Abdel Alkuor [Tue, 5 Mar 2024 04:20:29 +0000 (23:20 -0500)]
leds: Add NCP5623 multi-led driver
NCP5623 is DC-DC multi-LEDs driver which has three PWMs that can be
programmed up to 32 steps giving 32768 colors hue.
NCP5623 driver supports gradual dimming upward/downward with programmable
delays. Also, the driver supports driving a single LED or multi-LED
like RGB.
Randy Dunlap [Thu, 29 Feb 2024 07:19:30 +0000 (23:19 -0800)]
leds: leds-mlxcpld: Fix struct mlxcpld_led_priv member name
Change "cled" to "cdev" to quieten kernel-doc warnings:
leds-mlxcpld.c:86: warning: Function parameter or struct member 'cdev' not described in 'mlxcpld_led_priv'
leds-mlxcpld.c:86: warning: Excess struct member 'cled' description in 'mlxcpld_led_priv'
Arnd Bergmann [Wed, 28 Feb 2024 09:38:26 +0000 (10:38 +0100)]
leds: Fix ifdef check for gpio_led_register_device()
gpio_led_register_device() is built whenever CONFIG_LEDS_GPIO_REGISTER is
enabled, and this may be used even when CONFIG_NEW_LEDS is turned off.
However, the stub declaration in the header is provided for all configs
without CONFIG_NEW_LEDS, resulting in a build failure:
drivers/leds/leds-gpio-register.c:24:1: error: redefinition of 'gpio_led_register_device'
24 | gpio_led_register_device(int id, const struct gpio_led_platform_data *pdata)
| ^
include/linux/leds.h:646:39: note: previous definition is here
Change the #ifdef check to match the definition.
Note: this apparently took years of randconfig builds to hit, since
a number of other drivers just 'select NEW_LEDS' anyway.
Krzysztof Kozlowski [Mon, 26 Feb 2024 07:37:13 +0000 (08:37 +0100)]
dt-bindings: leds: qcom-lpg: Narrow nvmem for other variants
Only few variants support PPG and store LPG per-channel data and pattern
in SDAM modules. Disallow the nvmem for other variants to make the
binding precise.
Krzysztof Kozlowski [Mon, 26 Feb 2024 07:37:12 +0000 (08:37 +0100)]
dt-bindings: leds: qcom-lpg: Drop redundant qcom,pm8550-pwm in if:then:
"qcom,pm8550-pwm" is compatible with "qcom,pm8350c-pwm" (latter used as
fallback), thus it is enough for the "if:then:" clause to check for the
presence of the fallback "qcom,pm8350c-pwm".
Ondrej Jirman [Sat, 17 Feb 2024 19:11:30 +0000 (20:11 +0100)]
leds: sgm3140: Add missing timer cleanup and flash gpio control
Enabling strobe and then setting brightness to 0 causes the driver to enter
invalid state after strobe end timer fires. We should cancel strobe mode
resources when changing brightness (aka torch mode).
The ExpressWire library introduced in commit 25ae5f5f4168 ("leds:
Introduce ExpressWire library") does not depend on NEW_LEDS, but without
this revert it would never get compiled if NEW_LEDS is not enabled.
Revert this commit to allow the library to be compiled.
Arnd Bergmann [Mon, 12 Feb 2024 11:15:02 +0000 (12:15 +0100)]
leds: qcom-lpg: Add QCOM_PBS dependency
The lpg driver fails to link now when the pbs driver is in a loadable module:
x86_64-linux-ld: drivers/leds/rgb/leds-qcom-lpg.o: in function `lpg_brightness_set':
leds-qcom-lpg.c:(.text+0xe7f): undefined reference to `qcom_pbs_trigger_event'
x86_64-linux-ld: drivers/leds/rgb/leds-qcom-lpg.o: in function `lpg_probe':
leds-qcom-lpg.c:(.text+0x16a5): undefined reference to `get_pbs_client_device'
Add a dependency to avoid the broken configuration. Apparently there is still
a use for lpg with pbs disabled entirely for certain chips, so allow both
but not LEDS_QCOM_LPG=y with QCOM_PBS=m.
