The -modesetting ddx has a totally broken idea of how atomic works:
- doesn't disable old connectors, assuming they get auto-disable like
with the legacy setcrtc
- assumes ASYNC_FLIP is wired through for the atomic ioctl
- not a single call to TEST_ONLY
Iow the implementation is a 1:1 translation of legacy ioctls to
atomic, which is a) broken b) pointless.
We already have bugs in both i915 and amdgpu-DC where this prevents us
from enabling neat features.
If anyone ever cares about atomic in X we can easily add a new atomic
level (req->value == 2) for X to get back the shiny toys.
Since these broken versions of -modesetting have been shipping,
there's really no other way to get out of this bind.
v2:
- add an informational dmesg output (Rob, Ajax)
- reorder after the DRIVER_ATOMIC check to avoid useless noise (Ilia)
- allow req->value > 2 so that X can do another attempt at atomic in
the future
v3: Go with paranoid, insist that the X should be first (suggested by
Rob)
Cc: Ilia Mirkin <imirkin@alum.mit.edu>
References: https://gitlab.freedesktop.org/xorg/xserver/issues/629
References: https://gitlab.freedesktop.org/xorg/xserver/merge_requests/180
References: abbc0697d5fb ("drm/fb: revert the i915 Actually configure untiled displays from master") Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> (v1) Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> (v1) Cc: Michel Dänzer <michel@daenzer.net> Cc: Alex Deucher <alexdeucher@gmail.com> Cc: Adam Jackson <ajax@redhat.com> Acked-by: Adam Jackson <ajax@redhat.com> Cc: Sean Paul <sean@poorly.run> Cc: David Airlie <airlied@linux.ie> Cc: Rob Clark <robdclark@gmail.com> Acked-by: Rob Clark <robdclark@gmail.com> Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20190905185318.31363-1-daniel.vetter@ffwll.ch Cc: Yves-Alexis Perez <corsac@debian.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 018d26fcd12a ("cgroup, netclassid: periodically release file_lock
on classid") added a second cond_resched to write_classid indirectly by
update_classid_task. Remove the one in write_classid.
Signed-off-by: Jiri Slaby <jslaby@suse.cz> Cc: Dmitry Yakunin <zeil@yandex-team.ru> Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 08a5bdde3812 ("mac80211: consider QoS Null frames for STA_NULLFUNC_ACKED")
Fixed a bug where we failed to take into account a
nullfunc frame can be either non-QoS or QoS. It turns out
there is at least one more bug in
ieee80211_sta_tx_notify(), introduced in
commit 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing"),
where we forgot to check for the QoS variant and so
assumed the QoS nullfunc frame never went out
Fix this by adding a helper ieee80211_is_any_nullfunc()
which consolidates the check for non-QoS and QoS nullfunc
frames. Replace existing compound conditionals and add a
couple more missing checks for QoS variant.
Commit 1f27dbd8265d ("platform/x86: GPD pocket fan: Allow somewhat
lower/higher temperature limits") changed the module-param sanity check
to accept temperature limits between 20 and 90 degrees celcius.
But the error message printed when the module params are outside this
range was not updated. This commit updates the error message to match
the new min and max value for the temp-limits.
Reported-by: Pavel Machek <pavel@denx.de> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Pavel Machek <pavel@denx.de> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 3c6fd1f07ed0 ("ALSA: hda: Add driver blacklist") added a
new blacklist for the devices that are known to have empty codecs, and
one of the entries was ASUS ROG Zenith II (PCI SSID 1043:874f).
However, it turned out that the very same PCI SSID is used for the
previous model that does have the valid HD-audio codecs and the change
broke the sound on it.
Since the empty codec problem appear on the certain AMD platform (PCI
ID 1022:1487), this patch changes the blacklist matching to both PCI
ID and SSID using pci_match_id(). Also, the entry that was removed by
the previous fix for ASUS ROG Zenigh II is re-added.
Some BIOS erroneously specifies write-combining BAR for intel-lpss-pci
in MTRR. This will cause the system to hang during boot. If possible,
this bug could be corrected with a firmware update.
This patch use devm_ioremap_uc to overwrite/ignore the MTRR settings
by forcing the use of strongly uncachable pages for intel-lpss.
The BIOS bug is present on Dell XPS 13 7390 2-in-1:
On ARCTURUS and RENOIR, powerplay is not supported yet.
When plug in or unplug power jack, ACPI event will issue.
Then kernel NULL pointer BUG will be triggered.
Check for NULL pointers before calling.
Signed-off-by: Aaron Ma <aaron.ma@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When starting shutdown in sctp_sf_do_dupcook_a(), get the value for
SHUTDOWN Cumulative TSN Ack from the new association, which is
reconstructed from the cookie, instead of the old association, which
the peer doesn't have anymore.
Otherwise the SHUTDOWN is either ignored or replied to with an ABORT
by the peer because CTSN Ack doesn't match the peer's Initial TSN.
Fixes: bdf6fa52f01b ("sctp: handle association restarts when the socket is closed.") Signed-off-by: Jere Leppänen <jere.leppanen@nokia.com> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The driver is designed to drop Rx packets and reclaim the buffers
when an allocation fails, and the network interface needs to safely
handle this packet loss. Therefore, an allocation failure of Rx
SKBs is relatively benign.
