2000-11-04 10:53:57

by Stanislav Brabec

[permalink] [raw]
Subject: ATAPI: audio CD stops playing (2.2.17 & 2.4.0-test?)

Description:
On 2.4.0-pre9 and 2.4.0-pre10:
Playing of some audio CD's stops in nearly regular places. Also pressing >> in CD software
panel in nearly all cases ends by stop.

All these stops are reported:
Nov 3 21:41:47 utx kernel: ATAPI device hdb:
Nov 3 21:41:47 utx kernel: Error: Illegal request -- (Sense key=0x05)
Nov 3 21:41:47 utx kernel: Invalid field in command packet -- (asc=0x24, ascq=0x00)
Nov 3 21:41:47 utx kernel: The failed "Play Audio MSF" packet command was:
Nov 3 21:41:47 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "

On 2.2.17:
Rejects to play audio CD on this hardware at all with report:
Oct 19 15:24:15 utx kernel: ATAPI device hdb:
Oct 19 15:24:15 utx kernel: Error: Illegal request -- (Sense key=0x05)
Oct 19 15:24:15 utx kernel: Invalid command operation code -- (asc=0x20, ascq=0x00)
Oct 19 15:24:15 utx kernel: The failed "Play Audio TrackIndex" packet command was:
Oct 19 15:24:15 utx kernel: "48 00 00 00 01 01 00 0d 01 00 00 00 "
Oct 19 15:24:26 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Oct 19 15:24:26 utx kernel: hdb: packet command error: error=0x50
Oct 19 15:24:26 utx kernel: ATAPI device hdb:
Oct 19 15:24:26 utx kernel: Error: Illegal request -- (Sense key=0x05)
Oct 19 15:24:26 utx kernel: Invalid command operation code -- (asc=0x20, ascq=0x00)
Oct 19 15:24:26 utx kernel: The failed "Play Audio TrackIndex" packet command was:
Oct 19 15:24:26 utx kernel: "48 00 00 00 01 01 00 0d 01 00 00 00 "

linux-2.2.16 works OK


HW&SW:
ATAPI CD-ROM Mitsumi FX400E (4x speed)

Cyrix686MX200


Compiler:
gcc-2.96.2


--
Stanislav Brabec


2000-11-04 14:19:19

by Jens Axboe

[permalink] [raw]
Subject: Re: ATAPI: audio CD stops playing (2.2.17 & 2.4.0-test?)

On Fri, Nov 03 2000, Stanislav Brabec wrote:
> Description:
> On 2.4.0-pre9 and 2.4.0-pre10:
> Playing of some audio CD's stops in nearly regular places. Also pressing
> >> in CD software panel in nearly all cases ends by stop.

Known problem, patch not submitted yet.

--
* Jens Axboe <[email protected]>
* SuSE Labs

2000-12-15 14:43:01

by Stanislav Brabec

[permalink] [raw]
Subject: ATAPI: audio CD still stops on >> (fast forward, 2.4.0-test12)

Jens Axboe wrote, Sat?Nov 4,?2000,?15:18:45?GMT):
> On Fri, Nov 03 2000, Stanislav Brabec wrote:
> > Description:
> > On 2.4.0-pre9 and 2.4.0-pre10:
> > Playing of some audio CD's stops in nearly regular places. Also pressing
> > >> in CD software panel in nearly all cases ends by stop.
>
> Known problem, patch not submitted yet.
>

- play CD audio correctly, don't stop after 12 minutes.

Patch in 2.4.0-test12 really fixes this problem.

But problem with >> (fast forward playng of short samples) still remains
on some audio CD's.

HW:
ATAPI CD-ROM Mitsumi FX400E (4x speed)
Cyrix686MX200

SW:
tcd/gtcd

Dec 15 12:17:25 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Dec 15 12:17:25 utx kernel: hdb: packet command error: error=0x50
Dec 15 12:17:25 utx kernel: ATAPI device hdb:
Dec 15 12:17:25 utx kernel: Error: Illegal request -- (Sense key=0x05)
Dec 15 12:17:25 utx kernel: Invalid field in command packet -- (asc=0x24, ascq=0x00)
Dec 15 12:17:25 utx kernel: The failed "Play Audio MSF" packet command was:
Dec 15 12:17:25 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "
Dec 15 12:17:41 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Dec 15 12:17:41 utx kernel: hdb: packet command error: error=0x50
Dec 15 12:17:41 utx kernel: ATAPI device hdb:
Dec 15 12:17:41 utx kernel: Error: Illegal request -- (Sense key=0x05)
Dec 15 12:17:41 utx kernel: Invalid field in command packet -- (asc=0x24, ascq=0x00)
Dec 15 12:17:41 utx kernel: The failed "Play Audio MSF" packet command was:
Dec 15 12:17:41 utx kernel: "47 00 00 00 1b 00 3c 3a ff 00 00 00 "
Dec 15 12:17:44 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Dec 15 12:17:44 utx kernel: hdb: packet command error: error=0xb0
Dec 15 12:17:44 utx kernel: ATAPI device hdb:
Dec 15 12:17:44 utx kernel: Error: Aborted command -- (Sense key=0x0b)
Dec 15 12:17:44 utx kernel: Play operation aborted -- (asc=0xb9, ascq=0x00)
Dec 15 12:17:44 utx kernel: The failed "Pause/Resume" packet command was:
Dec 15 12:17:44 utx kernel: "4b 00 00 00 00 00 00 00 00 00 00 00 "
Dec 15 12:17:45 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Dec 15 12:17:45 utx kernel: hdb: packet command error: error=0x50
Dec 15 12:17:45 utx kernel: ATAPI device hdb:
Dec 15 12:17:45 utx kernel: Error: Illegal request -- (Sense key=0x05)
Dec 15 12:17:45 utx kernel: Invalid field in command packet -- (asc=0x24, ascq=0x00)
Dec 15 12:17:45 utx kernel: The failed "Play Audio MSF" packet command was:
Dec 15 12:17:45 utx kernel: "47 00 00 00 22 00 3c 3a ff 00 00 00 "
Dec 15 12:17:46 utx kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error }
Dec 15 12:17:46 utx kernel: hdb: packet command error: error=0x50
Dec 15 12:17:46 utx kernel: ATAPI device hdb:
Dec 15 12:17:46 utx kernel: Error: Illegal request -- (Sense key=0x05)
Dec 15 12:17:46 utx kernel: Invalid field in command packet -- (asc=0x24, ascq=0x00)
Dec 15 12:17:46 utx kernel: The failed "Play Audio MSF" packet command was:
Dec 15 12:17:46 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "

--
Stanislav Brabec

2000-12-16 14:30:45

by Jens Axboe

[permalink] [raw]
Subject: Re: ATAPI: audio CD still stops on >> (fast forward, 2.4.0-test12)

On Fri, Dec 15 2000, Stanislav Brabec wrote:
> - play CD audio correctly, don't stop after 12 minutes.
>
> Patch in 2.4.0-test12 really fixes this problem.

Good

> But problem with >> (fast forward playng of short samples) still remains
> on some audio CD's.
> Dec 15 12:17:25 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "
^^

This is the same case that Miles reported, it's very odd how that 8th
byte gets screwed somehow... But I know about this, I just haven't tracked
this down yet.

--
* Jens Axboe <[email protected]>
* SuSE Labs

2000-12-18 18:24:08

by Norbert Warmuth

[permalink] [raw]
Subject: Re: ATAPI: audio CD still stops on >> (fast forward, 2.4.0-test12)

On Sat, 16 Dec 2000, Jens Axboe wrote:
> > But problem with >> (fast forward playng of short samples) still remains
> > on some audio CD's.
> > Dec 15 12:17:25 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "
> ^^
> This is the same case that Miles reported, it's very odd how that 8th
> byte gets screwed somehow... But I know about this, I just haven't tracked
> this down yet.

At least Stanislav's problem is a userland problem. Sometimes tcd/gtcd
(the software Stanislav uses to play CDs) miscalculates frame values. A
patch to tcd is available at http://bugs.gnome.org/db/33/33600.html.

Regards,
Norbert

2000-12-18 18:27:30

by Jens Axboe

[permalink] [raw]
Subject: Re: ATAPI: audio CD still stops on >> (fast forward, 2.4.0-test12)

On Mon, Dec 18 2000, Norbert Warmuth wrote:
> On Sat, 16 Dec 2000, Jens Axboe wrote:
> > > But problem with >> (fast forward playng of short samples) still remains
> > > on some audio CD's.
> > > Dec 15 12:17:25 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "
> > ^^
> > This is the same case that Miles reported, it's very odd how that 8th
> > byte gets screwed somehow... But I know about this, I just haven't tracked
> > this down yet.
>
> At least Stanislav's problem is a userland problem. Sometimes tcd/gtcd
> (the software Stanislav uses to play CDs) miscalculates frame values. A
> patch to tcd is available at http://bugs.gnome.org/db/33/33600.html.

Ah interesting, I _bet_ this is also what everybody else is seeing!

--
* Jens Axboe <[email protected]>
* SuSE Labs

2000-12-20 00:02:51

by Stanislav Brabec

[permalink] [raw]
Subject: Re: ATAPI: audio CD still stops on >> (fast forward, 2.4.0-test12)

Jens Axboe wrote (Mon?Dec 18?2000,?18:56:13?GMT):
> On Mon, Dec 18 2000, Norbert Warmuth wrote:
> > On Sat, 16 Dec 2000, Jens Axboe wrote:
> > > > But problem with >> (fast forward playng of short samples) still remains
> > > > on some audio CD's.
> > > > Dec 15 12:17:25 utx kernel: "47 00 00 00 02 00 3c 3a ff 00 00 00 "
> > > ^^
> > > This is the same case that Miles reported, it's very odd how that 8th
> > > byte gets screwed somehow... But I know about this, I just haven't tracked
> > > this down yet.
> >
> > At least Stanislav's problem is a userland problem. Sometimes tcd/gtcd
> > (the software Stanislav uses to play CDs) miscalculates frame values. A
> > patch to tcd is available at http://bugs.gnome.org/db/33/33600.html.
>
> Ah interesting, I _bet_ this is also what everybody else is seeing!
>

I have tested this patch and there is my result:


Linux kernel patch solves the problem of stopping while playing after
2min 40sec.

gnome-media/tcd/linux-cdrom.c with fix http://bugs.gnome.org/db/33/33600.html
solves the problem of fast-fwd. Even on older kernels ffwd works OK.

Both these were two independent bugs, one in kernel, second in tcd.


I have looked into GNOME CVS and nobody have yet comitted this patch
into CVS, so I am doing it.

Can anybody close the bug in GNOME bug tracking system?


SW:
Linux & (g)tcd

HW:
Cyrix686MX200
ATAPI CD-ROM Mitsumi FX400E (4x speed)
CD Banco de Gaia / Last Train to Lhasa (CD 1) / track 1

--
Stanislav Brabec