Three more variants of the Bananapi M2 Plus have been introduced. One
with the H5 instead of the H3, another with the H2+ instead, and the
last with the H3 but with WiFi and eMMC removed.
All these variants use the same board. This patch splits out the
non-SoC-specific parts of the device tree, so that they can be shared
among all the variants. The original Bananapi M2 Plus has been renamed
to Bananapi M2 Plus H3.
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
The external RTL8211E RGMII Ethernet PHY is configured via external
resistors to use the address 0x1. The 0x0 address is a broadcast address
for this family of PHYs, and should not be used explicitly.
Fixes: 8c7ba536e7 ("ARM: sun8i: bananapi-m2-plus: Enable dwmac-sun8i")
Fixes: 4904337fe3 ("ARM: dts: sunxi: Restore EMAC changes (boards)")
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
The size of the register should be the size of the whole memory block,
not just the registers, that are needed.
Signed-off-by: Philipp Rossak <embed3d@gmail.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
The H5 has a Mali-450 GPU with 4 Pixel Processor cores.
Interestingly, while the datasheet lists an interrupt line for the GPU's
PMU, the hardware block itself doesn't seem to have it. Reads from the
PMU address range all return zero, and writes are ignored.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Orangepi Zero Plus 2 is an open-source single-board computer, available in two
Allwinner SOC variants, H3 and H5. We add support for H3 variant here, as the
H5 is already supported by sun50i-h5-orangepi-zero-plus2.dts.
H3 Orangepi Zero Plus 2 has:
- Quad-core Cortex-A7
- 512MB DDR3
- microSD slot and 8GB eMMC
- Debug TTL UART
- HDMI
- Wifi + BT
- OTG + power supply
Signed-off-by: Diego Rondini <diego.rondini@kynetics.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The H5 SoC have a SID controller that looks like the one in A64, the
cells are in the same offset but doesn't contain the same data (thermal
sensor calibration for example).
Add a binding for it.
Signed-off-by: Emmanuel Vadot <manu@freebsd.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
There is currently a warning when building the Kryo cpufreq driver into
the kernel image:
WARNING: vmlinux.o(.text+0x8aa424): Section mismatch in reference from
the function qcom_cpufreq_kryo_probe() to the function
.init.text:qcom_cpufreq_kryo_get_msm_id()
The function qcom_cpufreq_kryo_probe() references
the function __init qcom_cpufreq_kryo_get_msm_id().
This is often because qcom_cpufreq_kryo_probe lacks a __init
annotation or the annotation of qcom_cpufreq_kryo_get_msm_id is wrong.
Remove the '__init' annotation from qcom_cpufreq_kryo_get_msm_id
so that there is no more mismatch warning.
Additionally, Nick noticed that the remove function was marked as
'__init' when it should really be marked as '__exit'.
Fixes: 46e2856b8e (cpufreq: Add Kryo CPU scaling driver)
Fixes: 5ad7346b4a (cpufreq: kryo: Add module remove and exit)
Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Cc: 4.18+ <stable@vger.kernel.org> # 4.18+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Store fbhelper and afb struct directly in vbox_private and use
drm_fb_helper_fbdev_setup to replace vbox_fbdev_init, note we cannot use
drm_fb_helper_fbdev_teardown since we use a private framebuffer for the
fbdev.
And replace vbox_driver_lastclose with drm_fb_helper_lastclose.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vbox_mode_valid always returns MODE_OK, which is also the default if no
mode_valid callback is defined.
vbox_best_single_encoder chains to drm_encoder_find, the drm-core will
call drm_encoder_find itself if there is no best_encoder call back.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Extend our planes atomic_check callbacks to be more thorough by calling
the drm_atomic_helper_check_plane_state helper.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace vbox_crtc_commit and vbox_crtc_disable with
vbox_crtc_atomic_[en|dis]able which are the preferred callbacks for
these for atomic drivers.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Atomic modesetting does not use the traditional dpms call backs, instead
we should check crtc_state->active.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Restore page-flip support now that the atomic conversion is complete.
