firefly-linux-kernel-4.4.55.git
8 years agoandroid: configs: Initial commit of Android config fragments
Rom Lemarchand [Wed, 6 Feb 2013 23:49:47 +0000 (15:49 -0800)]
android: configs: Initial commit of Android config fragments

- Add 2 files that contain the minimal and recommended kernel config
options respectively.
- Add a README to explain their purpose and how to use them to
generate a device config compatible with Android.

Change-Id: I3a4883f3b04d2820e90ceb3c4d02390d6458d6ce
Signed-off-by: Rom Lemarchand <romlem@google.com>
8 years agoandroid: skip building drivers as modules
Amit Pundir [Tue, 29 Dec 2015 07:29:31 +0000 (12:59 +0530)]
android: skip building drivers as modules

Few Android drivers e.g. uid_cputime and PPPoLAC/oPNS,
keyreset/combo, cpufreq_interactive etc.. fail to build
as kernel modules. Instead of fixing the build lets
make these drivers non-modular (switch config to "bool"
from "tristate" in Kconfig) since Android doesn't
support building kernel modules anyway.

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agowakeup: Add the guard condition for len in pm_get_active_wakeup_sources
Ruchi Kandoi [Tue, 10 Nov 2015 18:53:55 +0000 (10:53 -0800)]
wakeup: Add the guard condition for len in pm_get_active_wakeup_sources

Check if the len is not greater than maximum to prevent buffer overflow.

Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
Change-Id: I575b0a72bb5448b68353408d71fa8b83420c9088

8 years agoandroid: drivers: workaround debugfs race in binder
Riley Andrews [Mon, 9 Nov 2015 21:16:32 +0000 (13:16 -0800)]
android: drivers: workaround debugfs race in binder

If a /d/binder/proc/[pid] entry is kept open after linux has
torn down the associated process, binder_proc_show can deference
an invalid binder_proc that has been stashed in the debugfs
inode.  Validate that the binder_proc ptr passed into binder_proc_show
has not been freed by looking for it within the global process list
whilst the global lock is held. If the ptr is not valid, print nothing.

Bug: 19587483

Change-Id: Idd5ad79f5648b7eed49d1ec75ae93f9e12a74ee9
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoandroid: binder: More offset validation
Arve Hjønnevåg [Mon, 9 Nov 2015 21:08:12 +0000 (13:08 -0800)]
android: binder: More offset validation

Make sure offsets don't point to overlapping flat_binder_object
structs.

Change-Id: I12c3757872e0f16dbd6e3b92fd214004cf87047e
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoUPSTREAM: KEYS: Fix keyring ref leak in join_session_keyring()
Yevgeny Pats [Tue, 19 Jan 2016 22:09:04 +0000 (22:09 +0000)]
UPSTREAM: KEYS: Fix keyring ref leak in join_session_keyring()

(cherry pick from commit 23567fd052a9abb6d67fe8e7a9ccdd9800a540f2)

This fixes CVE-2016-0728.

If a thread is asked to join as a session keyring the keyring that's already
set as its session, we leak a keyring reference.

This can be tested with the following program:

#include <stddef.h>
#include <stdio.h>
#include <sys/types.h>
#include <keyutils.h>

int main(int argc, const char *argv[])
{
int i = 0;
key_serial_t serial;

serial = keyctl(KEYCTL_JOIN_SESSION_KEYRING,
"leaked-keyring");
if (serial < 0) {
perror("keyctl");
return -1;
}

if (keyctl(KEYCTL_SETPERM, serial,
   KEY_POS_ALL | KEY_USR_ALL) < 0) {
perror("keyctl");
return -1;
}

for (i = 0; i < 100; i++) {
serial = keyctl(KEYCTL_JOIN_SESSION_KEYRING,
"leaked-keyring");
if (serial < 0) {
perror("keyctl");
return -1;
}
}

return 0;
}

If, after the program has run, there something like the following line in
/proc/keys:

3f3d898f I--Q---   100 perm 3f3f0000     0     0 keyring   leaked-keyring: empty

with a usage count of 100 * the number of times the program has been run,
then the kernel is malfunctioning.  If leaked-keyring has zero usages or
has been garbage collected, then the problem is fixed.

Reported-by: Yevgeny Pats <yevgeny@perception-point.io>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Acked-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: James Morris <james.l.morris@oracle.com>
Change-Id: I6052fa14ab54e32878ac9895356f1049bb96a138

8 years agogoldfish_bus: switch GOLDFISH_BUS config back to bool
Amit Pundir [Tue, 29 Dec 2015 07:09:11 +0000 (12:39 +0530)]
goldfish_bus: switch GOLDFISH_BUS config back to bool

AOSP commit a61e4d271cba "Enable adb with android-pipe
in IA image" made GOLDFISH_BUS modular (i.e switched it
to "tristate" in Kconfig from "bool"), which is not
buildable. Instead of fixing the build and supporting
it in modular form, which AOSP doesn't agree with anyway,
lets restore it to the original state ("bool") in the Kconfig.

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agogoldfish: pipe: fix platform_no_drv_owner.cocci warnings
Fengguang Wu [Sat, 26 Dec 2015 14:55:02 +0000 (20:25 +0530)]
goldfish: pipe: fix platform_no_drv_owner.cocci warnings

No need to set .owner here. The core will do it.

 Remove .owner field if calls are used which set it automatically

Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci

CC: Lizhe Liu <lizhe.liu@intel.com>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agoEnable adb with android-pipe in IA image
Lizhe Liu [Tue, 13 Oct 2015 05:27:20 +0000 (13:27 +0800)]
Enable adb with android-pipe in IA image

To enable adb for x86/x86_64 Brillo emulator, we bring
drivers/platform/goldfish/ in kernel/common (branch android-3.18)
up to date with that in kernel/goldfish
(branch android-goldfish-3.10), by porting the following patches:

044d26f goldfish: Enable ACPI-based enumeration for android pipe
acf92a5 goldfish_pipe: Pass physical addresses to the device if supported
77559b0 [MIPS] Enable platform support for Goldfish virtual devices
1bebc76 platform: goldfish: pipe: don't log when dropping PIPE_ERROR_AGAIN
3c56d07 platform: goldfish: pipe: add devicetree bindings
ca8dafc android_pipe: Pin pages to memory while copying and other cleanups
b765d47 android_pipe: don't be clever with #define offsets
7119108 goldfish: refactor goldfish platform configs
dc02035 goldfish: fix kernel panic when using multiple adb connection

Change-Id: Ic4f2f5e43ba2a70831d6a12a370417984f784dbc
Signed-off-by: Lizhe Liu <lizhe.liu@intel.com>
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Signed-off-by: Jason Hu <jia-cheng.hu@intel.com>
Signed-off-by: Yu Ning <yu.ning@intel.com>
8 years agoepoll: use freezable blocking call
Colin Cross [Mon, 6 May 2013 23:50:16 +0000 (23:50 +0000)]
epoll: use freezable blocking call

Avoid waking up every thread sleeping in an epoll_wait call during
suspend and resume by calling a freezable blocking call.  Previous
patches modified the freezer to avoid sending wakeups to threads
that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because
it doesn't hold any locks or release any resources when interrupted
that might be needed by another freezing task or a kernel driver
during suspend, and is a common site where idle userspace tasks are
blocked.

Change-Id: I848d08d28c89302fd42bbbdfa76489a474ab27bf
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agoFix for in kernel emergency remount when loop mounts are used
Christian Poetzsch [Fri, 24 Jul 2015 15:42:58 +0000 (16:42 +0100)]
Fix for in kernel emergency remount when loop mounts are used

adb reboot calls /proc/sysrq-trigger to force an emergency remount (ro) of all
mounted disks. This is executed in the order of the time the mount was originally
done. Because we have a test system which loop mount images from an extra
partition, we see errors cause the loop mounted partitions gets remounted after
this physical partition was set to read only already.

Fix this by reversing the order of the emergency remount. This will remount the
disk first which have been mounted last.

