linux-toradex.git/lib/kobject.c, branch v4.0-rc1

kobject: fix NULL pointer derefernce in kobj_child_ns_ops

2014-11-07T18:52:19+00:00

We will hit NULL pointer dereference if we call
platform_device_register_simple or platform_device_add at very early
stage. I have observed following crash when called platform_device_add
from "init_irq" hook of machine_desc. This patch fixes this issue and
let system handle this case gracefully instead of kernel panic.

[0.000000] Unable to handle kernel NULL pointer dereference at
virtual address 0000000c
[0.000000] pgd = c0004000
[0.000000] [0000000c] *pgd=00000000
[0.000000] Internal error: Oops: 5 [#1] PREEMPT ARM
[0.000000] Modules linked in:
[0.000000] CPU: 0 PID: 0 Comm: swapper Tainted: G        W 3.17.0-rc6-00198-ga1603f1-dirty #319
[0.000000] task: c05b23f0 ti: c05a8000 task.ti: c05a8000
[0.000000] PC is at kobject_namespace+0x18/0x58
[0.000000] LR is at kobject_add_internal+0x90/0x2ec
[snip]
[0.000000] [] (kobject_namespace) from [] (kobject_add_internal+0x90/0x2ec)
[0.000000] [] (kobject_add_internal) from [] (kobject_add+0x4c/0x98)
[0.000000] [] (kobject_add) from [] (device_add+0xe8/0x51c)
[0.000000] [] (device_add) from [] (platform_device_add+0xb4/0x214)
[0.000000] [] (platform_device_add) from [] (platform_device_register_full+0xb8/0xdc)
[0.000000] [] (platform_device_register_full) from [] (exynos_init_irq+0x90/0x9c)
[0.000000] [] (exynos_init_irq) from [] (init_IRQ+0x2c/0x78)
[0.000000] [] (init_IRQ) from [] (start_kernel+0x22c/0x378)
[0.000000] [] (start_kernel) from [<40008070>] (0x40008070)
[0.000000] Code: e590000c e3500000 0a00000e e5903014 (e593300c)

Signed-off-by: Pankaj Dubey 
Signed-off-by: Greg Kroah-Hartman

sysfs, kobject: add sysfs wrapper for kernfs_enable_ns()

2014-02-08T00:08:57+00:00

Currently, kobject is invoking kernfs_enable_ns() directly.  This is
fine now as sysfs and kernfs are enabled and disabled together.  If
sysfs is disabled, kernfs_enable_ns() is switched to dummy
implementation too and everything is fine; however, kernfs will soon
have its own config option CONFIG_KERNFS and !SYSFS && KERNFS will be
possible, which can make kobject call into non-dummy
kernfs_enable_ns() with NULL kernfs_node pointers leading to an oops.

Introduce sysfs_enable_ns() which is a wrapper around
kernfs_enable_ns() so that it can be made a noop depending only on
CONFIG_SYSFS regardless of the planned CONFIG_KERNFS.

Signed-off-by: Tejun Heo 
Reported-by: Fengguang Wu 
Signed-off-by: Greg Kroah-Hartman

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

2014-01-31T04:08:20+00:00

Pull btrfs updates from Chris Mason:
 "This is a pretty big pull, and most of these changes have been
  floating in btrfs-next for a long time.  Filipe's properties work is a
  cool building block for inheriting attributes like compression down on
  a per inode basis.

  Jeff Mahoney kicked in code to export filesystem info into sysfs.

  Otherwise, lots of performance improvements, cleanups and bug fixes.

  Looks like there are still a few other small pending incrementals, but
  I wanted to get the bulk of this in first"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (149 commits)
  Btrfs: fix spin_unlock in check_ref_cleanup
  Btrfs: setup inode location during btrfs_init_inode_locked
  Btrfs: don't use ram_bytes for uncompressed inline items
  Btrfs: fix btrfs_search_slot_for_read backwards iteration
  Btrfs: do not export ulist functions
  Btrfs: rework ulist with list+rb_tree
  Btrfs: fix memory leaks on walking backrefs failure
  Btrfs: fix send file hole detection leading to data corruption
  Btrfs: add a reschedule point in btrfs_find_all_roots()
  Btrfs: make send's file extent item search more efficient
  Btrfs: fix to catch all errors when resolving indirect ref
  Btrfs: fix protection between walking backrefs and root deletion
  btrfs: fix warning while merging two adjacent extents
  Btrfs: fix infinite path build loops in incremental send
  btrfs: undo sysfs when open_ctree() fails
  Btrfs: fix snprintf usage by send's gen_unique_name
  btrfs: fix defrag 32-bit integer overflow
  btrfs: sysfs: list the NO_HOLES feature
  btrfs: sysfs: don't show reserved incompat feature
  btrfs: call permission checks earlier in ioctls and return EPERM
  ...

kobject: export kobj_sysfs_ops

2014-01-28T21:19:24+00:00

struct kobj_attribute implements the baseline attribute functionality
that can be used all over the place. We should export the ops associated
with it.