However, the output of the warning message occurs with a high
scheduling priority that can cause excessive jitter/latency for
other high priority processing.
This commit suppresses the warning messages to prevent scheduling
problems while retaining the failure count in the statistics of
the network interface.
Signed-off-by: Doug Berger <opendmb@gmail.com> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The driver is designed to drop Rx packets and reclaim the buffers
when an allocation fails, and the network interface needs to safely
handle this packet loss. Therefore, an allocation failure of Rx
SKBs is relatively benign.
However, the output of the warning message occurs with a high
scheduling priority that can cause excessive jitter/latency for
other high priority processing.
This commit suppresses the warning messages to prevent scheduling
problems while retaining the failure count in the statistics of
the network interface.
Signed-off-by: Doug Berger <opendmb@gmail.com> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
0day reports over and over on an powerpc randconfig with clang:
lib/mpi/generic_mpih-mul1.c:37:13: error: invalid use of a cast in a
inline asm context requiring an l-value: remove the cast or build with
-fheinous-gnu-extensions
Remove the superfluous casts, which have been done previously for x86
and arm32 in commit dea632cadd12 ("lib/mpi: fix build with clang") and
commit 7b7c1df2883d ("lib/mpi/longlong.h: fix building with 32-bit
x86").
Switching tracers include instruction patching. To prevent that a
instruction is patched while it's read the instruction patching is done
in stop_machine 'context'. This also means that any function called
during stop_machine must not be traced. Thus add 'notrace' to all
functions called within stop_machine.
Fixes: 1ec2772e0c3c ("s390/diag: add a statistic for diagnose calls") Fixes: 38f2c691a4b3 ("s390: improve wait logic of stop_machine") Fixes: 4ecf0a43e729 ("processor: get rid of cpu_relax_yield") Signed-off-by: Philipp Rudo <prudo@linux.ibm.com> Signed-off-by: Vasily Gorbik <gor@linux.ibm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
We use a spinlock while we are reading and accessing the destination address for a server.
We need to also use this spinlock to protect when we are modifying this address from
reconn_set_ipaddr().
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix the rsnd_ssi_stop function to skip disabling the individual SSIs of
a multi-SSI setup, as the actual stop is performed by rsnd_ssiu_stop_gen2
- the same logic as in rsnd_ssi_start. The attempt to disable these SSIs
was harmless, but caused a "status check failed" message to be printed
for every SSI in the multi-SSI setup.
The disabling of interrupts is still performed, as they are enabled for
all SSIs in rsnd_ssi_init, but care is taken to not accidentally set the
EN bit for an SSI where it was not set by rsnd_ssi_start.
The master SSI of a multi-SSI setup was attached both to the
RSND_MOD_SSI slot and the RSND_MOD_SSIP slot of the rsnd_dai_stream.
This is not correct wrt. the meaning of being "parent" in the rest of
the SSI code, where it seems to indicate an SSI that provides clock and
word sync but is not transmitting/receiving audio data.
Not treating the multi-SSI master as parent allows removal of various
special cases to the rsnd_ssi_is_parent conditions introduced in commit a09fb3f28a60 ("ASoC: rsnd: Fix parent SSI start/stop in multi-SSI mode").
It also fixes the issue that operations performed via rsnd_dai_call()
were performed twice for the master SSI. This caused some "status check
failed" spam when stopping a multi-SSI stream as the driver attempted to
stop the master SSI twice.
In fine adjustement mode, which is the current default, the sub-second
increment register is the number of nanoseconds that will be added to
the clock when the accumulator overflows. At each clock cycle, the
value of the addend register is added to the accumulator.
Currently, we use 20ns = 1e09ns / 50MHz as this value whatever the
frequency of the ptp clock actually is.
The adjustment is then done on the addend register, only incrementing
every X clock cycles X being the ratio between 50MHz and ptp_clock_rate
(addend = 2^32 * 50MHz/ptp_clock_rate).
This causes the following issues :
- In case the frequency of the ptp clock is inferior or equal to 50MHz,
the addend value calculation will overflow and the default
addend value will be set to 0, causing the clock to not work at
all. (For instance, for ptp_clock_rate = 50MHz, addend = 2^32).
- The resolution of the timestamping clock is limited to 20ns while it
is not needed, thus limiting the accuracy of the timestamping to
20ns.
Fix this by setting sub-second increment to 2e09ns / ptp_clock_rate.
It will allow to reach the minimum possible frequency for
ptp_clk_ref, which is 5MHz for GMII 1000Mps Full-Duplex by setting the
sub-second-increment to a higher value. For instance, for 25MHz, it
gives ssinc = 80ns and default_addend = 2^31.
It will also allow to use a lower value for sub-second-increment, thus
improving the timestamping accuracy with frequencies higher than
100MHz, for instance, for 200MHz, ssinc = 10ns and default_addend =
2^31.
v1->v2:
- Remove modifications to the calculation of default addend, which broke
compatibility with clock frequencies for which 2000000000 / ptp_clk_freq
is not an integer.
- Modify description according to discussions.
Signed-off-by: Julien Beraud <julien.beraud@orolia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are 2 registers to write to enable a ptp ref clock coming from the
fpga.
