Re: Possibly wrong BIO usage in ide_multwrite

[Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>]

On Monday 05 of January 2004 20:37, Frederik Deweerdt wrote:
> On Mon, 2004-01-05 at 17:12, Bartlomiej Zolnierkiewicz wrote:
> > - hangs during reading /proc/ide/<cdrom>/identify on some drives
> >   (workaround is now known thanks to debugging done by Andi+BenH+Andre)
>
> could you explain about this workaround? i've searched the archives
> without finding anything.

Because it was discovered recently and it is not the proper fix. :-)

 drivers/ide/ide-taskfile.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)

diff -puN drivers/ide/ide-taskfile.c~ide-tf-identify-fix drivers/ide/ide-taskfile.c
--- linux-2.6.1-rc1/drivers/ide/ide-taskfile.c~ide-tf-identify-fix	2004-01-04 16:35:53.094766576 +0100
+++ linux-2.6.1-rc1-root/drivers/ide/ide-taskfile.c	2004-01-04 16:45:03.240131768 +0100
@@ -797,9 +797,12 @@ check_status:
 	if (!OK_STAT(stat, good_stat, BAD_R_STAT)) {
 		if (stat & (ERR_STAT | DRQ_STAT))
 			return DRIVER(drive)->error(drive, __FUNCTION__, stat);
-		/* BUSY_STAT: No data yet, so wait for another IRQ. */
-		ide_set_handler(drive, &task_in_intr, WAIT_WORSTCASE, NULL);
-		return ide_started;
+		/* Workaround for some ATAPI drives which set only BSY bit. */
+		if (drive->media != ide_cdrom) {
+			/* BUSY_STAT: No data yet, so wait for another IRQ. */
+			ide_set_handler(drive, &task_in_intr, WAIT_WORSTCASE, NULL);
+			return ide_started;
+		}
 	}
 
 	/*

_