Fixes: 214110175679 ("leds: rgb: leds-qcom-lpg: Add support for PPG through single SDAM") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Jean Delvare <jdelvare@suse.de> Link: https://lore.kernel.org/r/20240212111526.829122-1-arnd@kernel.org Signed-off-by: Lee Jones <lee@kernel.org>
Aren Moynihan [Tue, 6 Feb 2024 18:13:17 +0000 (13:13 -0500)]
leds: rgb: leds-group-multicolor: Allow LEDs to stay on in suspend
If none of the managed LEDs enable LED_CORE_SUSPENDRESUME, then we
shouldn't need to set it here. This makes it possible to use multicolor
groups with GPIO LEDs that enable retain-state-suspended in the device
tree.
Commit d5e01266e7f5 ("leds: trigger: netdev: add additional specific link
speed mode") in the various changes, reworked the way to set the LINKUP
mode in commit cee4bd16c319 ("leds: trigger: netdev: Recheck
NETDEV_LED_MODE_LINKUP on dev rename") and moved it to a generic function.
This changed the logic where, in the previous implementation the dev
from the trigger event was used to check if the carrier was ok, but in
the new implementation with the generic function, the dev in
trigger_data is used instead.
This is problematic and cause a possible kernel panic due to the fact
that the dev in the trigger_data still reference the old one as the
new one (passed from the trigger event) still has to be hold and saved
in the trigger_data struct (done in the NETDEV_REGISTER case).
On calling of get_device_state(), an invalid net_dev is used and this
cause a kernel panic.
To handle this correctly, move the call to get_device_state() after the
new net_dev is correctly set in trigger_data (in the NETDEV_REGISTER
case) and correctly parse the new dev.
Fixes: d5e01266e7f5 ("leds: trigger: netdev: add additional specific link speed mode") Cc: stable@vger.kernel.org Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Link: https://lore.kernel.org/r/20240203235413.1146-1-ansuelsmth@gmail.com Signed-off-by: Lee Jones <lee@kernel.org>
Marijn Suijten [Sun, 4 Feb 2024 17:24:20 +0000 (18:24 +0100)]
leds: qcom-lpg: Add PM660L configuration and compatible
Inherit PM660L PMIC LPG/triled block configuration from downstream
drivers and DT sources, consisting of a triled block with automatic
trickle charge control and source selection, three colored led channels
belonging to the synchronized triled block and one loose PWM channel.
Stefan Kalscheuer [Sun, 4 Feb 2024 15:07:26 +0000 (16:07 +0100)]
leds: spi-byte: Use devm_led_classdev_register_ext()
Use extended classdev registration to generate generic device names from
color and function enums instead of reading only the label from the
device tree.
Heiner Kallweit [Wed, 31 Jan 2024 14:30:53 +0000 (15:30 +0100)]
leds: trigger: Stop exporting trigger_list
Commit 682e98564ffb ("leds: trigger: panic: Simplify
led_trigger_set_panic") removed the last external user of variable
trigger_list. So stop exporting it. If in future a need should arise
again to access this variable, I think we better add some accessor
instead of exporting the variable directly.
Arnd Bergmann [Tue, 9 Jan 2024 09:06:40 +0000 (10:06 +0100)]
leds: Make flash and multicolor dependencies unconditional
Along the same lines as making devm_led_classdev_register() declared
extern unconditional, do the same thing for the two sub-classes
that have similar stubs.
The users of these interfaces go to great lengths to allow building
with both the generic leds API and the extended version, but realistically
there is not much use in this, so just simplify it to always rely
on it and remove the confusing fallback logic.
Arnd Bergmann [Tue, 9 Jan 2024 09:06:39 +0000 (10:06 +0100)]
leds: Remove led_init_default_state_get() and devm_led_classdev_register_ext() stubs
These two functions have stub implementations that are called when
NEW_LEDS and/or LEDS_CLASS are disabled, theorerically allowing drivers
to optionally use the LED subsystem.
However, this has never really worked because a built-in driver is
unable to link against these functions if the LED class is in a loadable
module. Heiner ran into this problem with a driver that newly gained
a LEDS_CLASS dependency and suggested using an IS_REACHABLE() check.