So instead of remounting in this order:
 /dev/sda1
 /dev/loop1
 /dev/loop2
we now remount in this order:
 /dev/loop2
 /dev/loop1
 /dev/sda1

Change-Id: I68fe7e16cc9400ab5278877af70c9ea1d9b57936
Signed-off-by: Christian Poetzsch <christian.potzsch@imgtec.com>
8 years agokbuild: Makefile.clean: make Kbuild and Makefile optional
Amit Pundir [Tue, 5 Jan 2016 12:06:31 +0000 (17:36 +0530)]
kbuild: Makefile.clean: make Kbuild and Makefile optional

AOSP commit b13ce9f4aa6f "ARM64: add option to build
Image.gz/dtb combo" broke archclean / mrproper build
targets and we run into:
----------
./scripts/Makefile.clean:14: arch/arm64/boot/amd/Makefile: No such file or directory
make[2]: *** No rule to make target `arch/arm64/boot/amd/Makefile'.  Stop.
make[1]: *** [arch/arm64/boot/amd] Error 2
make: *** [archclean] Error 2
----------

This patch skip the missing Kbuild/Makefile reporting
error. It does the job (i.e cleanup dts/*/*.dtb and
do not spit out missing file error messages as well).

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agokbuild: make it possible to specify the module output dir
Rom Lemarchand [Fri, 6 Feb 2015 00:07:59 +0000 (16:07 -0800)]
kbuild: make it possible to specify the module output dir

Make modinst_dir user-defined on the command line.

This allows to do things like:
make MODLIB=output/ modinst_dir=. modules_install

to ensure all the .ko are in the output/ directory.

Change-Id: I2bc007eea27ee744d35289e26e4a8ac43ba04151
Signed-off-by: Rom Lemarchand <romlem@android.com>
8 years agoext4: Add support for FIDTRIM, a best-effort ioctl for deep discard trim
JP Abgrall [Wed, 23 Jul 2014 23:55:07 +0000 (16:55 -0700)]
ext4: Add support for FIDTRIM, a best-effort ioctl for deep discard trim

* What
This provides an interface for issuing an FITRIM which uses the
secure discard instead of just a discard.
Only the eMMC command is "secure", and not how the FS uses it:
due to the fact that the FS might reassign a region somewhere else,
the original deleted data will not be affected by the "trim" which only
handles un-used regions.
So we'll just call it "deep discard", and note that this is a
"best effort" cleanup.

* Why
Once in a while, We want to be able to cleanup most of the unused blocks
after erasing a bunch of files.
We don't want to constantly secure-discard via a mount option.

From an eMMC spec perspective, it tells the device to really get rid of
all the data for the specified blocks and not just put them back into the
pool of free ones (unlike the normal TRIM). The eMMC spec says the
secure trim handling must make sure the data (and metadata) is not available
anymore. A simple TRIM doesn't clear the data, it just puts blocks in the
free pool.
JEDEC Standard No. 84-A441
  7.6.9 Secure Erase
  7.6.10 Secure Trim

From an FS perspective, it is acceptable to leave some data behind.
 - directory entries related to deleted files
 - databases entries related to deleted files
 - small-file data stored in inode extents
 - blocks held by the FS waiting to be re-used (mitigated by sync).
 - blocks reassigned by the FS prior to FIDTRIM.

Change-Id: I676a1404a80130d93930c84898360f2e6fb2f81e
Signed-off-by: Geremy Condra <gcondra@google.com>
Signed-off-by: JP Abgrall <jpa@google.com>
8 years agohardlockup: detect hard lockups without NMIs using secondary cpus
Colin Cross [Fri, 11 Jan 2013 21:51:48 +0000 (13:51 -0800)]
hardlockup: detect hard lockups without NMIs using secondary cpus

Emulate NMIs on systems where they are not available by using timer
interrupts on other cpus.  Each cpu will use its softlockup hrtimer
to check that the next cpu is processing hrtimer interrupts by
verifying that a counter is increasing.

This patch is useful on systems where the hardlockup detector is not
available due to a lack of NMIs, for example most ARM SoCs.
Without this patch any cpu stuck with interrupts disabled can
cause a hardware watchdog reset with no debugging information,
but with this patch the kernel can detect the lockup and panic,
which can result in useful debugging info.

Change-Id: Ia5faf50243e19c1755201212e04c8892d929785a
Signed-off-by: Colin Cross <ccross@android.com>
8 years agoANDROID: rtc-palmas: correct for bcd year
Mark Salyzyn [Wed, 30 Dec 2015 17:26:15 +0000 (09:26 -0800)]
ANDROID: rtc-palmas: correct for bcd year

Replace bcd2bin and bin2bcd with one that maps years 1970 to 2129
in a pattern that works with the underlying hardware.

The only transition that does not work correctly for this rtc clock
is the transition from 2099 to 2100, it proceeds to 2000. The rtc
clock retains and transitions the year correctly in all other
circumstances.

Signed-off-by: Mark Salyzyn <salyzyn@google.com>
Bug: 26346842
Change-Id: Ie527700190b1ae4b4bc3c12279d875aa5985b168

8 years agow1: ds2482: Manage SLPZ pin sleep state
Todd Poynor [Fri, 31 Aug 2012 06:09:14 +0000 (23:09 -0700)]
w1: ds2482: Manage SLPZ pin sleep state

Place SLPZ pin in sleep state at system suspend time if a GPIO is
provided by board platform data.

Change-Id: I93c61fa0ae474e968e0f909209c9bfcaafe3dd2c
Signed-off-by: Todd Poynor <toddpoynor@google.com>
8 years agofuse: Freeze client on suspend when request sent to userspace
Todd Poynor [Wed, 24 Aug 2011 22:01:30 +0000 (15:01 -0700)]
fuse: Freeze client on suspend when request sent to userspace

Suspend attempts can abort when the FUSE daemon is already frozen
and a client is waiting uninterruptibly for a response, causing
freezing of tasks to fail.

Use the freeze-friendly wait API, but disregard other signals.

Change-Id: Icefb7e4bbc718ccb76bf3c04daaa5eeea7e0e63c
Signed-off-by: Todd Poynor <toddpoynor@google.com>
8 years agoserial_core: Add wake_peer uart operation
San Mehat [Thu, 30 Jul 2009 03:21:28 +0000 (20:21 -0700)]
serial_core: Add wake_peer uart operation

Add wake_peer which is called before starting UART TX. The idea here
is to provide a mechanism where we can wakeup our peer before sending
data.

Change-Id: I42e0779b635f64ca99184b45d5b028de80197491
Signed-off-by: San Mehat <san@google.com>
8 years agoRevert "ARM: Blacklist GCC 4.8.0 to GCC 4.8.2 - PR58854"
Ruchi Kandoi [Mon, 26 Jan 2015 18:24:14 +0000 (10:24 -0800)]
Revert "ARM: Blacklist GCC 4.8.0 to GCC 4.8.2 - PR58854"

This reverts commit 7fc150543c73de71859631c8a6b17e3067fe7617.

8 years agouserfaultfd: Add missing vma_merge parameter
Dmitry Shmidt [Mon, 28 Dec 2015 18:53:45 +0000 (10:53 -0800)]
userfaultfd: Add missing vma_merge parameter

Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoFROMLIST: x86: mm: support ARCH_MMAP_RND_BITS.
dcashman [Tue, 12 Jan 2016 17:51:16 +0000 (09:51 -0800)]
FROMLIST: x86: mm: support ARCH_MMAP_RND_BITS.

(cherry picked from commit https://lkml.org/lkml/2015/12/21/339)

x86: arch_mmap_rnd() uses hard-coded values, 8 for 32-bit and 28 for
64-bit, to generate the random offset for the mmap base address.
This value represents a compromise between increased ASLR
effectiveness and avoiding address-space fragmentation. Replace it
with a Kconfig option, which is sensibly bounded, so that platform
developers may choose where to place this compromise. Keep default
values as new minimums.

Bug: 24047224
Signed-off-by: Daniel Cashman <dcashman@android.com>
Signed-off-by: Daniel Cashman <dcashman@google.com>
Change-Id: Ic38735a8de2943843a73b5c20855ccfa92513422

8 years agoFROMLIST: arm64: mm: support ARCH_MMAP_RND_BITS.
dcashman [Tue, 12 Jan 2016 17:47:53 +0000 (09:47 -0800)]
FROMLIST: arm64: mm: support ARCH_MMAP_RND_BITS.

(cherry picked from commit https://lkml.org/lkml/2015/12/21/340)

arm64: arch_mmap_rnd() uses STACK_RND_MASK to generate the
random offset for the mmap base address.  This value represents a
compromise between increased ASLR effectiveness and avoiding
address-space fragmentation. Replace it with a Kconfig option, which
is sensibly bounded, so that platform developers may choose where to
place this compromise. Keep default values as new minimums.

