Re: [patch] libata: add ioctls to support SMART

[Jeff Garzik <jgarzik@pobox.com>]

Brad Campbell wrote:
> John W. Linville wrote:
> 
>> Support for HDIO_DRIVE_CMD and HDIO_DRIVE_TASK in libata.  Useful for
>> supporting SMART w/ unmodified smartctl and smartd userland binaries.
>>
>> Not happy w/ loop after failed ata_qc_new_init(), but needed because 
>> smartctl
>> and smartd did not retry after failure.  Likely need an option to wait 
>> for
>> available qc?  Also not sure all the error return codes are correct...
>>
> 
> YYYYYYYYYYEeeeeeeeeeeeeeeeeeeeeeeeeeehhhhhhhhhhaaaaaaaaaaaaaaaaaaa!!!!!!!!
> 
> I know it's a bit kludgy and does not really fit the philosophy of 
> libata but it works and it lets me keep an eye on my drives *now*.
> 
> Although just for good measure I'll probably unmount and stop my raid 
> arrays before I use it on the disks. Whats it like for locking on a busy 
> system?

I wouldn't trust it on a busy system yet -- it submits the command to 
the device without checking if there is a command already outstanding.

The patch could _definitely_ corrupt data or lock your hardware, since 
it bypasses the SCSI mechanism that ensures that only one command is 
executing at a time.

	Jeff