F16 broken, rawhide ok? [Re: loop-related regression on Fedora 16

[Jim Meyering]

Brian C. Lane wrote:
> On Wed, Dec 21, 2011 at 07:30:43PM +0100, Jim Meyering wrote:
>> On Fedora 16 (3.1.5-6.fc16.x86_64), the two t8000*loop.sh tests have
>> been failing for some time.  I've finally investigated and found it is
>> due to this:
>
> It may be related to commit 1eb0cc30 where loop_get_partition_range was
> added. It reads from '/sys/module/loop/parameters/max_part' which on my
> F15 system (kernel 2.6.41.4-1.fc15.x86_64) returns 0.
>
> Partitioned loop devices certainly do work, so I think this is an error.
> I also am not sure what will happen with systems where loop is not a
> module.

Hi Brian,

I'm pretty sure this is an F16 kernel-related bug.
When I create a loop device like this on F16,

    truncate -s10m a && t=$(losetup --show -f a)

and print its ext_range value,

    cat /sys/devices/virtual/block/$(basename $t)/ext_range

I get "1", which means not partitionable, according to disk_max_parts
in include/linux/genhd.h which just returns gendisk.minors, described
like this:

struct gendisk {
        /* major, first_minor and minors are input parameters only,
         * don't use directly.  Use disk_devt() and disk_max_parts().
         */
        int major;                      /* major number of driver */
        int first_minor;
        int minors;                     /* maximum number of minors, =1 for

However, when I do the same on rawhide (3.2.0-0.rc7.git4.1.fc17.x86_64),
I get a reasonable value: 256.

I've just filed a BZ for this:
  http://bugzilla.redhat.com/771641