Re: [RFC]hpt366/ide-probe reset drive on probe error.

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

On Saturday 09 May 2009 14:09:05 Karl Hiramoto wrote:
> Sergei Shtylyov wrote:

[...]

> >>> Care to revise your patch?
> >
> >    I see no point in "revising" this hack now...
> 
> Why not reset the drive if it does not respond?  OK, you could blame 
> this on a bugy board, or redboot that does not reset the drive on warm 
> boot if it is busy.

Please read the whole discussion between Sergei & me before getting
discouraged by our occasional use of the vivid language. ;)

Your idea is good but the current implementation needs to be reworked
into a form which doesn't duplicate the code and which would work also
for other hardware setups.

The nice starting point is sanitizing return values of do_probe()
and probe_for_drive(), and then making sure that they are propagated
correctly to higher level.  Once this is done we can use those values
in ide_probe_port() to reset and re-probe the port if necessary.

[ I think that the port reset itself would be best realized by reusing
  the existing code from do_reset1() (plese see the code for SRST near
  the end of this function). ]

Thanks,
Bart