GitHub/LineageOS/android_kernel_samsung_universal7580.git
6 years agoANDROID: sdcardfs: Use seq_puts over seq_printf
Daniel Rosenberg [Wed, 22 Mar 2017 00:27:40 +0000 (17:27 -0700)]
ANDROID: sdcardfs: Use seq_puts over seq_printf

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: I3795ec61ce61e324738815b1ce3b0e09b25d723f

6 years agoANDROID: sdcardfs: Use to kstrout
Daniel Rosenberg [Fri, 17 Mar 2017 02:32:59 +0000 (19:32 -0700)]
ANDROID: sdcardfs: Use to kstrout

Switch from deprecated simple_strtoul to kstrout

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: If18bd133b4d2877f71e58b58fc31371ff6613ed5

6 years agoANDROID: sdcardfs: Use pr_[...] instead of printk
Daniel Rosenberg [Fri, 17 Mar 2017 00:46:13 +0000 (17:46 -0700)]
ANDROID: sdcardfs: Use pr_[...] instead of printk

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: Ibc635ec865750530d32b87067779f681fe58a003

6 years agoANDROID: sdcardfs: remove unneeded null check
Daniel Rosenberg [Tue, 21 Mar 2017 23:29:13 +0000 (16:29 -0700)]
ANDROID: sdcardfs: remove unneeded null check

As pointed out by checkpatch, these functions already
handle null inputs, so the checks are not needed.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: I189342f032dfcefee36b27648bb512488ad61d20

6 years agoANDROID: sdcardfs: Fix style issues with comments
Daniel Rosenberg [Fri, 17 Mar 2017 02:33:35 +0000 (19:33 -0700)]
ANDROID: sdcardfs: Fix style issues with comments

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: I8791ef7eac527645ecb9407908e7e5ece35b8f80

6 years agoANDROID: sdcardfs: Fix formatting
Daniel Rosenberg [Fri, 17 Mar 2017 00:42:58 +0000 (17:42 -0700)]
ANDROID: sdcardfs: Fix formatting

This fixes various spacing and bracket related issues
pointed out by checkpatch.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: I6e248833a7a04e3899f3ae9462d765cfcaa70c96

6 years agoANDROID: sdcardfs: correct order of descriptors
Daniel Rosenberg [Tue, 21 Mar 2017 23:28:27 +0000 (16:28 -0700)]
ANDROID: sdcardfs: correct order of descriptors

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: Ia6d16b19c8c911f41231d2a12be0740057edfacf

6 years agoANDROID: sdcardfs: Fix gid issue
Daniel Rosenberg [Mon, 13 Mar 2017 22:34:03 +0000 (15:34 -0700)]
ANDROID: sdcardfs: Fix gid issue

We were already calculating most of these values,
and erroring out because the check was confused by this.
Instead of recalculating, adjust it as needed.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 36160015
Change-Id: I9caf3e2fd32ca2e37ff8ed71b1d392f1761bc9a9

6 years agoANDROID: sdcardfs: Use tabs instead of spaces in multiuser.h
Daniel Rosenberg [Mon, 13 Mar 2017 20:53:54 +0000 (13:53 -0700)]
ANDROID: sdcardfs: Use tabs instead of spaces in multiuser.h

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35331000
Change-Id: Ic7801914a7dd377e270647f81070020e1f0bab9b

6 years agoANDROID: sdcardfs: Remove uninformative prints
Daniel Rosenberg [Sat, 11 Mar 2017 02:58:25 +0000 (18:58 -0800)]
ANDROID: sdcardfs: Remove uninformative prints

At best these prints do not provide useful information, and
at worst, some allow userspace to abuse the kernel log.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 36138424
Change-Id: I812c57cc6a22b37262935ab77f48f3af4c36827e

6 years agoANDROID: sdcardfs: move path_put outside of spinlock
Daniel Rosenberg [Fri, 10 Mar 2017 21:54:30 +0000 (13:54 -0800)]
ANDROID: sdcardfs: move path_put outside of spinlock

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35643557
Change-Id: Ib279ebd7dd4e5884d184d67696a93e34993bc1ef

6 years agoANDROID: sdcardfs: Use case insensitive hash function
Daniel Rosenberg [Fri, 10 Mar 2017 20:39:42 +0000 (12:39 -0800)]
ANDROID: sdcardfs: Use case insensitive hash function

Case insensitive comparisons don't help us much if
we hash to different buckets...

Signed-off-by: Daniel Rosenberg <drosen@google.com>
bug: 36004503
Change-Id: I91e00dbcd860a709cbd4f7fd7fc6d855779f3285

6 years agoANDROID: sdcardfs: declare MODULE_ALIAS_FS
Daniel Rosenberg [Fri, 10 Mar 2017 04:59:18 +0000 (20:59 -0800)]
ANDROID: sdcardfs: declare MODULE_ALIAS_FS

From commit ee616b78aa87 ("Wrapfs: declare MODULE_ALIAS_FS")

Signed-off-by: Daniel Rosenberg <drosen@google.com>
bug: 35766959
Change-Id: Ia4728ab49d065b1d2eb27825046f14b97c328cba

6 years agoANDROID: sdcardfs: Get the blocksize from the lower fs
Daniel Rosenberg [Fri, 10 Mar 2017 02:12:16 +0000 (18:12 -0800)]
ANDROID: sdcardfs: Get the blocksize from the lower fs

This changes sdcardfs to be more in line with the
getattr in wrapfs, which calls the lower fs's getattr
to get the block size

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 34723223
Change-Id: I1c9e16604ba580a8cdefa17f02dcc489d7351aed

6 years agoANDROID: sdcardfs: Use d_invalidate instead of drop_recurisve
Daniel Rosenberg [Thu, 9 Mar 2017 01:45:46 +0000 (17:45 -0800)]
ANDROID: sdcardfs: Use d_invalidate instead of drop_recurisve

drop_recursive did not properly remove stale dentries.
Instead, we use the vfs's d_invalidate, which does the proper cleanup.

Additionally, remove the no longer used drop_recursive, and
fixup_top_recursive that that are no longer used.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Change-Id: Ibff61b0c34b725b024a050169047a415bc90f0d8

6 years agoANDROID: sdcardfs: Switch to internal case insensitive compare
Daniel Rosenberg [Thu, 9 Mar 2017 01:20:02 +0000 (17:20 -0800)]
ANDROID: sdcardfs: Switch to internal case insensitive compare

There were still a few places where we called into a case
insensitive lookup that was not defined by sdcardfs.
Moving them all to the same place will allow us to switch
the implementation in the future.

Additionally, the check in fixup_perms_recursive did not
take into account the length of both strings, causing
extraneous matches when the name we were looking for was
a prefix of the child name.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Change-Id: I45ce768cd782cb4ea1ae183772781387c590ecc2

6 years agoANDROID: sdcardfs: Use spin_lock_nested
Daniel Rosenberg [Thu, 9 Mar 2017 01:11:51 +0000 (17:11 -0800)]
ANDROID: sdcardfs: Use spin_lock_nested

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 36007653
Change-Id: I805d5afec797669679853fb2bb993ee38e6276e4

6 years agoANDROID: sdcardfs: Replace get/put with d_lock
Daniel Rosenberg [Thu, 2 Mar 2017 23:11:27 +0000 (15:11 -0800)]
ANDROID: sdcardfs: Replace get/put with d_lock

dput cannot be called with a spin_lock. Instead,
we protect our accesses by holding the d_lock.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35643557
Change-Id: I22cf30856d75b5616cbb0c223724f5ab866b5114

6 years agoANDROID: sdcardfs: rate limit warning print
Daniel Rosenberg [Fri, 3 Mar 2017 02:07:21 +0000 (18:07 -0800)]
ANDROID: sdcardfs: rate limit warning print

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35848445
Change-Id: Ida72ea0ece191b2ae4a8babae096b2451eb563f6

6 years agoANDROID: sdcardfs: Fix case insensitive lookup
Daniel Rosenberg [Thu, 2 Mar 2017 01:04:41 +0000 (17:04 -0800)]
ANDROID: sdcardfs: Fix case insensitive lookup

The previous case insensitive lookup relied on the
entry being present in the dcache. This instead uses
iterate_dir to find the correct case.