This is the reverse approach, removing the stub entirely to acknowledge
that it is pointless in its current form, and that not having it avoids
misleading developers into thinking that they can rely on it.
This survived around 1000 randconfig builds to validate that any callers
of the interface already have the correct Kconfig dependency already,
with the exception of the one that Heiner just added.
I don't see why we iterate over all triggers to find the panic trigger.
We *are* the panic trigger. Therefore we also know that the panic
trigger doesn't have an activate() hook. So we can simplify the code
significantly.
Heiner Kallweit [Mon, 15 Jan 2024 21:46:23 +0000 (22:46 +0100)]
leds: triggers: default-on: Add module alias for module auto-loading
A bigger number of board device tree files, plus few drivers, set
default-on as default trigger for LED's. Therefore add an alias for
module auto-loading.
Heiner Kallweit [Sat, 13 Jan 2024 16:00:11 +0000 (17:00 +0100)]
leds: trigger: audio: Set module alias for module auto-loading
This a follow-up to 5edf7f11313d ("leds: trigger: Load trigger modules
on-demand if used as default trigger") and sets an alias for the audio
triggers.
Florian Eckert [Wed, 10 Jan 2024 13:34:10 +0000 (14:34 +0100)]
Documentation: leds: Update led-trigger-tty ABI description
The 'led-trigger-tty' uses the same naming in the ABI documentation as
the 'led-trigger-netdev'. Which leads to the following warning when
building the documentation.
Warning: /sys/class/leds/<led>/rx is defined 2 times:
Documentation/ABI/testing/sysfs-class-led-trigger-tty:7
Documentation/ABI/testing/sysfs-class-led-trigger-netdev:49
Warning: /sys/class/leds/<led>/tx is defined 2 times:
Documentation/ABI/testing/sysfs-class-led-trigger-tty:15
Documentation/ABI/testing/sysfs-class-led-trigger-netdev:34
Renaming the 'What' path by prefixing it with 'tty_' solves this problem.
Christian Marangi [Thu, 11 Jan 2024 16:04:54 +0000 (17:04 +0100)]
leds: trigger: netdev: Display only supported link speed attribute
With the addition of more link speed mode to the netdev trigger, it was
pointed out that there may be a problem with bloating the attribute list
with modes that won't ever be supported by the trigger as the attached
device name doesn't support them.
To clear and address this problem, change the logic where these
additional trigger modes are listed.
Since the netdev trigger REQUIRE a device name to be set, attach to the
device name change function additional logic to parse the supported link
speed modes using ethtool APIs and show only the supported link speed
modes attribute.
Link speed attribute are refreshed on device_name set and on
NETDEV_CHANGE events.
This only apply to the link speed modes and every other mode is still
provided by default.
Heiner Kallweit [Fri, 22 Dec 2023 21:32:28 +0000 (22:32 +0100)]
leds: class: If no default trigger is given, make hw_control trigger the default trigger
If a hw_control_trigger is defined, it's usually desirable to make it
the default trigger. Therefore make it the default trigger, except
the driver explicitly set a default trigger.
Heiner Kallweit [Thu, 21 Dec 2023 22:19:17 +0000 (23:19 +0100)]
leds: trigger: Load trigger modules on-demand if used as default trigger
Even if a trigger is set as default trigger for a LED device,
the respective trigger module (if built as module) isn't automatically
loaded by the kernel if the LED device is registered. I think we can
do better. Try to load the module asynchronously by alias
ledtrig:<trigger name>. This requires that such an alias is added to
relevant triggers.
Anjelique Melendez [Thu, 21 Dec 2023 18:58:36 +0000 (10:58 -0800)]
leds: rgb: leds-qcom-lpg: Include support for PPG with dedicated LUT SDAM
On PMICs such as PM8350C, the pattern lookup table (LUT) is stored in a
separate SDAM from the one where the lpg per-channel data is stored.
Add support for PPG with a dedicated LUT SDAM while maintaining backward
compatibility for those targets that use only a single SDAM.