Bug: 24047224
Signed-off-by: Daniel Cashman <dcashman@android.com>
Signed-off-by: Daniel Cashman <dcashman@google.com>
Change-Id: I0be0bf8b1ed412863f248323e2d86b1df5bf21c6

8 years agoFROMLIST: arm: mm: support ARCH_MMAP_RND_BITS.
dcashman [Tue, 12 Jan 2016 17:44:35 +0000 (09:44 -0800)]
FROMLIST: arm: mm: support ARCH_MMAP_RND_BITS.

(cherry picked from commit https://lkml.org/lkml/2015/12/21/341)

arm: arch_mmap_rnd() uses a hard-code value of 8 to generate the
random offset for the mmap base address.  This value represents a
compromise between increased ASLR effectiveness and avoiding
address-space fragmentation. Replace it with a Kconfig option, which
is sensibly bounded, so that platform developers may choose where to
place this compromise. Keep 8 as the minimum acceptable value.

Bug: 24047224
Signed-off-by: Daniel Cashman <dcashman@android.com>
Signed-off-by: Daniel Cashman <dcashman@google.com>
Change-Id: I438ae2dae939ea7f311246832cb789afdd6cba4e

8 years agoFROMLIST: mm: mmap: Add new /proc tunable for mmap_base ASLR.
dcashman [Tue, 12 Jan 2016 17:18:57 +0000 (09:18 -0800)]
FROMLIST: mm: mmap: Add new /proc tunable for mmap_base ASLR.

(cherry picked from commit https://lkml.org/lkml/2015/12/21/337)

ASLR  only uses as few as 8 bits to generate the random offset for the
mmap base address on 32 bit architectures. This value was chosen to
prevent a poorly chosen value from dividing the address space in such
a way as to prevent large allocations. This may not be an issue on all
platforms. Allow the specification of a minimum number of bits so that
platforms desiring greater ASLR protection may determine where to place
the trade-off.

Bug: 24047224
Signed-off-by: Daniel Cashman <dcashman@android.com>
Signed-off-by: Daniel Cashman <dcashman@google.com>
Change-Id: Ibf9ed3d4390e9686f5cc34f605d509a20d40e6c2

8 years agomm: private anonymous memory build fixes for 4.4
Amit Pundir [Thu, 10 Dec 2015 12:54:51 +0000 (18:24 +0530)]
mm: private anonymous memory build fixes for 4.4

Update vma_merge() call in private anonymous memory prctl,
introduced in AOSP commit ee8c5f78f09a
"mm: add a field to store names for private anonymous memory",
so as to align with changes from upstream commit 19a809afe2fe
"userfaultfd: teach vma_merge to merge across vma->vm_userfaultfd_ctx".

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agomm: add a field to store names for private anonymous memory
Colin Cross [Tue, 27 Oct 2015 23:42:08 +0000 (16:42 -0700)]
mm: add a field to store names for private anonymous memory

Userspace processes often have multiple allocators that each do
anonymous mmaps to get memory.  When examining memory usage of
individual processes or systems as a whole, it is useful to be
able to break down the various heaps that were allocated by
each layer and examine their size, RSS, and physical memory
usage.

This patch adds a user pointer to the shared union in
vm_area_struct that points to a null terminated string inside
the user process containing a name for the vma.  vmas that
point to the same address will be merged, but vmas that
point to equivalent strings at different addresses will
not be merged.

Userspace can set the name for a region of memory by calling
prctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME, start, len, (unsigned long)name);
Setting the name to NULL clears it.

The names of named anonymous vmas are shown in /proc/pid/maps
as [anon:<name>] and in /proc/pid/smaps in a new "Name" field
that is only present for named vmas.  If the userspace pointer
is no longer valid all or part of the name will be replaced
with "<fault>".

The idea to store a userspace pointer to reduce the complexity
within mm (at the expense of the complexity of reading
/proc/pid/mem) came from Dave Hansen.  This results in no
runtime overhead in the mm subsystem other than comparing
the anon_name pointers when considering vma merging.  The pointer
is stored in a union with fieds that are only used on file-backed
mappings, so it does not increase memory usage.

Includes fix from Jed Davis <jld@mozilla.com> for typo in
prctl_set_vma_anon_name, which could attempt to set the name
across two vmas at the same time due to a typo, which might
corrupt the vma list.  Fix it to use tmp instead of end to limit
the name setting to a single vma at a time.

Change-Id: I9aa7b6b5ef536cd780599ba4e2fba8ceebe8b59f
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoadd extra free kbytes tunable
Rik van Riel [Thu, 1 Sep 2011 19:26:50 +0000 (15:26 -0400)]
add extra free kbytes tunable

Add a userspace visible knob to tell the VM to keep an extra amount
of memory free, by increasing the gap between each zone's min and
low watermarks.

This is useful for realtime applications that call system
calls and have a bound on the number of allocations that happen
in any short time period.  In this application, extra_free_kbytes
would be left at an amount equal to or larger than than the
maximum number of allocations that happen in any burst.

It may also be useful to reduce the memory use of virtual
machines (temporarily?), in a way that does not cause memory
fragmentation like ballooning does.

[ccross]
Revived for use on old kernels where no other solution exists.
The tunable will be removed on kernels that do better at avoiding
direct reclaim.

Change-Id: I765a42be8e964bfd3e2886d1ca85a29d60c3bb3e
Signed-off-by: Rik van Riel<riel@redhat.com>
Signed-off-by: Colin Cross <ccross@android.com>
8 years agomm: vmscan: Add a debug file for shrinkers
Rebecca Schultz Zavin [Fri, 5 Oct 2012 20:54:59 +0000 (13:54 -0700)]
mm: vmscan: Add a debug file for shrinkers

This patch adds a debugfs file called "shrinker" when read this calls
all the shrinkers in the system with nr_to_scan set to zero and prints
the result.  These results are the number of objects the shrinkers have
available and can thus be used an indication of the total memory
that would be availble to the system if a shrink occurred.

Change-Id: Ied0ee7caff3d2fc1cb4bb839aaafee81b5b0b143
Signed-off-by: Rebecca Schultz Zavin <rebecca@android.com>
8 years agoFROMLIST: pstore-ram: add Device Tree bindings
Greg Hackmann [Sat, 24 Oct 2015 00:25:42 +0000 (17:25 -0700)]
FROMLIST: pstore-ram: add Device Tree bindings

ramoops is one of the remaining places where ARM vendors still rely on
board-specific shims.  Device Tree lets us replace those shims with
generic code.

These bindings mirror the ramoops module parameters, with two small
differences:

(1) dump_oops becomes an optional "no-dump-oops" property, since ramoops
    sets dump_oops=1 by default.

(2) mem_type=1 becomes the more self-explanatory "unbuffered" property.

(am from https://lkml.org/lkml/2016/1/7/750)

Change-Id: I2140199a861d50fc2bcbbe85b16bf17fb9ccaa1d
Signed-off-by: Greg Hackmann <ghackmann@google.com>
8 years agopstore/ram: Add ramoops_console_write_buf api
Arve Hjønnevåg [Sat, 3 May 2014 03:23:21 +0000 (20:23 -0700)]
pstore/ram: Add ramoops_console_write_buf api

Allow writing into the ramoops console buffer.

Change-Id: Iff0d69b562e4dae33ea7f8d19412227bebb17e47
Signed-off-by: Arve Hjønnevåg <arve@android.com>
8 years agopstore: Update Documentation/android.txt
John Stultz [Tue, 9 Apr 2013 17:30:26 +0000 (10:30 -0700)]
pstore: Update Documentation/android.txt

Update Documentation/android.txt to reference PSTORE_CONSOLE
and PSTORE_RAM instead of ANDROID_RAM_CONSOLE

Change-Id: I2c56e73f8c65c3ddbe6ddbf1faadfacb42a09575
Reported-by: Jon Medhurst (Tixy) <tixy@linaro.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoinitramfs: Add skip_initramfs command line option
Rom Lemarchand [Mon, 6 Jul 2015 23:50:33 +0000 (16:50 -0700)]
initramfs: Add skip_initramfs command line option

Add a skip_initramfs option to allow choosing whether to boot using
the initramfs or not at runtime.