One that enables the usage of the clock from the fpga for emac0 and emac1
as a ptp ref clock, and the other to allow signals from the fpga to reach
emac0 and emac1.
Currently, if the dwmac-socfpga has phymode set to PHY_INTERFACE_MODE_MII,
PHY_INTERFACE_MODE_GMII, or PHY_INTERFACE_MODE_SGMII, both registers will
be written and the ptp ref clock will be set as coming from the fpga.
Separate the 2 register writes to only enable signals from the fpga to
reach emac0 or emac1 when ptp ref clock is not coming from the fpga.
Signed-off-by: Julien Beraud <julien.beraud@orolia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
i2400mu_bus_bm_wait_for_ack() invokes usb_get_urb(), which increases the
refcount of the "notif_urb".
When i2400mu_bus_bm_wait_for_ack() returns, local variable "notif_urb"
becomes invalid, so the refcount should be decreased to keep refcount
balanced.
The issue happens in all paths of i2400mu_bus_bm_wait_for_ack(), which
forget to decrease the refcnt increased by usb_get_urb(), causing a
refcnt leak.
Fix this issue by calling usb_put_urb() before the
i2400mu_bus_bm_wait_for_ack() returns.
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
If we don't find any pcm, pcm will point at address at an offset from
the the list head and not a meaningful structure. Fix this by returning
correct pcm if found and NULL if not. Found with coccinelle.
The HDMI?_SEL register maps up to four stereo SSI data lanes onto the
sdata[0..3] inputs of the HDMI output block. The upper half of the
register contains four blocks of 4 bits, with the most significant
controlling the sdata3 line and the least significant the sdata0 line.
The shift calculation has an off-by-one error, causing the parent SSI to
be mapped to sdata3, the first multi-SSI child to sdata0 and so forth.
As the parent SSI transmits the stereo L/R channels, and the HDMI core
expects it on the sdata0 line, this causes no audio to be output when
playing stereo audio on a multichannel capable HDMI out, and
multichannel audio has permutated channels.
Fix the shift calculation to map the parent SSI to sdata0, the first
child to sdata1 etc.
The parent SSI of a multi-SSI setup must be fully setup, started and
stopped since it is also part of the playback/capture setup. So only
skip the SSI (as per commit 203cdf51f288 ("ASoC: rsnd: SSI parent cares
SWSP bit") and commit 597b046f0d99 ("ASoC: rsnd: control SSICR::EN
correctly")) if the SSI is parent outside of a multi-SSI setup.
Currently the calculation of max packet size limit for IN endpoints is
too restrictive. This prevents a matching of a capable hardware endpoint
during configuration. Below is the minimum recommended HW configuration
to support a particular endpoint setup from the databook:
For OUT endpoints, the databook recommended the minimum RxFIFO size to
be at least 3x MaxPacketSize + 3x setup packets size (8 bytes each) +
clock crossing margin (16 bytes).
For IN endpoints, the databook recommended the minimum TxFIFO size to be
at least 3x MaxPacketSize for endpoints that support burst. If the
endpoint doesn't support burst or when the device is operating in USB
2.0 mode, a minimum TxFIFO size of 2x MaxPacketSize is recommended.
Base on these recommendations, we can calculate the MaxPacketSize limit
of each endpoint. This patch revises the IN endpoint MaxPacketSize limit
and also sets the MaxPacketSize limit for OUT endpoints.
As mentioned slightly out of patch context in the code, there
is no reset routine for the chip. On boards where the chip is
supplied by a fixed regulator, it might not even be resetted
during (e.g. watchdog) reboot and can be in any state.
If the device is probed with VAG enabled, the driver's probe
routine will generate a loud pop sound when ANA_POWER is
being programmed. Avoid this by properly disabling just the
VAG bit and waiting the required power down time.
After successfully running the IPC msgque test once, subsequent runs
result in a test failure:
$ sudo ./run_kselftest.sh
TAP version 13
1..1
# selftests: ipc: msgque
# Failed to get stats for IPC queue with id 0
# Failed to dump queue: -22
# Bail out!
# # Pass 0 Fail 0 Xfail 0 Xpass 0 Skip 0 Error 0
not ok 1 selftests: ipc: msgque # exit=1
The dump_queue() function loops through the possible message queue index
values using calls to msgctl(kern_id, MSG_STAT, ...) where kern_id
represents the index value. The first time the test is ran, the initial
index value of 0 is valid and the test is able to complete. The index
value of 0 is not valid in subsequent test runs and the loop attempts to
try index values of 1, 2, 3, and so on until a valid index value is
found that corresponds to the message queue created earlier in the test.
The msgctl() syscall returns -1 and sets errno to EINVAL when invalid
index values are used. The test failure is caused by incorrectly
comparing errno to -EINVAL when cycling through possible index values.
Fix invalid test failures on subsequent runs of the msgque test by
correctly comparing errno values to a non-negated EINVAL.
The pseries platform uses the PCI_PROBE_DEVTREE method of PCI probing
which reads "assigned-addresses" of every PCI device and initializes
the device resources. However if the property is missing or zero sized,
then there is no fallback of any kind and the PCI resources remain
undiscovered, i.e. pdev->resource[] array remains empty.
This adds a fallback which parses the "reg" property in pretty much same
way except it marks resources as "unset" which later make Linux assign
those resources proper addresses.
