Now that we can get there in RCU mode, we shouldn't play with
nd->path.dentry->d_inode - it's not guaranteed to be stable.
Use nd->inode instead.
Reported-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The USB Battery Charging spec (BC1.2) suggests a dedicated
charging port can deliver from 0.5 to 5.0A at between 4.75 and 5.25
volts.
To choose the "correct" current voltage setting requires a trial
and error approach: try to draw current and see if the voltage drops
too low.
Even with a configured Standard Downstream Port, it may not be possible
to reliably pull 500mA - depending on cable quality and source
quality I have reports of charging failure due to the voltage dropping
too low.
To address both these concerns, this patch introduce incremental
current setting.
The current pull from VBUS is increased in steps of 20mA every 100ms
until the target is reached or until the measure voltage drops below
4.75V. If the voltage does go too low, the target current is reduced
by 20mA and kept there.
This applies to currents selected automatically, or to values
set via sysfs. So setting a large value will cause the maximum
available to be used - up to the limit of 1.7A imposed by the
hardware.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
This allows AC charging to be turned off, much like usb charging.
"continuous" mode is not available though.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Add a 'continuous' option for usb charging which enables
the "linear" charging mode of the twl4030.
Linear charging does a good job with not-so-reliable power sources.
Auto mode does not work well as it switches off when voltage drops
momentarily. Care must be taken not to over-charge.
It was used with a bike hub dynamo since a year or so. In that case
there are automatically charging stops when the cyclist needs a break.
Original-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
'off' or 'auto' to
/sys/class/power/twl4030_usb/mode
will now enable or disable charging from USB port. Normally this is
enabled on 'plug' and disabled on 'unplug'.
Unplug will still disable charging. 'plug' will only enable it if
'auto' if selected.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
'max_current' sysfs attributes are created which allow the
max to be set.
Whenever a current source changes, the default is restored.
This will be followed by a uevent, so user-space can decide to
update again.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
The twl4030 charger has two current sources, 'USB' and 'AC'
(presumably "Accessory Charger" because it isn't Alternating Current).
If 'AC' is providing current, we should set the current limit
differently to when it isn't (and so USB is used).
So split 'cur' into 'usb_cur' and 'ac_cur' and use accordingly.
Now we must review the current setting on any interrupt or USB
event which might indicate that the charger-source has changed.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
If CONFIG_SH_DMAE_BASE (which is required for DMA engine support for
legacy SH, SH/R-Mobile, and R-Car Gen1, but not for R-Car Gen2) is not
enabled, but CONFIG_RCAR_DMAC (for R-Car Gen2 DMA engine support) is,
and the DTS doesn't provide a "dmas" property for a device,
dma_request_slave_channel_compat() incorrectly succeeds, and returns a
DMA channel.
However, when trying to use that DMA channel later, it fails with:
rcar-dmac e6700000.dma-controller: rcar_dmac_prep_slave_sg: bad parameter: len=1, id=-22
(Fortunately most drivers can handle this failure, and fall back to
PIO)
The reason for this is that a NULL legacy filter function is used, which
actually means "all channels are OK", not "do not match".
If CONFIG_SH_DMAE_BASE is enabled (like in shmobile_defconfig, which
supports other SoCs besides R-Car Gen2), shdma_chan_filter() correctly
returns false, as no available channel on R-Car Gen2 matches a
shdma-base channel.
If the DTS does provide a "dmas" property, dma_request_slave_channel()
succeeds, and legacy filter-based matching is not used.
To fix this, change shdma_chan_filter from being NULL to a dummy
function that always returns false, like is done on other platforms.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
This patch adds DEV_TO_DEV support for i.MX SDMA driver to support data
transfer between two peripheral FIFOs.
The per_2_per script requires two peripheral addresses and two DMA
requests, and it need to check the src addr and dst addr is in the SPBA
bus space or in the AIPS bus space.
Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The twl4030 allows control of the incoming current.
