and trivial correction to Kconfig text

[Stefan Richter]

Change the text of the ATA_BMDMA prompt to indicate that it lets one access a class of
device drivers (rather than only add a certain feature to libata).  The divisions of
the "ATA SFF support" menu now read
  - SFF controllers with custom DMA interface,
  - SFF controllers with bus master DMA  (was: "ATA BMDMA support"),
      - SATA SFF controllers with BMDMA,
      - PATA SFF controllers with BMDMA,
  - PIO-only SFF controllers.
This hopefully helps convince users of "make oldconfig" to follow the
recommendation to switch that option on.

Also let the help text of this prompt say more precisely what it is about:  It adds a
little bit of support code to libata that is required by many but not all SFF drivers.
This option alone does not add support of respective controllers though; according
drivers are chosen at subsequent dependent Kconfig prompts. --- And fix a typo.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
---
Supersedes patch "libata-sff: trivial corrections to Kconfig help text".

 drivers/ata/Kconfig |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/ata/Kconfig b/drivers/ata/Kconfig
index 73f8833..31e288b 100644
--- a/drivers/ata/Kconfig
+++ b/drivers/ata/Kconfig
@@ -164,14 +164,14 @@ config SATA_SX4
 	  If unsure, say N.
 
 config ATA_BMDMA
-	bool "ATA BMDMA support"
+	bool "SFF controllers with bus master DMA"
 	default y
 	help
-	  This option adds support for SFF ATA controllers with BMDMA
-	  capability.  BMDMA stands for bus-master DMA and the
-	  de-facto DMA interface for SFF controllers.
+	  This option is required by drivers of SFF ATA controllers with
+	  BMDMA capability.  BMDMA stands for bus master DMA and is the
+	  de facto DMA interface for SFF controllers.
 
-	  If unuser, say Y.
+	  If unsure, say Y.
 
 if ATA_BMDMA
 
-- 
Stefan Richter
-=====-==-=- -=-= ====-
http://arcgraph.de/sr/

Re: [PATCH] libata-sff: clarification and trivial correction to Kconfig text

[Tejun Heo]

On 05/30/2010 04:00 PM, Stefan Richter wrote:
> Change the text of the ATA_BMDMA prompt to indicate that it lets one access a class of
> device drivers (rather than only add a certain feature to libata).  The divisions of
> the "ATA SFF support" menu now read
>   - SFF controllers with custom DMA interface,
>   - SFF controllers with bus master DMA  (was: "ATA BMDMA support"),
>       - SATA SFF controllers with BMDMA,
>       - PATA SFF controllers with BMDMA,
>   - PIO-only SFF controllers.
> This hopefully helps convince users of "make oldconfig" to follow the
> recommendation to switch that option on.
> 
> Also let the help text of this prompt say more precisely what it is about:  It adds a
> little bit of support code to libata that is required by many but not all SFF drivers.
> This option alone does not add support of respective controllers though; according
> drivers are chosen at subsequent dependent Kconfig prompts. --- And fix a typo.
> 
> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

Re: [PATCH] libata-sff: trivial corrections to Kconfig help text

[Jeff Garzik]

On 05/29/2010 06:47 AM, Stefan Richter wrote:
> Signed-off-by: Stefan Richter<stefanr@s5r6.in-berlin.de>
> ---
>   drivers/ata/Kconfig |    6 +++---
>   1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/ata/Kconfig b/drivers/ata/Kconfig
> index 73f8833..aa85a98 100644
> --- a/drivers/ata/Kconfig
> +++ b/drivers/ata/Kconfig
> @@ -168,10 +168,10 @@ config ATA_BMDMA
>   	default y
>   	help
>   	  This option adds support for SFF ATA controllers with BMDMA
> -	  capability.  BMDMA stands for bus-master DMA and the
> -	  de-facto DMA interface for SFF controllers.
> +	  capability.  BMDMA stands for bus-master DMA and is the
> +	  de facto DMA interface for SFF controllers.
>
> -	  If unuser, say Y.
> +	  If unsure, say Y.
>
>   if ATA_BMDMA

applied