Re: MFC of "Unbreak Promise SATAII/150 controllers" ?

Thierry Herbelot wrote:
> Hello,
>
> My main server runs Releng-6 and has SATA disks driven by a Promise SATA 
> controller.
>
> Recent releng-6 kernels (6.1-Beta) lose access to the disks with messages 
> like :
> ad10: req=0xc2771e10 SETFEATURES SET TRANSFER MODE semaphore timeout !! DANGER 
> Will Robinson !!
> (the disks themselves are new and have been checked recently : they have been 
> converted to a g_mirror setup).
>
> After searching in cvs archives, I have found this commit for -Current :
> http://lists.freebsd.org/pipermail/cvs-src/2006-February/059647.html
>
>   Modified files:
>     sys/dev/ata          ata-all.c ata-all.h ata-chipset.c 
>   Log:
>   Unbreak Promise SATAII/150 controllers caused by the DMA dump changes.
>
> a quick check with a recent (yesterday's) -current kernel (running in the 
> place of releng_6) seems to show that -current is fixed (at least a 'ls 
> -lR /big_part' runs as expected).
>
> Is there any plan to MFC this commit ?
>   
Yes, and I got ready for it just as the announce of BETA2 builds hit 
developers. It'll get in as soon as I have a spare minute to catch up as 
there now a a few more things to MFC...


-Søren