cdrom_is_mrw() can incorrectly think that a drive is Mt Rainier
capable, because if forgets to check if the "GET CONFIGURATION"
command returns the MRW feature number. According to the MMC spec,
the drive shall return all feature numbers >= the starting feature
number, so even if the drive doesn't support Mt Rainier, it can return
some data that makes cdrom_is_mrw() incorrectly think the drive is MRW
capable.
This problem stops me from mounting DVD+RW discs in R/W mode on my
laptop, because it makes cdrom_open_write() call
cdrom_mrw_open_write() which fails because the drive isn't really MRW
capable.
The fix is to make sure the returned feature number is the correct one
for Mt Rainier.
Signed-off-by: Peter Osterlund <[email protected]>
---
linux-petero/drivers/cdrom/cdrom.c | 2 ++
1 files changed, 2 insertions(+)
diff -puN drivers/cdrom/cdrom.c~mrw-fix drivers/cdrom/cdrom.c
--- linux/drivers/cdrom/cdrom.c~mrw-fix 2004-10-25 22:43:15.711347640 +0200
+++ linux-petero/drivers/cdrom/cdrom.c 2004-10-25 22:43:15.716346880 +0200
@@ -546,6 +546,8 @@ int cdrom_is_mrw(struct cdrom_device_inf
return ret;
mfd = (struct mrw_feature_desc *)&buffer[sizeof(struct feature_header)];
+ if (be16_to_cpu(mfd->feature_code) != CDF_MRW)
+ return 1;
*write = mfd->write;
if ((ret = cdrom_mrw_probe_pc(cdi))) {
_
--
Peter Osterlund - [email protected]
http://w1.894.telia.com/~u89404340
On Mon, Oct 25 2004, Peter Osterlund wrote:
> cdrom_is_mrw() can incorrectly think that a drive is Mt Rainier
> capable, because if forgets to check if the "GET CONFIGURATION"
> command returns the MRW feature number. According to the MMC spec,
> the drive shall return all feature numbers >= the starting feature
> number, so even if the drive doesn't support Mt Rainier, it can return
> some data that makes cdrom_is_mrw() incorrectly think the drive is MRW
> capable.
>
> This problem stops me from mounting DVD+RW discs in R/W mode on my
> laptop, because it makes cdrom_open_write() call
> cdrom_mrw_open_write() which fails because the drive isn't really MRW
> capable.
>
> The fix is to make sure the returned feature number is the correct one
> for Mt Rainier.
>
> Signed-off-by: Peter Osterlund <[email protected]>
Acked-by: Jens Axboe <[email protected]>
> ---
>
> linux-petero/drivers/cdrom/cdrom.c | 2 ++
> 1 files changed, 2 insertions(+)
>
> diff -puN drivers/cdrom/cdrom.c~mrw-fix drivers/cdrom/cdrom.c
> --- linux/drivers/cdrom/cdrom.c~mrw-fix 2004-10-25 22:43:15.711347640 +0200
> +++ linux-petero/drivers/cdrom/cdrom.c 2004-10-25 22:43:15.716346880 +0200
> @@ -546,6 +546,8 @@ int cdrom_is_mrw(struct cdrom_device_inf
> return ret;
>
> mfd = (struct mrw_feature_desc *)&buffer[sizeof(struct feature_header)];
> + if (be16_to_cpu(mfd->feature_code) != CDF_MRW)
> + return 1;
> *write = mfd->write;
>
> if ((ret = cdrom_mrw_probe_pc(cdi))) {
> _
>
> --
> Peter Osterlund - [email protected]
> http://w1.894.telia.com/~u89404340
>
--
Jens Axboe