Patchwork cdrom: support devices that have check_events but not media_changed

login
register
mail settings
Submitter Simon Arlott
Date Jan. 30, 2011, 1:31 a.m.
Message ID <4D44BF64.4060205@simon.arlott.org.uk>
Download mbox | patch
Permalink /patch/80993/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

Simon Arlott - Jan. 30, 2011, 1:31 a.m.
93aae17af1172c40c6f74b7294e93a90c3cfaa5d replaced the media_changed op
with the check_events op in drivers/scsi/sr.c

All users that check for the CDC_MEDIA_CHANGED capability try both
the check_events op and the media_changed op, but register_cdrom()
was requiring media_changed.

This patch fixes the capability checking and removes a redundant
check that media_changed != NULL.

The cdrom_select_disc ioctl is also using the two operations, so
they should be required for CDC_SELECT_DISC too.

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Tejun Heo <tj@kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
---
 drivers/cdrom/cdrom.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
Tejun Heo - Jan. 31, 2011, 10:12 a.m.
Hello,

On Sun, Jan 30, 2011 at 01:31:16AM +0000, Simon Arlott wrote:
> 93aae17af1172c40c6f74b7294e93a90c3cfaa5d replaced the media_changed op
> with the check_events op in drivers/scsi/sr.c
> 
> All users that check for the CDC_MEDIA_CHANGED capability try both
> the check_events op and the media_changed op, but register_cdrom()
> was requiring media_changed.
> 
> This patch fixes the capability checking and removes a redundant
> check that media_changed != NULL.
> 
> The cdrom_select_disc ioctl is also using the two operations, so
> they should be required for CDC_SELECT_DISC too.

First of all, thanks a lot for fixing this.  Just minor nits below.

> diff --git a/drivers/cdrom/cdrom.c b/drivers/cdrom/cdrom.c
> index 14033a3..4716d76 100644
> --- a/drivers/cdrom/cdrom.c
> +++ b/drivers/cdrom/cdrom.c
> @@ -409,7 +409,9 @@ int register_cdrom(struct cdrom_device_info *cdi)
>  	}
>  
>  	ENSURE(drive_status, CDC_DRIVE_STATUS );
> -	ENSURE(media_changed, CDC_MEDIA_CHANGED);
> +	if (cdo->check_events == NULL && cdo->media_changed == NULL) {
> +		*change_capability = ~(CDC_MEDIA_CHANGED | CDC_SELECT_DISC);
> +	}

Can we lose the unnecessary {}?

>  	ENSURE(tray_move, CDC_CLOSE_TRAY | CDC_OPEN_TRAY);
>  	ENSURE(lock_door, CDC_LOCK);
>  	ENSURE(select_speed, CDC_SELECT_SPEED);
> @@ -1471,7 +1473,7 @@ int cdrom_media_changed(struct cdrom_device_info *cdi)
>  	/* This talks to the VFS, which doesn't like errors - just 1 or 0.  
>  	 * Returning "0" is always safe (media hasn't been changed). Do that 
>  	 * if the low-level cdrom driver dosn't support media changed. */ 
> -	if (cdi == NULL || cdi->ops->media_changed == NULL)
> +	if (cdi == NULL)

This change is probably not necessary as cdrom_media_changed() is only
called from drivers which implement media_changed.  I don't think it's
gonna break anything but the whole media_changed stuff is scheduled
for removal anyway, so I think it would be better to keep the changes
minimal.

Thank you.
Sergei Shtylyov - Jan. 31, 2011, 11:22 a.m.
Hello.

On 30-01-2011 4:31, Simon Arlott wrote:

> 93aae17af1172c40c6f74b7294e93a90c3cfaa5d replaced the media_changed op

    Please specify the commit summary in parens, as asked by Linus.

> with the check_events op in drivers/scsi/sr.c

> All users that check for the CDC_MEDIA_CHANGED capability try both
> the check_events op and the media_changed op, but register_cdrom()
> was requiring media_changed.

> This patch fixes the capability checking and removes a redundant
> check that media_changed != NULL.

> The cdrom_select_disc ioctl is also using the two operations, so
> they should be required for CDC_SELECT_DISC too.

> Signed-off-by: Simon Arlott<simon@fire.lp0.eu>
> Cc: Jens Axboe<axboe@kernel.dk>
> Cc: Tejun Heo<tj@kernel.org>
> Cc: Kay Sievers<kay.sievers@vrfy.org>
[...]

> diff --git a/drivers/cdrom/cdrom.c b/drivers/cdrom/cdrom.c
> index 14033a3..4716d76 100644
> --- a/drivers/cdrom/cdrom.c
> +++ b/drivers/cdrom/cdrom.c
> @@ -409,7 +409,9 @@ int register_cdrom(struct cdrom_device_info *cdi)
>   	}
>
>   	ENSURE(drive_status, CDC_DRIVE_STATUS );
> -	ENSURE(media_changed, CDC_MEDIA_CHANGED);
> +	if (cdo->check_events == NULL && cdo->media_changed == NULL) {
> +		*change_capability = ~(CDC_MEDIA_CHANGED | CDC_SELECT_DISC);
> +	}

    {} are not needed here. I think checkstatus.pl should detect this...

WBR, Sergei
--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/cdrom/cdrom.c b/drivers/cdrom/cdrom.c
index 14033a3..4716d76 100644
--- a/drivers/cdrom/cdrom.c
+++ b/drivers/cdrom/cdrom.c
@@ -409,7 +409,9 @@  int register_cdrom(struct cdrom_device_info *cdi)
 	}
 
 	ENSURE(drive_status, CDC_DRIVE_STATUS );
-	ENSURE(media_changed, CDC_MEDIA_CHANGED);
+	if (cdo->check_events == NULL && cdo->media_changed == NULL) {
+		*change_capability = ~(CDC_MEDIA_CHANGED | CDC_SELECT_DISC);
+	}
 	ENSURE(tray_move, CDC_CLOSE_TRAY | CDC_OPEN_TRAY);
 	ENSURE(lock_door, CDC_LOCK);
 	ENSURE(select_speed, CDC_SELECT_SPEED);
@@ -1471,7 +1473,7 @@  int cdrom_media_changed(struct cdrom_device_info *cdi)
 	/* This talks to the VFS, which doesn't like errors - just 1 or 0.  
 	 * Returning "0" is always safe (media hasn't been changed). Do that 
 	 * if the low-level cdrom driver dosn't support media changed. */ 
-	if (cdi == NULL || cdi->ops->media_changed == NULL)
+	if (cdi == NULL)
 		return 0;
 	if (!CDROM_CAN(CDC_MEDIA_CHANGED))
 		return 0;