Change-Id: If30428fa748c1d4d3d7b9d97c1f781de5e4558c3
Signed-off-by: Rom Lemarchand <romlem@google.com>
8 years agoof: Fix build warnings
John Stultz [Thu, 19 Nov 2015 21:45:41 +0000 (13:45 -0800)]
of: Fix build warnings

In commit d6cb004d80 (of: fix CONFIG_CMDLINE_EXTEND),
the types of some variables in early_init_dt_scan_chosen()
were modified, which results in build warnings.

This patch resets the unsigned long to an int, and re-adds
the const to the char*.

Change-Id: Ie60ae92b4552e453cf477dd83f42838b3f95975e
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoof: fix CONFIG_CMDLINE_EXTEND
Colin Cross [Thu, 7 Mar 2013 03:10:29 +0000 (19:10 -0800)]
of: fix CONFIG_CMDLINE_EXTEND

strlcat takes the size of the buffer, not the number of characters
to concatenate.  If the size of the device tree command line p is
larger than the CONFIG_CMDLINE string data, then strcat(data, p, l)
will hit a BUG_ON because strlen(data) > l.

Replace the second strlcat with a strncpy plus a manual null
termination.

Also rearrange the code to reduce indent depth to make it more
readable, and replace data with a char *cmdline to avoid extra
casts.

Signed-off-by: Colin Cross <ccross@android.com>
8 years agoARM64: copy CONFIG_CMDLINE_EXTEND from ARM
Colin Cross [Thu, 3 Apr 2014 01:02:15 +0000 (18:02 -0700)]
ARM64: copy CONFIG_CMDLINE_EXTEND from ARM

Copy the config choice for CONFIG_CMDLINE_EXTEND from
arch/arm/Kconfig, including CONFIG_CMDLINE_FROM_BOOTLOADER
as the default.  These will be used by drivers/of/fdt.c.

Change-Id: I8416038498ddf8fc1e99ab06109825eb1492aa7f
Signed-off-by: Colin Cross <ccross@android.com>
8 years agoof: Support CONFIG_CMDLINE_EXTEND config option
Doug Anderson [Fri, 3 Feb 2012 06:58:28 +0000 (22:58 -0800)]
of: Support CONFIG_CMDLINE_EXTEND config option

The old logic assumes CMDLINE_FROM_BOOTLOADER vs. CMDLINE_FORCE and
ignores CMDLINE_EXTEND.  Here's the old logic:

- CONFIG_CMDLINE_FORCE=true
    CONFIG_CMDLINE
- dt bootargs=non-empty:
    dt bootargs
- dt bootargs=empty, @data is non-empty string
    @data is left unchanged
- dt bootargs=empty, @data is empty string
    CONFIG_CMDLINE (or "" if that's not defined)

The new logic is now documented in of_fdt.h and is copied here for
reference:

- CONFIG_CMDLINE_FORCE=true
    CONFIG_CMDLINE
- CONFIG_CMDLINE_EXTEND=true, @data is non-empty string
    @data + dt bootargs (even if dt bootargs are empty)
- CONFIG_CMDLINE_EXTEND=true, @data is empty string
    CONFIG_CMDLINE + dt bootargs (even if dt bootargs are empty)
- CMDLINE_FROM_BOOTLOADER=true, dt bootargs=non-empty:
    dt bootargs
- CMDLINE_FROM_BOOTLOADER=true, dt bootargs=empty, @data is non-empty string
    @data is left unchanged
- CMDLINE_FROM_BOOTLOADER=true, dt bootargs=empty, @data is empty string
    CONFIG_CMDLINE (or "" if that's not defined)

Signed-off-by: Doug Anderson <dianders@chromium.org>
CC: devicetree-discuss@lists.ozlabs.org
CC: Grant Likely <grant.likely@secretlab.ca>
CC: Benjamin Herrenschmidt <benh@kernel.crashing.org>
CC: Rob Herring <rob.herring@calxeda.com>
Change-Id: I40ace250847f813358125dfcaa8998fd32cf7ea3
Signed-off-by: Colin Cross <ccross@android.com>
8 years agoARM: decompressor: Flush tlb before swiching domain 0 to client mode
Arve Hjønnevåg [Sat, 1 Dec 2012 01:05:40 +0000 (17:05 -0800)]
ARM: decompressor: Flush tlb before swiching domain 0 to client mode

If the bootloader used a page table that is incompatible with domain 0
in client mode, and boots with the mmu on, then swithing domain 0 to
client mode causes a fault if we don't flush the tlb after updating
the page table pointer.

v2: Add ISB before loading dacr.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
8 years agoARM64: add option to build Image.gz/dtb combo
Alex Ray [Mon, 17 Mar 2014 20:44:01 +0000 (13:44 -0700)]
ARM64: add option to build Image.gz/dtb combo

Allows a defconfig to set a list of dtbs to concatenate with an
Image.gz to create a Image.gz-dtb.

Includes 8adb162 arm64: Fix correct dtb clean-files location

Change-Id: I0b462322d5c970f1fdf37baffece7ad058099f4a
Signed-off-by: Alex Ray <aray@google.com>
8 years agoARM: convert build of appended dtb zImage to list of dtbs
Colin Cross [Wed, 17 Apr 2013 23:58:36 +0000 (16:58 -0700)]
ARM: convert build of appended dtb zImage to list of dtbs

Allow CONFIG_BUILD_ARM_APPENDED_DTB_IMAGE_NAMES to specify
a space separated list of dtbs to append to the zImage,
and name the resulting file zImage-dtb

Change-Id: Ied5d0bafbd1d01fc1f109c15c4283de7029903c9
Signed-off-by: Colin Cross <ccross@android.com>
8 years agoARM: add config option to build zImage/dtb combo
Erik Gilling [Mon, 25 Mar 2013 22:04:41 +0000 (15:04 -0700)]
ARM: add config option to build zImage/dtb combo

Allows a defconfig to set a default dtb to concatenate with a zImage
to create a zImage-dtb.<dtb name>

Signed-off-by: Erik Gilling <konkers@android.com>
Change-Id: I34b643b1c49228fbae88a56e46c93c478089620d

8 years agoARM: Fix dtb list when DTB_IMAGE_NAMES is empty
Benoit Goby [Fri, 8 Nov 2013 23:24:19 +0000 (15:24 -0800)]
ARM: Fix dtb list when DTB_IMAGE_NAMES is empty

In the 3.10 kernel, dtb-y is not defined in Makefile.boot anymore
but in dts/Makefile, so it needs to be included too.

Change-Id: I6d6fccf933709bcb6220ce8f12b4b9e2a7c40d63
Signed-off-by: Benoit Goby <benoit@android.com>
8 years agoarm64: pass return address to dma_common_contiguous_remap
Jin Qian [Wed, 30 Sep 2015 01:57:35 +0000 (18:57 -0700)]
arm64: pass return address to dma_common_contiguous_remap

Added return address to show caller function in /proc/vmallocinfo

Change-Id: Ieb0bbf6ec82b561cea6ff18f0516744050dfc269

8 years agoarm64 Kconfig: Move LIST_POISON to a safe value
Jeff Vander Stoep [Tue, 18 Aug 2015 18:15:53 +0000 (11:15 -0700)]
arm64 Kconfig: Move LIST_POISON to a safe value

Move the poison pointer offset to 0xdead000000000000, a
recognized value that is not mappable by user-space exploits.

Change-Id: I558441a26a7c8390aa087f32c4cbe980de8c8ce3
Signed-off-by: Thierry Strudel <tstrudel@google.com>
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
8 years agoarch: arm64: force -fno-pic
Greg Hackmann [Tue, 2 Dec 2014 00:13:30 +0000 (16:13 -0800)]
arch: arm64: force -fno-pic

The aarch64-linux-android- toolchain enables -fpic by default.  -fpic
isn't needed for the kernel and breaks CONFIG_JUMP_LABEL, so turn it
off.

Change-Id: I685da1dc60e4cf1e9abcfb56e03654675ac02a0c
Signed-off-by: Greg Hackmann <ghackmann@google.com>
8 years agoarm64: process: dump memory around registers when displaying regs
Greg Hackmann [Wed, 10 Sep 2014 00:36:05 +0000 (17:36 -0700)]
arm64: process: dump memory around registers when displaying regs

A port of 8608d7c4418c75841c562a90cddd9beae5798a48 to ARM64.  Both the
original code and this port are limited to dumping kernel addresses, so
don't bother if the registers are from a userspace process.