This has an effect when:
1. a hypervisor failed to assign any resource for a device;
2. /chosen/linux,pci-probe-only=0 is in the DT so the system may try
assigning a resource.
Neither is likely to happen under PowerVM.
Ning Bo reported an abnormal 2-second gap when booting Kata container [1].
The unconditional timeout was caused by VSOCK_DEFAULT_CONNECT_TIMEOUT of
connecting from the client side. The vhost vsock client tries to connect
an initializing virtio vsock server.
The abnormal flow looks like:
host-userspace vhost vsock guest vsock
============== =========== ============
connect() --------> vhost_transport_send_pkt_work() initializing
| vq->private_data==NULL
| will not be queued
V
schedule_timeout(2s)
vhost_vsock_start() <--------- device ready
set vq->private_data
wait for 2s and failed
connect() again vq->private_data!=NULL recv connecting pkt
Details:
1. Host userspace sends a connect pkt, at that time, guest vsock is under
initializing, hence the vhost_vsock_start has not been called. So
vq->private_data==NULL, and the pkt is not been queued to send to guest
2. Then it sleeps for 2s
3. After guest vsock finishes initializing, vq->private_data is set
4. When host userspace wakes up after 2s, send connecting pkt again,
everything is fine.
As suggested by Stefano Garzarella, this fixes it by additional kicking the
send_pkt worker in vhost_vsock_start once the virtio device is started. This
makes the pending pkt sent again.
After this patch, kata-runtime (with vsock enabled) boot time is reduced
from 3s to 1s on a ThunderX2 arm64 server.
The recent commit 0d84c3e6a5b2 ("mmc: core: Convert to
mmc_poll_for_busy() for erase/trim/discard") makes use of the
->card_busy() op for SD cards. This uncovered that the ->card_busy() op
in the Meson SDIO driver was never working right:
while polling the busy status with ->card_busy()
meson_mx_mmc_card_busy() reads only one of the two MESON_MX_SDIO_IRQC
register values 0x1f001f10 or 0x1f003f10. This translates to "three out
of four DAT lines are HIGH" and "all four DAT lines are HIGH", which
is interpreted as "the card is busy".
It turns out that no situation can be observed where all four DAT lines
are LOW, meaning the card is not busy anymore. Upon further research the
3.10 vendor driver for this controller does not implement the
->card_busy() op.
Remove the ->card_busy() op from the meson-mx-sdio driver since it is
not working. At the time of writing this patch it is not clear what's
needed to make the ->card_busy() implementation work with this specific
controller hardware. For all use-cases which have previously worked the
MMC_CAP_WAIT_WHILE_BUSY flag is now taking over, even if we don't have
a ->card_busy() op anymore.
Fixes: ed80a13bb4c4c9 ("mmc: meson-mx-sdio: Add a driver for the Amlogic Meson8 and Meson8b SoCs") Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20200416183513.993763-3-martin.blumenstingl@googlemail.com Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Meson SDIO controller uses the DAT0 lane for hardware busy
detection. Set MMC_CAP_WAIT_WHILE_BUSY accordingly. This fixes
the following error observed with Linux 5.7 (pre-rc-1):
mmc1: Card stuck being busy! __mmc_poll_for_busy
blk_update_request: I/O error, dev mmcblk1, sector 17111080 op
0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
Fixes: ed80a13bb4c4c9 ("mmc: meson-mx-sdio: Add a driver for the Amlogic Meson8 and Meson8b SoCs") Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20200416183513.993763-2-martin.blumenstingl@googlemail.com Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
MSM sd host controller is capable of HW busy detection of device busy
signaling over DAT0 line. And it requires the R1B response for commands
that have this response associated with them.
So set the below two host capabilities for qcom SDHC.
- MMC_CAP_WAIT_WHILE_BUSY
- MMC_CAP_NEED_RSP_BUSY
Recent development of the mmc core in regards to this, revealed this as
being a potential bug, hence the stable tag.
BIOS writers have begun the practice of setting 40 ohm eMMC driver strength
even though the eMMC may not support it, on the assumption that the kernel
will validate the value against the eMMC (Extended CSD DRIVER_STRENGTH
[offset 197]) and revert to the default 50 ohm value if 40 ohm is invalid.
This is done to avoid changing the value for different boards.
Putting aside the merits of this approach, it is clear the eMMC's mask
of supported driver strengths is more reliable than the value provided
by BIOS. Add validation accordingly.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Fixes: 51ced59cc02e ("mmc: sdhci-pci: Use ACPI DSM to get driver strength for some Intel devices") Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20200422111629.4899-1-adrian.hunter@intel.com Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For some reason the Host Control2 register of the Xenon SDHCI controller
sometimes reports the bit representing 1.8V signaling as 0 when read
after it was written as 1. Subsequent read reports 1.
This causes the sdhci_start_signal_voltage_switch function to report
1.8V regulator output did not become stable
When CONFIG_PM is enabled, the host is suspended and resumend many
times, and in each resume the switch to 1.8V is called, and so the
kernel log reports this message annoyingly often.
Do an empty read of the Host Control2 register in Xenon's
.voltage_switch method to circumvent this.
This patch fixes this particular problem on Turris MOX.
