Prepare the panel before it's enabled and un-prepare after disable, this
will make sure that the regulators are switched on and off correctly.
Tested it on APQ8064 based IFC6410 with panel.
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The dsi bridge ops call drm_panel functions to set up the connected
drm_panel. Add checks to make sure these aren't called when we're
connected to an external bridge.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
There are platforms where the DSI output can be connected to another
encoder bridge chip (DSI to HDMI, DSI to LVDS etc).
Add support for external bridge support to the dsi driver. We assume that
the external bridge chip would be of the type drm_bridge. The dsi driver's
internal drm_bridge (msm_dsi->bridge) is linked to the external bridge's
drm_bridge struct.
In the case we're connected to an external bridge, we don't need to create
and manage a connector within our driver, it's the bridge driver's
responsibility to create one.
v2:
- Move the external bridge attaching stuff to dsi manager to make things
cleaner.
- Force the bridge to connect to a video mode encoder for now (the dsi
mode flags may have not been populated by modeset_init)
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Create a helper msm_dsi_device_connected() which checks whether we have a
device connected to the dsi host or not. This check gets messy when we
have support external bridges too. Having an inline function makes it
more legible.
For now, the check only consists of msm_dsi->panel being non-NULL. Later,
this will check if we have an external bridge or not.
This helper isn't used in dsi_connector related code as that's specific
to only when a drm_panel is connected.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
We currently support only panels connected to dsi output. We're going to
also support external bridge chips now.
Change 'panel_node' to 'device_node' in the struct msm_dsi_host and
'panel_flags' to 'device_flags' in msm_dsi. This makes things sound a
bit more generic.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Platforms containing only DSI video mode devices don't need a TE gpio.
Make TE gpio optional.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
We currently get the output connected to LVDS by looking for a phandle
called 'qcom,lvds-panel' under the mdp DT node.
Use the more standard of_graph approach to create an lvds output port,
and retrieve the panel node from the port's endpoint data.
v3
- Fix return value checks of of_graph_* calls.
Tested-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The dsi host looks for the connected panel node by parsing for a child
named 'panel'. This hierarchy isn't very flexible. The connected
panel is forced to be a child to the dsi host, and hence, a mipi dsi
device. This isn't suitable for dsi devices that don't use mipi dsi
as their control bus.
Follow the of_graph approach of creating ports and endpoints to
represent the connections between the dsi host and the panel connected
to it. In our case, the dsi host will only have one output port, linked
to the panel's input port.
Update DT binding documentation with device graph usage info.
v3:
- Fix return value checks of of_graph_* calls.
- Don't make port a mandatory DT property
- Fix defer check when no panel node specified
- Rename parse_dt func to align with other dsi_host funcs
Reviewed-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Decrement device node refcount if of_get_child_by_name is successfully
called.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
AHB clock should be enabled before accessing registers during
enable/disable_vblank(). Since these 2 callbacks are called in
atomic context while clk_prepare may cause thread sleep, a work
is scheduled to control vblanks.
v2: fixup spinlock initialization
Signed-off-by: Hai Li <hali@codeaurora.org>
[add comment about cancel_work_sync() before drm_irq_uninstall()]
Signed-off-by: Rob Clark <robdclark@gmail.com>
msm8x74v1 has different MDP5 version (v1.0) from msm8x74v2 (v1.2).
Add a separate config data to support msm8x74v1.
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
This change is to add planes which use DMA pipes for MDP5.
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>
[slight comment adjust to s/Construct public planes/Construct video
planes/ since DMA planes are public planes too, they just can't scale
or CSC]
Signed-off-by: Rob Clark <robdclark@gmail.com>
MDP planes can be implemented using different type of HW pipes,
RGB/VIG/DMA pipes for MDP5 and RGB/VG/DMA pipes for MDP4. Each type
of pipe has different HW capabilities such as scaling, color space
conversion, decimation... Add a variable in plane data structure
to specify the difference of each plane which comes from mdp5_cfg data
and use it to differenciate the plane operation.
V1: Initial change
V2: Fix a typo in mdp4_kms.h
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Add packed YUV422 and planar YUV420 formats to MDP supported
formats.