Change-Id: Idc76804c54efaaeb70311cbb500c54db6dac4525
Signed-off-by: Greg Hackmann <ghackmann@google.com>
8 years agoarm64: check for upper PAGE_SHIFT bits in pfn_valid()
Greg Hackmann [Thu, 28 Aug 2014 21:00:10 +0000 (14:00 -0700)]
arm64: check for upper PAGE_SHIFT bits in pfn_valid()

pfn_valid() returns a false positive when the lower (64 - PAGE_SHIFT)
bits match a valid pfn but some of the upper bits are set.  This caused
a kernel panic in kpageflags_read() when a userspace utility parsed
/proc/*/pagemap, neglected to discard the upper flag bits, and tried to
lseek()+read() from the corresponding offset in /proc/kpageflags.

A valid pfn will never have the upper PAGE_SHIFT bits set, so simply
check for this before passing the pfn to memblock_is_memory().

Change-Id: Ief5d8cd4dd93cbecd545a634a8d5885865cb5970
Signed-off-by: Greg Hackmann <ghackmann@google.com>
8 years agoARM: fault: assume no context when IRQs are disabled during data abort.
JP Abgrall [Mon, 29 Apr 2013 23:07:00 +0000 (16:07 -0700)]
ARM: fault: assume no context when IRQs are disabled during data abort.

Bail out early if IRQs are disabled in do_page_fault or else
  [14415.157266] BUG: sleeping function called from invalid context at arch/arm/mm/fault.c:301

Russell King's idea from
  http://comments.gmane.org/gmane.linux.ports.arm.omap/59256

Signed-off-by: JP Abgrall <jpa@google.com>
8 years agoARM: Fix "Make low-level printk work" to use a separate config option
Arve Hjønnevåg [Sat, 15 Jun 2013 02:54:40 +0000 (19:54 -0700)]
ARM: Fix "Make low-level printk work" to use a separate config option

Change-Id: I5ca8db61b595adc642a07ea187bd41fd7636840e
Signed-off-by: Arve Hjønnevåg <arve@android.com>
8 years agopanic: Add board ID to panic output
Nishanth Menon [Wed, 17 Aug 2011 23:31:58 +0000 (18:31 -0500)]
panic: Add board ID to panic output

At times, it is necessary for boards to provide some additional information
as part of panic logs. Provide information on the board hardware as part
of panic logs.

It is safer to print this information at the very end in case something
bad happens as part of the information retrieval itself.

To use this, set global mach_panic_string to an appropriate string in the
board file.

Change-Id: Id12cdda87b0cd2940dd01d52db97e6162f671b4d
Signed-off-by: Nishanth Menon <nm@ti.com>
8 years agoARM: add option to flush console before reboot
Dima Zavin [Tue, 23 Aug 2011 22:56:50 +0000 (15:56 -0700)]
ARM: add option to flush console before reboot

If the console_lock was held while the system was rebooted, the messages
in the temporary logbuffer would not have propogated to all the console
drivers.

This force releases the console lock if it failed to be acquired.

Change-Id: I193dcf7b968be17966833e50b8b8bc70d5d9fe89
Signed-off-by: Dima Zavin <dima@android.com>
8 years agoARM: Make low-level printk work
Tony Lindgren [Mon, 9 May 2005 21:10:26 +0000 (14:10 -0700)]
ARM: Make low-level printk work

Makes low-level printk work.

Signed-off-by: Tony Lindgren <tony@atomide.com>
8 years agoOptionally flush entire dcache from v6_dma_flush_range
Arve Hjønnevåg [Wed, 10 Jun 2009 03:17:45 +0000 (20:17 -0700)]
Optionally flush entire dcache from v6_dma_flush_range

If CACHE_FLUSH_RANGE_LIMIT is defined, then the entire dcache will
be flushed if the requested range is larger than this limit.

Change-Id: I29277d645a9d6716b1952cf3b870c78496261dd0
Signed-off-by: Arve Hjønnevåg <arve@android.com>
8 years agoarm: dcc_tty: fix armv6 dcc tty build failure
John Stultz [Tue, 27 Jan 2015 19:08:06 +0000 (00:38 +0530)]
arm: dcc_tty: fix armv6 dcc tty build failure

Fix spinlock declaration and tty_insert/flip arguments.

If ARM DCC tty driver is enabled then we run into following build failure:
----------
  CC      drivers/char/dcc_tty.o
drivers/char/dcc_tty.c:29:36: error: ‘SPIN_LOCK_UNLOCKED’ undeclared here (not in a function)
drivers/char/dcc_tty.c: In function ‘dcc_poll_locked’:
drivers/char/dcc_tty.c:83:4: warning: passing argument 1 of ‘tty_insert_flip_string’ from incompatible pointer type [enabled by default]
In file included from drivers/char/dcc_tty.c:23:0:
include/linux/tty_flip.h:32:19: note: expected ‘struct tty_port *’ but argument is of type ‘struct tty_struct *’
drivers/char/dcc_tty.c:84:4: warning: passing argument 1 of ‘tty_flip_buffer_push’ from incompatible pointer type [enabled by default]
In file included from drivers/char/dcc_tty.c:23:0:
include/linux/tty_flip.h:13:13: note: expected ‘struct tty_port *’ but argument is of type ‘struct tty_struct *’
make[2]: *** [drivers/char/dcc_tty.o] Error 1
----------

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agoarmv6 dcc tty driver
Arve Hjønnevåg [Sun, 2 Dec 2007 02:34:14 +0000 (18:34 -0800)]
armv6 dcc tty driver

Signed-off-by: Brian Swetland <swetland@google.com>
Signed-off-by: Arve Hjønnevåg <arve@android.com>
8 years agoprocess: Add display of memory around registers when displaying regs.
San Mehat [Tue, 25 Aug 2009 23:52:22 +0000 (16:52 -0700)]
process: Add display of memory around registers when displaying regs.

This is extremely useful in diagnosing remote crashes, and is based heavily
on original work by <md@google.com>.

Signed-off-by: San Mehat <san@google.com>
Cc: Michael Davidson <md@google.com>
[ARM] process: Use uber-safe probe_kernel_address() to read mem when dumping.

This prevents the dump from taking pagefaults / external aborts.

Change-Id: I8df76e8638780f94fb1bd7ea4471e3f7b01df950
Signed-off-by: San Mehat <san@google.com>
8 years agoion: fix page pool cache policy
Amit Pundir [Mon, 18 Jan 2016 19:47:30 +0000 (01:17 +0530)]
ion: fix page pool cache policy

Fix redundant "buffer->private_flags & ION_PRIV_FLAG_SHRINKER_FREE"
checks in if(!cached ...) condition block.

AOSP Change-Id: I98ee8902df0c80135dddfa998c4ca4c2bb44e40e,
"ion: Handle the memory mapping correctly on x86", is broken
on android-3.18+ kernels. It conflicts with upstream commit
53a91c68fa7b, "staging: ion: Add private buffer flag to skip
page pooling on free", and break the ION_PRIV_FLAG_SHRINKER_FREE
private flag check logic.

Change-Id: I9cee4bcc3545cf92e07c21c2b42d27cf88da3316
Reported-by: chenfeng <puck.chen@hisilicon.com>
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agostaging: ion: Add X86 dependency for ION_POOL_CACHE_POLICY
Daniel Rosenberg [Thu, 1 Oct 2015 03:44:16 +0000 (20:44 -0700)]
staging: ion: Add X86 dependency for ION_POOL_CACHE_POLICY

ION_POOL_CACHE_POLICY uses x86 specific commands.
Only allow it to be used for x86.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Change-Id: Ic66e175fa41486b0b5f77872f2193add77cbe66f
(cherry picked from commit 43f4056cbec017c4414d74b032e9c14e009eabb0)

8 years agoion: Handle the memory mapping correctly on x86
Vinil Cheeramvelil [Wed, 8 Jul 2015 02:35:06 +0000 (10:35 +0800)]
ion: Handle the memory mapping correctly on x86

This patch modifies the ion page pool code to address
limitation in x86 PAT. When one physical page is mapped
to multiple virtual pages, the same cache policy
should be used. Add set_memory_wc/uc call to avoid aliases.
If not, all mappings will be cached(write back).

