nobody else will see that struct file after return from ->release()
anyway; just leave ->f_op as is and let __fput() do that fops_put().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
If the underlying dentry doesn't have ->d_revalidate(), there's no need to
force dropping out of RCU mode. All we need for that is to make freeing
ecryptfs_dentry_info RCU-delayed.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
software_resume is being called after deferred_probe_initcall in
drivers base. If the probing of the device that contains the resume
image is deferred, and the system has been instructed to wait for
it to show up, this wait will occur in software_resume. This causes
a deadlock.
Move software_resume into late_initcall_sync so that it happens
after all the other late_initcalls.
Signed-off-by: Russ Dill <Russ.Dill@ti.com>
Acked-by: Pavel Machek <Pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
hdmi_setup_audio_infoframe() does not set up pin and infoframe if there
is no connected sink. If a sink is connected while audio playback is
already in progress, the pin and infoframe will not be properly set up,
causing no audio or wrongly mapped audio.
On Intel Haswell codecs the hdmi_setup_audio_infoframe() is already
called again from hdmi_present_sense() when an ELD appears because
transcoder:port mapping may have changed.
Make the call non-Haswell-specific so that audio will be properly set up
if the playback was started before a sink was connected.
Tested on non-Haswell Intel HDMI codec by plugging sink in during
playback.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In a recent commit:
commit f455578dd9
Author: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Date: Mon Aug 12 14:14:53 2013 -0300
mtd: nand: pxa3xx: Remove hardcoded mtd name
There's no advantage in using a hardcoded name for the mtd device.
Instead use the provided by the platform_device.
The MTD name was changed to use the one provided by the platform_device.
However, this can be problematic as some users want to set partitions
using the kernel parameter 'mtdparts', where the name is needed.
Therefore, to avoid regressions in users relying in 'mtdparts' we revert
the change and use the previous one 'pxa3xx_nand-0'.
While at it, let's put a big comment and prevent this change from happening
ever again.
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Recent AMD HDMI codecs (revision ID 3 and later, 0x100300 as reported by
procfs codec#0) have a configurable ramp-up/down functionality.
The documentation ( http://www.x.org/docs/AMD/AMD_HDA_verbs_v2.pdf )
specifies that 180 ("180/256 =~ 0.7") is recommended for PCM and 0 for
non-PCM.
Apply the recommended values according to provided S/PDIF AES0 settings
since ramp-up/down does not make sense for non-PCM.
v2: adapted to hdmi_ops infrastructure
* More note from Anssi:
actually, re-reading mails reveals that Olivier didn't find the
expected difference with this setting, except for "maybe slightly
slower startup with AES0=6" (i.e. value 0, which is unexpected).
So maybe
a) it makes too unnoticiable a difference, or
b) only affects certain hardware (card and/or sink), or
c) ramp-up/down is only triggered with the MUTE bit of
ATI_VERB_SET_MULTICHANNEL_xx which is also rev3+ specific,
but is not presently used by the driver,
or something else.
So there's a significant chance setting ramp rate is useless for us ATM,
but probably does not do actual harm either.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Tested-by: Olivier Langlois <olivier@trillion01.com> # v1
Signed-off-by: Takashi Iwai <tiwai@suse.de>
ATI/AMD HDMI codecs do not include standard HDA HDMI HBR support (which
is required for bitstreaming DTS-HD and Dolby TrueHD), instead they have
custom verbs for checking and enabling it.
Add support for the ATI/AMD HDMI HBR verbs.
The specification is available at:
http://www.x.org/docs/AMD/AMD_HDA_verbs_v2.pdf
v2: adapted to hdmi_ops infrastructure
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Tested-by: Peter Frühberger <fritsch@xbmc.org> # v1
Signed-off-by: Takashi Iwai <tiwai@suse.de>
ATI/AMD HDMI/DP codecs do not include standard HDA ELD (EDID-like data)
support.
In place of providing access to an ELD buffer, various vendor-specific
verbs are provided to provide the relevant information. Revision ID 3
and later (0x100300 as reported by procfs codec#X) have support for
providing more information than the previous revisions (but only if
supported by the display driver).
Generate ELD from the information provided by the vendor-specific verbs
on ATI/AMD codecs.
The specification is available at:
http://www.x.org/docs/AMD/AMD_HDA_verbs_v2.pdf
v2: moved code to hda_eld.c and cleaned it up
v3: adapted to hdmi_ops infrastructure
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Tested-by: Peter Frühberger <fritsch@xbmc.org> # v2
Tested-by: Olivier Langlois <olivier@trillion01.com> # v2
Signed-off-by: Takashi Iwai <tiwai@suse.de>
ATI/AMD codecs do not support all the standard HDA HDMI/DP functions,
instead various vendor-specific verbs are provided.