Since the mode parameter to vbox_crtc_set_base_and_mode() now never
is NULL call drm_atomic_crtc_needs_modeset() to check if we need to
check for input-mapping changes, to avoid doing unnecesarry work on
a flip. And hookup the drm_atomic_helper_page_flip helper to implement
the page_flip callback.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now that the state objects are wired up, we can:
1) Move to the final atomic handlers
2) Wire up atomic set_config helper
3) Switch to drm_mode_config_helper_suspend/resume for suspend/resume
4) Enable atomic modesetting ioctl
This is all done in one commit because doing this piecemeal leads to
an intermediate state which triggers WARN_ONs in the atomic code because
e.g. plane->fb is still being set.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Once we are fully atomic plane->fb will always be NULL and we also
should not access things like crtc->enabled and crt->[hw]mode.
Now that we've wired up the state object handlers, we always have a
plane_state and crtc_state so change the code referencing plane->fb and
crtc->* to use the data from the plane_state and crt_state instead.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Wire up state object handlers for the crtc-s and the planes, call
drm_mode_config_reset() after creating all the crtc-s and encoders and
remove the legacy drm_helper_disable_unused_functions() call.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use drm_plane_helpers for the primary plane and replace our custom
mode_set callback with drm_helper_crtc_mode_set.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In preparation for atomic conversion, let's use the transitional atomic
helpers drm_plane_helper_update/disable.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When setting a mode we not only pass the mode to the hypervisor,
but also information on how to map / translate input coordinates
for the emulated USB tablet. This input-mapping may change when
the mode on *another* crtc changes.
This means that sometimes we must do a modeset on other crtc-s then
the one being changed to update the input-mapping. Including crtc-s
which may be disabled inside the guest (shown as a black window
on the host unless closed by the user).
With atomic modesetting the mode-info of disabled crtcs gets zeroed
yet we need it when updating the input-map to avoid resizing the
window as a side effect of a mode_set on another crtc.
This commit adds caching of the mode info into out private vbox_crtc
struct so that we always have the info at hand when we need it.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm_mode_page_flip_ioctl() cannot deal with the in between phase of
the transitioning to atomic modeset support. Once we start using
drm_helper_crtc_mode_set(), we start setting plane->state on the primary
plane. But we are not fully atomic yet so then set both plane-state->fb
and plane->fb.
If both plane-state->fb and plane->fb are set drm_mode_page_flip_ioctl()
gets confused and stops calling drm_framebuffer_get() on the new fb while
still calling drm_framebuffer_put() on the old fb.
The current page_flip implementation expects drm_mode_page_flip_ioctl()
to take care of both and once we switch to drm_atomic_helper_page_flip()
that will expect neither to be done, taking care of both itself.
So for the transition we need to remove page_flip support and then after
the transition is complete and we set DRIVER_ATOMIC in our driver_features,
we can start using drm_atomic_helper_page_flip().
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All the encoder_helper_funcs are optional, and even setting the
drm_encoder_helper_funcs vtable itself is optional and may be left out
when not using any of the helper funcs, so lets drop all of this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This cleanups 2 things:
1) The first time we loop over the crtc-s, to compare framebuffers, fb1 may
get set to NULL by the fb1 = CRTC_FB(crtci); statement and then we call
to_vbox_framebuffer() on it. The result of this call is only used for
an address comparison, so we don't end up dereferencing the bad pointer,
but still it is better to not do this.
2) Since we already figure out the first crtc with a fb in the first loop
and store that in fb1, there is no need to loop over the crtc-s again just
to find the first crtc with a fb again.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The correct way to handle errors returned by regualtor_get() and friends is
to propagate the error since that means that an regulator was specified,
but something went wrong when requesting it.
For handling optional regulators, e.g. when the device has an internal
vref, regulator_get_optional() should be used to avoid getting the dummy
regulator that the regulator core otherwise provides.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Until now the "channel" number is identical to how the channel is
identified inside the (FIFO) registers. In our case we have eight
channels and the hardware also has eight inputs.
However, there are two special inputs:
- channel 6 can select between the SAR_ADC_CH6 pad and the chip's
internal temperature sensor
- channel 7 can select between SAR_ADC_CH7 and VSS, VDD / 4, VDD / 2,
VDD * 3 / 4 and VDD.