Change-Id: I98ee8902df0c80135dddfa998c4ca4c2bb44e40e
Signed-off-by: Zhebin Jin <zhebin.jin@intel.com>
Signed-off-by: Vinil Cheeramvelil <vinil.cheeramvelil@intel.com>
8 years agoion : In carveout heap, change minimum allocation order from 12 to PAGE_SHIFT, After...
Rajmal Menariya [Wed, 26 Aug 2015 13:02:52 +0000 (21:02 +0800)]
ion : In carveout heap, change minimum allocation order from 12 to PAGE_SHIFT, After this change each bit in bitmap (genalloc - General purpose special memory pool)

represents one page size memory.

Cc:sprd-ind-kernel-group@googlegroups.com
Cc:sanjeev.yadav@spreadtrum.com
Cc:Colin Cross <ccross@android.com>
cc:John Stultz <john.stultz@linaro.org>

Change-Id: I1172285dac014ecd2f1a965604b697d6739c1726
Signed-off-by: Rajmal Menariya <rajmal.menariya@spreadtrum.com>
8 years agoSELinux: build fix for 4.1
Amit Pundir [Thu, 29 Oct 2015 19:17:53 +0000 (00:47 +0530)]
SELinux: build fix for 4.1

Commit "SELinux: Enable setting security contexts on rootfs inodes."
cherry-picked in experimental/android-4.1 used a now obsolete flag
SE_SBLABELSUPP. Rename it to SBLABEL_MNT as intended by upstream
commit 12f348b9dcf6 "SELinux: rename SE_SBLABELSUPP to SBLABEL_MNT",
otherwise we run into following build error:

  CC      security/selinux/hooks.o
security/selinux/hooks.c: In function ‘sb_finish_set_opts’:
security/selinux/hooks.c:459:19: error: ‘SE_SBLABELSUPP’ undeclared (first use in this function)
   sbsec->flags |= SE_SBLABELSUPP;
                   ^
security/selinux/hooks.c:459:19: note: each undeclared identifier is reported only once for each function it appears in
make[2]: *** [security/selinux/hooks.o] Error 1

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agoselinux: Android kernel compatibility with M userspace
Jeff Vander Stoep [Thu, 22 Oct 2015 16:30:40 +0000 (09:30 -0700)]
selinux: Android kernel compatibility with M userspace

NOT intended for new Android devices - this commit is unnecessary
for a target device that does not have a previous M variant.

DO NOT upstream. Android only.

Motivation:

This commit mitigates a mismatch between selinux kernel and
selinux userspace. The selinux ioctl white-listing binary policy
format that was accepted into Android M differs slightly from what
was later accepted into the upstream kernel. This leaves Android
master branch kernels incompatible with Android M releases. This
patch restores backwards compatibility. This is important because:

1. kernels may be updated on a different cycle than the rest of the
   OS e.g. security patching.
2. Android M bringup may still be ongoing for some devices. The
   same kernel should work for both M and master.

Backwards compatibility is achieved by checking for an Android M
policy characteristic during initial policy read and converting to
upstream policy format. The inverse conversion is done for policy
write as required for CTS testing.

Bug: 22846070
Change-Id: I2f1ee2eee402f37cf3c9df9f9e03c1b9ddec1929
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
8 years agoSELinux: Enable setting security contexts on rootfs inodes.
Stephen Smalley [Tue, 23 Jul 2013 21:38:41 +0000 (17:38 -0400)]
SELinux: Enable setting security contexts on rootfs inodes.

rootfs (ramfs) can support setting of security contexts
by userspace due to the vfs fallback behavior of calling
the security module to set the in-core inode state
for security.* attributes when the filesystem does not
provide an xattr handler.  No xattr handler required
as the inodes are pinned in memory and have no backing
store.

This is useful in allowing early userspace to label individual
files within a rootfs while still providing a policy-defined
default via genfs.

Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: Eric Paris <eparis@redhat.com>
8 years agosecurity: Add proper checks for Android specific capability checks
Tushar Behera [Mon, 26 Mar 2012 11:24:15 +0000 (16:54 +0530)]
security: Add proper checks for Android specific capability checks

Commit b641072 ("security: Add AID_NET_RAW and AID_NET_ADMIN capability
check in cap_capable().") introduces additional checks for AID_NET_xxx
macros. Since the header file including those macros are conditionally
included, the checks should also be conditionally executed.

Change-Id: Iaec5208d5b95a46b1ac3f2db8449c661e803fa5b
Signed-off-by: Tushar Behera <tushar.behera@linaro.org>
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org>
8 years agouid_cputime: skip power reporting per uid for now
Amit Pundir [Mon, 14 Dec 2015 06:26:35 +0000 (11:56 +0530)]
uid_cputime: skip power reporting per uid for now

AOSP's cpufreq_stats patch to report power/current
per cpufreq is broken at the moment so skip power
reporting for now.

Change-Id: I07779511f51ff3a9303dc98a3b71ac0a5882a4e9
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agouid_cputime: Check for the range while removing range of UIDs.
Ruchi Kandoi [Sat, 24 Oct 2015 00:49:11 +0000 (17:49 -0700)]
uid_cputime: Check for the range while removing range of UIDs.

Checking if the uid_entry->uid matches the uid intended to be removed will
prevent deleting unwanted uid_entry.
Type cast the key for the hashtable to the same size, as when they were
inserted. This will make sure that we can find the uid_entry we want.

Bug: 25195548
Change-Id: I567942123cfb20e4b61ad624da19ec4cc84642c1
Signed-off: Ruchi kandoi <kandoiruchi@google.com>

8 years agouid_cputime: Iterates over all the threads instead of processes.
Ruchi Kandoi [Fri, 31 Jul 2015 17:17:54 +0000 (10:17 -0700)]
uid_cputime: Iterates over all the threads instead of processes.

Bug: 22833116
Change-Id: I775a18f61bd2f4df2bec23d01bd49421d0969f87
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agouid_cputime: fix cputime overflow
Jin Qian [Tue, 14 Jul 2015 01:16:55 +0000 (18:16 -0700)]
uid_cputime: fix cputime overflow

Converting cputime_t to usec caused overflow when the value is greater
than 1 hour. Use msec and convert to unsigned long long to support bigger
range.

Bug: 22461683

Change-Id: I853fe3e8e7dbf0d3e2cc5c6f9688a5a6e1f1fb3e
Signed-off-by: Jin Qian <jinqian@google.com>
8 years agouid_cputime: Avoids double accounting of process stime, utime and cpu_power in task...
Ruchi Kandoi [Fri, 26 Jun 2015 21:19:21 +0000 (14:19 -0700)]
uid_cputime: Avoids double accounting of process stime, utime and cpu_power in task exit.

This avoids the race where a particular process is terminating and we
read the show_uid_stats. At this time since the task_struct still exists
and we will account for the terminating process as one of the active
task, where as the stats would have been added in the task exit
callback.

Bug: 22064385
Change-Id: Id2ae04b33fcd230eda9683a41b6019d4dd8f5d85
Signed-off-by: Jin Qian <jinqian@google.com>
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agouid_cputime: Extends the cputime functionality to report power per uid
Ruchi Kandoi [Fri, 17 Apr 2015 23:52:54 +0000 (16:52 -0700)]
uid_cputime: Extends the cputime functionality to report power per uid

/proc/uid_cputime/show_uid_stats shows a third field power for each of
the uids. It represents the power in the units (uAusec)

Change-Id: I52fdc5e59647e9dc97561a26d56f462a2689ba9c
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agoproc: uid_cputime: fix show_uid_stat permission
Jin Qian [Tue, 12 May 2015 00:57:52 +0000 (17:57 -0700)]
proc: uid_cputime: fix show_uid_stat permission

Change-Id: Ice9084e39da599261df0be6dc305b817b50cfbbf
Signed-off-by: Jin Qian <jinqian@google.com>
8 years agoproc: uid_cputime: create uids from kuids
Amit Pundir [Tue, 14 Apr 2015 19:10:21 +0000 (00:40 +0530)]
proc: uid_cputime: create uids from kuids

Create uids from kuids using from_kuid_munged(),
otherwise we run into following build error and warnings:
--------------------
  CC      drivers/misc/uid_cputime.o
drivers/misc/uid_cputime.c: In function ‘uid_stat_show’:
drivers/misc/uid_cputime.c:90:36: error: incompatible type for argument 1 of ‘find_or_register_uid’
drivers/misc/uid_cputime.c:54:26: note: expected ‘uid_t’ but argument is of type ‘kuid_t’
drivers/misc/uid_cputime.c:94:4: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘kuid_t’ [-Wformat]
drivers/misc/uid_cputime.c: In function ‘process_notifier’:
drivers/misc/uid_cputime.c:194:6: error: incompatible types when assigning to type ‘uid_t’ from type ‘kuid_t’
make[2]: *** [drivers/misc/uid_cputime.o] Error 1
--------------------

Change-Id: Ifecb98001f7fe2fac74d1ef3e1abd03d43fc9059
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
(cherry picked from commit b0f4decae627cf2d74e6f72c7ecb939c77d48625)

8 years agoproc: uid: Adds accounting for the cputimes per uid.
jinqian [Wed, 11 Mar 2015 17:44:50 +0000 (10:44 -0700)]
proc: uid: Adds accounting for the cputimes per uid.