Part of this control is a 'CGAIN' setting which doubles
the range for half the precision. This control affects
several different current setting, so all need to be updated
at once when CGAIN is changed.
With this patch, all of these current setting are managed
by the driver, but most are left at their default settings.
The current drawn is set to 500mA if the allow_usb module parameter is
set, and to 100mA otherwise.
More fine control will appear in later patches.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
We will need this calculation in other places, so
create functions to map between register value and uA value.
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
The usb phy driver already determines when VBUS is available,
so repeating the test in the charger driver is pointless duplication.
On probe, process the last event from the phy, and from then on,
do whatever the phy tells us without double-checking.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Now that twl4030_bci_probe can safely return -EPROBE_DEFER,
do so when devm_usb_get_phy_by_node returns that error.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Drivers using module_platform_driver_probe cannot return
EPROBE_DEFER from the probe function, which makes them rather useless
these days...
Convert to module_platform_driver() so EPROBE_DEFER can be used.
Signed-off-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
The twl4030 usb phy needs to be active while we are using
the USB VBUS as a current source for charging.
In particular, the usb3v1 regulator must be enabled and the
PHY_PWR_PHYPWD bit must be set to keep the phy powered.
commit ab37813f40
twl4030_charger: Allow charger to control the regulator that feeds it
gave the charger control over the regulator, but didn't resolve
the PHY_PWR_PHYPWD issue.
Now that both of these are controlled by runtime_pm in
phy-twl4030-usb, we can simply take a runtime_pm reference to the USB
phy whenever the charger wants to use it as a current source.
So this patch reverts the above commit, and adds the necessary
runtime_pm calls.
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
* pci/irq:
PCI/MSI: Free legacy IRQ when enabling MSI/MSI-X
PCI: Add helpers to manage pci_dev->irq and pci_dev->irq_managed
PCI, x86: Implement pcibios_alloc_irq() and pcibios_free_irq()
PCI: Add pcibios_alloc_irq() and pcibios_free_irq()
* pci/misc:
PCI: Remove unused "pci_probe" flags
PCI: Add VPD function 0 quirk for Intel Ethernet devices
PCI: Add dev_flags bit to access VPD through function 0
PCI / ACPI: Fix pci_acpi_optimize_delay() comment
PCI: Remove a broken link in quirks.c
PCI: Remove useless redundant code
PCI: Simplify pci_find_(ext_)capability() return value checks
PCI: Move PCI_FIND_CAP_TTL to pci.h and use it in quirks
PCI: Add pcie_downstream_port() (true for Root and Switch Downstream Ports)
PCI: Fix pcie_port_device_resume() comment
PCI: Shift PCI_CLASS_NOT_DEFINED consistently with other classes
PCI: Revert aeb30016fe ("PCI: add Intel USB specific reset method")
PCI: Fix TI816X class code quirk
PCI: Fix generic NCR 53c810 class code quirk
PCI: Use PCI_CLASS_SERIAL_USB instead of bare number
PCI: Add quirk for Intersil/Techwell TW686[4589] AV capture cards
PCI: Remove Intel Cherrytrail D3 delays
* pci/resource:
PCI: Call pci_read_bridge_bases() from core instead of arch code
* pci/virtualization:
PCI: Restore ACS configuration as part of pci_restore_state()
* pci/host-designware:
PCI: designware: Don't complain missing *config* reg space if va_cfg0 is set
* pci/host-xgene:
PCI: xgene: Add support for a 64-bit prefetchable memory window
arm64: dts: Add APM X-Gene PCIe 64-bit prefetchable window
PCI: xgene: Drop owner assignment from platform_driver
* pci/host-xilinx:
PCI: xilinx: Check for MSI interrupt flag before handling as INTx
This indicates the device coordinates should be directly mapped to screen.
This is valid since scaling/flipping/rotation should be done by configuring
the MXT device.
It also flags to Android using INPUT_PROP_DIRECT that the device should be
treated as a touch screen by default, and removes the necessity for a
default IDC file.
