linux-stable.git/block/partitions, branch v3.11.2

partitions/msdos: enumerate also AIX LVM partitions

2013-07-09T17:33:28+00:00

Graft AIX partitions enumeration into partitions/msdos.c

There is already a AIX disks detection logic in msdos.c.  When an AIX disk
has been found, and if configured to, call the aix partitions recognizer.
This avoids removal of AIX disks protection from msdos.c, avoids code
duplication, and ensures that AIX partitions enumeration is called before
plain msdos partitions enumeration.

Signed-off-by: Philippe De Muyter 
Cc: Karel Zak 
Cc: Jens Axboe 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

partitions: add aix lvm partition support files

2013-07-09T17:33:28+00:00

Add partitions/aix.h and partitions/aix.c.

AIX LVM permits to make "logical volumes" which are made of multiple
slices of multiple disks.  The new code allows only access to the
"logical volumes" which are made of one slice on the probed disk, a
slice being a contiguous disk area.  The code also detects "logical
volumes" made of multiple slices on the probed disk, but can not
describe them to the partition layer, because the partition layer
generic code does not support that.  When such non-contiguous "logical
volumes" are detected, a diagnostic message is printed.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Philippe De Muyter 
Cc: Karel Zak 
Cc: Jens Axboe 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

partitions/msdos.c: end-of-line whitespace and semicolon cleanup

2013-07-09T17:33:28+00:00

Signed-off-by: Philippe De Muyter 
Cc: Karel Zak 
Cc: Jens Axboe 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

partitions/efi.c: replace useless kzalloc's by kmalloc's

2013-04-30T06:34:25+00:00

In alloc_read_gpt_entries and alloc_read_gpt_header, the kzalloc'ated
zones are either totally overwritten by the following read_lba call,
or freed.  As kmalloc is cheaper than kzalloc, use kmalloc.

Signed-off-by: Philippe De Muyter 
Cc: Matt Domsch 
Cc: Panagiotis Issaris 
Cc: Andrew Morton 
Signed-off-by: Jens Axboe

block/partitions: optimize memory allocation in check_partition()

2013-02-28T03:10:21+00:00

Currently, sizeof(struct parsed_partitions) may be 64KB in 32bit arch, so
it is easy to trigger page allocation failure by check_partition,
especially in hotplug block device situation(such as, USB mass storage,
MMC card, ...), and Felipe Balbi has observed the failure.

This patch does below optimizations on the allocation of struct
parsed_partitions to try to address the issue:

- make parsed_partitions.parts as pointer so that the pointed memory can
  fit in 32KB buffer, then approximate 32KB memory can be saved

- vmalloc the buffer pointed by parsed_partitions.parts because 32KB is
  still a bit big for kmalloc

- given that many devices have the partition count limit, so only
  allocate disk_max_parts() partitions instead of 256 partitions always

Signed-off-by: Ming Lei 
Reported-by: Felipe Balbi 
Cc: Jens Axboe 
Reviewed-by: Yasuaki Ishimatsu 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

block/partitions/mac.c: obey the state->limit constraint

2013-02-28T03:10:21+00:00

It isn't necessary to read the information of partitions whose number is
equal and more than state->limit since only maximum state->limit
partitions will be added inside rescan_partitions().

That is also what other kind of partitions are doing.

Signed-off-by: Ming Lei 
Cc: Jens Axboe 
Cc: Yasuaki Ishimatsu 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

block/partitions/efi.c: ensure that the GPT header is at least the size of the structure.

2013-02-28T03:10:21+00:00

UEFI 2.3.1D will include a change to the spec language mandating that a
GPT header must be greater than *or equal to* the size of the defined
structure.  While verifying that this would work on Linux, I discovered
that we're not actually checking the minimum bound at all.

The result of this is that when we verify the checksum, it's possible that
on a malformed header (with header_size of 0), we won't actually verify
any data.

[akpm@linux-foundation.org: fix printk warning]
Signed-off-by: Peter Jones 
Acked-by: Matt Fleming 
Cc: Jens Axboe 
Cc: Stephen Warren 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

block/partition/msdos: detect AIX formatted disks even without 55aa

2013-02-28T03:10:21+00:00

AIX formatted disks do not always have the MSDOS 55aa signature.
This happens e.g. for unbootable AIX disks.

Up to now, such disks were not recognized as AIX disks, because of the
missing 55aa.  Fix that by inverting the two tests.  Let's first
check for the AIX magic strings, and only if that fails check for
the MSDOS magic word.

Signed-off-by: Philippe De Muyter 
Cc: Andreas Mohr 
Cc: OGAWA Hirofumi 
Cc: Jens Axboe 
Cc: Olaf Hering 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

Merge branch 'for-3.8/drivers' of git://git.kernel.dk/linux-block

2012-12-17T21:39:11+00:00

Pull block driver update from Jens Axboe:
 "Now that the core bits are in, here are the driver bits for 3.8.  The
  branch contains:

   - A huge pile of drbd bits that were dumped from the 3.7 merge
     window.  Following that, it was both made perfectly clear that
     there is going to be no more over-the-wall pulls and how the
     situation on individual pulls can be improved.

   - A few cleanups from Akinobu Mita for drbd and cciss.

   - Queue improvement for loop from Lukas.  This grew into adding a
     generic interface for waiting/checking an even with a specific
     lock, allowing this to be pulled out of md and now loop and drbd is
     also using it.

   - A few fixes for xen back/front block driver from Roger Pau Monne.

   - Partition improvements from Stephen Warren, allowing partiion UUID
     to be used as an identifier."

* 'for-3.8/drivers' of git://git.kernel.dk/linux-block: (609 commits)
  drbd: update Kconfig to match current dependencies
  drbd: Fix drbdsetup wait-connect, wait-sync etc... commands
  drbd: close race between drbd_set_role and drbd_connect
  drbd: respect no-md-barriers setting also when changed online via disk-options
  drbd: Remove obsolete check
  drbd: fixup after wait_even_lock_irq() addition to generic code
  loop: Limit the number of requests in the bio list
  wait: add wait_event_lock_irq() interface
  xen-blkfront: free allocated page
  xen-blkback: move free persistent grants code
  block: partition: msdos: provide UUIDs for partitions
  init: reduce PARTUUID min length to 1 from 36
  block: store partition_meta_info.uuid as a string
  cciss: use check_signature()
  cciss: cleanup bitops usage
  drbd: use copy_highpage
  drbd: if the replication link breaks during handshake, keep retrying
  drbd: check return of kmalloc in receive_uuids
  drbd: Broadcast sync progress no more often than once per second
  drbd: don't try to clear bits once the disk has failed
  ...

partitions: enable EFI/GPT support by default

2012-12-06T13:34:58+00:00

The Kconfig currently enables MSDOS partitions by default because they
are assumed to be essential, but it's necessary to enable "advanced
partition selection" in order to get GPT support. IMO GPT partitions
are becoming common enought to deserve the same treatment MSDOS
partitions get.

(Side note: I got bit by a disk that had MSDOS and GPT partition
tables, but for some reason the MSDOS table was different from the
GPT one. I was stupid enought to disable "advanced partition
selection" in my .config, which disabled GPT partitioning and made
my btrfs pool unbootable because it couldn't find the partitions)

Signed-off-by: Diego Calleja 
Signed-off-by: Jens Axboe