Adds proc files /proc/uid_cputime/show_uid_stat and
/proc/uid_cputime/remove_uid_range.

show_uid_stat lists the total utime and stime for the active as well as
terminated processes for each of the uids.

Writing a range of uids to remove_uid_range will delete the accounting
for all the uids within that range.

Change-Id: I21d9210379da730b33ddc1a0ea663c8c9d2ac15b

8 years agofixup! proc: make oom adjustment files user read-only
Dan Willemsen [Wed, 18 Mar 2015 18:22:44 +0000 (11:22 -0700)]
fixup! proc: make oom adjustment files user read-only

Fix the build by removing the duplicate line that uses the obsolete INF
macro.

Signed-off-by: Dan Willemsen <dwillemsen@nvidia.com>
8 years agoproc: make oom adjustment files user read-only
Rom Lemarchand [Sat, 7 Mar 2015 17:38:05 +0000 (09:38 -0800)]
proc: make oom adjustment files user read-only

Make oom_adj and oom_score_adj user read-only.

Bug: 19636629
Change-Id: I055bb172d5b4d3d856e25918f3c5de8edf31e4a3
Signed-off-by: Rom Lemarchand <romlem@google.com>
8 years agoproc: smaps: Allow smaps access for CAP_SYS_RESOURCE
San Mehat [Wed, 16 Sep 2009 19:39:10 +0000 (12:39 -0700)]
proc: smaps: Allow smaps access for CAP_SYS_RESOURCE

Signed-off-by: San Mehat <san@google.com>
8 years agoprctl: reset PR_SET_TIMERSLACK_PID value to avoid conflict
Amit Pundir [Wed, 30 Sep 2015 10:31:08 +0000 (16:01 +0530)]
prctl: reset PR_SET_TIMERSLACK_PID value to avoid conflict

PR_SET_TIMERSLACK_PID value keep colliding with that of
newer prctls in mainline (e.g. first with PR_SET_THP_DISABLE,
and again with PR_MPX_ENABLE_MANAGEMENT).

So reset PR_SET_TIMERSLACK_PID to a large number so as to
avoid conflict in the near term while it is out of mainline
tree.

Corresponding Change-Id up for review in platform/system/core
is Icd8c658c8eb62136dc26c2c4c94f7782e9827cdb

Change-Id: I061b25473acc020c13ee22ecfb32336bc358e76a
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agoprctl: make PR_SET_TIMERSLACK_PID pid namespace aware
Micha Kalfon [Tue, 13 Jan 2015 10:12:57 +0000 (12:12 +0200)]
prctl: make PR_SET_TIMERSLACK_PID pid namespace aware

Make PR_SET_TIMERSLACK_PID consider pid namespace and resolve the
target pid in the caller's namespace. Otherwise, calls from pid
namespace other than init would fail or affect the wrong task.

Change-Id: I1da15196abc4096536713ce03714e99d2e63820a
Signed-off-by: Micha Kalfon <micha@cellrox.com>
Acked-by: Oren Laadan <orenl@cellrox.com>
8 years agoprctl: fix misplaced PR_SET_TIMERSLACK_PID case
Micha Kalfon [Tue, 13 Jan 2015 09:47:20 +0000 (11:47 +0200)]
prctl: fix misplaced PR_SET_TIMERSLACK_PID case

The case clause for the PR_SET_TIMERSLACK_PID option was placed inside
the an internal switch statement for PR_MCE_KILL (see commits 37a591d4
and 8ae872f1) . This commit moves it to the right place.

Change-Id: I63251669d7e2f2aa843d1b0900e7df61518c3dea
Signed-off-by: Micha Kalfon <micha@cellrox.com>
Acked-by: Oren Laadan <orenl@cellrox.com>
8 years agoprctl: adds the capable(CAP_SYS_NICE) check to PR_SET_TIMERSLACK_PID.
Ruchi Kandoi [Sat, 14 Jun 2014 00:03:01 +0000 (17:03 -0700)]
prctl: adds the capable(CAP_SYS_NICE) check to PR_SET_TIMERSLACK_PID.

Adds a capable() check to make sure that arbitary apps do not change the
timer slack for other apps.

Bug: 15000427
Change-Id: I558a2551a0e3579c7f7e7aae54b28aa9d982b209
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agoprctl: adds PR_SET_TIMERSLACK_PID for setting timer slack of an arbitrary thread.
Ruchi Kandoi [Fri, 18 Apr 2014 21:07:28 +0000 (14:07 -0700)]
prctl: adds PR_SET_TIMERSLACK_PID for setting timer slack of an arbitrary thread.

Second argument is similar to PR_SET_TIMERSLACK, if non-zero then the
slack is set to that value otherwise sets it to the default for the thread.

Takes PID of the thread as the third argument.

This allows power/performance management software to set timer slack for
other threads according to its policy for the thread (such as when the
thread is designated foreground vs. background activity)

Change-Id: I744d451ff4e60dae69f38f53948ff36c51c14a3f
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agoUPSTREAM: memcg: Only free spare array when readers are done
Martijn Coenen [Thu, 14 Jan 2016 13:33:52 +0000 (14:33 +0100)]
UPSTREAM: memcg: Only free spare array when readers are done

A spare array holding mem cgroup threshold events is kept around
to make sure we can always safely deregister an event and have an
array to store the new set of events in.

In the scenario where we're going from 1 to 0 registered events, the
pointer to the primary array containing 1 event is copied to the spare
slot, and then the spare slot is freed because no events are left.
However, it is freed before calling synchronize_rcu(), which means
readers may still be accessing threshold->primary after it is freed.

Fixed by only freeing after synchronize_rcu().

Signed-off-by: Martijn Coenen <maco@google.com>
8 years agocgroup: refactor allow_attach handler for 4.4
Amit Pundir [Mon, 21 Dec 2015 10:31:10 +0000 (16:01 +0530)]
cgroup: refactor allow_attach handler for 4.4

Refactor *allow_attach() handler to align it with the changes
from mainline commit 1f7dd3e5a6e4 "cgroup: fix handling of
multi-destination migration from subtree_control enabling".

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agocgroup: fix cgroup_taskset_for_each call in allow_attach() for 4.1
Dmitry Shmidt [Wed, 28 Oct 2015 01:18:04 +0000 (18:18 -0700)]
cgroup: fix cgroup_taskset_for_each call in allow_attach() for 4.1

Change-Id: I05013f6e76c30b0ece3671f9f2b4bbdc626cd35c
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoinclude: linux: cgroup: Fix compiler warning
Daniel Rosenberg [Mon, 12 Jan 2015 22:21:52 +0000 (14:21 -0800)]
include: linux: cgroup: Fix compiler warning

Signed-off-by: Daniel Rosenberg <drosen@google.com>
8 years agoFix generic cgroup subsystem permission checks
Christian Poetzsch [Wed, 3 Jun 2015 11:07:33 +0000 (12:07 +0100)]
Fix generic cgroup subsystem permission checks

In 53b5e2f generic cgroup subsystem permission checks have been added.
When this is been done within procs_write an empty taskset is added to
the tasks css set. When a task later on migrates to a new group we see a
dmesg warning cause the mg_node isn't empty (cgroup.c:2086). Cause this
happens all the time this spams dmesg.

I am not really familiar with this code, but it looks to me like adding
the taskset is just a temporary action in this context. Therefore this
taskset should be removed after the actual check. This is what this fix
does.

This problem was seen and the fix tested on x86 using l-mr1 and master.