Co-developed-by: Guru Das Srinagesh <quic_gurus@quicinc.com> Signed-off-by: Guru Das Srinagesh <quic_gurus@quicinc.com> Signed-off-by: Anjelique Melendez <quic_amelende@quicinc.com> Link: https://lore.kernel.org/r/20231221185838.28440-7-quic_amelende@quicinc.com Signed-off-by: Lee Jones <lee@kernel.org>
Anjelique Melendez [Thu, 21 Dec 2023 18:58:34 +0000 (10:58 -0800)]
leds: rgb: leds-qcom-lpg: Add support for PPG through single SDAM
In some PMICs like pmi632, the pattern look up table (LUT) and LPG
configuration is stored in a single SDAM module instead of LUT
peripheral. This feature is called PPG. PPG uses Qualcomm Programmable
Boot Sequencer (PBS) inorder to trigger pattern sequences for PMICs.
Heiner Kallweit [Thu, 21 Dec 2023 17:31:48 +0000 (18:31 +0100)]
leds: trigger: netdev: Skip setting baseline state in activate if hw-controlled
The current codes uses the sw_control path in set_baseline_state() when
called from netdev_trig_activate() even if we're hw-controlled. This
may result in errors when led_set_brightness() is called because we may
not have set_brightness led ops (if hw doesn't support setting a "LED"
to ON). In addition this path may schedule trigger_data->work which
doesn't make sense when being hw-controlled.
Therefore set trigger_data->hw_control = true before calling
set_device_name() from netdev_trig_activate(). In this call chain we
have to prevent set_baseline_state() from being called, because this
would call hw_control_set(). Use led_cdev->trigger_data == NULL as
indicator for being called from netdev_trig_activate().
Arnd Bergmann [Tue, 13 Feb 2024 16:55:50 +0000 (17:55 +0100)]
leds: ktd2692: Add GPIOLIB dependency
The expresswire module requires gpiolib, so anything selecting it
also needs this dependency:
WARNING: unmet direct dependencies detected for LEDS_EXPRESSWIRE
Depends on [n]: NEW_LEDS [=y] && GPIOLIB [=n]
Selected by [y]:
- LEDS_KTD2692 [=y] && NEW_LEDS [=y] && LEDS_CLASS_FLASH [=y] && OF [=y]
Fixes: e59a15af7aa6 ("leds: ktd2692: Convert to use ExpressWire library") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Link: https://lore.kernel.org/r/20240213165602.2230970-1-arnd@kernel.org Signed-off-by: Lee Jones <lee@kernel.org>
Duje Mihanović [Sat, 10 Feb 2024 16:16:17 +0000 (17:16 +0100)]
backlight: ktd2801: Make timing struct static
The struct containing the KTD2801 timing can be made static as it's not
referenced outside the KTD2801 driver. Do this to prevent sparse
complaints.
Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202402100625.M0RkJhMh-lkp@intel.com/ Signed-off-by: Duje Mihanović <duje.mihanovic@skole.hr> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Link: https://lore.kernel.org/r/20240210-ktd2801-static-v1-1-90ad2e2e8483@skole.hr Signed-off-by: Lee Jones <lee@kernel.org>
Duje Mihanović [Thu, 25 Jan 2024 15:30:56 +0000 (16:30 +0100)]
backlight: Add Kinetic KTD2801 Backlight support
KTD2801 is a LED backlight driver IC found in samsung,coreprimevelte.
The brightness can be set using PWM or the ExpressWire protocol. Add
support for the KTD2801.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Duje Mihanović <duje.mihanovic@skole.hr> Link: https://lore.kernel.org/r/20240125-ktd2801-v5-4-e22da232a825@skole.hr Signed-off-by: Lee Jones <lee@kernel.org>
KTD2801 is a LED backlight driver IC found in samsung,coreprimevelte.