Signed-off-by: Jeff Mahoney 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Josef Bacik 
Signed-off-by: Chris Mason

kobject: Fix source code comment spelling

2014-01-08T23:36:18+00:00

Signed-off-by: Bart Van Assche 
Signed-off-by: Greg Kroah-Hartman

Revert "kobject: introduce kobj_completion"

2014-01-05T03:56:40+00:00

This reverts commit eee031649707db3c9920d9498f8d03819b74fc23.

Jeff writes:
	I have no objections to reverting it. There were concerns from
	Al Viro that it'd be tough to get right by callers and I had
	assumed it got dropped after that. I had planned on using it in
	my btrfs sysfs exports patchset but came up with a better way.

Cc: Jeff Mahoney 
Cc: Al Viro 
Signed-off-by: Greg Kroah-Hartman

kernfs: s/sysfs_dirent/kernfs_node/ and rename its friends accordingly

2013-12-11T23:28:36+00:00

kernfs has just been separated out from sysfs and we're already in
full conflict mode.  Nothing can make the situation any worse.  Let's
take the chance to name things properly.

This patch performs the following renames.

* s/sysfs_elem_dir/kernfs_elem_dir/
* s/sysfs_elem_symlink/kernfs_elem_symlink/
* s/sysfs_elem_attr/kernfs_elem_file/
* s/sysfs_dirent/kernfs_node/
* s/sd/kn/ in kernfs proper
* s/parent_sd/parent/
* s/target_sd/target/
* s/dir_sd/parent/
* s/to_sysfs_dirent()/rb_to_kn()/
* misc renames of local vars when they conflict with the above

Because md, mic and gpio dig into sysfs details, this patch ends up
modifying them.  All are sysfs_dirent renames and trivial.  While we
can avoid these by introducing a dummy wrapping struct sysfs_dirent
around kernfs_node, given the limited usage outside kernfs and sysfs
proper, I don't think such workaround is called for.

This patch is strictly rename only and doesn't introduce any
functional difference.

- mic / gpio renames were missing.  Spotted by kbuild test robot.

Signed-off-by: Tejun Heo 
Cc: Neil Brown 
Cc: Linus Walleij 
Cc: Ashutosh Dixit 
Cc: kbuild test robot 
Signed-off-by: Greg Kroah-Hartman

kobject: fix memory leak in kobject_set_name_vargs

2013-12-09T02:19:15+00:00

If the call to kvasprintf fails then the old name of the object will be leaked,
this patch fixes the bug by restoring the old name before returning ENOMEM.

Signed-off-by: Maurizio Lombardi 
Signed-off-by: Greg Kroah-Hartman

kobject: remove kset from sysfs immediately in kset_unregister()

2013-12-08T05:20:11+00:00

There's no "unlink from sysfs" interface for ksets, so I think callers of
kset_unregister() expect the kset to be removed from sysfs immediately,
without waiting for the last reference to be released.

This patch makes the sysfs removal happen immediately, so the caller may
create a new kset with the same name as soon as kset_unregister() returns.
Without this, every caller has to call "kobject_del(&kset->kobj)" first
unless it knows it will never create a new kset with the same name.

This sometimes shows up on module unload and reload, where the reload fails
because it tries to create a kobject with the same name as one from the
original load that still exists.  CONFIG_DEBUG_KOBJECT_RELEASE=y makes this
problem easier to hit.

Signed-off-by: Bjorn Helgaas 
Signed-off-by: Greg Kroah-Hartman

kobject: delay kobject release for random time

2013-12-08T05:14:13+00:00

When CONFIG_DEBUG_KOBJECT_RELEASE=y, delay kobject release functions for a
random time between 1 and 8 seconds, which effectively changes the order in
which they're called.

Signed-off-by: Bjorn Helgaas 
Signed-off-by: Greg Kroah-Hartman