Signed-off-by: Nick Dyer <nick.dyer@itdev.co.uk>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
The CHG/interrupt line is momentarily set (approximately 100 ms) as an
input after power-up or reset for diagnostic purposes. This may cause
spurious interrupts, so disable interrupt handler during this period.
Signed-off-by: Nick Dyer <nick.dyer@itdev.co.uk>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
input_dev may be NULL if mxt_initialize_input_device fails. But pm ops is
still available and suspend/resume assume
input_dev is not NULL. To fix this issue, we add a check if (!input_dev).
Signed-off-by: Pan Xinhui <xinhuix.pan@intel.com>
Signed-off-by: Nick Dyer <nick.dyer@itdev.co.uk>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
The hardcoded 0x83 CTRL setting overrides other settings in that byte,
enabling extra reporting that may not be useful on a particular platform.
Implement improved suspend mechanism via deep sleep. By writing zero to
both the active and idle cycle times the maXTouch device can be put into a
deep sleep mode, using minimal power. It is necessary to issue a calibrate
command after the chip has spent any time in deep sleep, however a soft
reset is unnecessary.
Use the old method on Chromebook Pixel via platform data option.
This patch also deals with the situation where the power configuration is
zero on probe, which would mean that the device never wakes up to execute
commands.
After a config download, the T7 power configuration may have changed so it
is necessary to re-read it.
Signed-off-by: Nick Dyer <nick.dyer@itdev.co.uk>
Acked-by: Benson Leung <bleung@chromium.org>
Acked-by: Yufeng Shen <miletus@chromium.org>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Without this initialization, gateways which actually announce up/down
bandwidth of 0/0 could be added. If these nodes get purged via
_batadv_purge_orig() later, the gw_node structure does not get removed
since batadv_gw_node_delete() updates the gw_node with up/down
bandwidth of 0/0, and the updating function then discards the change
and does not free gw_node.
This results in leaking the gw_node structures, which references other
structures: gw_node -> orig_node -> orig_node_ifinfo -> hardif. When
removing the interface later, the open reference on the hardif may cause
hangs with the infamous "unregister_netdevice: waiting for mesh1 to
become free. Usage count = 1" message.
Signed-off-by: Simon Wunderlich <simon@open-mesh.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
The tt_local_entry deletion performed in batadv_tt_local_remove() was neither
protecting against simultaneous deletes nor checking whether the element was
still part of the list before calling hlist_del_rcu().
Replacing the hlist_del_rcu() call with batadv_hash_remove() provides adequate
protection via hash spinlocks as well as an is-element-still-in-hash check to
avoid 'blind' hash removal.
Fixes: 068ee6e204 ("batman-adv: roaming handling mechanism redesign")
Reported-by: alfonsname@web.de
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
batadv_softif_vlan_get() may return NULL which has to be verified
by the caller.
Fixes: 35df3b298f ("batman-adv: fix TT VLAN inconsistency on VLAN re-add")
Reported-by: Ryan Thompson <ryan@eero.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
When a node running DAT receives an ARP request from the LAN for the
first time, it is likely that this node will request the ARP entry
through the distributed ARP table (DAT) in the mesh.
Once a DAT reply is received the asking node must check if the MAC
address for which the IP address has been asked is local. If it is, the
node must drop the ARP reply bceause the client should have replied on
its own locally.
Forwarding this reply means fooling any L2 bridge (e.g. Ethernet
switches) lying between the batman-adv node and the LAN. This happens
because the L2 bridge will think that the client sending the ARP reply
lies somewhere in the mesh, while this node is sitting in the same LAN.
Reported-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
The audit watch parent count was imbalanced, adding an unnecessary layer of
watch parent references. Decrement the additional parent reference when a
watch is reused, already having a reference to the parent.
audit_find_parent() gets a reference to the parent, if the parent is
already known. This additional parental reference is not needed if the
watch is subsequently found by audit_add_to_parent(), and consumed if
the watch does not already exist, so we need to put the parent if the
watch is found, and do nothing if this new watch is added to the parent.
