DT-enabled Dove has moved from ARCH_DOVE in mach-dove to MACH_DOVE
in mach-mvebu. As non-DT ARCH_DOVE will stay to rot for a while, add a new
DT-only MACH_DOVE to thermal Kconfig.
This was originally supposed to go in via "ARM: dove: prepare new Dove DT Kconfig"
patch from Sebastian Hesselbarth for 3.15, but slipped through the cracks.
I've tested on CuBox that without this patch you can't compile
dove_thermal into a mach-mvebu based kernel, and with this patch I can
build the driver and it works as expected run-time.
v2: non-ascii char creeped in somehow
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Eduardo Valentin <edubezval@gmail.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Acked-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Now ext4_has_inline_data() is used in wide spread codepaths. So we need
to make it as a inline function to avoid burning some CPU cycles.
Change in text size:
text data bss dec hex filename
before: 326110 19258 5528 350896 55ab0 fs/ext4/ext4.o
after: 326227 19258 5528 351013 55b25 fs/ext4/ext4.o
I use the following script to measure the CPU usage.
#!/bin/bash
shm_base='/dev/shm'
img=${shm_base}/ext4-img
mnt=/mnt/loop
e2fsprgs_base=$HOME/e2fsprogs
mkfs=${e2fsprgs_base}/misc/mke2fs
fsck=${e2fsprgs_base}/e2fsck/e2fsck
sudo umount $mnt
dd if=/dev/zero of=$img bs=4k count=3145728
${mkfs} -t ext4 -O inline_data -F $img
sudo mount -t ext4 -o loop $img $mnt
# start testing...
testdir="${mnt}/testdir"
mkdir $testdir
cd $testdir
echo "start testing..."
for ((cnt=0;cnt<100;cnt++)); do
for ((i=0;i<5;i++)); do
for ((j=0;j<5;j++)); do
for ((k=0;k<5;k++)); do
for ((l=0;l<5;l++)); do
mkdir -p $i/$j/$k/$l
echo "$i-$j-$k-$l" > $i/$j/$k/$l/testfile
done
done
done
done
ls -R $testdir > /dev/null
rm -rf $testdir/*
done
The result of `perf top -G -U` is as below.
vanilla:
13.92% [ext4] [k] ext4_do_update_inode
9.36% [ext4] [k] __ext4_get_inode_loc
4.07% [ext4] [k] ftrace_define_fields_ext4_writepages
3.83% [ext4] [k] __ext4_handle_dirty_metadata
3.42% [ext4] [k] ext4_get_inode_flags
2.71% [ext4] [k] ext4_mark_iloc_dirty
2.46% [ext4] [k] ftrace_define_fields_ext4_direct_IO_enter
2.26% [ext4] [k] ext4_get_inode_loc
2.22% [ext4] [k] ext4_has_inline_data
[...]
After applied the patch, we don't see ext4_has_inline_data() because it
has been inlined and perf couldn't sample it. Although it doesn't mean
that the CPU cycles can be saved but at least the overhead of function
calls can be eliminated. So IMHO we'd better inline this function.
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This is the traditional way of obtaining a device driver's register
address space. The aim of this driver is to supply controller specific
information to the ST Thermal Core.
Signed-off-by: Ajit Pal Singh <ajitpal.singh@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
This core is shared by both ST's 'memory mapped' and
'system configuration register' based Thermal controllers.
Signed-off-by: Ajit Pal Singh <ajitpal.singh@st.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
There is no kind of file which does not supply a page reading function.
Signed-off-by: Zhang Zhen <zhenzhang.zhang@huawei.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Since init_failed flag was moved to struct hda_intel, its access in
the commit [4da63c6f: ALSA: hda - Fix broken PM due to incomplete
i915 initialization] is also replaced with hda->init_failed
appropriately.
When the initialization of Intel HDMI controller fails due to missing
i915 kernel symbols (e.g. HD-audio is built in while i915 is module),
the driver discontinues the probe. However, since the probe was done
asynchronously, the driver object still remains, thus the relevant PM
ops are still called at suspend/resume. This results in the bad access
to the incomplete audio card object, eventually leads to Oops or stall
at PM.
This patch adds the missing checks of chip->init_failed flag at each
PM callback in order to fix the problem above.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=79561
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The commit [247bc037: PM / Sleep: Mitigate race between the freezer
and request_firmware()] introduced the finer state control, but it
also leads to a new bug; for example, a bug report regarding the
firmware loading of intel BT device at suspend/resume:
https://bugzilla.novell.com/show_bug.cgi?id=873790
The root cause seems to be a small window between the process resume
and the clear of usermodehelper lock. The request_firmware() function
checks the UMH lock and gives up when it's in UMH_DISABLE state. This
is for avoiding the invalid f/w loading during suspend/resume phase.