Signed-off-by: Daniel Rosenberg <drosen@google.com
bug: 35633782
Change-Id: I556f7090773468c1943c89a5e2aa07f746ba49c5

6 years agoRevert "sdcardfs: limit stacking depth"
alexax66 [Sat, 9 Sep 2017 13:07:13 +0000 (13:07 +0000)]
Revert "sdcardfs: limit stacking depth"

This reverts commit 1d9ff6b31bd797409085ddeb606d8654979be7dc.

6 years agoRevert "sdcardfs: Fix build for linux-3.10.y"
alexax66 [Sat, 9 Sep 2017 13:04:47 +0000 (13:04 +0000)]
Revert "sdcardfs: Fix build for linux-3.10.y"

This reverts commit 86cb714606711a06109eb7c973dc585f993202c9.

6 years agosdcardfs: Fix build for linux-3.10.y
Parth Bhatia [Sun, 21 May 2017 15:03:58 +0000 (15:03 +0000)]
sdcardfs: Fix build for linux-3.10.y

* Fix compilation after 6637ecd

6 years agosdcardfs: limit stacking depth
Andrew Chant [Wed, 8 Feb 2017 23:33:48 +0000 (15:33 -0800)]
sdcardfs: limit stacking depth

Limit filesystem stacking to prevent stack overflow.

Bug: 32761463
Change-Id: I8b1462b9c0d6c7f00cf110724ffb17e7f307c51e
Signed-off-by: Andrew Chant <achant@google.com>
6 years agoANDROID: sdcardfs: Don't bother deleting freelist
Daniel Rosenberg [Wed, 22 Feb 2017 22:41:58 +0000 (14:41 -0800)]
ANDROID: sdcardfs: Don't bother deleting freelist

There is no point deleting entries from dlist, as
that is a temporary list on the stack from which
contains only entries that are being deleted.

Not all code paths set up dlist, so those that
don't were performing invalid accesses in
hash_del_rcu. As an additional means to prevent
any other issue, we null out the list entries when
we allocate from the cache.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35666680
Change-Id: Ibb1e28c08c3a600c29418d39ba1c0f3db3bf31e5

6 years agoANDROID: sdcardfs: Add missing path_put
Daniel Rosenberg [Fri, 17 Feb 2017 01:55:22 +0000 (17:55 -0800)]
ANDROID: sdcardfs: Add missing path_put

"ANDROID: sdcardfs: Add GID Derivation to sdcardfs" introduced
an unbalanced pat_get, leading to storage space not being freed
after deleting a file until rebooting. This adds the missing path_put.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 34691169
Change-Id: Ia7ef97ec2eca2c555cc06b235715635afc87940e

6 years agoANDROID: sdcardfs: Fix incorrect hash
Daniel Rosenberg [Wed, 15 Feb 2017 04:47:17 +0000 (20:47 -0800)]
ANDROID: sdcardfs: Fix incorrect hash

This adds back the hash calculation removed as part of
the previous patch, as it is in fact necessary.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 35307857
Change-Id: Ie607332bcf2c5d2efdf924e4060ef3f576bf25dc

6 years agoANDROID: sdcardfs: Switch strcasecmp for internal call
Daniel Rosenberg [Wed, 1 Feb 2017 04:07:51 +0000 (20:07 -0800)]
ANDROID: sdcardfs: Switch strcasecmp for internal call

This moves our uses of strcasecmp over to an internal call so we can
easily change implementations later if we so desire. Additionally,
we leverage qstr's where appropriate to save time on comparisons.

Change-Id: I32fdc4fd0cd3b7b735dcfd82f60a2516fd8272a5
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: switch to full_name_hash and qstr
Daniel Rosenberg [Sat, 28 Jan 2017 03:35:08 +0000 (19:35 -0800)]
ANDROID: sdcardfs: switch to full_name_hash and qstr

Use the kernel's string hash function instead of rolling
our own. Additionally, save a bit of calculation by using
the qstr struct in place of strings.

Change-Id: I0bbeb5ec2a9233f40135ad632e6f22c30ffa95c1
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: Add GID Derivation to sdcardfs
Daniel Rosenberg [Wed, 25 Jan 2017 21:48:45 +0000 (13:48 -0800)]
ANDROID: sdcardfs: Add GID Derivation to sdcardfs

This changes sdcardfs to modify the user and group in the
underlying filesystem depending on its usage. Ownership is
set by Android user, and package, as well as if the file is
under obb or cache. Other files can be labeled by extension.
Those values are set via the configfs interace.

To add an entry,
mkdir -p [configfs root]/sdcardfs/extensions/[gid]/[ext]

Bug: 34262585
Change-Id: I4e030ce84f094a678376349b1a96923e5076a0f4
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: Remove redundant operation
Daniel Rosenberg [Fri, 27 Jan 2017 04:10:34 +0000 (20:10 -0800)]
ANDROID: sdcardfs: Remove redundant operation

We call get_derived_permission_new unconditionally, so we don't need
to call update_derived_permission_lock, which does the same thing.

Change-Id: I0748100828c6af806da807241a33bf42be614935
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: add support for user permission isolation
Daniel Rosenberg [Sun, 22 Jan 2017 23:32:49 +0000 (15:32 -0800)]
ANDROID: sdcardfs: add support for user permission isolation

This allows you to hide the existence of a package from
a user by adding them to an exclude list. If a user
creates that package's folder and is on the exclude list,
they will not see that package's id.

Bug: 34542611
Change-Id: I9eb82e0bf2457d7eb81ee56153b9c7d2f6646323
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: Refactor configfs interface
Daniel Rosenberg [Sat, 21 Jan 2017 08:35:26 +0000 (00:35 -0800)]
ANDROID: sdcardfs: Refactor configfs interface

This refactors the configfs code to be more easily extended.
It will allow additional files to be added easily.

Bug: 34542611
Bug: 34262585
Change-Id: I73c9b0ae5ca7eb27f4ebef3e6807f088b512d539
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: Allow non-owners to touch
Daniel Rosenberg [Fri, 20 Jan 2017 23:19:13 +0000 (15:19 -0800)]
ANDROID: sdcardfs: Allow non-owners to touch

This modifies the permission checks in setattr to
allow for non-owners to modify the timestamp of
files to things other than the current time.
This still requires write access, as enforced by
the permission call, but relaxes the requirement
that the caller must be the owner, allowing those
with group permissions to change it as well.

Bug: 11118565
Change-Id: Ied31f0cce2797675c7ef179eeb4e088185adcbad
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: implement vm_ops->page_mkwrite
Daniel Rosenberg [Fri, 24 Feb 2017 23:41:48 +0000 (15:41 -0800)]
ANDROID: sdcardfs: implement vm_ops->page_mkwrite

This comes from the wrapfs patch
3dfec0ffe5e2 Wrapfs: implement vm_ops->page_mkwrite

Some file systems (e.g., ext4) require it.  Reported by Ted Ts'o.

Signed-off-by: Erez Zadok <ezk@cs.sunysb.edu>
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 34133558
Change-Id: I1a389b2422c654a6d3046bb8ec3e20511aebfa8e

6 years agoANDROID: sdcardfs: support direct-IO (DIO) operations
Daniel Rosenberg [Fri, 24 Feb 2017 23:49:45 +0000 (15:49 -0800)]
ANDROID: sdcardfs: support direct-IO (DIO) operations

This comes from the wrapfs patch
2e346c83b26e Wrapfs: support direct-IO (DIO) operations

Signed-off-by: Li Mengyang <li.mengyang@stonybrook.edu>
Signed-off-by: Erez Zadok <ezk@cs.sunysb.edu>
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 34133558
Change-Id: I3fd779c510ab70d56b1d918f99c20421b524cdc4

6 years agosdcardfs: Flag files as non-mappable
fluxi [Fri, 21 Oct 2016 20:57:35 +0000 (22:57 +0200)]
sdcardfs: Flag files as non-mappable

WARNING: it's not full commit, it just update sdcardfs

Implement Samsung's FMODE_NONMAPPABLE flag from
sdcardfs version 2.1.4 as we hit a BUG on ext4:

[   49.655037]@0 Kernel BUG at ffffffc0001deeec [verbose debug info unavailable]
[   49.655045]@0 Internal error: Oops - BUG: 0 [#1] PREEMPT SMP
[   49.655052]@0 Modules linked in:
[   49.655061]@0 CPU: 0 PID: 283 Comm: kworker/u8:7 Tainted: G        W      3.18.20-perf-g3be2054-00086-ga8307fb #1
[   49.655070]@0 Hardware name: Qualcomm Technologies, Inc. MSM 8996 v3 + PMI8996 MTP (DT)
[   49.655077]@0 Workqueue: writeback bdi_writeback_workfn (flush-8:0)
[   49.655096]@0 task: ffffffc174ba8b00 ti: ffffffc174bb4000 task.ti: ffffffc174bb4000
[   49.655108]@0 PC is at mpage_prepare_extent_to_map+0x198/0x218
[   49.655116]@0 LR is at mpage_prepare_extent_to_map+0x110/0x218
[   49.655121]@0 pc : [<ffffffc0001deeec>] lr : [<ffffffc0001dee64>] pstate: 60000145
[   49.655126]@0 sp : ffffffc174bb7800
[   49.655130]@0 x29: ffffffc174bb7800 x28: ffffffc174bb7880
[   49.655140]@0 x27: 000000000000000d x26: ffffffc1245505e8
[   49.655149]@0 x25: 0000000000000000 x24: 0000000000003400
[   49.655160]@0 x23: ffffffffffffffff x22: 0000000000000000
[   49.655172]@0 x21: ffffffc174bb7888 x20: ffffffc174bb79e0
[   49.655182]@0 x19: ffffffbdc4ee7b80 x18: 0000007f92872000
[   49.655191]@0 x17: 0000007f959b6424 x16: ffffffc00016d1ac
[   49.655201]@0 x15: 0000007f9285d158 x14: ffffffc1734796e8
[   49.655210]@0 x13: ffffffbdc1ffa4c0 x12: ffffffbdc4ee7b80
[   49.655220]@0 x11: 0000000000000100 x10: 0000000000000000
[   49.655229]@0 x9 : 0000000000000000 x8 : ffffffc0b444e210
[   49.655237]@0 x7 : 0000000000000000 x6 : ffffffc0b444e1e0
[   49.655246]@0 x5 : 0000000000000000 x4 : 0000000000000001
[   49.655254]@0 x3 : 0000000000000000 x2 : 400000000002003d
[   49.655263]@0 x1 : ffffffbdc4ee7b80 x0 : 400000000002003d
[   49.655271]@0
[   49.656502]@0 Process kworker/u8:7 (pid: 283, stack limit = 0xffffffc174bb4058)
[   49.656509]@0 Call trace:
[   49.656514]@0 [<ffffffc0001deeec>] mpage_prepare_extent_to_map+0x198/0x218
[   49.656526]@0 [<ffffffc0001e28d0>] ext4_writepages+0x270/0xa58
[   49.656533]@0 [<ffffffc00012982c>] do_writepages+0x24/0x40
[   49.656541]@0 [<ffffffc000180160>] __writeback_single_inode+0x40/0x114
[   49.656549]@0 [<ffffffc000180e50>] writeback_sb_inodes+0x1dc/0x34c
[   49.656555]@0 [<ffffffc00018103c>] __writeback_inodes_wb+0x7c/0xc4
[   49.656560]@0 [<ffffffc000181224>] wb_writeback+0x110/0x1a8
[   49.656565]@0 [<ffffffc000181344>] wb_check_old_data_flush+0x88/0x98
[   49.656571]@0 [<ffffffc00018156c>] bdi_writeback_workfn+0xf4/0x1fc
[   49.656576]@0 [<ffffffc0000b14f8>] process_one_work+0x1e0/0x300
[   49.656585]@0 [<ffffffc0000b1e14>] worker_thread+0x318/0x438
[   49.656590]@0 [<ffffffc0000b5da0>] kthread+0xe0/0xec
[   49.656598]@0 Code: f9400260 f9400a63 1ad92063 37580040 (e7f001f2)
[   49.656604]@0 ---[ end trace cbed09f772fd630d ]---

Change-Id: I931da7cb3841db1f130dba298a7d256b6f02d1bc

6 years agoANDROID: sdcardfs: Fix locking issue with permision fix up
Daniel Rosenberg [Tue, 27 Dec 2016 20:36:29 +0000 (12:36 -0800)]
ANDROID: sdcardfs: Fix locking issue with permision fix up

Don't use lookup_one_len so we can grab the spinlock that
protects d_subdirs.

Bug: 30954918
Change-Id: I0c6a393252db7beb467e0d563739a3a14e1b5115
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Use per mount permissions
Daniel Rosenberg [Thu, 27 Oct 2016 03:27:20 +0000 (20:27 -0700)]
sdcardfs: Use per mount permissions

This switches sdcardfs over to using permission2.
Instead of mounting several sdcardfs instances onto
the same underlaying directory, you bind mount a
single mount several times, and remount with the
options you want. These are stored in the private
mount data, allowing you to maintain the same tree,
but have different permissions for different mount
points.

Warning functions have been added for permission,
as it should never be called, and the correct
behavior is unclear.

Change-Id: I841b1d70ec60cf2b866fa48edeb74a0b0f8334f5
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Add gid and mask to private mount data
Daniel Rosenberg [Thu, 27 Oct 2016 00:36:05 +0000 (17:36 -0700)]
sdcardfs: Add gid and mask to private mount data

Adds support for mount2, remount2, and the functions
to allocate/clone/copy the private data

The next patch will switch over to actually using it.

Change-Id: I8a43da26021d33401f655f0b2784ead161c575e3
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: User new permission2 functions
Daniel Rosenberg [Wed, 26 Oct 2016 23:48:45 +0000 (16:48 -0700)]
sdcardfs: User new permission2 functions

Change-Id: Ic7e0fb8fdcebb31e657b079fe02ac834c4a50db9
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Move directory unlock before touch
Daniel Rosenberg [Mon, 26 Sep 2016 21:48:22 +0000 (14:48 -0700)]
sdcardfs: Move directory unlock before touch

This removes a deadlock under low memory conditions.
filp_open can call lookup_slow, which will attempt to
lock the parent.

Change-Id: I940643d0793f5051d1e79a56f4da2fa8ca3d8ff7
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: fix external storage exporting incorrect uid
alvin_liang [Sun, 14 May 2017 12:30:30 +0000 (12:30 +0000)]
sdcardfs: fix external storage exporting incorrect uid

Symptom: App cannot write into per-app folder
Root Cause: sdcardfs exports incorrect uid
Solution: fix uid
Project: All
Note:
Test done by RD: passed

6 years agosdcardfs: Added top to sdcardfs_inode_info
Daniel Rosenberg [Wed, 18 May 2016 23:57:10 +0000 (16:57 -0700)]
sdcardfs: Added top to sdcardfs_inode_info

Adding packages to the package list and moving files
takes a large amount of locks, and is currently a
heavy operation. This adds a 'top' field to the
inode_info, which points to the inode for the top
most directory whose owner you would like to match.

On permission checks and get_attr, we look up the
owner based on the information at top. When we change
a package mapping, we need only modify the information
in the corresponding top inode_info's. When renaming,
we must ensure top is set correctly in all children.
This happens when an app specific folder gets moved
outside of the folder for that app.

Change-Id: Ib749c60b568e9a45a46f8ceed985c1338246ec6c
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Switch package list to RCU
Daniel Rosenberg [Tue, 10 May 2016 20:42:43 +0000 (13:42 -0700)]
sdcardfs: Switch package list to RCU

Switched the package id hashmap to use RCU.

Change-Id: I9fdcab279009005bf28536247d11e13babab0b93
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Fix locking for permission fix up
Daniel Rosenberg [Tue, 16 Aug 2016 22:19:26 +0000 (15:19 -0700)]
sdcardfs: Fix locking for permission fix up

Iterating over d_subdirs requires taking d_lock.
Removed several unneeded locks.

Change-Id: I5b1588e54c7e6ee19b756d6705171c7f829e2650
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Check for other cases on path lookup
Daniel Rosenberg [Wed, 27 Apr 2016 22:31:29 +0000 (15:31 -0700)]
sdcardfs: Check for other cases on path lookup

This fixes a bug where the first lookup of a
file or folder created under a different view
would not be case insensitive. It will now
search through for a case insensitive match
if the initial lookup fails.