If the parent wasn't already known, it is created with a refcount of 1
and added to the audit_watch_group, then incremented by one to be
subsequently consumed by the newly created watch in
audit_add_to_parent().
The rule points to the watch, not to the parent, so the rule's refcount
gets bumped, not the parent's.
See LKML, 2015-07-16
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
The audit watch count was imbalanced, adding an unnecessary layer of watch
references. Only add the second reference when it is added to a parent.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
If a GPIO driver uses gpiochip_add_pin_range() (which is usually the
case for GPIO/PFC combos), the GPIO hogging mechanism configured from DT
doesn't work:
requesting hog GPIO led1-high (chip sh73a0_pfc, offset 20) failed
The actual error code is -517 == -EPROBE_DEFER.
The problem is that PFC+GPIO registration is handled in multiple steps:
1. pinctrl_register(),
2. gpiochip_add(),
3. gpiochip_add_pin_range().
Configuration of the hogs is handled in gpiochip_add():
gpiochip_add
of_gpiochip_add
of_gpiochip_scan_hogs
gpiod_hog
gpiochip_request_own_desc
__gpiod_request
chip->request
pinctrl_request_gpio
pinctrl_get_device_gpio_range
However, at this point the GPIO controller hasn't been added to
pinctrldev_list yet, so the range can't be found, and the operation fails
with -EPROBE_DEFER.
To fix this, add a "gpio-ranges" property to the gpio device node, so
the ranges are added by of_gpiochip_add_pin_range(), which is called by
of_gpiochip_add() before the call to of_gpiochip_scan_hogs().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
If a GPIO driver uses gpiochip_add_pin_range() (which is usually the
case for GPIO/PFC combos), the GPIO hogging mechanism configured from DT
doesn't work:
requesting hog GPIO lcd0 (chip r8a7740_pfc, offset 176) failed
The actual error code is -517 == -EPROBE_DEFER.
The problem is that PFC+GPIO registration is handled in multiple steps:
1. pinctrl_register(),
2. gpiochip_add(),
3. gpiochip_add_pin_range().
Configuration of the hogs is handled in gpiochip_add():
gpiochip_add
of_gpiochip_add
of_gpiochip_scan_hogs
gpiod_hog
gpiochip_request_own_desc
__gpiod_request
chip->request
pinctrl_request_gpio
pinctrl_get_device_gpio_range
However, at this point the GPIO controller hasn't been added to
pinctrldev_list yet, so the range can't be found, and the operation fails
with -EPROBE_DEFER.
To fix this, add a "gpio-ranges" property to the gpio device node, so
the range is added by of_gpiochip_add_pin_range(), which is called by
of_gpiochip_add() before the call to of_gpiochip_scan_hogs().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
If a GPIO driver uses gpiochip_add_pin_range() (which is usually the
case for GPIO/PFC combos), the GPIO hogging mechanism configured from DT
doesn't work:
requesting hog GPIO led1-high (chip r8a73a4_pfc, offset 28) failed
The actual error code is -517 == -EPROBE_DEFER.
The problem is that PFC+GPIO registration is handled in multiple steps:
1. pinctrl_register(),
2. gpiochip_add(),
3. gpiochip_add_pin_range().
Configuration of the hogs is handled in gpiochip_add():
gpiochip_add
of_gpiochip_add
of_gpiochip_scan_hogs
gpiod_hog
gpiochip_request_own_desc
__gpiod_request
chip->request
pinctrl_request_gpio
pinctrl_get_device_gpio_range
However, at this point the GPIO controller hasn't been added to
pinctrldev_list yet, so the range can't be found, and the operation fails
with -EPROBE_DEFER.
To fix this, add a "gpio-ranges" property to the gpio device node, so
the ranges are added by of_gpiochip_add_pin_range(), which is called by
of_gpiochip_add() before the call to of_gpiochip_scan_hogs().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
If there are gced dirty pages and normal dirty pages in the mapping
of one inode, we might writeback them alternately with discontinuous
block address, resulting in low performance.