The problem is, however, that usermodehelper_enable() is called at the
end of thaw_processes(). Thus, a thawed process in between can kick
off the f/w loader code path (in this case, via btusb_setup_intel())
even before the call of usermodehelper_enable(). Then
usermodehelper_read_trylock() returns an error and request_firmware()
spews WARN_ON() in the end.
This oneliner patch fixes the issue just by setting to UMH_FREEZING
state again before restarting tasks, so that the call of
request_firmware() will be blocked until the end of this function
instead of returning an error.
Fixes: 247bc03742 (PM / Sleep: Mitigate race between the freezer and request_firmware())
Link: https://bugzilla.novell.com/show_bug.cgi?id=873790
Cc: 3.4+ <stable@vger.kernel.org> # 3.4+
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The legacy-style definition of the hwmod addr space is no longer
required as AM33xx/AM43xx are DT-boot only, and the minimal mailbox
DT nodes have been added, so clean up this data.
Acked-by: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The legacy-style definition of the hwmod addr space is no longer
required after the addition of the OMAP4 mailbox DT node, so
clean up this data.
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Benoît Cousson <bcousson@baylibre.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
OMAP2 devices are devicetree boot only, and the legacy mode
of mailbox device creation should no longer be used, so remove
the mailbox attribute data and the hwmod addr space used for
creating mailboxes in legacy mode.
Acked-by: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The legacy platform device for mailbox should not be created for
a DT boot, so adjust the platform device initialization logic
appropriately.
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Add the hwmod data for the 13 instances of the system mailbox
IP in DRA7 SoC. The patch is needed for performing a soft-reset
while configuring the respective mailbox instance, otherwise is
a non-essential change for functionality. The modules are smart
idled on reset, and the IP module mode is hardware controlled.
Cc: Rajendra Nayak <rnayak@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
DRA7xx has 13 system mailboxes, and is present on both the
DRA72x and DRA74x family of SoCs. Add the DT nodes for all
these 13 mailboxes. Except for mailbox 1, all other mailboxes
do not have interrupts mapped into the MPU GIC by default.
All the mailboxes have been disabled and the interrupts
property information is left out intentionally for now,
because of the dependencies against the crossbar driver.
These mailboxes can be enabled when a usecase arises
and the crossbar driver dependencies are met.
NOTE: The mailbox 1 has different number of mailbox fifos
and IP interrupts compared to the remaining 12 mailboxes.
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The mailbox DT node for AM4372 is enabled and is corrected to
remove some properties that have crept in by mistake.
Fixes: 9e3269b (ARM: dts: AM4372: Add L2, EDMA, mailbox, MMC and SHAM nodes)
Cc: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The mailbox DT node data has been added for AM33xx device.
The mailbox IP in AM33xx is similar to the version found in
OMAP4+ devices.
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The mailbox DT node data has been added for OMAP44xx
devices.
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The number of mailbox fifos and users (IP interrupts) are added
to the Mailbox DT nodes on OMAP2420, OMAP2430, OMAP3, and OMAP5
family of SoCs through the DT properties "ti,mbox-num-fifos" and
"ti,mbox-num-users" properties. This data represents the same data
that used to be represented in hwmod attribute data through the
.num_fifos and .num_users fields previously.
Signed-off-by: Suman Anna <s-anna@ti.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Add a driver to handle the shutdown of the Atmel SoCs. This code used to be
(and still is) in arch/arm/mach-at91. We didn't remove it yet so that we can
convert all the boards to using this driver, before removing it entirely in a
separate patch.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Implement the reset behaviour of the various AT91 SoCS in drivers/power/reset.
It used to be (and still is) located in arch/arm/mach-at91, and in order to
preserve bisectability is not removed yet, but every board should be converted
to use this driver instead.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
If an IRQ has been configured for wakeup via enable_irq_wake(), the
driver who has done that must be prepared for receiving interrupts
after suspend_device_irqs() has returned, so there is no need to
"suspend" such IRQs. Moreover, if drivers using enable_irq_wake()
actually want to receive interrupts after suspend_device_irqs() has
returned, they need to add IRQF_NO_SUSPEND to the IRQ flags while
requesting the IRQs, which shouldn't be necessary (it also goes a bit
too far, as IRQF_NO_SUSPEND causes the IRQ to be ignored by
suspend_device_irqs() all the time regardless of whether or not it
has been configured for signaling wakeup).
For the above reasons, make __disable_irq() ignore IRQ descriptors
with IRQD_WAKEUP_STATE set when its suspend argument is true which
effectively causes them to behave like IRQs with IRQF_NO_SUSPEND
set.