Bug:28024488
Change-Id: I4ff9ce297b9f2f9864b47540e740fd491c545229
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: override umask on mkdir and create
Daniel Rosenberg [Wed, 13 Apr 2016 23:38:34 +0000 (16:38 -0700)]
sdcardfs: override umask on mkdir and create

The mode on files created on the lower fs should
not be affected by the umask of the calling
task's fs_struct. Instead, we create a copy
and modify it as needed. This also lets us avoid
the string shenanigans around .nomedia files.

Bug: 27992761
Change-Id: Ia3a6e56c24c6e19b3b01c1827e46403bb71c2f4c
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agoANDROID: sdcardfs: fix itnull.cocci warnings
Julia Lawall [Wed, 1 Jun 2016 17:28:49 +0000 (10:28 -0700)]
ANDROID: sdcardfs: fix itnull.cocci warnings

List_for_each_entry has the property that the first argument is always
bound to a real list element, never NULL, so testing dentry is not needed.

Generated by: scripts/coccinelle/iterators/itnull.cocci

Change-Id: I51033a2649eb39451862b35b6358fe5cfe25c5f5
Cc: Daniel Rosenberg <drosen@google.com>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Guenter Roeck <groeck@chromium.org>
6 years agosdcardfs: Truncate packages_gid.list on overflow
Daniel Rosenberg [Fri, 8 Jul 2016 21:15:14 +0000 (14:15 -0700)]
sdcardfs: Truncate packages_gid.list on overflow

packages_gid.list was improperly returning the wrong
count. Use scnprintf instead, and inform the user that
the list was truncated if it is.

Bug: 30013843
Change-Id: Ida2b2ef7cd86dd87300bfb4c2cdb6bfe2ee1650d
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agovfs: change d_canonical_path to take two paths
Daniel Rosenberg [Fri, 22 Apr 2016 07:00:14 +0000 (00:00 -0700)]
vfs: change d_canonical_path to take two paths

bug: 23904372
Change-Id: I4a686d64b6de37decf60019be1718e1d820193e6
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: remove unneeded __init and __exit
Daniel Rosenberg [Mon, 28 Mar 2016 23:00:34 +0000 (16:00 -0700)]
sdcardfs: remove unneeded __init and __exit

Change-Id: I2a2d45d52f891332174c3000e8681c5167c1564f

6 years agosdcardfs: Remove unused code
Daniel Rosenberg [Mon, 28 Mar 2016 22:00:20 +0000 (15:00 -0700)]
sdcardfs: Remove unused code

Change-Id: Ie97cba27ce44818ac56cfe40954f164ad44eccf6

6 years agosdcardfs: remove effectless config option
Daniel Rosenberg [Wed, 23 Mar 2016 23:39:30 +0000 (16:39 -0700)]
sdcardfs: remove effectless config option

CONFIG_SDCARD_FS_CI_SEARCH only guards a define for
LOOKUP_CASE_INSENSITIVE, which is never used in the
kernel. Remove both, along with the option matching
that supports it.

Change-Id: I363a8f31de8ee7a7a934d75300cc9ba8176e2edf
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Add support for d_canonicalize
Daniel Rosenberg [Fri, 12 Feb 2016 00:53:36 +0000 (16:53 -0800)]
sdcardfs: Add support for d_canonicalize

Change-Id: I5d6f0e71b8ca99aec4b0894412f1dfd1cfe12add
Signed-off-by: Daniel Rosenberg <drosen@google.com>
6 years agosdcardfs: Bring up to date with Android M permissions:
Daniel Rosenberg [Thu, 4 Feb 2016 05:08:21 +0000 (21:08 -0800)]
sdcardfs: Bring up to date with Android M permissions:

In M, the workings of sdcardfs were changed significantly.
This brings sdcardfs into line with the changes.

Change-Id: I10e91a84a884c838feef7aa26c0a2b21f02e052e

6 years agosdcardfs: Changed type-cast in packagelist management
Daniel Campello [Wed, 24 Jun 2015 22:14:54 +0000 (15:14 -0700)]
sdcardfs: Changed type-cast in packagelist management

Change-Id: Ic8842de2d7274b7a5438938d2febf5d8da867148

6 years agosdcardfs: port to 3.10
Daniel Campello [Fri, 19 Jun 2015 23:11:40 +0000 (16:11 -0700)]
sdcardfs: port to 3.10

Change-Id: I832a14cee3fcbf47ee6e5da2943a90f9dea5b60a

6 years agoInitial port of sdcardfs
Daniel Campello [Fri, 19 Jun 2015 21:31:25 +0000 (14:31 -0700)]
Initial port of sdcardfs

Change-Id: I5b5772a2bbff9f3a7dda641644630a7b8afacec0

 Conflicts:
include/linux/namei.h

fixed-by: vlw <vlwwwwww@gmail.com>

6 years agofs: remove Samsung's sdcardfs
alexax66 [Sun, 14 May 2017 08:44:28 +0000 (08:44 +0000)]
fs: remove Samsung's sdcardfs

6 years agofs: sdfat: Fix compilation for 32-bit targets
Paul Keith [Wed, 14 Mar 2018 13:05:38 +0000 (14:05 +0100)]
fs: sdfat: Fix compilation for 32-bit targets

Change-Id: I9a9f3e253001bfbb3a209bd16d2741c95c99f46b

6 years agofs: sdfat: Update to version 2.0.6-lineage
Paul Keith [Wed, 14 Mar 2018 11:52:08 +0000 (12:52 +0100)]
fs: sdfat: Update to version 2.0.6-lineage

* Samsung version G960USQU1ARBG with -lineage patches applied

Change-Id: I0432751926085aa249b377a418728854618929e5
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Add -lineage extraversion to SDFAT_VERSION
Paul Keith [Fri, 2 Mar 2018 04:30:42 +0000 (05:30 +0100)]
fs: sdfat: Add -lineage extraversion to SDFAT_VERSION

* It has diverged enough to add this to differentiate it

Change-Id: I5e43ee01c785acbc5292c6c115a4e083eeeb36a6
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Add config option to register sdFAT for VFAT
Paul Keith [Fri, 2 Mar 2018 04:10:27 +0000 (05:10 +0100)]
fs: sdfat: Add config option to register sdFAT for VFAT

Change-Id: I72ba7a14b56175535884390e8601960b5d8ed1cf
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Add config option to register sdFAT for exFAT
Paul Keith [Fri, 2 Mar 2018 03:51:53 +0000 (04:51 +0100)]
fs: sdfat: Add config option to register sdFAT for exFAT

Change-Id: Id57abf0a4bd0b433fecc622eecb383cd4ea29d17
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Allow disabling sdfat
Paul Keith [Mon, 19 Feb 2018 19:56:40 +0000 (20:56 +0100)]
fs: sdfat: Allow disabling sdfat

Change-Id: If508804ba4d3536a98c70eb871771d26b628ad50
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Fix compilation without debugging
Paul Keith [Mon, 19 Feb 2018 19:52:10 +0000 (20:52 +0100)]
fs: sdfat: Fix compilation without debugging

* And make WARNON debugging optional

Change-Id: Id59e908c8a60ded1238d3fd010f3d96cdde95f40
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: exfat: Allow disabling exfat
Paul Keith [Mon, 19 Feb 2018 17:16:18 +0000 (18:16 +0100)]
fs: exfat: Allow disabling exfat

Change-Id: If8e8f3a0b7962617ac806c6fb64dbf463b906f59
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: sdfat: Fix compilation on Linux 3.4
Paul Keith [Mon, 19 Feb 2018 17:04:06 +0000 (18:04 +0100)]
fs: sdfat: Fix compilation on Linux 3.4

Change-Id: I3a500f03f399abc9af9586e80419d75aca5b4320
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agofs: Add sdfat
Paul Keith [Mon, 19 Feb 2018 16:31:10 +0000 (17:31 +0100)]
fs: Add sdfat

* Samsung package version: G950FXXU1CRAP (yes, really)

Change-Id: Id866574b34d4434fd4955fac154c9684210abebd
Signed-off-by: Paul Keith <javelinanddart@gmail.com>
6 years agoremove samsung sdfat
Stricted [Sat, 31 Mar 2018 22:23:26 +0000 (00:23 +0200)]
remove samsung sdfat

6 years agoV4L/DVB (13571): v4l: Adding Digital Video Timings APIs
Muralidharan Karicheri [Thu, 19 Nov 2009 15:00:31 +0000 (12:00 -0300)]
V4L/DVB (13571): v4l: Adding Digital Video Timings APIs

This adds the above APIs to the v4l2 core. This is based on version v1.2
of the RFC titled "V4L - Support for video timings at the input/output interface"
Following new ioctls are added:-

        - VIDIOC_ENUM_DV_PRESETS
        - VIDIOC_S_DV_PRESET
        - VIDIOC_G_DV_PRESET
        - VIDIOC_QUERY_DV_PRESET
        - VIDIOC_S_DV_TIMINGS
        - VIDIOC_G_DV_TIMINGS

Please refer to the RFC for the details. This code was tested using vpfe
capture driver on TI's DM365. Following is the test configuration used :-

Blu-Ray HD DVD source -> TVP7002 -> DM365 (VPFE) ->DDR

A draft version of the TVP7002 driver (currently being reviewed in the mailing
list) was used that supports V4L2_DV_1080I60 & V4L2_DV_720P60 presets.