Signed-off-by: Stephane Viau <sviau@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Newer MDP5 uses 2 shared memory pool clients for certain YUV formats.
For example, if VIG0 is used to fetch data in YUYV format, it will use
VIG0_Y for Y component, and VIG0_Cr for UV packed.
Signed-off-by: Wentao Xu <wentaox@codeaurora.org>
[rebase]
Signed-off-by: Stephane Viau <sviau@codeaurora.org>
This makes it easy to determine if a format is YUV. The old
method of using chroma sample type incorrectly marks YUV444 as
RGB format.
Signed-off-by: Wentao Xu <wentaox@codeaurora.org>
[rebase]
Signed-off-by: Stephane Viau <sviau@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Since 39b2bbe3d7 (gpio: add flags argument to gpiod_get*() functions)
which appeared in v3.17-rc1, the gpiod_get* functions take an additional
parameter that allows to specify direction and initial value for output.
Use this to simplify the driver. Furthermore this is one caller less
that stops us making the flags argument to gpiod_get*() mandatory.
Acked-by: Alexandre Courbot <acourbot@nvidia.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reset DSI PHY silently changes its PLL registers to reset status,
which will make cached status in clock driver invalid and result
in wrong output rate of link clocks. The current restore mechanism
in DSI PLL does not cover all the cases. This change is to recover
PLL status after PHY reset to match HW status with cached status
in clock driver.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The dsi_destroy() function was called in two cases by the dsi_init() function
during error handling even if the passed variable contained a null pointer.
* This implementation detail could be improved by adjustments for jump
targets according to the Linux coding style convention.
* Drop an unnecessary initialisation for the variable "msm_dsi" then.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
[add couple missing ERR_PTR()'s]
Signed-off-by: Rob Clark <robdclark@gmail.com>
The dsi_destroy() function tests whether its argument is NULL and then
returns immediately. Thus the test around the call is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Rob Clark <robdclark@gmail.com>
This change takes advantage of a HW feature that synchronize
flush operation on CTL1 to CTL0, to keep dual DSI pipes in
sync.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
In MDP5, CTL contains information of the whole pipeline whose
output goes down to a display interface. In various cases, one
interface may require 2 CRTCs, but only one CTL. Some interfaces
also require to use certain CTLs.
Instead of allocating CTL for each active CRTC, this change is to
associate a CTL with each interface.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
This change is to add properties alpha/zpos/blend_mode to mdp5 plane
for alpha blending operation to generate the blended output.
v1: Initial change
v2: Change "premultilied" property to enum (Rob's comment)
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>
[Don't actually expose alpha/premultiplied props to userspace yet
pending a chance for discussion and some userspace to exercise it]
Signed-off-by: Rob Clark <robdclark@gmail.com>
This was a hold-over from the pre-atomic days and legacy userspace that
only understood CRTCs. Fortunately we don't have any properties, so
this doesn't change anything. But before we start growing some plane
properties, we should fix this.
Signed-off-by: Rob Clark <robdclark@gmail.com>
DSI PHY errors are falsely reported whenever a dsi error occurs. This is
because DSI_DLN0_PHY_ERR isn't only used as a status register, but also
used to mask PHY errors. Currently, we end up reading the mask bits too
and therefore always report errors.
Ignore the register mask bits and check for only the status/clear bits.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The maximum output width of one pipeline depends on the LayerMixer's
capability. It may be different on each target. Also, MDP5 doesn't
have vertical limitation in one frame, as long as the pixel clock
can be supported.
This change obtains the maximum LM resolution from configuration
table and treat it as the whole pipe's limitation for MDP5. The size
limit on MDP4 is not changed.
Signed-off-by: Hai Li <hali@codeaurora.org>
For all of these devices, msm89xy was the lead chip, so standardize the
compatible names to align with convention used by rest of the qcom/msm
drivers.
Signed-off-by: Rob Clark <robdclark@gmail.com>
This change adds the MDP and HDMI support for msm8x94.
Note that HDMI PHY registers are not being accessed anymore from
the driver.
