Re: [PATCH] drivers/block/scsi_ioctl.c, Video DVD playback support

[Jens Axboe <axboe@suse.de>]

On Tue, Jan 25 2005, Elias da Silva wrote:
> On Tuesday 25 January 2005 13:45, you wrote:
> [snip]
> : If I'm not mistaken, Peter Jones has posted a few iterations of such an
> : fs some months ago.
> 
> Thank you. I will check this...
> 
> : > Do we have a clear understanding that this fs would only
> : > be a benefit if *All* the different ways to access the device would
> : > use the same policy enforcement and consistently allow or
> : > disallow certain operations regardless of the access method?
> : 
> : The command restriction table _only_ works through the SG_IO path, which
> : does include CDROM_SEND_PACKET as well since it is layered on top of
> : SG_IO. It doesn't control various driver ioctl exported interfaces, they
> : would need to add a callback to verify_command() for permission checks.
> 
> Hmm... what exactly does that mean? Who is ment by "_they_ would need..."?

It refers back to 'various driver ioctl' earlier, so it refers to the
driver itself.

-- 
Jens Axboe