Re: 2.6.19-rc1 regression: unable to read dvd's

On Thu, Oct 12 2006, Jens Axboe wrote:
On Thu, Oct 12 2006, Mike Galbraith wrote:
On Thu, 2006-10-12 at 10:28 +0000, Mike Galbraith wrote:
On Thu, 2006-10-12 at 08:53 +0200, Jens Axboe wrote:
Test case, please.

Xine.

I just built it from scratch(the one that comes with SuSE 10.1 is
useless for DVDs), and tried it in 2.6.19-rc1 after verifying that it
worked fine in 2.6.17.

s/17/18

hdd: BENQ DVD DD DW1625, ATAPI CD/DVD-ROM drive
hdd: ATAPI 63X DVD-ROM DVD-R CD-R/RW drive, 2048kB Cache, UDMA(33)

hdd: read_intr: Drive wants to transfer data the wrong way!
hdd: read_intr: Drive wants to transfer data the wrong way!

We're having secret handshake troubles?

[pid 8348] stat64("/dev/dvd", {st_dev=makedev(0, 13), st_ino=3750, st_mode=S_IFBLK|0640, st_nlink=1, st_uid=0, st_gid=6, st_blksize=4096, st_blocks=0, st_rdev=makedev(22, 64), st_atime=2006/10/12-10:03:04, st_mtime=2006/10/12-10:00:12, st_ctime=2006/10/12-10:00:17}) = 0
[pid 8348] open("/dev/dvd", O_RDONLY|O_LARGEFILE <unfinished ...>
[pid 8348] <... open resumed> ) = 8
[pid 8348] fstat64(8, {st_dev=makedev(0, 13), st_ino=3750, st_mode=S_IFBLK|0640, st_nlink=1, st_uid=0, st_gid=6, st_blksize=4096, st_blocks=0, st_rdev=makedev(22, 64), st_atime=2006/10/12-10:03:04, st_mtime=2006/10/12-10:00:12, st_ctime=2006/10/12-10:00:17}) = 0
[pid 8348] ioctl(8, DVD_READ_STRUCT <unfinished ...>
[pid 8348] <... ioctl resumed> , 0xbfc792a4) = 0
[pid 8348] ioctl(8, DVD_AUTH <unfinished ...>
[pid 8348] <... ioctl resumed> , 0xbfc7916c) = 0
[pid 8348] ioctl(8, DVD_AUTH <unfinished ...>
[pid 8348] <... ioctl resumed> , 0xbfc7916c) = 0
[pid 8348] ioctl(8, DVD_AUTH <unfinished ...>
[pid 8348] <... ioctl resumed> , 0xbfc7916c) = 0
[pid 8348] ioctl(8, DVD_AUTH <unfinished ...>
[pid 8348] <... ioctl resumed> , 0xbfc79170) = 0
[pid 8348] close(8) = 0
[pid 8348] write(2, "libdvdread: Could not open /dev/"..., 52 <unfinished ...>
[pid 8348] <... write resumed> ) = 52
[pid 8348] write(2, "libdvdread: Can\'t open /dev/dvd "..., 44 <unfinished ...>
[pid 8348] <... write resumed> ) = 44
[pid 8348] write(1, "libdvdnav: vm: faild to open/rea"..., 42 <unfinished ...>
[pid 8348] <... write resumed> ) = 42
[pid 8348] write(1, "libdvdnav: Using dvdnav version "..., 62 <unfinished ...>
[pid 8348] <... write resumed> ) = 62
[pid 8348] write(2, "libdvdread: Using libdvdcss vers"..., 57 <unfinished ...>
[pid 8348] <... write resumed> ) = 57

It's the rq-cmd-type patch, it must be causing a disturbance for some of
the internal cd commands. I bet it's the same thing affecting the report
on broken dvd identification on ppc from Olaf.

Mike/Alex/Olaf, can you give this a spin? Totally untested here, but I
think it should fix it.

diff --git a/drivers/ide/ide-cd.c b/drivers/ide/ide-cd.c
index 69bbb62..e7513e5 100644
--- a/drivers/ide/ide-cd.c
+++ b/drivers/ide/ide-cd.c
@@ -597,7 +597,7 @@ static void cdrom_prepare_request(ide_dr
struct cdrom_info *cd = drive->driver_data;

ide_init_drive_cmd(rq);
- rq->cmd_type = REQ_TYPE_BLOCK_PC;
+ rq->cmd_type = REQ_TYPE_ATA_PC;
rq->rq_disk = cd->disk;
}

@@ -2023,7 +2023,8 @@ ide_do_rw_cdrom (ide_drive_t *drive, str
}
info->last_block = block;
return action;
- } else if (rq->cmd_type == REQ_TYPE_SENSE) {
+ } else if (rq->cmd_type == REQ_TYPE_SENSE ||
+ rq->cmd_type == REQ_TYPE_ATA_PC) {
return cdrom_do_packet_command(drive);
} else if (blk_pc_request(rq)) {
return cdrom_do_block_pc(drive, rq);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 26f7856..d370d2c 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -157,6 +157,7 @@ enum rq_cmd_type_bits {
REQ_TYPE_ATA_CMD,
REQ_TYPE_ATA_TASK,
REQ_TYPE_ATA_TASKFILE,
+ REQ_TYPE_ATA_PC,
};

/*

--
Jens Axboe

-
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/

Relevant Pages

  • Re: 2.6.19-rc1 regression: unable to read dvds
    ... useless for DVDs), and tried it in 2.6.19-rc1 after verifying that it ... hdd: read_intr: Drive wants to transfer data the wrong way! ...
    (Linux-Kernel)
  • Re: 2.6.19-rc1 regression: unable to read dvds
    ... useless for DVDs), and tried it in 2.6.19-rc1 after verifying that it ... hdd: read_intr: Drive wants to transfer data the wrong way! ...
    (Linux-Kernel)
  • Re: 2.6.19-rc1 regression: unable to read dvds
    ... useless for DVDs), and tried it in 2.6.19-rc1 after verifying that it ... hdd: read_intr: Drive wants to transfer data the wrong way! ...
    (Linux-Kernel)
  • Re: [PATCH rc6] block: Fix CDROMEJECT to work in more cases
    ... > On Tue, 20 Dec 2005, Jens Axboe wrote: ... There are two sides to this - looking at the request allocations pools, ... But a request that originates from that particular pool (in this ... not issue any request unless it needed to transfer data from a device. ...
    (Linux-Kernel)
  • editing DVD-R or DVD+R
    ... I have an RCA DRC8030N 'set top box' which can record to DVD discs or to an ... It has very limited editing capability for recorded DVDs and ... essentially no editing for recordings to the HDD. ... looking for a way to take one of these home recorded DVDs and edit it on my ...
    (rec.video.dvd.tech)