Date: Tue, 2 Dec 1997 22:08:48 -0600 (CST) From: "Daniel M. Eischen" <deischen@iworks.interworks.org> To: sos@FreeBSD.dk Cc: freebsd-hackers@FreeBSD.ORG, freebsd-mobile@FreeBSD.ORG Subject: Re: Fix for slow reacting ATAPI CD-ROM Message-ID: <199712030408.WAA11699@iworks.InterWorks.org>
next in thread | raw e-mail | index | archive | help
To follow up on my original posting:
> > I have a new ChemUSA Chembook and everything works fine
> > except for the CD-ROM drive, which times out waiting
> > for the command phase in atapi_wait_cmd(). I was able
> > to get it to work by increasing the wait count to 3000.
> >
> > /* Wait for DRQ from 50 usec to 3 msec for slow devices */
> > int cnt = ata->intrcmd ? 10000 : ata->slow ? 3000 : 50;
> > int ireason = 0, phase = 0;
> >
> > + cnt = 3000;
> > +
> > /* Wait for command phase. */
> > for (; cnt>0; cnt-=10) {
> > ireason = inb (ata->port + AR_IREASON);
> >
> > Obviously, this drive isn't detected as a slow device (I
> > can post the boot messages later if needed). What I wanted
> > to do was to add a config flag for setting this, but it doesn't
> > look like that is sent down to the atapi driver.
> >
> > Question: Can we change atapi_attach() to accept a 4th argument
> > and use this to pass config flags from wdattach() in
> > wd.c? Is there another method I should be using to
> > handle atapi device quirks?
>
> Hmm, I think that atapi_attach should be educated on how to treat
> this device, test for model make etc, and then set the slow bit...
> This is the place to do quirks for atapi CDROM device.
>
I was able to lower the wait time to 75 usec for my CD-ROM drive,
but anything <= 70 usecs would not work.
According to the comments in atapi.h:
#define AT_DRQT_MPROC 0 /* microprocessor DRQ - 3 msec delay */
#define AT_DRQT_INTR 1 /* interrupt DRQ - 10 msec delay */
#define AT_DRQT_ACCEL 2 /* accelerated DRQ - 50 usec delay */
my CD-ROM drive should have a 50usec delay. The code in atapi.c
uses these numbers without any tolerance. I assume these numbers
come from some sort of ATAPI spec, but shouldn't we allow some
tolerance? Increasing these waits will not affect drives that
conform to the times, so I don't see why they can't be larger.
Dan Eischen
deischen@iworks.InterWorks.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199712030408.WAA11699>