When programming the registers to read for example the temperature
sensor we have to select FIFO channel 6, set the correct bit which muxes
channel 6 to the temperature sensor and then start the ADC measurement
for channel 6 as usual.
When we add support for the temperature sensor the driver has to know
about that it has to use FIFO channel 6 to measure using the chip's
internal temperature sensor. However, in that case the iio_chan_spec
channel will not be 6 because this is already used for the SAR_ADC_CH6
pad input. Thus we use iio_chan_spec's address field to store the FIFO
channel number for each channel.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
In the future we may support two different channel sets:
- one which includes the voltage pads and the temperature sensor output
(for Meson8, Meson8b and Meson8m2)
- one which only includes the voltage pads (GXBB, GXL, GXM and AXG)
Channel 7 has a special function on all of these platforms. However,
since we will have different channel array definitions we want our code
to always use whatever channels struct iio_dev uses.
No functional changes for now. This is the preparation for adding
temperature sensor support to this driver.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
When channels are registered, the hardware channel number is not the
actual iio channel number.
This is because the driver is probed with a certain number of accessible
channels. Some pins are routed and some not, depending on the description of
the board in the DT.
Because of that, channels 0,1,2,3 can correspond to hardware channels
2,3,4,5 for example.
In the buffered triggered case, we need to do the translation accordingly.
Fixed the channel number to stop reading the wrong channel.
Fixes: 0e589d5fb ("ARM: AT91: IIO: Add AT91 ADC driver.")
Cc: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
When doing simple conversions, the driver did not acknowledge the DRDY irq.
If this irq status is not acked, it will be left pending, and as soon as a
trigger is enabled, the irq handler will be called, it doesn't know why
this status has occurred because no channel is pending, and then it will go
int a irq loop and board will hang.
To avoid this situation, read the LCDR after a raw conversion is done.
Fixes: 0e589d5fb ("ARM: AT91: IIO: Add AT91 ADC driver.")
Cc: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Commit 053ffe3c8c ("iio: adc: meson-saradc: squash and share the
common adc platform data") put all the data which is needed at runtime
from struct meson_sar_adc_data to a new struct meson_sar_adc_param so
we can re-use the platform specific configuration without having to
duplicate everything just to change the name.
The only place where struct meson_sar_adc_data is now needed is the
_probe function which has to pass the name to the iio_dev. All other
functions only need access to struct meson_sar_adc_param. This means we
can simplify struct meson_sar_adc_priv.
The goal of this patch is to make the code a bit easier to read since
this removes one level of nesting. No functional changes intended.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This simplifies our _probe function by using of_device_get_match_data
instead of open-coding it. No functional changes.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This define is of no use because the driver is avoiding shifting bits
by itself but using FIELD_GET/FIELD_PREP (which are using bit masks)
instead. There is already a MESON_SAR_ADC_DELTA_10_TS_C_MASK bit mask so
MESON_SAR_ADC_DELTA_10_TS_C_SHIFT was redundant.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Add device tree support for ROHM BH1750 series ambient light sensors.
Signed-off-by: ryang <decatf@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
-----BEGIN PGP SIGNATURE-----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=QKIn
-----END PGP SIGNATURE-----
Merge tag 'dma-mapping-4.19-3' of git://git.infradead.org/users/hch/dma-mapping
Christoph writes:
"dma mapping fix for 4.19-rc6
fix a missing Kconfig symbol for commits introduced in 4.19-rc"
* tag 'dma-mapping-4.19-3' of git://git.infradead.org/users/hch/dma-mapping:
dma-mapping: add the missing ARCH_HAS_SYNC_DMA_FOR_CPU_ALL declaration
The count of errors is picked up from bits 52:38 of the machine check
bank status register. But this is the count of *corrected* errors. If an
uncorrected error is being logged, the h/w sets this field to 0. Which
means that when edac_mc_handle_error() is called, the EDAC core will
carefully add zero to the appropriate uncorrected error counts.