Signed-off-by: Stephane Viau <sviau@codeaurora.org>
[rename compatible s/8x94/8994/ since preference is to not trust the
marketing folks who invent chip #'s but instead name things after the
lead chip.. we should rename some 80XY to 89XY to standardize on the
lead chip but leave that for another patch. Also, update dt bindings
doc]
Signed-off-by: Rob Clark <robdclark@gmail.com>
->reset() currently only accesses HDMI core registers, and yet it
is located in hdmi_phy*. Since no PHY registers are being
accessed during ->reset(), it would be better to bring that
function in hdmi core module where HDMI core registers are
usually being accessed.
This will also help for msm8x94 for which no PHY registers
accesses are done (->phy_init == NULL) but the HDMI PHY reset
from HDMI core still needs to be done.
Note:
SW_RESET_PLL bit is not written in hdmi_phy_8x60_reset(); this
write should not affect anything if the corresponding field is
not writable.
Signed-off-by: Stephane Viau <sviau@codeaurora.org>
[fixed warning about unused 'phy' in hpd_enable() while merging]
Signed-off-by: Rob Clark <robdclark@gmail.com>
DSI controller on msm8x94 is version 1.3, which requires different
power supplies and works with 20nm DSI PHY. This change is to add
the basic support for this version.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Some targets use pinctrl framework to configure some
pins. This change allows DSI driver to set default and
sleep pinctrl status.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The current term of *dual panel* in DSI driver code causes confusion.
It is supposed to indicate the panel using two DSI links. Rename it
to *dual DSI*.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
The source PLL to be used by each DSI PHY should be decided by
DSI manager based on dual DSI information, while the register
programming to select PLL is different from one type of PHY to
another. This change adds the H/W difference to PHY configuration
and updates the interface between DSI manager and PHY.
With this change, PLL selection can be supported on different
targets.
Signed-off-by: Hai Li <hali@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
Add HDMI HDCP support including HDCP PartI/II/III authentication.
V1: Initial Change
V2: Address Bjorn&Rob's comments
Refactor the authentication process to use single work instead
of multiple work for different authentication stages.
V3: Update to align with qcom SCM api.
Signed-off-by: Jilai Wang <jilaiw@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
This patch consists of 2 libfc fixes causing rare crashes, one iscsi one
causing a potential hang on shutdown, an I/O blocksize issue which caused a
regression and a memory leak in scsi-mq.
Signed-off-by: James Bottomley <JBottomley@Odin.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQEcBAABAgAGBQJVzsOiAAoJEDeqqVYsXL0M4wQIAIUe1SGv66OcAQLBcxxJUL2T
i+Ph2RE/er1iyzJepN56i77Mn2hnBgnQtmB/ibnjQfUsz/zo5PANjIfy+eFcG53G
qcb8l7a/BFCH3JHWXL7rJYN9G64sirADDL6SDLpX1JFsL21bAGdcEgbmefysvDmr
qFkiGH0Ty9YH58W+6j1pzQhh437rRgcM1KuY08sJsbKmyCVdzG5ketzBkONmBcTh
OTfPQjL32L4KR3THDUbpCiK6YAUtDvHjVB51lwoiB1ER7Ke+E+nqlCuxhXvZfMoD
lGfvWgwaQnoN1fun6c85zcqnl72kymropWWiJUhhPPjeZEj8sgn/eaaAAST9mlQ=
=Fztd
-----END PGP SIGNATURE-----
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"This has two libfc fixes for bugs causing rare crashes, one iscsi fix
for a potential hang on shutdown, and a fix for an I/O blocksize issue
which caused a regression"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
sd: Fix maximum I/O size for BLOCK_PC requests
libfc: Fix fc_fcp_cleanup_each_cmd()
libfc: Fix fc_exch_recv_req() error path
libiscsi: Fix host busy blocking during connection teardown
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iQEcBAABAgAGBQJVzmylAAoJEL/70l94x66D7r0IAKd8oclVTdbo8RxR1Hg2zZev
ytTm2Mjd0kgSqhTaBBUgyE900/cznYpT1xJq1/5Wwc+FP1J1QBzsDemtrQlEZIBh
Zi4b7zm37K1ai7xWs6oLaXieVjiyX8vuUGO6saBw1n/ZLURgPjVzTmQMxdnYtyFX
