Add warning notes for known quirkly ata chipsets

[Mark Kirkwood]

Mark Kirkwood wrote:
> Do folks think that amending the ata(4) man page, and adding warnings 
> beside  known quirky or broken chipsets is a good idea? I'm thinking of 
> Si 3112 here in particular(!) - the aim being to discourage folks from 
> buying 'em based on the fact they appear in ata(4).
> 
> I'm happy to draft up something - altho I'd be pretty basic like:

A tentative patch, adding a BUGS section instead of lumping it into 
NOTES (see attached).

regards

Mark
-------------- next part --------------
*** ata.4.orig	Tue Mar  7 13:00:17 2006
--- ata.4	Tue Mar  7 15:21:12 2006
***************
*** 135,141 ****
  .It ServerWorks:
  ROSB4, CSB5, CSB6.
  .It Silicon Image:
! SiI0680, SiI3112, SiI3114, SiI3512.
  .It SiS:
  SIS180, SIS181, SIS182, SIS5513, SIS530, SIS540, SIS550, SIS620, SIS630, SIS630S, SIS633, SIS635, SIS730, SIS733, SIS735, SIS745, SIS961, SIS962, SIS963, SIS964, SIS965.
  .It VIA:
--- 135,141 ----
  .It ServerWorks:
  ROSB4, CSB5, CSB6.
  .It Silicon Image:
! SiI0680, SiI3112 (see BUGS), SiI3114, SiI3512.
  .It SiS:
  SIS180, SIS181, SIS182, SIS5513, SIS530, SIS540, SIS550, SIS620, SIS630, SIS630S, SIS633, SIS635, SIS730, SIS733, SIS735, SIS745, SIS961, SIS962, SIS963, SIS964, SIS965.
  .It VIA:
***************
*** 216,221 ****
--- 216,225 ----
  This is useful in hotswap scenarios
  where disks should always show up as the same numbered device,
  and not depend on attach order.
+ .Sh BUGS
+ There have been reports of data corruption and erratic behavior with some 
+ controllers which seem to be related to deficiencies with the chipset design 
+ rather than the driver code. It is recommended to avoid these: SiI3112.
  .Sh SEE ALSO
  .Xr ataraid 4 ,
  .Xr atacontrol 8 ,