This commit addresses these missing functions:
- standard channel mapping support
- standard infoframe configuration support
ATI/AMD provides their own verbs that allow the following:
- setting CA for infoframe
- setting down-mix information for infoframe
- channel pair remapping
- individual channel remapping (revision ID 3+, 0x100300+)
The documentation for the verbs has now been released by AMD:
http://www.x.org/docs/AMD/AMD_HDA_verbs_v2.pdf
Add support for the ATI/AMD specific verbs and use them instead of the
generic methods on ATI/AMD codecs. This allows multi-channel PCM audio
to work.
Channel remapping is restricted to pairwise mapping on codecs with
revision ID 2 (0x100200 as reported by procfs codec#X) or lower. This
means cards up to Radeon HD7670 as far as I know. This will not affect
standard multi-channel modes since these codecs support automatic
FC-LFE swapping for HDMI.
ATI/AMD codecs do not advertise all of their supported rates, formats
and channel counts, therefore that information is forced accordingly so
that all HDMI 1.x PCM parameters are marked as supported.
Support for multiple ports is also added to patch_atihdmi so that
0x1002aa01 codecs with multiple ports will work properly when switched
back to that patch.
v2: splitted ELD emulation to a separate patch, tlv fixes
v3: adapted to the new hdmi_ops infrastructure, fixed rev3+ vendor id
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Tested-by: Peter Frühberger <fritsch@xbmc.org> # v2
Tested-by: Olivier Langlois <olivier@trillion01.com> # v2+rev3fix
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Upcoming AMD multichannel support requires many customized operations
(channel mapping, ELD, HBR) but can otherwise share most of its code
with the generic patch.
Add a local struct hdmi_ops containing customizable HDMI-specific
callbacks and move the current code to those callbacks. Functionality is
unaltered.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Shifting page->index on 32 bit systems was overflowing, causing
data corruption of > 4GB files. Fix this by casting it first.
https://launchpad.net/bugs/1243636
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reported-by: Lars Duesing <lars.duesing@camelotsweb.de>
Cc: stable@vger.kernel.org # v3.11+
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Move all kmemleak calls into hook functions, and make it so
that all hooks (both inside and outside of #ifdef CONFIG_SLUB_DEBUG)
call the appropriate kmemleak routines. This allows for kmemleak
to be configured independently of slub debug features.
It also fixes a bug where kmemleak was only partially enabled in some
configurations.
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Roman Bobniev <Roman.Bobniev@sonymobile.com>
Signed-off-by: Tim Bird <tim.bird@sonymobile.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now, bufctl is not proper name to this array.
So change it.
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now, virt_to_page(page->s_mem) is same as the page,
because slab use this structure for management.
So remove useless statement.
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now, there are a few field in struct slab, so we can overload these
over struct page. This will save some memory and reduce cache footprint.
After this change, slabp_cache and slab_size no longer related to
a struct slab, so rename them as freelist_cache and freelist_size.
These changes are just mechanical ones and there is no functional change.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now, free in struct slab is same meaning as inuse.
So, remove both and replace them with active.
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
It's useless now, so remove it.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now, we changed the management method of free objects of the slab and
there is no need to use special value, BUFCTL_END, BUFCTL_FREE and
BUFCTL_ACTIVE. So remove them.
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Current free objects management method of the slab is weird, because
it touch random position of the array of kmem_bufctl_t when we try to
get free object. See following example.
struct slab's free = 6
kmem_bufctl_t array: 1 END 5 7 0 4 3 2
To get free objects, we access this array with following pattern.
6 -> 3 -> 7 -> 2 -> 5 -> 4 -> 0 -> 1 -> END
If we have many objects, this array would be larger and be not in the same
cache line. It is not good for performance.
We can do same thing through more easy way, like as the stack.
Only thing we have to do is to maintain stack top to free object. I use
free field of struct slab for this purpose. After that, if we need to get
an object, we can get it at stack top and manipulate top pointer.
That's all. This method already used in array_cache management.
Following is an access pattern when we use this method.
struct slab's free = 0
kmem_bufctl_t array: 6 3 7 2 5 4 0 1
To get free objects, we access this array with following pattern.
0 -> 1 -> 2 -> 3 -> 4 -> 5 -> 6 -> 7
This may help cache line footprint if slab has many objects, and,
in addition, this makes code much much simpler.
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
If we use 'struct page' of first page as 'struct slab', there is no
advantage not to use __GFP_COMP. So use __GFP_COMP flag for all the cases.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
This is trivial change, just use well-defined macro.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
With build-time size checking, we can overload the RCU head over the LRU
of struct page to free pages of a slab in rcu context. This really help to
implement to overload the struct slab over the struct page and this
eventually reduce memory usage and cache footprint of the SLAB.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
We can get cachep using page in struct slab_rcu, so remove it.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
We can get nodeid using address translation, so this field is not useful.
Therefore, remove it.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
Now there is no user colouroff, so remove it.
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Pekka Enberg <penberg@iki.fi>