A loopback video capture application was used for testing these APIs. This calls
following IOCTLS :-

 -  verify the new v4l2_input capabilities flag added
 -  Enumerate available presets using VIDIOC_ENUM_DV_PRESETS
 -  Set one of the supported preset using VIDIOC_S_DV_PRESET
 -  Get current preset using VIDIOC_G_DV_PRESET
 -  Detect current preset using VIDIOC_QUERY_DV_PRESET
 -  Using stub functions in tvp7002, verify VIDIOC_S_DV_TIMINGS
    and VIDIOC_G_DV_TIMINGS ioctls are received at the sub device.
 -  Tested on 64bit platform by Hans Verkuil

Change-Id: Idf1492474fb5cbd2c0608f01437fcde6f5ed86fa
Signed-off-by: Muralidharan Karicheri <m-karicheri2@ti.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Reviewed-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
6 years agodrivers: gpu: arm: t72x: r7p0: fix "platform_name" assignment
Stricted [Fri, 30 Mar 2018 23:20:25 +0000 (01:20 +0200)]
drivers: gpu: arm: t72x: r7p0: fix "platform_name" assignment

6 years agosilence firmware warnings
Stricted [Thu, 11 May 2017 08:00:21 +0000 (10:00 +0200)]
silence firmware warnings

6 years agofs: Fix the compile w/o ROOT Restriction enabled
Apavayan Sinha [Tue, 21 Feb 2017 02:17:04 +0000 (07:47 +0530)]
fs: Fix the compile w/o ROOT Restriction enabled

Change-Id: I2cb687617eb1228ad7f62da25449afeb11cdf701

6 years agosysrq: Emergency Remount R/O in reverse order
Amir Goldstein [Mon, 4 Aug 2014 16:29:32 +0000 (19:29 +0300)]
sysrq: Emergency Remount R/O in reverse order

This change fixes a problem where reboot on Android panics the kernel
almost every time when file systems are mounted over loop devices.

Android reboot command does:
- sync
- echo u > /proc/sysrq-trigger
- syscall_reboot

The problem is with sysrq emergency remount R/O trying to remount-ro
in wrong order.
since /data is re-mounted ro before loop devices, loop device
remount-ro fails to flush the journal and panics the kernel:

  EXT4-fs (loop0): Remounting filesystem read-only
  EXT4-fs (loop0): previous I/O error to superblock detected
  loop: Write error at byte offset 0, length 4096.
  Buffer I/O error on device loop0, logical block 0
  lost page write due to I/O error on loop0
  Kernel panic - not syncing: EXT4-fs panic from previous error

The fix is quite simple. In do_emergency_remount(), use
list_for_each_entry_reverse() on sb list instead of list_for_each_entry().
It makes a lot of sense to umount the file systems in reverse order in
which they were added to sb list.

Change-Id: I4370e39b5873bd16ade5d5f9ddb2704beb02a2bb
Signed-off-by: Amir Goldstein <amir@cellrox.com>
Acked-by: Oren Laadan <orenl@cellrox.com>
Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agoarm64: vdso: Define sigtramp offset if needed
Christopher N. Hesse [Mon, 22 Feb 2016 20:32:13 +0000 (21:32 +0100)]
arm64: vdso: Define sigtramp offset if needed

Something breaks vdso offset generation currently, this needs to be
fixed properly.

Change-Id: I3f56a78eb83e61e97c4ffa11a7e865c326084c0a
Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agoKbuild: Add missing videodev2 exynos headers.
Andreas Schneider [Fri, 21 Nov 2014 07:17:02 +0000 (08:17 +0100)]
Kbuild: Add missing videodev2 exynos headers.

Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agodrivers: media: fix compiling
Apavayan Sinha [Sun, 24 Jul 2016 10:21:34 +0000 (15:51 +0530)]
drivers: media: fix compiling

Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agodrivers: muic: fix the compile
Apavayan Sinha [Sun, 24 Jul 2016 09:51:31 +0000 (15:21 +0530)]
drivers: muic: fix the compile

Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agosecurity: Fix tima compile
Apavayan Sinha [Sun, 24 Jul 2016 09:32:04 +0000 (15:02 +0530)]
security: Fix tima compile

Signed-off-by: Apavayan Sinha <info@apavayan.com>
6 years agoimport SM-A510F_EUR_NN_Opensource.zip
Stricted [Wed, 19 Sep 2018 02:06:44 +0000 (04:06 +0200)]
import SM-A510F_EUR_NN_Opensource.zip

A510FXXU4CQH2

10 years agoLinux 3.10.61
Greg Kroah-Hartman [Fri, 21 Nov 2014 17:23:22 +0000 (09:23 -0800)]
Linux 3.10.61

10 years agomm: memcg: handle non-error OOM situations more gracefully
Johannes Weiner [Wed, 16 Oct 2013 20:46:59 +0000 (13:46 -0700)]
mm: memcg: handle non-error OOM situations more gracefully

commit 4942642080ea82d99ab5b653abb9a12b7ba31f4a upstream.

Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full
callstack on OOM") assumed that only a few places that can trigger a
memcg OOM situation do not return VM_FAULT_OOM, like optional page cache
readahead.  But there are many more and it's impractical to annotate
them all.

First of all, we don't want to invoke the OOM killer when the failed
allocation is gracefully handled, so defer the actual kill to the end of
the fault handling as well.  This simplifies the code quite a bit for
added bonus.

Second, since a failed allocation might not be the abrupt end of the
fault, the memcg OOM handler needs to be re-entrant until the fault
finishes for subsequent allocation attempts.  If an allocation is
attempted after the task already OOMed, allow it to bypass the limit so
that it can quickly finish the fault and invoke the OOM killer.

Reported-by: azurIt <azurit@pobox.sk>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomm: memcg: do not trap chargers with full callstack on OOM
Johannes Weiner [Thu, 12 Sep 2013 22:13:44 +0000 (15:13 -0700)]
mm: memcg: do not trap chargers with full callstack on OOM

commit 3812c8c8f3953921ef18544110dafc3505c1ac62 upstream.

The memcg OOM handling is incredibly fragile and can deadlock.  When a
task fails to charge memory, it invokes the OOM killer and loops right
there in the charge code until it succeeds.  Comparably, any other task
that enters the charge path at this point will go to a waitqueue right
then and there and sleep until the OOM situation is resolved.  The problem
is that these tasks may hold filesystem locks and the mmap_sem; locks that
the selected OOM victim may need to exit.

For example, in one reported case, the task invoking the OOM killer was
about to charge a page cache page during a write(), which holds the
i_mutex.  The OOM killer selected a task that was just entering truncate()
and trying to acquire the i_mutex:

OOM invoking task:
  mem_cgroup_handle_oom+0x241/0x3b0
  mem_cgroup_cache_charge+0xbe/0xe0
  add_to_page_cache_locked+0x4c/0x140
  add_to_page_cache_lru+0x22/0x50
  grab_cache_page_write_begin+0x8b/0xe0
  ext3_write_begin+0x88/0x270
  generic_file_buffered_write+0x116/0x290
  __generic_file_aio_write+0x27c/0x480
  generic_file_aio_write+0x76/0xf0           # takes ->i_mutex
  do_sync_write+0xea/0x130
  vfs_write+0xf3/0x1f0
  sys_write+0x51/0x90
  system_call_fastpath+0x18/0x1d

OOM kill victim:
  do_truncate+0x58/0xa0              # takes i_mutex
  do_last+0x250/0xa30
  path_openat+0xd7/0x440
  do_filp_open+0x49/0xa0
  do_sys_open+0x106/0x240
  sys_open+0x20/0x30
  system_call_fastpath+0x18/0x1d

The OOM handling task will retry the charge indefinitely while the OOM
killed task is not releasing any resources.

