[Libburn] lite-on 52327s

[Tiago Cogumbreiro]

A Ter, 2004-01-13 =E0s 00:12, Derek Foreman escreveu:
> On Mon, 12 Jan 2004, Tiago Cogumbreiro wrote:
>=20
> > A Seg, 2004-01-12 =E0s 21:37, Derek Foreman escreveu:
> > > On Mon, 12 Jan 2004, Tiago Cogumbreiro wrote:
> > >
> > > > A Seg, 2004-01-12 =E0s 18:26, Derek Foreman escreveu:
>=20
> > > The first one died while I was blanking with libburn when it still =
had the
> > > "flood the device with test unit ready commands" bug.  Which, while
> > > stupid, really shouldn't damage hardware.
> > Bug? I implemented that part :D
> > That's the way to get the drive's progress. It's on the mmc specs... =
oh
> > well... I just check the cvs and it still has that part, at least the
> > public one... this will of course mean no blanking progress meter
>=20
> Oh there was a bug alright...
>=20
> one of the two commands you issued (1000 times per second, usleep is in
> microseconds :) would fail on my drive with a sense buffer that the
> command transport stuff read as a "retryable error".  it was immediatel=
y
> retried, and failed again.
>=20
> Basically, it would send test_unit_ready commands as fast as the proces=
sor
> could get them to the drive until the burning process was complete.
>=20
> The new code disables automatic retries for test_unit_ready, so it stil=
l
> polls the drive (once per second.  maybe we'll go higher than that, but
> certainly not as fast as we were), it just doesn't flood it. :)
>=20
eheeh that issue about usleep being 1/1000000 of a sec and not 1/1000 :P
I am such a noob :D
>=20
> ps.  oh yeah, I also made raw96r work. :)
Cool!