Re: Testing CAM wrapper for ata(4) controller drivers

Alexander Motin schrieb am 2009-12-06:
> Alexander Best wrote:
> > i've stumbled over a problem with ATA_CAM. when i accessing my
> > dvd-drive i'm
> > not able to access any other cam members (ada* e.g.).

> > when i issue the following command:

> > `cdrecord dev=2,1,0 blank=all'

> > binaries such as ls, top or vmstat simply stall until cdrecord
> > finishes.

> > this is the output of `camcontrol devlist -v':

> > scbus2 on ata2 bus 0:
> > <HDS722516VLAT80 V34OA63A>         at scbus2 target 0 lun 0
> > (ada0,pass0)
> > <HL-DT-ST DVDRAM GSA-H10N JL12>    at scbus2 target 1 lun 0
> > (pass1,cd0)
> > <>                                 at scbus2 target -1 lun -1 ()

> It was always bad idea to put CD and disk, which need to be used
> simultaneously, to the same PATA channel. PATA channel can execute
> only
> one command at a time, without using special technique, named overlap
> (part of TCQ), which is usually not supported by devices. When your
> blank command occupied CD, disk became inaccessible, until it
> complete.

thanks for the explanation. unfortunately my mothboard only comes with a
single PATA bus, but i'm trying to completely move from a SATA/PATA mixed
setup to a pure SATA setup this month. that should solve the issue then. i
just thought this was an issue introduced by ATA_CAM.

i would also like to ask: what will become of burncd in the future? will it be
adjusted to work with CAM? will it be removed/moved to ports? or will we
replace it with cdrecord/something else?

cheers.
alex