Date: Mon, 22 Mar 1999 16:02:20 +0200 (EET) From: "Jukka A. Ukkonen" <jau@jau.tmt.tele.fi> To: FreeBSD-gnats-submit@freebsd.org Subject: kern/10728: WCD driver does not recognize a recovered atapi/ide error Message-ID: <199903221402.QAA01971@jau.tmt.tele.fi>
index | next in thread | raw e-mail
>Number: 10728
>Category: kern
>Synopsis: WCD driver does not recognize a recovered atapi/ide error
>Confidential: no
>Severity: non-critical
>Priority: medium
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Mon Mar 22 06:10:00 PST 1999
>Closed-Date:
>Last-Modified:
>Originator: Jukka A. Ukkonen
>Release: FreeBSD 2.2.8-RELEASE i386
>Organization:
Sonera Ltd. / Internet Services
>Environment:
FreeBSD 2.2.8-RELEASE based AMD/i486 with an ATAPI/IDE CD drive.
>Description:
Occasionally the CD drives report a recovered error when the
requested data is delayed. The driver code failed to recognize
this situation and returned EIO in vain. By checking `dmesg'
output the driver generated quite often log entries like these...
wcd0: i/o error, status=51<ready,opdone,check>, error=30
wcd0: i/o error, status=51<ready,opdone,check>, error=30
wcd0: i/o error, status=51<ready,opdone,check>, error=30
This tended to make e.g. workman sometimes fail to probe the CD
status, and which time workman assumes there is no longer a CD
in the drive. When the next status probe succeeded, workman then
automatically started playing a disc with `automatically play this
disc' set even though the disc had been previously manually stopped.
>How-To-Repeat:
Simply start workman on a atapi/ide cd-drive like Mitsumi-FX400
which occasionally report delayed completion of requested operations,
insert a CD with `automatically play this disc' set, manually stop
the disc, and let the disc stay in the drive.
Sometimes it takes hours for this to happen, sometimes it happens
in a matter of minutes. Anyhow when the problem occurs, there will
be a new log entry for a recovered error, and workman will start
playing the disc all by itself.
>Fix:
This is a tentative patch, but at least workman no longer re-starts
playing a stopped disc all by itself.
I simply made the assumption that a delayed but recovered response
from the drive is quite as much all right as any normal successful
response.
--- /sys/i386/isa/wcd.c.old Sun Mar 14 22:24:49 1999
+++ /sys/i386/isa/wcd.c Sun Mar 14 22:55:12 1999
@@ -620,7 +620,8 @@
result = atapi_request_wait (t->ata, t->unit, cmd,
a1, a2, a3, a4, a5, a6, a7, a8, a9, 0, 0, 0, 0, 0, 0,
addr, count);
- if (result.code) {
+ if (result.code
+ && ((result.error & AER_SKEY) != AER_SK_RECOVERED_ERROR)) {
wcd_error (t, result);
return (EIO);
}
@@ -962,7 +963,8 @@
ATAPI_TEST_UNIT_READY, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0);
}
- if (result.code) {
+ if (result.code
+ && ((result.error & AER_SKEY) != AER_SK_RECOVERED_ERROR)) {
wcd_error (t, result);
return (EIO);
}
@@ -1084,7 +1086,8 @@
return (0);
}
- if (result.code) {
+ if (result.code
+ && ((result.error & AER_SKEY) != AER_SK_RECOVERED_ERROR)) {
wcd_error (t, result);
return (EIO);
}
>Release-Note:
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199903221402.QAA01971>