cvs commit: src/sys/dev/ata ata-all.c ata-all.h ata-dma.c ata-pci.c

Yar Tikhiy yar at comp.chem.msu.su
Thu Apr 21 19:22:25 PDT 2005


On Thu, Apr 21, 2005 at 07:13:28AM -0700, Doug Ambrisko wrote:
> Yar Tikhiy writes:
> | On Wed, Apr 20, 2005 at 11:51:32PM +0000, Doug Ambrisko wrote:
> | > ambrisko    2005-04-20 23:51:32 UTC
> | > 
> | >   FreeBSD src repository
> | > 
> | >   Modified files:        (Branch: RELENG_4)
> | >     sys/dev/ata          ata-all.c ata-all.h ata-dma.c ata-pci.c 
> | >   Log:
> | >   Add in Promise and Intel SATA support based on FreeBSD 5.X ATAng.
> | >   This gets the HW to work but not to deal with SATA errors.  So if
> | >   the drive has a read error it will probably lock up.  This is
> | >   a small subset of my patches that do deal with these errors and
> | >   other issues in FreeBSD and ata-raid.
> | >   
> | >   Sponsored by:   IronPort Systems
> | >   
> | >   Revision   Changes    Path
> | >   1.50.2.49  +168 -36   src/sys/dev/ata/ata-all.c
> | >   1.26.2.15  +166 -13   src/sys/dev/ata/ata-all.h
> | >   1.35.2.39  +95 -17    src/sys/dev/ata/ata-dma.c
> | >   1.32.2.22  +164 -18   src/sys/dev/ata/ata-pci.c
> | 
> | I myself was bitten by RELENG_4 ATA support lagging behind modern
> | hardware, which made me move to RELENG_5 on some systems at last.
> | However, isn't your change too big and possibly destabilizing for
> | the conservative RELENG_4 branch?  I'm judging by the number of
> | changed lines only.  And honestly, I'm having trouble understanding
> | from the commit message whether your change introduces or corrects
> | the bug in SATA error detection.
> 
> It makes some Promise HW work that wouldn't probe at all.  For the
> Intel stuff adds device ID's.  Most of the changed lines are a
> mechanical type thing which is using the new indirection macros so
> we can probe HW like that in the Promise cards.  The core ATA code
> didn't change.  Addition code was required to make the Promise HW work.
> 
> It is based on a subset of stuff at:
> 	http://www.ambrisko.com/doug/ata/
> which has been running on lots of machines.  This patchset which is
> approaching 140k has lots of various panic (ata & FreeBSD) fixes and 
> enhanced SATA & ata-raid support that prevents systems from freezing, 
> crashing and automatic drive recovery in mirrors.  I don't plan to be
> merge all of that in, just the basic HW support so people can install.
> 
> It looks like in ata mkIII there is better support of SATA & RAID stuff 
> going in so that looks encouraging.

Thank you for the explanation!  Now I see that the only issue about
your change is that it will let some feeble folks avoid their destiny
of running RELENG_5 or CURRENT on their cool brand-new production
systems ;-)

-- 
Yar


More information about the cvs-src mailing list