yf37rPvksnyzyctv+D9ZvdhrpD7Xd3NFNoCOSiukkeZkjb97JabDRrzpTlVmj4wu
KNReYCN+iA6jZe5tEZHzCGplVrEMfHdAcoRc3GVz3oecPVZojX/NLzwlw97iN/2z
mm5SVOlxbvCO7sqEQXo/db91xlP3E6Q1QGuDE21NboClbNeinC/uFJMFzpVInSI=
=AD69
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull KVM fixes from Paolo Bonzini:
"Just two very small & simple patches"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: x86: Use adjustment in guest cycles when handling MSR_IA32_TSC_ADJUST
KVM: x86: zero IDT limit on entry to SMM
Merge fixes from Andrew Morton:
"11 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
Update maintainers for DRM STI driver
mm: cma: mark cma_bitmap_maxno() inline in header
zram: fix pool name truncation
memory-hotplug: fix wrong edge when hot add a new node
.mailmap: Andrey Ryabinin has moved
ipc/sem.c: update/correct memory barriers
mm/hwpoison: fix panic due to split huge zero page
ipc,sem: remove uneeded sem_undo_list lock usage in exit_sem()
ipc,sem: fix use after free on IPC_RMID after a task using same semaphore set exits
mm/hwpoison: fix fail isolate hugetlbfs page w/ refcount held
mm/hwpoison: fix page refcount of unknown non LRU page
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=IHq6
-----END PGP SIGNATURE-----
Merge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Pull clock fix from Stephen Boyd:
"A one-liner for a regression found in the PXA clock driver"
* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
clk: pxa: pxa3xx: fix CKEN register access
Add Vincent Abriou and myself as maintainers.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Cc: Vincent Abriou <vincent.abriou@st.com>
Cc: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
cma_bitmap_maxno() was marked as static and not static inline, which can
cause warnings about this function not being used if this file is included
in a file that does not call that function, and violates the conventions
used elsewhere. The two options are to move the function implementation
back to mm/cma.c or make it inline here, and it's simple enough for the
latter to make sense.
Signed-off-by: Gregory Fong <gregory.0xf0@gmail.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
zram_meta_alloc() constructs a pool name for zs_create_pool() call as
snprintf(pool_name, sizeof(pool_name), "zram%d", device_id);
However, it defines pool name buffer to be only 8 bytes long (minus
trailing zero), which means that we can have only 1000 pool names: zram0
-- zram999.
With CONFIG_ZSMALLOC_STAT enabled an attempt to create a device zram1000
can fail if device zram100 already exists, because snprintf() will
truncate new pool name to zram100 and pass it debugfs_create_dir(),
causing:
debugfs dir <zram100> creation failed
zram: Error creating memory pool
... and so on.
Fix it by passing zram->disk->disk_name to zram_meta_alloc() instead of
divice_id. We construct zram%d name earlier and keep it as a ->disk_name,
no need to snprintf() it again.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When we add a new node, the edge of memory may be wrong.
e.g. system has 4 nodes, and node3 is movable, node3 mem:[24G-32G],
1. hotremove the node3,
2. then hotadd node3 with a part of memory, mem:[26G-30G],
3. call hotadd_new_pgdat()
free_area_init_node()
get_pfn_range_for_nid()
4. it will return wrong start_pfn and end_pfn, because we have not
update the memblock.
This patch also fixes a BUG_ON during hot-addition, please see
http://marc.info/?l=linux-kernel&m=142961156129456&w=2
Signed-off-by: Xishi Qiu <qiuxishi@huawei.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Gu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
sem_lock() did not properly pair memory barriers:
!spin_is_locked() and spin_unlock_wait() are both only control barriers.
The code needs an acquire barrier, otherwise the cpu might perform read
operations before the lock test.
As no primitive exists inside <include/spinlock.h> and since it seems
noone wants another primitive, the code creates a local primitive within
ipc/sem.c.