A similar scenario can happen when the kernel OOM killer for a memcg is
disabled and a userspace task is in charge of resolving OOM situations.
In this case, ALL tasks that enter the OOM path will be made to sleep on
the OOM waitqueue and wait for userspace to free resources or increase
the group's limit.  But a userspace OOM handler is prone to deadlock
itself on the locks held by the waiting tasks.  For example one of the
sleeping tasks may be stuck in a brk() call with the mmap_sem held for
writing but the userspace handler, in order to pick an optimal victim,
may need to read files from /proc/<pid>, which tries to acquire the same
mmap_sem for reading and deadlocks.

This patch changes the way tasks behave after detecting a memcg OOM and
makes sure nobody loops or sleeps with locks held:

1. When OOMing in a user fault, invoke the OOM killer and restart the
   fault instead of looping on the charge attempt.  This way, the OOM
   victim can not get stuck on locks the looping task may hold.

2. When OOMing in a user fault but somebody else is handling it
   (either the kernel OOM killer or a userspace handler), don't go to
   sleep in the charge context.  Instead, remember the OOMing memcg in
   the task struct and then fully unwind the page fault stack with
   -ENOMEM.  pagefault_out_of_memory() will then call back into the
   memcg code to check if the -ENOMEM came from the memcg, and then
   either put the task to sleep on the memcg's OOM waitqueue or just
   restart the fault.  The OOM victim can no longer get stuck on any
   lock a sleeping task may hold.

Debugged by Michal Hocko.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: azurIt <azurit@pobox.sk>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomm: memcg: rework and document OOM waiting and wakeup
Johannes Weiner [Thu, 12 Sep 2013 22:13:43 +0000 (15:13 -0700)]
mm: memcg: rework and document OOM waiting and wakeup

commit fb2a6fc56be66c169f8b80e07ed999ba453a2db2 upstream.

The memcg OOM handler open-codes a sleeping lock for OOM serialization
(trylock, wait, repeat) because the required locking is so specific to
memcg hierarchies.  However, it would be nice if this construct would be
clearly recognizable and not be as obfuscated as it is right now.  Clean
up as follows:

1. Remove the return value of mem_cgroup_oom_unlock()

2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock().

3. Pull the prepare_to_wait() out of the memcg_oom_lock scope.  This
   makes it more obvious that the task has to be on the waitqueue
   before attempting to OOM-trylock the hierarchy, to not miss any
   wakeups before going to sleep.  It just didn't matter until now
   because it was all lumped together into the global memcg_oom_lock
   spinlock section.

4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope.
   It is proctected by the hierarchical OOM-lock.

5. The memcg_oom_lock spinlock is only required to propagate the OOM
   lock in any given hierarchy atomically.  Restrict its scope to
   mem_cgroup_oom_(trylock|unlock).

6. Do not wake up the waitqueue unconditionally at the end of the
   function.  Only the lockholder has to wake up the next in line
   after releasing the lock.

   Note that the lockholder kicks off the OOM-killer, which in turn
   leads to wakeups from the uncharges of the exiting task.  But a
   contender is not guaranteed to see them if it enters the OOM path
   after the OOM kills but before the lockholder releases the lock.
   Thus there has to be an explicit wakeup after releasing the lock.

7. Put the OOM task on the waitqueue before marking the hierarchy as
   under OOM as that is the point where we start to receive wakeups.
   No point in listening before being on the waitqueue.

8. Likewise, unmark the hierarchy before finishing the sleep, for
   symmetry.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomm: memcg: enable memcg OOM killer only for user faults
Johannes Weiner [Thu, 12 Sep 2013 22:13:42 +0000 (15:13 -0700)]
mm: memcg: enable memcg OOM killer only for user faults

commit 519e52473ebe9db5cdef44670d5a97f1fd53d721 upstream.

System calls and kernel faults (uaccess, gup) can handle an out of memory
situation gracefully and just return -ENOMEM.

Enable the memcg OOM killer only for user faults, where it's really the
only option available.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agox86: finish user fault error path with fatal signal
Johannes Weiner [Thu, 12 Sep 2013 22:13:40 +0000 (15:13 -0700)]
x86: finish user fault error path with fatal signal

commit 3a13c4d761b4b979ba8767f42345fed3274991b0 upstream.

The x86 fault handler bails in the middle of error handling when the
task has a fatal signal pending.  For a subsequent patch this is a
problem in OOM situations because it relies on pagefault_out_of_memory()
being called even when the task has been killed, to perform proper
per-task OOM state unwinding.

Shortcutting the fault like this is a rather minor optimization that
saves a few instructions in rare cases.  Just remove it for
user-triggered faults.

Use the opportunity to split the fault retry handling from actual fault
errors and add locking documentation that reads suprisingly similar to
ARM's.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoarch: mm: pass userspace fault flag to generic fault handler
Johannes Weiner [Thu, 12 Sep 2013 22:13:39 +0000 (15:13 -0700)]
arch: mm: pass userspace fault flag to generic fault handler

commit 759496ba6407c6994d6a5ce3a5e74937d7816208 upstream.

Unlike global OOM handling, memory cgroup code will invoke the OOM killer
in any OOM situation because it has no way of telling faults occuring in
kernel context - which could be handled more gracefully - from
user-triggered faults.

Pass a flag that identifies faults originating in user space from the
architecture-specific fault handlers to generic code so that memcg OOM
handling can be improved.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoarch: mm: do not invoke OOM killer on kernel fault OOM
Johannes Weiner [Thu, 12 Sep 2013 22:13:38 +0000 (15:13 -0700)]
arch: mm: do not invoke OOM killer on kernel fault OOM

commit 871341023c771ad233620b7a1fb3d9c7031c4e5c upstream.

Kernel faults are expected to handle OOM conditions gracefully (gup,
uaccess etc.), so they should never invoke the OOM killer.  Reserve this
for faults triggered in user context when it is the only option.

Most architectures already do this, fix up the remaining few.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoarch: mm: remove obsolete init OOM protection
Johannes Weiner [Thu, 12 Sep 2013 22:13:36 +0000 (15:13 -0700)]
arch: mm: remove obsolete init OOM protection

commit 94bce453c78996cc4373d5da6cfabe07fcc6d9f9 upstream.

The memcg code can trap tasks in the context of the failing allocation
until an OOM situation is resolved.  They can hold all kinds of locks
(fs, mm) at this point, which makes it prone to deadlocking.

This series converts memcg OOM handling into a two step process that is
started in the charge context, but any waiting is done after the fault
stack is fully unwound.

Patches 1-4 prepare architecture handlers to support the new memcg
requirements, but in doing so they also remove old cruft and unify
out-of-memory behavior across architectures.

Patch 5 disables the memcg OOM handling for syscalls, readahead, kernel
faults, because they can gracefully unwind the stack with -ENOMEM.  OOM
handling is restricted to user triggered faults that have no other
option.

Patch 6 reworks memcg's hierarchical OOM locking to make it a little
more obvious wth is going on in there: reduce locked regions, rename
locking functions, reorder and document.

Patch 7 implements the two-part OOM handling such that tasks are never
trapped with the full charge stack in an OOM situation.

This patch:

Back before smart OOM killing, when faulting tasks were killed directly on
allocation failures, the arch-specific fault handlers needed special
protection for the init process.

Now that all fault handlers call into the generic OOM killer (see commit
609838cfed97: "mm: invoke oom-killer from remaining unconverted page
fault handlers"), which already provides init protection, the
arch-specific leftovers can be removed.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: azurIt <azurit@pobox.sk>
Acked-by: Vineet Gupta <vgupta@synopsys.com> [arch/arc bits]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomm: invoke oom-killer from remaining unconverted page fault handlers
Johannes Weiner [Mon, 8 Jul 2013 22:59:50 +0000 (15:59 -0700)]
mm: invoke oom-killer from remaining unconverted page fault handlers

commit 609838cfed972d49a65aac7923a9ff5cbe482e30 upstream.

A few remaining architectures directly kill the page faulting task in an
out of memory situation.  This is usually not a good idea since that
task might not even use a significant amount of memory and so may not be
the optimal victim to resolve the situation.