Signed-off-by: Marek Behún <marek.behun@nic.cz> Fixes: 8d876bf472db ("mmc: sdhci-xenon: wait 5ms after set 1.8V...") Cc: stable@vger.kernel.org # v4.16+ Link: https://lore.kernel.org/r/20200420080444.25242-1-marek.behun@nic.cz Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Open-coding a timeout loop invariably leads to errors with handling
the timeout properly in one corner case or another. In the case of
cqhci we might report "CQE stuck on" even if it wasn't stuck on.
You'd just need this sequence of events to happen in cqhci_off():
1. Call ktime_get().
2. Something happens to interrupt the CPU for > 100 us (context switch
or interrupt).
3. Check time and; set "timed_out" to true since > 100 us.
4. Read CQHCI_CTL.
5. Both "reg & CQHCI_HALT" and "timed_out" are true, so break.
6. Since "timed_out" is true, falsely print the error message.
Rather than fixing the polling loop, use readx_poll_timeout() like
many people do. This has been time tested to handle the corner cases.
Fixes: a4080225f51d ("mmc: cqhci: support for command queue enabled host") Signed-off-by: Douglas Anderson <dianders@chromium.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20200413162717.1.Idece266f5c8793193b57a1ddb1066d030c6af8e0@changeid Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[BUG]
One run of btrfs/063 triggered the following lockdep warning:
============================================
WARNING: possible recursive locking detected
5.6.0-rc7-custom+ #48 Not tainted
--------------------------------------------
kworker/u24:0/7 is trying to acquire lock: ffff88817d3a46e0 (sb_internal#2){.+.+}, at: start_transaction+0x66c/0x890 [btrfs]
but task is already holding lock: ffff88817d3a46e0 (sb_internal#2){.+.+}, at: start_transaction+0x66c/0x890 [btrfs]
other info that might help us debug this:
Possible unsafe locking scenario:
When we have an inode with a prealloc extent that starts at an offset
lower than the i_size and there is another prealloc extent that starts at
an offset beyond i_size, we can end up losing part of the first prealloc
extent (the part that starts at i_size) and have an implicit hole if we
fsync the file and then have a power failure.
Consider the following example with comments explaining how and why it
happens.
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt
# Create our test file with 2 consecutive prealloc extents, each with a
# size of 128Kb, and covering the range from 0 to 256Kb, with a file
# size of 0.
$ xfs_io -f -c "falloc -k 0 128K" /mnt/foo
$ xfs_io -c "falloc -k 128K 128K" /mnt/foo
# Fsync the file to record both extents in the log tree.
$ xfs_io -c "fsync" /mnt/foo
# Now do a redudant extent allocation for the range from 0 to 64Kb.
# This will merely increase the file size from 0 to 64Kb. Instead we
# could also do a truncate to set the file size to 64Kb.
$ xfs_io -c "falloc 0 64K" /mnt/foo
# Fsync the file, so we update the inode item in the log tree with the
# new file size (64Kb). This also ends up setting the number of bytes
# for the first prealloc extent to 64Kb. This is done by the truncation
# at btrfs_log_prealloc_extents().
# This means that if a power failure happens after this, a write into
# the file range 64Kb to 128Kb will not use the prealloc extent and
# will result in allocation of a new extent.
$ xfs_io -c "fsync" /mnt/foo
# Now set the file size to 256K with a truncate and then fsync the file.
# Since no changes happened to the extents, the fsync only updates the
# i_size in the inode item at the log tree. This results in an implicit
# hole for the file range from 64Kb to 128Kb, something which fsck will
# complain when not using the NO_HOLES feature if we replay the log
# after a power failure.
$ xfs_io -c "truncate 256K" -c "fsync" /mnt/foo
So instead of always truncating the log to the inode's current i_size at
btrfs_log_prealloc_extents(), check first if there's a prealloc extent
that starts at an offset lower than the i_size and with a length that
crosses the i_size - if there is one, just make sure we truncate to a
size that corresponds to the end offset of that prealloc extent, so
that we don't lose the part of that extent that starts at i_size if a
power failure happens.
A test case for fstests follows soon.
Fixes: 31d11b83b96f ("Btrfs: fix duplicate extents after fsync of file with prealloc extents") CC: stable@vger.kernel.org # 4.14+ Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix the SELinux netlink_send hook to properly handle multiple netlink
messages in a single sk_buff; each message is parsed and subject to
SELinux access control. Prior to this patch, SELinux only inspected
the first message in the sk_buff.
Cc: stable@vger.kernel.org Reported-by: Dmitry Vyukov <dvyukov@google.com> Reviewed-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Under some circumstances, i.e. when test is still running and about to
time out and user runs, for example,
grep -H . /sys/module/dmatest/parameters/*
the iterations parameter is not respected and test is going on and on until
user gives
echo 0 > /sys/module/dmatest/parameters/run
This is not what expected.
The history of this bug is interesting. I though that the commit 2d88ce76eb98 ("dmatest: add a 'wait' parameter")
is a culprit, but looking closer to the code I think it simple revealed the
broken logic from the day one, i.e. in the commit 0a2ff57d6fba ("dmaengine: dmatest: add a maximum number of test iterations")
which adds iterations parameter.
So, to the point, the conditional of checking the thread to be stopped being
first part of conjunction logic prevents to check iterations. Thus, we have to
always check both conditions to be able to stop after given iterations.