With regards to -stable:
The change of sem_wait_array() is a bugfix, the change to sem_lock() is a
nop (just a preprocessor redefinition to improve the readability). The
bugfix is necessary for all kernels that use sem_wait_array() (i.e.:
starting from 3.10).
Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
Reported-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Kirill Tkhai <ktkhai@parallels.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: <stable@vger.kernel.org> [3.10+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Bug:
------------[ cut here ]------------
kernel BUG at mm/huge_memory.c:1957!
invalid opcode: 0000 [#1] SMP
Modules linked in: snd_hda_codec_hdmi i915 rpcsec_gss_krb5 snd_hda_codec_realtek snd_hda_codec_generic nfsv4 dns_re
CPU: 2 PID: 2576 Comm: test_huge Not tainted 4.2.0-rc5-mm1+ #27
Hardware name: Dell Inc. OptiPlex 7020/0F5C5X, BIOS A03 01/08/2015
task: ffff880204e3d600 ti: ffff8800db16c000 task.ti: ffff8800db16c000
RIP: split_huge_page_to_list+0xdb/0x120
Call Trace:
memory_failure+0x32e/0x7c0
madvise_hwpoison+0x8b/0x160
SyS_madvise+0x40/0x240
? do_page_fault+0x37/0x90
entry_SYSCALL_64_fastpath+0x12/0x71
Code: ff f0 41 ff 4c 24 30 74 0d 31 c0 48 83 c4 08 5b 41 5c 41 5d c9 c3 4c 89 e7 e8 e2 58 fd ff 48 83 c4 08 31 c0
RIP split_huge_page_to_list+0xdb/0x120
RSP <ffff8800db16fde8>
---[ end trace aee7ce0df8e44076 ]---
Testcase:
#define _GNU_SOURCE
#include <stdlib.h>
#include <stdio.h>
#include <sys/mman.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/types.h>
#include <errno.h>
#include <string.h>
#define MB 1024*1024
int main(void)
{
char *mem;
posix_memalign((void **)&mem, 2 * MB, 200 * MB);
madvise(mem, 200 * MB, MADV_HWPOISON);
free(mem);
return 0;
}
Huge zero page is allocated if page fault w/o FAULT_FLAG_WRITE flag.
The get_user_pages_fast() which called in madvise_hwpoison() will get
huge zero page if the page is not allocated before. Huge zero page is a
tranparent huge page, however, it is not an anonymous page.
memory_failure will split the huge zero page and trigger
BUG_ON(is_huge_zero_page(page));
After commit 98ed2b0052 ("mm/memory-failure: give up error handling
for non-tail-refcounted thp"), memory_failure will not catch non anon
thp from madvise_hwpoison path and this bug occur.
Fix it by catching non anon thp in memory_failure in order to not split
huge zero page in madvise_hwpoison path.
After this patch:
Injecting memory failure for page 0x202800 at 0x7fd8ae800000
MCE: 0x202800: non anonymous thp
[...]
[akpm@linux-foundation.org: remove second split, per Wanpeng]
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After we acquire the sma->sem_perm lock in exit_sem(), we are protected
against a racing IPC_RMID operation. Also at that point, we are the last
user of sem_undo_list. Therefore it isn't required that we acquire or use
ulp->lock.
Signed-off-by: Herton R. Krzesinski <herton@redhat.com>
Acked-by: Manfred Spraul <manfred@colorfullife.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Rafael Aquini <aquini@redhat.com>
CC: Aristeu Rozanski <aris@redhat.com>
Cc: David Jeffery <djeffery@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Hugetlbfs pages will get a refcount in get_any_page() or
madvise_hwpoison() if soft offlining through madvise. The refcount which
is held by the soft offline path should be released if we fail to isolate
hugetlbfs pages.
Fix it by reducing the refcount for both isolation success and failure.
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: <stable@vger.kernel.org> [3.9+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After trying to drain pages from pagevec/pageset, we try to get reference
count of the page again, however, the reference count of the page is not
reduced if the page is still not on LRU list.
Fix it by adding the put_page() to drop the page reference which is from
__get_any_page().
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: <stable@vger.kernel.org> [3.9+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