The brightness can be set using PWM or the ExpressWire protocol. Add
a DT binding for the KTD2801.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Duje Mihanović <duje.mihanovic@skole.hr> Link: https://lore.kernel.org/r/20240125-ktd2801-v5-3-e22da232a825@skole.hr Signed-off-by: Lee Jones <lee@kernel.org>
Duje Mihanović [Thu, 25 Jan 2024 15:30:54 +0000 (16:30 +0100)]
leds: ktd2692: Convert to use ExpressWire library
The KTD2692 uses the ExpressWire protocol implemented in the newly
introduced ExpressWire library. Convert the driver to use the library.
Suggested-by: Daniel Thompson <daniel.thompson@linaro.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Duje Mihanović <duje.mihanovic@skole.hr> Link: https://lore.kernel.org/r/20240125-ktd2801-v5-2-e22da232a825@skole.hr Signed-off-by: Lee Jones <lee@kernel.org>
Duje Mihanović [Thu, 25 Jan 2024 15:30:53 +0000 (16:30 +0100)]
leds: Introduce ExpressWire library
The ExpressWire protocol is shared between at least KTD2692 and KTD2801
with slight differences such as timings and the former not having a
defined set of pulses for enabling the protocol (possibly because it
does not support PWM unlike KTD2801). Despite these differences the
ExpressWire handling code can be shared between the two, so in
preparation for adding KTD2801 support introduce a library implementing
this protocol.
Suggested-by: Daniel Thompson <daniel.thompson@linaro.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Duje Mihanović <duje.mihanovic@skole.hr> Link: https://lore.kernel.org/r/20240125-ktd2801-v5-1-e22da232a825@skole.hr Signed-off-by: Lee Jones <lee@kernel.org>
Linus Torvalds [Sun, 21 Jan 2024 22:01:12 +0000 (14:01 -0800)]
Merge tag 'bcachefs-2024-01-21' of https://evilpiepirate.org/git/bcachefs
Pull more bcachefs updates from Kent Overstreet:
"Some fixes, Some refactoring, some minor features:
- Assorted prep work for disk space accounting rewrite
- BTREE_TRIGGER_ATOMIC: after combining our trigger callbacks, this
makes our trigger context more explicit
- A few fixes to avoid excessive transaction restarts on
multithreaded workloads: fstests (in addition to ktest tests) are
now checking slowpath counters, and that's shaking out a few bugs
- Assorted tracepoint improvements
- Starting to break up bcachefs_format.h and move on disk types so
they're with the code they belong to; this will make room to start
documenting the on disk format better.
- A few minor fixes"
* tag 'bcachefs-2024-01-21' of https://evilpiepirate.org/git/bcachefs: (46 commits)
bcachefs: Improve inode_to_text()
bcachefs: logged_ops_format.h
bcachefs: reflink_format.h
bcachefs; extents_format.h
bcachefs: ec_format.h
bcachefs: subvolume_format.h
bcachefs: snapshot_format.h
bcachefs: alloc_background_format.h
bcachefs: xattr_format.h
bcachefs: dirent_format.h
bcachefs: inode_format.h
bcachefs; quota_format.h
bcachefs: sb-counters_format.h
bcachefs: counters.c -> sb-counters.c
bcachefs: comment bch_subvolume
bcachefs: bch_snapshot::btime
bcachefs: add missing __GFP_NOWARN
bcachefs: opts->compression can now also be applied in the background
bcachefs: Prep work for variable size btree node buffers
bcachefs: grab s_umount only if snapshotting
...
Linus Torvalds [Sun, 21 Jan 2024 19:14:40 +0000 (11:14 -0800)]
Merge tag 'timers-core-2024-01-21' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer updates from Thomas Gleixner:
"Updates for time and clocksources:
- A fix for the idle and iowait time accounting vs CPU hotplug.
The time is reset on CPU hotplug which makes the accumulated
systemwide time jump backwards.