This also allows IRQs configured for wakeup via enable_irq_wake()
to work as wakeup interrupts for the "freeze" (suspend-to-idle)
sleep mode automatically just like for any other sleep states.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Li Aubrey <aubrey.li@linux.intel.com>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: One Thousand Gnomes <gnomes@lxorguk.ukuu.org.uk>
Link: http://lkml.kernel.org/r/4679574.kGUnqAuNl9@vostro.rjw.lan
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This panel is used by Atmel's SAMA5D3 Evaluation Kits (sama5d3xek) and
supported by the simple-panel driver.
Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Currently punch hole code on files with direct/indirect mapping has some
problems which may lead to a data loss. For example (from Jan Kara):
fallocate -n -p 10240000 4096
will punch the range 10240000 - 12632064 instead of the range 1024000 -
10244096.
Also the code is a bit weird and it's not using infrastructure provided
by indirect.c, but rather creating it's own way.
This patch fixes the issues as well as making the operation to run 4
times faster from my testing (punching out 60GB file). It uses similar
approach used in ext4_ind_truncate() which takes advantage of
ext4_free_branches() function.
Also rename the ext4_free_hole_blocks() to something more sensible, like
the equivalent we have for extent mapped files. Call it
ext4_ind_remove_space().
This has been tested mostly with fsx and some xfstests which are testing
punch hole but does not require unwritten extents which are not
supported with direct/indirect mapping. Not problems showed up even with
1024k block size.
CC: stable@vger.kernel.org
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Commit 27dd438542 ("ext4: introduce reserved space") reserves 2% of
the file system space to make sure metadata allocations will always
succeed. Given that, tracking the reservation of metadata blocks is
no longer necessary.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The EXT4FS_DEBUG is a *very* developer specific #ifdef designed for
ext4 developers only. (You have to modify fs/ext4/ext4.h to enable
it.)
Rearrange how we initialize data structures to avoid calling
ext4_count_free_clusters() until the multiblock allocator has been
initialized.
This also allows us to only call ext4_count_free_clusters() once, and
simplifies the code somewhat.
(Thanks to Chen Gang <gang.chen.5i5j@gmail.com> for pointing out a
!CONFIG_SMP compile breakage in the original patch.)
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
The check for the blacklist in hci_le_conn_complete_evt() should be when
we know that we have an actual successful connection (ev->status being
non-zero). This patch fixes this ordering.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
For CPU's that don't have a cycle counter, or something equivalent
which can be used for random_get_entropy(), random_get_entropy() will
always return 0. In that case, substitute with the saved interrupt
registers to add a bit more unpredictability.
Some folks have suggested hashing all of the registers
unconditionally, but this would increase the overhead of
add_interrupt_randomness() by at least an order of magnitude, and this
would very likely be unacceptable.
The changes in this commit have been benchmarked as mostly unaffecting
the overhead of add_interrupt_randomness() if the entropy counter is
present, and doubling the overhead if it is not present.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: Jörn Engel <joern@logfs.org>
This patch introduces a derating factor to struct hwrng for
the random bits going into the kernel input pool, and a common
default derating for drivers which do not specify one.
Signed-off-by: Torsten Duwe <duwe@suse.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Acked-by: H. Peter Anvin <hpa@zytor.com>
This can be viewed as the in-kernel equivalent of hwrngd;
like FUSE it is a good thing to have a mechanism in user land,
but for some reasons (simplicity, secrecy, integrity, speed)
it may be better to have it in kernel space.
This patch creates a thread once a hwrng registers, and uses
the previously established add_hwgenerator_randomness() to feed
its data to the input pool as long as needed. A derating factor
is used to bias the entropy estimation and to disable this
mechanism entirely when set to zero.
Signed-off-by: Torsten Duwe <duwe@suse.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Acked-by: H. Peter Anvin <hpa@zytor.com>
This patch adds an interface to the random pool for feeding entropy
in-kernel.
Signed-off-by: Torsten Duwe <duwe@suse.de>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Acked-by: H. Peter Anvin <hpa@zytor.com>
Use more efficient fast_mix() function. Thanks to George Spelvin for
doing the leg work to find a more efficient mixing function.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: George Spelvin <linux@horizon.com>
For architectures that don't have cycle counters, the algorithm for
deciding when to avoid giving entropy credit due to back-to-back timer
interrupts didn't make any sense, since we were checking every 64
interrupts. Change it so that we only give an entropy credit if the
majority of the interrupts are not based on the timer.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: George Spelvin <linux@horizon.com>
In xfer_secondary_pull(), check to make sure we need to pull from the
secondary pool before checking and potentially updating the
last_pulled time.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: George Spelvin <linux@horizon.com>