Since 2.6.29's 1c0fe6e ("mm: invoke oom-killer from page fault") there
is a hook that architecture page fault handlers are supposed to call to
invoke the OOM killer and let it pick the right task to kill.  Convert
the remaining architectures over to this hook.

To have the previous behavior of simply taking out the faulting task the
vm.oom_kill_allocating_task sysctl can be set to 1.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Vineet Gupta <vgupta@synopsys.com> [arch/arc bits]
Cc: James Hogan <james.hogan@imgtec.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chen Liqin <liqin.chen@sunplusct.com>
Cc: Lennox Wu <lennox.wu@gmail.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agonet: sctp: fix skb_over_panic when receiving malformed ASCONF chunks
Daniel Borkmann [Thu, 9 Oct 2014 20:55:31 +0000 (22:55 +0200)]
net: sctp: fix skb_over_panic when receiving malformed ASCONF chunks

commit 9de7922bc709eee2f609cd01d98aaedc4cf5ea74 upstream.

Commit 6f4c618ddb0 ("SCTP : Add paramters validity check for
ASCONF chunk") added basic verification of ASCONF chunks, however,
it is still possible to remotely crash a server by sending a
special crafted ASCONF chunk, even up to pre 2.6.12 kernels:

skb_over_panic: text:ffffffffa01ea1c3 len:31056 put:30768
 head:ffff88011bd81800 data:ffff88011bd81800 tail:0x7950
 end:0x440 dev:<NULL>
 ------------[ cut here ]------------
kernel BUG at net/core/skbuff.c:129!
[...]
Call Trace:
 <IRQ>
 [<ffffffff8144fb1c>] skb_put+0x5c/0x70
 [<ffffffffa01ea1c3>] sctp_addto_chunk+0x63/0xd0 [sctp]
 [<ffffffffa01eadaf>] sctp_process_asconf+0x1af/0x540 [sctp]
 [<ffffffff8152d025>] ? _read_unlock_bh+0x15/0x20
 [<ffffffffa01e0038>] sctp_sf_do_asconf+0x168/0x240 [sctp]
 [<ffffffffa01e3751>] sctp_do_sm+0x71/0x1210 [sctp]
 [<ffffffff8147645d>] ? fib_rules_lookup+0xad/0xf0
 [<ffffffffa01e6b22>] ? sctp_cmp_addr_exact+0x32/0x40 [sctp]
 [<ffffffffa01e8393>] sctp_assoc_bh_rcv+0xd3/0x180 [sctp]
 [<ffffffffa01ee986>] sctp_inq_push+0x56/0x80 [sctp]
 [<ffffffffa01fcc42>] sctp_rcv+0x982/0xa10 [sctp]
 [<ffffffffa01d5123>] ? ipt_local_in_hook+0x23/0x28 [iptable_filter]
 [<ffffffff8148bdc9>] ? nf_iterate+0x69/0xb0
 [<ffffffff81496d10>] ? ip_local_deliver_finish+0x0/0x2d0
 [<ffffffff8148bf86>] ? nf_hook_slow+0x76/0x120
 [<ffffffff81496d10>] ? ip_local_deliver_finish+0x0/0x2d0
 [<ffffffff81496ded>] ip_local_deliver_finish+0xdd/0x2d0
 [<ffffffff81497078>] ip_local_deliver+0x98/0xa0
 [<ffffffff8149653d>] ip_rcv_finish+0x12d/0x440
 [<ffffffff81496ac5>] ip_rcv+0x275/0x350
 [<ffffffff8145c88b>] __netif_receive_skb+0x4ab/0x750
 [<ffffffff81460588>] netif_receive_skb+0x58/0x60

This can be triggered e.g., through a simple scripted nmap
connection scan injecting the chunk after the handshake, for
example, ...

  -------------- INIT[ASCONF; ASCONF_ACK] ------------->
  <----------- INIT-ACK[ASCONF; ASCONF_ACK] ------------
  -------------------- COOKIE-ECHO -------------------->
  <-------------------- COOKIE-ACK ---------------------
  ------------------ ASCONF; UNKNOWN ------------------>

... where ASCONF chunk of length 280 contains 2 parameters ...

  1) Add IP address parameter (param length: 16)
  2) Add/del IP address parameter (param length: 255)

... followed by an UNKNOWN chunk of e.g. 4 bytes. Here, the
Address Parameter in the ASCONF chunk is even missing, too.
This is just an example and similarly-crafted ASCONF chunks
could be used just as well.

The ASCONF chunk passes through sctp_verify_asconf() as all
parameters passed sanity checks, and after walking, we ended
up successfully at the chunk end boundary, and thus may invoke
sctp_process_asconf(). Parameter walking is done with
WORD_ROUND() to take padding into account.

In sctp_process_asconf()'s TLV processing, we may fail in
sctp_process_asconf_param() e.g., due to removal of the IP
address that is also the source address of the packet containing
the ASCONF chunk, and thus we need to add all TLVs after the
failure to our ASCONF response to remote via helper function
sctp_add_asconf_response(), which basically invokes a
sctp_addto_chunk() adding the error parameters to the given
skb.

When walking to the next parameter this time, we proceed
with ...

  length = ntohs(asconf_param->param_hdr.length);
  asconf_param = (void *)asconf_param + length;

... instead of the WORD_ROUND()'ed length, thus resulting here
in an off-by-one that leads to reading the follow-up garbage
parameter length of 12336, and thus throwing an skb_over_panic
for the reply when trying to sctp_addto_chunk() next time,
which implicitly calls the skb_put() with that length.

Fix it by using sctp_walk_params() [ which is also used in
INIT parameter processing ] macro in the verification *and*
in ASCONF processing: it will make sure we don't spill over,
that we walk parameters WORD_ROUND()'ed. Moreover, we're being
more defensive and guard against unknown parameter types and
missized addresses.

Joint work with Vlad Yasevich.

Fixes: b896b82be4ae ("[SCTP] ADDIP: Support for processing incoming ASCONF_ACK chunks.")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Vlad Yasevich <vyasevich@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Josh Boyer <jwboyer@fedoraproject.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agonet: sctp: fix panic on duplicate ASCONF chunks
Daniel Borkmann [Thu, 9 Oct 2014 20:55:32 +0000 (22:55 +0200)]
net: sctp: fix panic on duplicate ASCONF chunks

commit b69040d8e39f20d5215a03502a8e8b4c6ab78395 upstream.

When receiving a e.g. semi-good formed connection scan in the
form of ...

  -------------- INIT[ASCONF; ASCONF_ACK] ------------->
  <----------- INIT-ACK[ASCONF; ASCONF_ACK] ------------
  -------------------- COOKIE-ECHO -------------------->
  <-------------------- COOKIE-ACK ---------------------
  ---------------- ASCONF_a; ASCONF_b ----------------->

... where ASCONF_a equals ASCONF_b chunk (at least both serials
need to be equal), we panic an SCTP server!

The problem is that good-formed ASCONF chunks that we reply with
ASCONF_ACK chunks are cached per serial. Thus, when we receive a
same ASCONF chunk twice (e.g. through a lost ASCONF_ACK), we do
not need to process them again on the server side (that was the
idea, also proposed in the RFC). Instead, we know it was cached
and we just resend the cached chunk instead. So far, so good.

Where things get nasty is in SCTP's side effect interpreter, that
is, sctp_cmd_interpreter():

While incoming ASCONF_a (chunk = event_arg) is being marked
!end_of_packet and !singleton, and we have an association context,
we do not flush the outqueue the first time after processing the
ASCONF_ACK singleton chunk via SCTP_CMD_REPLY. Instead, we keep it
queued up, although we set local_cork to 1. Commit 2e3216cd54b1
changed the precedence, so that as long as we get bundled, incoming
chunks we try possible bundling on outgoing queue as well. Before
this commit, we would just flush the output queue.

Now, while ASCONF_a's ASCONF_ACK sits in the corked outq, we
continue to process the same ASCONF_b chunk from the packet. As
we have cached the previous ASCONF_ACK, we find it, grab it and
do another SCTP_CMD_REPLY command on it. So, effectively, we rip
the chunk->list pointers and requeue the same ASCONF_ACK chunk
another time. Since we process ASCONF_b, it's correctly marked
with end_of_packet and we enforce an uncork, and thus flush, thus
crashing the kernel.