Change-Id: I9894d39e8b5692ef65149002b07e65a84a33ffea
Signed-off-by: Christian Poetzsch <christian.potzsch@imgtec.com>
8 years agocgroup: Fix issues in allow_attach callback
Rom Lemarchand [Wed, 27 May 2015 00:00:44 +0000 (17:00 -0700)]
cgroup: Fix issues in allow_attach callback

- Return -EINVAL when cgroups support isn't enabled
- Add allow_attach callback in CPU cgroups

Change-Id: Id3360b4a39919524fc4b6fcbd44fa2050009f000
Signed-off-by: Rom Lemarchand <romlem@android.com>
8 years agocgroup: memcg: pass correct argument to subsys_cgroup_allow_attach
Amit Pundir [Wed, 1 Apr 2015 08:28:39 +0000 (13:58 +0530)]
cgroup: memcg: pass correct argument to subsys_cgroup_allow_attach

Pass correct argument to subsys_cgroup_allow_attach(), which
expects 'struct cgroup_subsys_state *' argument but we pass
'struct cgroup *' instead which doesn't seem right.

This fixes following 'incompatible pointer type' compiler warning:
----------
  CC      mm/memcontrol.o
mm/memcontrol.c: In function ‘mem_cgroup_allow_attach’:
mm/memcontrol.c:5052:2: warning: passing argument 1 of ‘subsys_cgroup_allow_attach’ from incompatible pointer type [enabled by default]
In file included from include/linux/memcontrol.h:22:0,
                 from mm/memcontrol.c:29:
include/linux/cgroup.h:953:5: note: expected ‘struct cgroup_subsys_state *’ but argument is of type ‘struct cgroup *’
----------

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agocgroup: Add generic cgroup subsystem permission checks
Colin Cross [Wed, 13 Jul 2011 02:53:24 +0000 (19:53 -0700)]
cgroup: Add generic cgroup subsystem permission checks

Rather than using explicit euid == 0 checks when trying to move
tasks into a cgroup via CFS, move permission checks into each
specific cgroup subsystem. If a subsystem does not specify a
'allow_attach' handler, then we fall back to doing our checks
the old way.

Use the 'allow_attach' handler for the 'cpu' cgroup to allow
non-root processes to add arbitrary processes to a 'cpu' cgroup
if it has the CAP_SYS_NICE capability set.

This version of the patch adds a 'allow_attach' handler instead
of reusing the 'can_attach' handler.  If the 'can_attach' handler
is reused, a new cgroup that implements 'can_attach' but not
the permission checks could end up with no permission checks
at all.

Change-Id: Icfa950aa9321d1ceba362061d32dc7dfa2c64f0c
Original-Author: San Mehat <san@google.com>
Signed-off-by: Colin Cross <ccross@android.com>
8 years agomemcg: add permission check
Rom Lemarchand [Fri, 7 Nov 2014 17:42:40 +0000 (09:42 -0800)]
memcg: add permission check

Use the 'allow_attach' handler for the 'mem' cgroup to allow
non-root processes to add arbitrary processes to a 'mem' cgroup
if it has the CAP_SYS_NICE capability set.

Bug: 18260435
Change-Id: If7d37bf90c1544024c4db53351adba6a64966250
Signed-off-by: Rom Lemarchand <romlem@android.com>
8 years agocgroup: refactor allow_attach function into common code
Rom Lemarchand [Fri, 7 Nov 2014 20:48:17 +0000 (12:48 -0800)]
cgroup: refactor allow_attach function into common code

move cpu_cgroup_allow_attach to a common subsys_cgroup_allow_attach.
This allows any process with CAP_SYS_NICE to move tasks across cgroups if
they use this function as their allow_attach handler.

Bug: 18260435
Change-Id: I6bb4933d07e889d0dc39e33b4e71320c34a2c90f
Signed-off-by: Rom Lemarchand <romlem@android.com>
8 years agowakeup_reason: use vsnprintf instead of snsprintf for vargs.
Ruchi Kandoi [Wed, 5 Aug 2015 23:54:53 +0000 (16:54 -0700)]
wakeup_reason: use vsnprintf instead of snsprintf for vargs.

Bug: 22368519
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agopower: wakeup_reason: fix suspend time reporting
Amit Pundir [Mon, 13 Apr 2015 21:08:20 +0000 (02:38 +0530)]
power: wakeup_reason: fix suspend time reporting

Suspend time reporting Change-Id: I2cb9a9408a5fd12166aaec11b935a0fd6a408c63
(Power: Report suspend times from last_suspend_time), is broken on 3.16+
kernels because get_xtime_and_monotonic_and_sleep_offset() hrtimer helper
routine is removed from kernel timekeeping.

The replacement helper routines ktime_get_update_offsets_{tick,now}()
are private to core kernel timekeeping so we can't use them, hence using
ktime_get() and ktime_get_boottime() instead and sampling the time twice.

Idea is to use Monotonic boottime offset to calculate total time spent
in last suspend state and CLOCK_MONOTONIC to calculate time spent in
last suspend-resume process.

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agowakeup: Add last wake up source logging for suspend abort reason.
Ruchi Kandoi [Wed, 8 Apr 2015 22:42:29 +0000 (15:42 -0700)]
wakeup: Add last wake up source logging for suspend abort reason.

There is a possibility that a wakeup source event is received after
the device prepares to suspend which might cause the suspend to abort.

This patch adds the functionality of reporting the last active wakeup
source which is currently not active but caused the suspend to abort reason
via the /sys/kernel/power/last_wakeup_reason file.

Change-Id: I1760d462f497b33e425f5565cb6cff5973932ec3
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
8 years agoPower: Report suspend times from last_suspend_time
jinqian [Wed, 25 Mar 2015 23:18:44 +0000 (16:18 -0700)]
Power: Report suspend times from last_suspend_time

This node epxorts two values separated by space.
From left to right:
1. time spent in suspend/resume process
2. time spent sleep in suspend state

Change-Id: I2cb9a9408a5fd12166aaec11b935a0fd6a408c63

8 years agopower: Remove HAS_WAKELOCK config and document WAKELOCK
Dylan Reid [Tue, 3 Mar 2015 17:38:28 +0000 (09:38 -0800)]
power: Remove HAS_WAKELOCK config and document WAKELOCK

Remove the HAS_WAKELOCK config as it doesn't seem to have been used in
the 3.10 or 3.14 kernels.

Add some Documentation to CONFIG_WAKELOCK so that it is selectable and
can be disabled is desired.

Signed-off-by: Dylan Reid <dgreid@chromium.org>
8 years agoMake suspend abort reason logging depend on CONFIG_PM_SLEEP
Lorenzo Colitti [Thu, 27 Nov 2014 06:12:10 +0000 (15:12 +0900)]
Make suspend abort reason logging depend on CONFIG_PM_SLEEP

This unbreaks the build on architectures such as um that do not
support CONFIG_PM_SLEEP.

Change-Id: Ia846ed0a7fca1d762ececad20748d23610e8544f
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
8 years agopower: Add check_wakeup_reason() to verify wakeup source irq
Dmitry Shmidt [Fri, 31 Oct 2014 23:05:46 +0000 (16:05 -0700)]
power: Add check_wakeup_reason() to verify wakeup source irq

Wakeup reason is set before driver resume handlers are called.
It is cleared before driver suspend handlers are called, on
PM_SUSPEND_PREPARE.

Change-Id: I04218c9b0c115a7877e8029c73e6679ff82e0aa4
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agopower: Adds functionality to log the last suspend abort reason.
Ruchi Kandoi [Wed, 29 Oct 2014 17:36:27 +0000 (10:36 -0700)]
power: Adds functionality to log the last suspend abort reason.

Extends the last_resume_reason to log suspend abort reason. The abort
reasons will have "Abort:" appended at the start to distinguish itself
from the resume reason.

Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
Change-Id: I3207f1844e3d87c706dfc298fb10e1c648814c5f

8 years agopower: Avoids bogus error messages for the suspend aborts.
Ruchi Kandoi [Wed, 15 Oct 2014 00:43:21 +0000 (17:43 -0700)]
power: Avoids bogus error messages for the suspend aborts.

Avoids printing bogus error message "tasks refusing to freeze", in cases
where pending wakeup source caused the suspend abort.

Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
Change-Id: I913ad290f501b31cd536d039834c8d24c6f16928