[KARMIC] SRU-PREVIEW lp#490290 Updates to cciss driver

Stefan Bader stefan.bader at canonical.com
Wed Jan 6 14:22:44 UTC 2010

Manoj Iyer wrote:
> smb/ogasawara,
> This is a preview of the patches that I cherry picked to fix 
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/490290 the test 
> kernel is available at http://people.canonical.com/~manjo/lp490290-karmic/ 
> and the kernel was tested by the originator of the bug and reported to 
> have fixed the problem.
> I could possibly not include the patches from David Brownell and James 
> Morris, and avoid touching files under fs/ arch/ drivers/ kernel/ etc, but 
> since the changes there are just that of constifying I don't see any harm, 
> these patches do touch the cciss files.

I would agree with Tim to disagree on this as a SRU. Clearly all of the patches
that touch other code do not belong here. And while a larger portion of the
remaining patches are possibly acceptable with the assumption that HP is
maintaining this driver and the affected hardware can be (and has been) tested
by them, there are a few within which re-arrange enough code to give a bad
feeling on this.
So it feels like the best solution would be to add this to l-b-m and make a new
sub-package for it.

> The following changes since commit c93b4ea71f6ca8beca25e3afde993e813713e3ac:
>    Stefan Bader (1):
>          UBUNTU: [Config] Disable CONFIG_USB_DEVICEFS
> are available in the git repository at:
>    ssh://zinc.canonical.com/srv/kernel.ubuntu.com/git/manjo/ubuntu-karmic.git lp490290
> Alexey Dobriyan (1):
>        const: make block_device_operations const
> Andrew Patterson (3):
>        cciss: Remove sysfs entries for logical drives on driver cleanup.
>        cciss: Use one scan thread per controller and fix hang during rmmod
>        cciss: Allow triggering of rescan of logical drive topology via sysfs entry
> David Brownell (1):
>        driver model: constify attribute groups
> Eric Dumazet (1):
>        cciss: memory leak in cciss_init_one()
> James Morris (1):
>        seq_file: constify seq_operations
> Randy Dunlap (1):
>        cciss: fix schedule_timeout() parameters
> Stephen M. Cameron (17):
>        cciss: Remove some unused code in rebuild_lun_table()
>        cciss: Dynamically allocate struct device for each logical drive as needed.
>        cciss: Rearrange logical drive sysfs code to make the "changing a disk" path work.
>        cciss: Handle failure of blk_init_queue gracefully in cciss_add_disk.
>        cciss: Handle cases when cciss_add_disk fails.
>        cciss: Handle special case for sysfs attributes of the first logical drive.
>        cciss: Clear all sysfs-exposed data for deleted logical drives.
>        cciss: Fix usage_count check in rebuild_lun_table when triggered via sysfs.
>        cciss: Fix excessive gendisk freeing bug on driver unload.
>        cciss: Silence noisy per-disk messages output by cciss_read_capacity
>        cciss: Preserve all 8 bytes of LUN ID for logical drives.
>        cciss: Don't check h->busy_initializing in cciss_open().
>        cciss: Add lunid attribute to each logical drive in /sys
>        cciss: fix some magic numbers in the raid-level decoding
>        cciss: Add a "raid_level" attribute to each logical drive in /sys
>        cciss: Add usage_count attribute to each logical drive in /sys
>        cciss: Dynamically allocate the drive_info_struct for each logical drive.

More information about the kernel-team mailing list