Since it wasn't visible before second commit appeared, I add a respective
Fixes tag.
Fixes: 2d88ce76eb98 ("dmatest: add a 'wait' parameter") Cc: Dan Williams <dan.j.williams@intel.com> Cc: Nicolas Ferre <nicolas.ferre@microchip.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com> Link: https://lore.kernel.org/r/20200424161147.16895-1-andriy.shevchenko@linux.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
nfs3_set_acl keeps track of the acl it allocated locally to determine if an acl
needs to be released at the end. This results in a memory leak when the
function allocates an acl as well as a default acl. Fix by releasing acls
that differ from the acl originally passed into nfs3_set_acl.
Fixes: b7fa0554cf1b ("[PATCH] NFS: Add support for NFSv3 ACLs") Reported-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
gcc-10 points out a few instances of suspicious integer arithmetic
leading to value truncation:
sound/isa/opti9xx/opti92x-ad1848.c: In function 'snd_opti9xx_configure':
sound/isa/opti9xx/opti92x-ad1848.c:322:43: error: overflow in conversion from 'int' to 'unsigned char' changes value from '(int)snd_opti9xx_read(chip, 3) & -256 | 240' to '240' [-Werror=overflow]
322 | (snd_opti9xx_read(chip, reg) & ~(mask)) | ((value) & (mask)))
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~
sound/isa/opti9xx/opti92x-ad1848.c:351:3: note: in expansion of macro 'snd_opti9xx_write_mask'
351 | snd_opti9xx_write_mask(chip, OPTi9XX_MC_REG(3), 0xf0, 0xff);
| ^~~~~~~~~~~~~~~~~~~~~~
sound/isa/opti9xx/miro.c: In function 'snd_miro_configure':
sound/isa/opti9xx/miro.c:873:40: error: overflow in conversion from 'int' to 'unsigned char' changes value from '(int)snd_miro_read(chip, 3) & -256 | 240' to '240' [-Werror=overflow]
873 | (snd_miro_read(chip, reg) & ~(mask)) | ((value) & (mask)))
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~
sound/isa/opti9xx/miro.c:1010:3: note: in expansion of macro 'snd_miro_write_mask'
1010 | snd_miro_write_mask(chip, OPTi9XX_MC_REG(3), 0xf0, 0xff);
| ^~~~~~~~~~~~~~~~~~~
These are all harmless here as only the low 8 bit are passed down
anyway. Change the macros to inline functions to make the code
more readable and also avoid the warning.
Strictly speaking those functions also need locking to make the
read/write pair atomic, but it seems unlikely that anyone would
still run into that issue.
Currently, system fails to boot because the legacy interrupt remapping
mode does not enable 128-bit IRTE (GA), which is required for x2APIC
support.
Fix by using AMD_IOMMU_GUEST_IR_LEGACY_GA mode when booting with
kernel option amd_iommu_intr=legacy instead. The initialization
logic will check GASup and automatically fallback to using
AMD_IOMMU_GUEST_IR_LEGACY if GA mode is not supported.
SBC4 specifies that WRITE SAME requests with the UNMAP bit set to zero
"shall perform the specified write operation to each LBA specified by the
command". Commit 2237498f0b5c ("target/iblock: Convert WRITE_SAME to
blkdev_issue_zeroout") modified the iblock backend to call
blkdev_issue_zeroout() when handling WRITE SAME requests with UNMAP=0 and a
zero data segment.
The iblock blkdev_issue_zeroout() call incorrectly provides a flags
parameter of 0 (bool false), instead of BLKDEV_ZERO_NOUNMAP. The bool
false parameter reflects the blkdev_issue_zeroout() API prior to commit ee472d835c26 ("block: add a flags argument to (__)blkdev_issue_zeroout")
which was merged shortly before 2237498f0b5c.
Link: https://lore.kernel.org/r/20200419163109.11689-1-ddiss@suse.de Fixes: 2237498f0b5c ("target/iblock: Convert WRITE_SAME to blkdev_issue_zeroout") Reviewed-by: Bart Van Assche <bvanassche@acm.org> Signed-off-by: David Disseldorp <ddiss@suse.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function qcom_iommu_device_probe() does not perform sufficient
error checking after executing devm_ioremap_resource(), which can
result in crashes if a critical error path is encountered.
Use follow_pfn() to get the PFN of a PFNMAP VMA instead of assuming that
vma->vm_pgoff holds the base PFN of the VMA. This fixes a bug where
attempting to do VFIO_IOMMU_MAP_DMA on an arbitrary PFNMAP'd region of
memory calculates garbage for the PFN.
Hilariously, this only got detected because the first "PFN" calculated
by vaddr_get_pfn() is PFN 0 (vma->vm_pgoff==0), and iommu_iova_to_phys()
uses PA==0 as an error, which triggers a WARN in vfio_unmap_unpin()
because the translation "failed". PFN 0 is now unconditionally reserved
on x86 in order to mitigate L1TF, which causes is_invalid_reserved_pfn()
to return true and in turns results in vaddr_get_pfn() returning success
for PFN 0. Eventually the bogus calculation runs into PFNs that aren't
reserved and leads to failure in vfio_pin_map_dma(). The subsequent
call to vfio_remove_dma() attempts to unmap PFN 0 and WARNs.