- Assorted fixes and improvements for clocksource/event drivers"
* tag 'timers-core-2024-01-21' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
tick-sched: Fix idle and iowait sleeptime accounting vs CPU hotplug
clocksource/drivers/ep93xx: Fix error handling during probe
clocksource/drivers/cadence-ttc: Fix some kernel-doc warnings
clocksource/drivers/timer-ti-dm: Fix make W=n kerneldoc warnings
clocksource/timer-riscv: Add riscv_clock_shutdown callback
dt-bindings: timer: Add StarFive JH8100 clint
dt-bindings: timer: thead,c900-aclint-mtimer: separate mtime and mtimecmp regs
Kent Overstreet [Tue, 16 Jan 2024 21:20:21 +0000 (16:20 -0500)]
bcachefs: opts->compression can now also be applied in the background
The "apply this compression method in the background" paths now use the
compression option if background_compression is not set; this means that
setting or changing the compression option will cause existing data to
be compressed accordingly in the background.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 16 Jan 2024 18:29:59 +0000 (13:29 -0500)]
bcachefs: Prep work for variable size btree node buffers
bcachefs btree nodes are big - typically 256k - and btree roots are
pinned in memory. As we're now up to 18 btrees, we now have significant
memory overhead in mostly empty btree roots.
And in the future we're going to start enforcing that certain btree node
boundaries exist, to solve lock contention issues - analagous to XFS's
AGIs.
Thus, we need to start allocating smaller btree node buffers when we
can. This patch changes code that refers to the filesystem constant
c->opts.btree_node_size to refer to the btree node buffer size -
btree_buf_bytes() - where appropriate.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
In __bch2_ioctl_subvolume_create(), we grab s_umount unconditionally
and unlock it at the end of the function. There is a comment
"why do we need this lock?" about the lock coming from
commit 42d237320e98 ("bcachefs: Snapshot creation, deletion")
The reason is that __bch2_ioctl_subvolume_create() calls
sync_inodes_sb() which enforce locked s_umount to writeback all dirty
nodes before doing snapshot works.
Fix it by read locking s_umount for snapshotting only and unlocking
s_umount after sync_inodes_sb().
Signed-off-by: Su Yue <glass.su@suse.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Colin Ian King [Tue, 16 Jan 2024 11:07:23 +0000 (11:07 +0000)]
bcachefs: remove redundant variable tmp
The variable tmp is being assigned a value but it isn't being
read afterwards. The assignment is redundant and so tmp can be
removed.
Cleans up clang scan build warning:
warning: Although the value stored to 'ret' is used in the enclosing
expression, the value is never actually read from 'ret'
[deadcode.DeadStores]
Signed-off-by: Colin Ian King <colin.i.king@gmail.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 16 Jan 2024 01:37:23 +0000 (20:37 -0500)]
bcachefs: Fix excess transaction restarts in __bchfs_fallocate()
drop_locks_do() should not be used in a fastpath without first trying
the do in nonblocking mode - the unlock and relock will cause excessive
transaction restarts and potentially livelocking with other threads that
are contending for the same locks.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Jan 2024 22:59:51 +0000 (17:59 -0500)]
bcachefs: Better journal tracepoints
Factor out bch2_journal_bufs_to_text(), and use it in the
journal_entry_full() tracepoint; when we can't get a journal reservation
we need to know the outstanding journal entry sizes to know if the
problem is due to excessive flushing.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Jan 2024 22:56:22 +0000 (17:56 -0500)]
bcachefs: Avoid flushing the journal in the discard path
When issuing discards, we may need to flush the journal if there's too
many buckets that can't be discarded until a journal flush.
But the heuristic was bad; we should be comparing the number of buckets
that need to flushes against the number of free buckets, not the number
of buckets we saw.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Jan 2024 19:15:26 +0000 (14:15 -0500)]
bcachefs: bch2_kthread_io_clock_wait() no longer sleeps until full amount
Drop t he loop in bch2_kthread_io_clock_wait(): this allows the code
that uses it to be woken up for other reasons, and fixes a bug where
rebalance wouldn't wake up when a scan was requested.
This raises the possibility of spurious wakeups, but callers should
always be able to handle that reasonably well.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Thu, 11 Jan 2024 04:47:04 +0000 (23:47 -0500)]
bcachefs: Reduce would_deadlock restarts
We don't have to take locks in any particular ordering - we'll make
forward progress just fine - but if we try to stick to an ordering, it
can help to avoid excessive would_deadlock transaction restarts.
This tweaks the reflink path to take extents btree locks in the right
order.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
projects / users / jedix / linux-maple.git / log