Signed-off-by: Tony Luck <tony.luck@intel.com>
[ Massage commit message. ]
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: stable@vger.kernel.org
Cc: Aristeu Rozanski <aris@redhat.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Link: http://lkml.kernel.org/r/20180928213934.19890-1-tony.luck@intel.com
Trivial fix to spelling mistake in function name flexrm_flip_header_toogle,
rename it to flexrm_flip_header_toggle.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
Add support for the QCS404 APPS Global block with IPC register at offset
8.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Vinod Koul <vkoul@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
In preparation to remove the node name pointer from struct device_node,
convert printf users to use the %pOFn format specifier.
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
The iomap page fault mechanism currently dirties the associated page
after the full block range of the page has been allocated. This
leaves the page susceptible to delayed allocations without ever
being set dirty on sub-page block sized filesystems.
For example, consider a page fault on a page with one preexisting
real (non-delalloc) block allocated in the middle of the page. The
first iomap_apply() iteration performs delayed allocation on the
range up to the preexisting block, the next iteration finds the
preexisting block, and the last iteration attempts to perform
delayed allocation on the range after the prexisting block to the
end of the page. If the first allocation succeeds and the final
allocation fails with -ENOSPC, iomap_apply() returns the error and
iomap_page_mkwrite() fails to dirty the page having already
performed partial delayed allocation. This eventually results in the
page being invalidated without ever converting the delayed
allocation to real blocks.
This problem is reliably reproduced by generic/083 on XFS on ppc64
systems (64k page size, 4k block size). It results in leaked
delalloc blocks on inode reclaim, which triggers an assert failure
in xfs_fs_destroy_inode() and filesystem accounting inconsistency.
Move the set_page_dirty() call from iomap_page_mkwrite() to the
actor callback, similar to how the buffer head implementation works.
The actor callback is called iff ->iomap_begin() returns success, so
ensures the page is dirtied as soon as possible after an allocation.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
One of the first steps of log recovery is to check for the special
case of a zeroed log. If the first cycle in the log is zero or the
tail portion of the log is zeroed, the head is set to the first
instance of cycle 0. xlog_find_zeroed() includes a sanity check that
enforces that the first cycle in the log must be 1 if the last cycle
is 0. While this is true in most cases, the check is not totally
valid because it doesn't consider the case where the filesystem
crashed after a partial/out of order log buffer completion that
wraps around the end of the physical log.
For example, consider a filesystem that has completed most of the
first cycle of the log, reaches the end of the physical log and
splits the next single log buffer write into two in order to wrap
around the end of the log. If these I/Os are reordered, the second
(wrapped) I/O completes and the first happens to fail, the log is
left in a state where the last cycle of the log is 0 and the first
cycle is 2. This causes the xlog_find_zeroed() sanity check to fail
and prevents the filesystem from mounting. This situation has been
reproduced on particular systems via repeated runs of generic/475.
This is an expected state that log recovery already knows how to
deal with, however. Since the log is still partially zeroed, the
head is detected correctly and points to a valid tail. The
subsequent stale block detection clears blocks beyond the head up to
the tail (within a maximum range), with the express purpose of
clearing such out of order writes. As expected, this removes the out
of order cycle 2 blocks at the physical start of the log.
In other words, the only thing that prevents a clean mount and
recovery of the filesystem in this scenario is the specific (last ==
0 && first != 1) sanity check in xlog_find_zeroed(). Since the log
head/tail are now independently validated via cycle, log record and
CRC checks, this highly specific first cycle check is of dubious
value. Remove it and rely on the higher level validation to
determine whether log content is sane and recoverable.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Verify the inode di_forkoff, lifted from xfs_repair's
process_check_inode_forkoff().
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
The iomap direct I/O code issues a single ->end_io call for the whole
I/O request, and if some of the extents cowered needed a COW operation
it will call xfs_reflink_end_cow over the whole range.
When we do AIO writes we drop the iolock after doing the initial setup,
but before the I/O completion. Between dropping the lock and completing
the I/O we can have a racing buffered write create new delalloc COW fork
extents in the region covered by the outstanding direct I/O write, and
thus see delalloc COW fork extents in xfs_reflink_end_cow. As
concurrent writes are fundamentally racy and no guarantees are given we
can simply skip those.
