Re: ide taskfile and cdrom hang

[Bartlomiej Zolnierkiewicz]

On Monday 02 of February 2004 22:44, Mark Haverkamp wrote:
> On Mon, 2004-02-02 at 13:35, Bartlomiej Zolnierkiewicz wrote:
> > On Monday 02 of February 2004 21:45, Mark Haverkamp wrote:
> > > On Mon, 2004-02-02 at 11:45, Bartlomiej Zolnierkiewicz wrote:
> > > > On Monday 02 of February 2004 19:46, Mark Haverkamp wrote:
> > > > > On Sun, 2004-02-01 at 12:48, Bartlomiej Zolnierkiewicz wrote:
> > >
> > > [ .... ]
> > >
> > > > > Thanks,
> > > > >
> > > > > hdparm -I /dev/hda didn't hang.  I got this on the console:
> > > > >
> > > > > hda: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
> > > > > hda: drive_cmd: error=0x04Aborted Command
> > > >
> > > > That's okay.
> > > > hdparm first tries WIN_IDENTIFY which can fail on ATAPI devices.
> > > >
> > > > > I added the patch that you provided and I still get the hang when I
> > > > > cat /proc/ide/hda/identify.  I put a printk In the code to make
> > > > > sure that it was going through the added code, and it is.
> > > >
> > > > So it is something different.  Can you give this patch a go?
> > > > We will know more about what's going on.
> > > >
> > > > Thanks,
> > > > --bart
> > >
> > > OK, Here it is:
> > >
> > > hda: (WAIT_NOT_BUSY) status=0xd0
> > > hda: (CHECK_STATUS) status=0xd0
> > > hda: (BUSY) status=0xd0
> > > hda: lost interrupt
> > > hda: (BUSY) status=0x50
> > > hda: lost interrupt
> > > hda: (BUSY) status=0x50
> > > hda: lost interrupt
> > > hda: (BUSY) status=0x50
> > > hda: lost interrupt
> > > hda: (BUSY) status=0x50
> >
> > Here we go with next (incremental) patch...
> >
> >  linux-2.6.2-rc2-root/drivers/ide/ide-taskfile.c |   12 ++++++++----
> >  1 files changed, 8 insertions(+), 4 deletions(-)
> >
> > diff -puN drivers/ide/ide-taskfile.c~ide_tf_identify_debug2
> > drivers/ide/ide-taskfile.c ---
> > linux-2.6.2-rc2/drivers/ide/ide-taskfile.c~ide_tf_identify_debug2	2004-02
> >-02 22:24:31.802461904 +0100 +++
> > linux-2.6.2-rc2-root/drivers/ide/ide-taskfile.c	2004-02-02
> > 22:33:54.987844712 +0100 @@ -786,14 +786,18 @@
> > EXPORT_SYMBOL(task_mulout_intr);
> >  static u8 wait_drive_not_busy(ide_drive_t *drive)
> >  {
> >  	ide_hwif_t *hwif = HWIF(drive);
> > -	int retries = 5;
> > +	int retries = 100;
> >  	u8 stat;
> > +
> >  	/*
> > -	 * (ks) Last sector was transfered, wait until drive is ready.
> > -	 * This can take up to 10 usec. We willl wait max 50 us.
> > +	 * Last sector was transfered, wait until drive is ready.
> > +	 * This can take up to 10 usec, but we will wait max 1 ms
> > +	 * (drive_cmd_intr() waits that long).
> >  	 */
> > -	while (((stat = hwif->INB(IDE_STATUS_REG)) & BUSY_STAT) && retries--)
> > +	while (((stat = hwif->INB(IDE_STATUS_REG)) & BUSY_STAT) && retries--) {
> > +		printk("%s: (UDELAY(10)) status=0x%02x\n", drive->name, stat);
> >  		udelay(10);
> > +	}
> >  	printk("%s: (WAIT_NOT_BUSY) status=0x%02x\n", drive->name, stat);
> >  	return stat;
> >  }
> >
> > _
>
> No hang this time.  Saw this on the console:
>
> hda: (UDELAY(10)) status=0xd8
> hda: (WAIT_NOT_BUSY) status=0x50
> hda: (CHECK_STATUS) status=0x50
> hda: (UDELAY(10)) status=0xd8
> hda: (WAIT_NOT_BUSY) status=0x50
> hda: (CHECK_STATUS) status=0x50

Now, can you comment out "(UDELAY(10))" printk and add printk for "retries"
variable after while {} loop.  I thought there will be more "(UDELAY(10))"
messages - but I forgot about delay introduced by printk() call :-).

--bart

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/