FDs can only be used on the ufile that created them, they cannot be mixed
to other ufiles. We are lacking a check to prevent it.
BUG: KASAN: null-ptr-deref in atomic64_sub_and_test include/asm-generic/atomic-instrumented.h:1547 [inline]
BUG: KASAN: null-ptr-deref in atomic_long_sub_and_test include/asm-generic/atomic-long.h:460 [inline]
BUG: KASAN: null-ptr-deref in fput_many+0x1a/0x140 fs/file_table.c:336
Write of size 8 at addr 0000000000000038 by task syz-executor179/284
Initialize ib_spec on the stack before using it, otherwise we will have
garbage values that will break creating default rules with invalid parsing
error.
Fixes: a37a1a428431 ("IB/mlx4: Add mechanism to support flow steering over IB links") Link: https://lore.kernel.org/r/20200413132235.930642-1-leon@kernel.org Signed-off-by: Alaa Hleihel <alaa@mellanox.com> Reviewed-by: Maor Gottlieb <maorg@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
GRH fields such as sgid_index, hop limit, et. are set in the QP context
when QP is created/modified.
Currently, when query QP is performed, we fill the GRH fields only if the
GRH bit is set in the QP context, but this bit is not set for RoCE. Adjust
the check so we will set all relevant data for the RoCE too.
Since this data is returned to userspace, the below is an ABI regression.
Fixes: d8966fcd4c25 ("IB/core: Use rdma_ah_attr accessor functions") Link: https://lore.kernel.org/r/20200413132028.930109-1-leon@kernel.org Signed-off-by: Aharon Landau <aharonl@mellanox.com> Reviewed-by: Maor Gottlieb <maorg@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
qlt_free_session_done() tries to post async PRLO / LOGO, and waits for the
completion of these async commands. If UNLOADING is set, this is doomed to
timeout, because the async logout command will never complete.
The only way to avoid waiting pointlessly is to fail posting these commands
in the first place if the driver is in UNLOADING state. In general,
posting any command should be avoided when the driver is UNLOADING.
With this patch, "rmmod qla2xxx" completes without noticeable delay.
Link: https://lore.kernel.org/r/20200421204621.19228-3-mwilck@suse.com Fixes: 45235022da99 ("scsi: qla2xxx: Fix driver unload by shutting down chip") Acked-by: Arun Easi <aeasi@marvell.com> Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com> Signed-off-by: Martin Wilck <mwilck@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The purpose of the UNLOADING flag is to avoid port login procedures to
continue when a controller is in the process of shutting down. It makes
sense to set this flag before starting session teardown.
Furthermore, use atomic test_and_set_bit() to avoid the shutdown being run
multiple times in parallel. In qla2x00_disable_board_on_pci_error(), the
test for UNLOADING is postponed until after the check for an already
disabled PCI board.
Link: https://lore.kernel.org/r/20200421204621.19228-2-mwilck@suse.com Fixes: 45235022da99 ("scsi: qla2xxx: Fix driver unload by shutting down chip") Reviewed-by: Arun Easi <aeasi@marvell.com> Reviewed-by: Daniel Wagner <dwagner@suse.de> Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com> Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com> Signed-off-by: Martin Wilck <mwilck@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When adding devices that don't have a scsi_dh on a BIO based multipath,
I was able to consistently hit the warning below and lock-up the system.
The problem is that __map_bio reads the flag before it potentially being
modified by choose_pgpath, and ends up using the older value.
The WARN_ON below is not trivially linked to the issue. It goes like
this: The activate_path delayed_work is not initialized for non-scsi_dh
devices, but we always set MPATHF_QUEUE_IO, asking for initialization.
That is fine, since MPATHF_QUEUE_IO would be cleared in choose_pgpath.
Nevertheless, only for BIO-based mpath, we cache the flag before calling
choose_pgpath, and use the older version when deciding if we should
initialize the path. Therefore, we end up trying to initialize the
paths, and calling the non-initialized activate_path work.
The dm-writecache reads metadata in the target constructor. However, when
we reload the target, there could be another active instance running on
the same device. This is the sequence of operations when doing a reload:
1. construct new target
2. suspend old target
3. resume new target
4. destroy old target
Metadata that were written by the old target between steps 1 and 2 would
not be visible by the new target.
Fix the data corruption by loading the metadata in the resume handler.
Also, validate block_size is at least as large as both the devices'
logical block size and only read 1 block from the metadata during
target constructor -- no need to read entirety of metadata now that it
is done during resume.
The error correction data is computed as if data and hash blocks
were concatenated. But hash block number starts from v->hash_start.
So, we have to calculate hash block number based on that.
Fixes: a739ff3f543af ("dm verity: add support for forward error correction") Cc: stable@vger.kernel.org Signed-off-by: Sunwook Eom <speed.eom@samsung.com> Reviewed-by: Sami Tolvanen <samitolvanen@google.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the kernel threads are not frozen in software_resume(), so
between dpm_suspend_start(PMSG_QUIESCE) and resume_target_kernel(),
system_freezable_power_efficient_wq can still try to submit SCSI
commands and this can cause a panic since the low level SCSI driver
(e.g. hv_storvsc) has quiesced the SCSI adapter and can not accept
any SCSI commands: https://lkml.org/lkml/2020/4/10/47
At first I posted a fix (https://lkml.org/lkml/2020/4/21/1318) trying
to resolve the issue from hv_storvsc, but with the help of
Bart Van Assche, I realized it's better to fix software_resume(),
since this looks like a generic issue, not only pertaining to SCSI.