Fix it by testing if the ASCONF_ACK is currently pending and if
that is the case, do not requeue it. When flushing the output
queue we may relink the chunk for preparing an outgoing packet,
but eventually unlink it when it's copied into the skb right
before transmission.

Joint work with Vlad Yasevich.

Fixes: 2e3216cd54b1 ("sctp: Follow security requirement of responding with 1 packet")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Josh Boyer <jwboyer@fedoraproject.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agonet: sctp: fix remote memory pressure from excessive queueing
Daniel Borkmann [Thu, 9 Oct 2014 20:55:33 +0000 (22:55 +0200)]
net: sctp: fix remote memory pressure from excessive queueing

commit 26b87c7881006311828bb0ab271a551a62dcceb4 upstream.

This scenario is not limited to ASCONF, just taken as one
example triggering the issue. When receiving ASCONF probes
in the form of ...

  -------------- INIT[ASCONF; ASCONF_ACK] ------------->
  <----------- INIT-ACK[ASCONF; ASCONF_ACK] ------------
  -------------------- COOKIE-ECHO -------------------->
  <-------------------- COOKIE-ACK ---------------------
  ---- ASCONF_a; [ASCONF_b; ...; ASCONF_n;] JUNK ------>
  [...]
  ---- ASCONF_m; [ASCONF_o; ...; ASCONF_z;] JUNK ------>

... where ASCONF_a, ASCONF_b, ..., ASCONF_z are good-formed
ASCONFs and have increasing serial numbers, we process such
ASCONF chunk(s) marked with !end_of_packet and !singleton,
since we have not yet reached the SCTP packet end. SCTP does
only do verification on a chunk by chunk basis, as an SCTP
packet is nothing more than just a container of a stream of
chunks which it eats up one by one.

We could run into the case that we receive a packet with a
malformed tail, above marked as trailing JUNK. All previous
chunks are here goodformed, so the stack will eat up all
previous chunks up to this point. In case JUNK does not fit
into a chunk header and there are no more other chunks in
the input queue, or in case JUNK contains a garbage chunk
header, but the encoded chunk length would exceed the skb
tail, or we came here from an entirely different scenario
and the chunk has pdiscard=1 mark (without having had a flush
point), it will happen, that we will excessively queue up
the association's output queue (a correct final chunk may
then turn it into a response flood when flushing the
queue ;)): I ran a simple script with incremental ASCONF
serial numbers and could see the server side consuming
excessive amount of RAM [before/after: up to 2GB and more].

The issue at heart is that the chunk train basically ends
with !end_of_packet and !singleton markers and since commit
2e3216cd54b1 ("sctp: Follow security requirement of responding
with 1 packet") therefore preventing an output queue flush
point in sctp_do_sm() -> sctp_cmd_interpreter() on the input
chunk (chunk = event_arg) even though local_cork is set,
but its precedence has changed since then. In the normal
case, the last chunk with end_of_packet=1 would trigger the
queue flush to accommodate possible outgoing bundling.

In the input queue, sctp_inq_pop() seems to do the right thing
in terms of discarding invalid chunks. So, above JUNK will
not enter the state machine and instead be released and exit
the sctp_assoc_bh_rcv() chunk processing loop. It's simply
the flush point being missing at loop exit. Adding a try-flush
approach on the output queue might not work as the underlying
infrastructure might be long gone at this point due to the
side-effect interpreter run.

One possibility, albeit a bit of a kludge, would be to defer
invalid chunk freeing into the state machine in order to
possibly trigger packet discards and thus indirectly a queue
flush on error. It would surely be better to discard chunks
as in the current, perhaps better controlled environment, but
going back and forth, it's simply architecturally not possible.
I tried various trailing JUNK attack cases and it seems to
look good now.

Joint work with Vlad Yasevich.

Fixes: 2e3216cd54b1 ("sctp: Follow security requirement of responding with 1 packet")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Josh Boyer <jwboyer@fedoraproject.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoKVM: x86: Don't report guest userspace emulation error to userspace
Nadav Amit [Tue, 16 Sep 2014 23:50:50 +0000 (02:50 +0300)]
KVM: x86: Don't report guest userspace emulation error to userspace

commit a2b9e6c1a35afcc0973acb72e591c714e78885ff upstream.

Commit fc3a9157d314 ("KVM: X86: Don't report L2 emulation failures to
user-space") disabled the reporting of L2 (nested guest) emulation failures to
userspace due to race-condition between a vmexit and the instruction emulator.
The same rational applies also to userspace applications that are permitted by
the guest OS to access MMIO area or perform PIO.

This patch extends the current behavior - of injecting a #UD instead of
reporting it to userspace - also for guest userspace code.

Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoSCSI: hpsa: fix a race in cmd_free/scsi_done
Tomas Henzl [Thu, 1 Aug 2013 13:14:00 +0000 (15:14 +0200)]
SCSI: hpsa: fix a race in cmd_free/scsi_done

commit 2cc5bfaf854463d9d1aa52091f60110fbf102a96 upstream.

When the driver calls scsi_done and after that frees it's internal
preallocated memory it can happen that a new job is enqueud before
the memory is freed. The allocation fails and the message
"cmd_alloc returned NULL" is shown.
Patch below fixes it by moving cmd->scsi_done after cmd_free.

Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Acked-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Cc: Masoud Sharbiani <msharbiani@twitter.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agonet/mlx4_en: Fix BlueFlame race
Eugenia Emantayev [Thu, 25 Jul 2013 16:21:23 +0000 (19:21 +0300)]
net/mlx4_en: Fix BlueFlame race

commit 2d4b646613d6b12175b017aca18113945af1faf3 upstream.

Fix a race between BlueFlame flow and stamping in post send flow.
Example:
SW: Build WQE 0 on the TX buffer, except the ownership bit
SW: Set ownership for WQE 0 on the TX buffer
SW: Ring doorbell for WQE 0
SW: Build WQE 1 on the TX buffer, except the ownership bit
SW: Set ownership for WQE 1 on the TX buffer
HW: Read WQE 0 and then WQE 1, before doorbell was rung/BF was done for WQE 1
HW: Produce CQEs for WQE 0 and WQE 1
SW: Process the CQEs, and stamp WQE 0 and WQE 1 accordingly (on the TX buffer)
SW: Copy WQE 1 from the TX buffer to the BF register - ALREADY STAMPED!
HW: CQE error with index 0xFFFF  - the BF WQE's control segment is STAMPED,
so the BF index is 0xFFFF. Error: Invalid Opcode.
As a result QP enters the error state and no traffic can be sent.

Solution:
When stamping - do not stamp last completed wqe.

Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Vinson Lee <vlee@twopensource.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoARM: Correct BUG() assembly to ensure it is endian-agnostic
Ben Dooks [Thu, 25 Jul 2013 13:38:03 +0000 (14:38 +0100)]
ARM: Correct BUG() assembly to ensure it is endian-agnostic

commit 63328070eff2f4fd730c86966a0dbc976147c39f upstream.

Currently BUG() uses .word or .hword to create the necessary illegal
instructions. However if we are building BE8 then these get swapped
by the linker into different illegal instructions in the text. This
means that the BUG() macro does not get trapped properly.

Change to using <asm/opcodes.h> to provide the necessary ARM instruction
building as we cannot rely on gcc/gas having the `.inst` instructions
which where added to try and resolve this issue (reported by Dave Martin
<Dave.Martin@arm.com>).

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Cc: Wang Nan <wangnan0@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoperf/x86/intel: Use proper dTLB-load-misses event on IvyBridge
Vince Weaver [Mon, 14 Jul 2014 19:33:25 +0000 (15:33 -0400)]
perf/x86/intel: Use proper dTLB-load-misses event on IvyBridge

commit 1996388e9f4e3444db8273bc08d25164d2967c21 upstream.

This was discussed back in February:

https://lkml.org/lkml/2014/2/18/956

But I never saw a patch come out of it.

On IvyBridge we share the SandyBridge cache event tables, but the
dTLB-load-miss event is not compatible.  Patch it up after
the fact to the proper DTLB_LOAD_MISSES.DEMAND_LD_MISS_CAUSES_A_WALK

Signed-off-by: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1407141528200.17214@vincent-weaver-1.umelst.maine.edu
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Hou Pengyang <houpengyang@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>