This patch introduces f2fs_write_cache_pages with codes copied from
write_cache_pages in mm/page-writeback.c.
In this function, we refactor flow with two steps:
1) writeback all cold type pages.
2) writeback all non-cold type pages.
By using this method, f2fs will writeback dirty pages with the same
temperature in bunch mode, it makes writeouted block being with
more continuous address, so they can be merged as much as possible
in f2fs bio cache, and also it will reduce the chance of submiting
small IO from block layer.
Test environment: 8g nokia sd card (very old sd card, but it shows
better effect when testing with this patch, and with a 32g kingston
sd card, I didn't see much more improvement).
Test step:
1. touch testfile;
2. truncate -s 512K testfile;
3. write all pages with odd index;
4. trigger gc by ioctl;
5. write all pages with even index;
6. time fsync testfile.
before:
real 0m0.402s
user 0m0.000s
sys 0m0.000s
after:
real 0m0.143s
user 0m0.004s
sys 0m0.004s
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch fixes to return correct error number of ->setxattr, which
is reported by xfstest tests/generic/026 as below:
generic/026 - output mismatch
--- tests/generic/026.out
+++ results/generic/026.out.bad
@@ -4,6 +4,6 @@
1 below acl max
acl max
1 above acl max
-chacl: cannot set access acl on "largeaclfile": Argument list too long
+chacl: cannot set access acl on "largeaclfile": Numerical result out of range
use 16 aces
use 17 aces
...
Ran: generic/026
Failures: generic/026
Failed 1 of 1 tests
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Previously, since 'commit 4531929e39 ("f2fs: move grabing orphan
pages out of protection region")' was committed, in write_orphan_inodes(),
we will grab all meta page in a batch before we use them under spinlock,
so that we can avoid large time delay of grabbing meta pages under
spinlock.
Now, 'commit d6c67a4fee ("f2fs: revmove spin_lock for
write_orphan_inodes")' remove the spinlock in write_orphan_inodes,
so there is no issue we describe above, we'd better recover to move
the grab operation to original place for readability.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
With cost-benifit method, background gc will consider old section with
fewer valid blocks as candidate victim, these old blocks in section will
be treated as cold data, and laterly will be moved into cold segment.
But if the gcing page is attached by user through buffered or mmaped
write, we should reset the page as non-cold one, because this page may
have more opportunity for further updating.
So fix to add clearing code for the missed 'mmap' case.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
When background gc is off, the only way to trigger gc is executing
a force gc in some operations who wants to grab space in disk.
The executing condition is limited: to execute force gc, we should
wait for the time when there is almost no more free section for LFS
allocation. This seems not reasonable for our user who wants to
control triggering gc by himself.
This patch introduces F2FS_IOC_GARBAGE_COLLECT interface for
triggering garbage collection by using ioctl. It provides our users
one more option to trigger gc.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch moves extent cache related code from data.c into extent_cache.c
since extent cache is independent feature, and its codes are not relate to
others in data.c, it's better for us to maintain them in separated place.
There is no functionality change, but several small coding style fixes
including:
* rename __drop_largest_extent to f2fs_drop_largest_extent for exporting;
* rename misspelled word 'untill' to 'until';
* remove unneeded 'return' in the end of f2fs_destroy_extent_tree().
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Since only parts of extents longer than F2FS_MIN_EXTENT_LEN will
be kept in extent cache after split, extents already shorter than
F2FS_MIN_EXTENT_LEN don't need to try split at all.
Signed-off-by: Fan Li <fanofcode.li@samsung.com>
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch fixes to update page flag (e.g. Uptodate/cold flag) in
->write_begin.
Otherwise, page will be non-uptodate when we try to write entire
page, and cold data flag in page will not be clean when gced page
is being rewritten.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
If an extent_tree entry has a zero reference count, we can drop it from the
cache in higher priority rather than currently referencing entries.
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In ->writepages, we use writepages mutex lock to serialize all block
address allocation and page submitting pairs from different inodes.