This can be easily reproduced with xfstests generic/208 in always_cow
mode.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
xchk_inode_flags2() currently treats any di_flags2 values that the
running kernel doesn't recognize as corruption, and calls
xchk_ino_set_corrupt() if they are set. However, it's entirely possible
that these flags were set in some newer kernel and are quite valid,
but ignored in this kernel.
(Validators don't care one bit about unknown di_flags2.)
Call xchk_ino_set_warning instead, because this may or may not actually
indicate a problem.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Remove duplicated include xfs_alloc.h
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
This function is only used to punch out delayed allocations on I/O
failure, which means we need to have read the extents earlier.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
When xfs_reflink_allocate_cow() allocates a transaction, it drops
the ILOCK to perform the operation. This Introduces a race condition
where another thread modifying the file can perform the COW
allocation operation underneath us. This result in the retry loop
finding an allocated block and jumping straight to the conversion
code. It does not, however, cancel the transaction it holds and so
this gets leaked. This results in a lockdep warning:
================================================
WARNING: lock held when returning to user space!
4.18.5 #1 Not tainted
------------------------------------------------
worker/6123 is leaving the kernel with locks still held!
1 lock held by worker/6123:
#0: 000000009eab4f1b (sb_internal#2){.+.+}, at: xfs_trans_alloc+0x17c/0x220
And eventually the filesystem deadlocks because it runs out of log
space that is reserved by the leaked transaction and never gets
released.
The logic flow in xfs_reflink_allocate_cow() is a convoluted mess of
gotos - it's no surprise that it has bug where the flow through
several goto jumps then fails to clean up context from a non-obvious
logic path. CLean up the logic flow and make sure every path does
the right thing.
Reported-by: Alexander Y. Fomichev <git.user@gmail.com>
Tested-by: Alexander Y. Fomichev <git.user@gmail.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=200981
Signed-off-by: Dave Chinner <dchinner@redhat.com>
[hch: slight refactor]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
We've had a few reports of lockdep tripping over memory reclaim
context vs filesystem freeze "deadlocks". They all have looked
to be false positives on analysis, but it seems that they are
being tripped because we take freeze references before we run
a GFP_KERNEL allocation for the struct xfs_trans.
We can avoid this false positive vector just by re-ordering the
operations in xfs_trans_alloc(). That is. we need allocate the
structure before we take the freeze reference and enter the GFP_NOFS
allocation context that follows the xfs_trans around. This prevents
lockdep from seeing the GFP_KERNEL allocation inside the transaction
context, and that prevents it from triggering the freeze level vs
alloc context vs reclaim warnings.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
The xfs_buf_log_item structure has a reference counter with slightly
tricky semantics. In the common case, a buffer is logged and
committed in a transaction, committed to the on-disk log (added to
the AIL) and then finally written back and removed from the AIL. The
bli refcount covers two potentially overlapping timeframes:
1. the bli is held in an active transaction
2. the bli is pinned by the log
The caveat to this approach is that the reference counter does not
purely dictate the lifetime of the bli. IOW, when a dirty buffer is
physically logged and unpinned, the bli refcount may go to zero as
the log item is inserted into the AIL. Only once the buffer is
written back can the bli finally be freed.
The above semantics means that it is not enough for the various
refcount decrementing contexts to release the bli on decrement to
zero. xfs_trans_brelse(), transaction commit (->iop_unlock()) and
unpin (->iop_unpin()) must all drop the associated reference and
make additional checks to determine if the current context is
responsible for freeing the item.
For example, if a transaction holds but does not dirty a particular
bli, the commit may drop the refcount to zero. If the bli itself is
clean, it is also not AIL resident and must be freed at this time.
The same is true for xfs_trans_brelse(). If the transaction dirties
a bli and then aborts or an unpin results in an abort due to a log
I/O error, the last reference count holder is expected to explicitly
remove the item from the AIL and release it (since an abort means
filesystem shutdown and metadata writeback will never occur).
This leads to fairly complex checks being replicated in a few
different places. Since ->iop_unlock() and xfs_trans_brelse() are
nearly identical, refactor the logic into a common helper that
implements and documents the semantics in one place. This patch does
not change behavior.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>