This method makes our delayed dirty pages of one inode being written
continously as many as possible.
But there is one problem that we did not submit current cached bio in
protection region of writepages mutex lock, so there is a small chance
that we submit the one of other thread's as below, resulting in
splitting more bios.
thread 1 thread 2
->writepages
lock(writepages)
->write_cache_pages
unlock(writepages)
lock(writepages)
->write_cache_pages
->f2fs_submit_merged_bio
->writepage
unlock(writepages)
fs_mark-6535 [002] .... 2242.270230: f2fs_submit_write_bio: dev = (1,0), WRITE_SYNC, DATA, sector = 5766152, size = 524288
fs_mark-6536 [000] .... 2242.270361: f2fs_submit_write_bio: dev = (1,0), WRITE_SYNC, DATA, sector = 5767176, size = 4096
fs_mark-6536 [000] .... 2242.270370: f2fs_submit_write_bio: dev = (1,0), WRITE_SYNC, NODE, sector = 8138112, size = 4096
fs_mark-6535 [002] .... 2242.270776: f2fs_submit_write_bio: dev = (1,0), WRITE_SYNC, DATA, sector = 5767184, size = 516096
This may really increase time of block layer works, and may cause
larger IO lantency.
This patch moves the submitting operation into region of writepages
mutex lock to avoid bio splits when concurrently writebacking is
intensive.
my test environment: virtual machine,
intel cpu i5 2500, 8GB size memory, 4GB size ramdisk
time fs_mark -t 16 -L 1 -s 524288 -S 1 -d /mnt/f2fs/
before:
real 0m4.244s
user 0m0.088s
sys 0m12.336s
after:
real 0m3.822s
user 0m0.072s
sys 0m10.760s
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
When testing with fs_mark, some blocks were written out as cold
data which were mixed with warm data, resulting in splitting more
bios.
This is because fs_mark will create file with random filename as
below:
559551ee~~~~~~~~15Z29OCC05JCKQP60JQ42MKV
559551ee~~~~~~~~NZAZ6X8OA8LHIIP6XD0L58RM
559551ef~~~~~~~~B15YDSWAK789HPSDZKYTW6WM
559551f1~~~~~~~~2DAE5DPS79785BUNTFWBEMP3
559551f1~~~~~~~~1MYDY0BKSQCJPI32Q8C514RM
559551f1~~~~~~~~YQOTMAOMN5CVRFOUNI026MP4
559551f3~~~~~~~~1WF42LPRTQJNPPGR3EINKMPE
559551f3~~~~~~~~8Y2NRK7CEPPAA02LY936PJPG
They are regarded as cold file since their filename are ended with
multimedia files' extension, but this should be wrong as we only
match the extension of filename, not the whole one.
In this patch, we try to fix the format of multimedia filename to:
"filename + '.' + extension", then we set cold file only its
filename matches the format.
So after this change, it will reduce the probability we set the
wrong cold file, also it helps a little for fs_mark's performance
on f2fs.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch adds missed trace file in maintainer-ship of f2fs,
so it completes the description of files maintained in f2fs,
and also it allows people to find correct mailing list by using
get_maintainer.pl when only patching the trace file of f2fs.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This makes the function check_dnode have a return type of bool
due to this particular function only ever returning either one
or zero as its return value and changes the name of the function
to is_alive in order to better explain this function's intended
work of checking if a dnode is still in use by the filesystem.
Signed-off-by: Nicholas Krause <xerofoify@gmail.com>
[Jaegeuk Kim: change the return value check for the renamed function]
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Because of the extent shrinker or other -ENOMEM scenarios, it cannot guarantee
that the largest extent would be cached in the tree all the time.
Instead of relying on extent_tree, we can simply check the cached one in extent
tree accordingly.
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
We don't need to handle the duplicate extent information.
The integrated rule is:
- update on-disk extent with largest one tracked by in-memory extent_cache
- destroy extent_tree for the truncation case
- drop per-inode extent_cache by shrinker
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