Cc: All applicable <stable@vger.kernel.org> Signed-off-by: Dexuan Cui <decui@microsoft.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The checks of the plugin buffer overflow in the previous fix by commit f2ecf903ef06 ("ALSA: pcm: oss: Avoid plugin buffer overflow")
are put in the wrong places mistakenly, which leads to the expected
(repeated) sound when the rate plugin is involved. Fix in the right
places.
Also, at those right places, the zero check is needed for the
termination node, so added there as well, and let's get it done,
finally.
This new Lenovo ThinkCenter has two front mics which can't be handled
by PA so far, so apply the fixup ALC283_FIXUP_HEADSET_MIC to change
the location for one of the mics.
btrfs_remove_block_group() invokes btrfs_lookup_block_group(), which
returns a local reference of the block group that contains the given
bytenr to "block_group" with increased refcount.
When btrfs_remove_block_group() returns, "block_group" becomes invalid,
so the refcount should be decreased to keep refcount balanced.
The reference counting issue happens in several exception handling paths
of btrfs_remove_block_group(). When those error scenarios occur such as
btrfs_alloc_path() returns NULL, the function forgets to decrease its
refcnt increased by btrfs_lookup_block_group() and will cause a refcnt
leak.
Fix this issue by jumping to "out_put_group" label and calling
btrfs_put_block_group() when those error scenarios occur.
CC: stable@vger.kernel.org # 4.4+ Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
qxl_release should not be accesses after qxl_push_*_ring_release() calls:
userspace driver can process submitted command quickly, move qxl_release
into release_ring, generate interrupt and trigger garbage collector.
It can lead to crashes in qxl driver or trigger memory corruption
in some kmalloc-192 slab object
Gerd Hoffmann proposes to swap the qxl_release_fence_buffer_objects() +
qxl_push_{cursor,command}_ring_release() calls to close that race window.
The DispID DTD pixel clock is documented as:
"00 00 00 h → FF FF FF h | Pixel clock ÷ 10,000 0.01 → 167,772.16 Mega Pixels per Sec"
Which seems to imply that we to add one to the raw value.
Reality seems to agree as there are tiled displays in the wild
which currently show a 10kHz difference in the pixel clock
between the tiles (one tile gets its mode from the base EDID,
the other from the DispID block).
... to protect the modification of mp->m_count done by it. Most of
the places that modify that thing also have namespace_lock held,
but not all of them can do so, so we really need mount_lock here.
Kudos to Piotr Krysiuk <piotras@gmail.com>, who'd spotted a related
bug in pivot_root(2) (fixed unnoticed in 5.3); search for other
similar turds has caught out this one.
Cc: stable@kernel.org Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Piotr Krysiuk <piotras@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While calculating overhead for internal journal, also check
that j_inum shouldn't be 0. Otherwise we get below error with
xfstests generic/050 with external journal (XXX_LOGDEV config) enabled.
It could be simply reproduced with loop device with an external journal
and marking blockdev as RO before mounting.
[ 3337.146838] EXT4-fs error (device pmem1p2): ext4_get_journal_inode:4634: comm mount: inode #0: comm mount: iget: illegal inode #
------------[ cut here ]------------
generic_make_request: Trying to write to read-only block-device pmem1p2 (partno 2)
WARNING: CPU: 107 PID: 115347 at block/blk-core.c:788 generic_make_request_checks+0x6b4/0x7d0
CPU: 107 PID: 115347 Comm: mount Tainted: G L --------- -t - 4.18.0-167.el8.ppc64le #1
NIP: c0000000006f6d44 LR: c0000000006f6d40 CTR: 0000000030041dd4
<...>
NIP [c0000000006f6d44] generic_make_request_checks+0x6b4/0x7d0
LR [c0000000006f6d40] generic_make_request_checks+0x6b0/0x7d0
<...>
Call Trace:
generic_make_request_checks+0x6b0/0x7d0 (unreliable)
generic_make_request+0x3c/0x420
submit_bio+0xd8/0x200
submit_bh_wbc+0x1e8/0x250
__sync_dirty_buffer+0xd0/0x210
ext4_commit_super+0x310/0x420 [ext4]
__ext4_error+0xa4/0x1e0 [ext4]
__ext4_iget+0x388/0xe10 [ext4]
ext4_get_journal_inode+0x40/0x150 [ext4]
ext4_calculate_overhead+0x5a8/0x610 [ext4]
ext4_fill_super+0x3188/0x3260 [ext4]
mount_bdev+0x778/0x8f0
ext4_mount+0x28/0x50 [ext4]
mount_fs+0x74/0x230
vfs_kern_mount.part.6+0x6c/0x250
do_mount+0x2fc/0x1280
sys_mount+0x158/0x180
system_call+0x5c/0x70
EXT4-fs (pmem1p2): no journal found
EXT4-fs (pmem1p2): can't get journal size
EXT4-fs (pmem1p2): mounted filesystem without journal. Opts: dax,norecovery
Fixes: 3c816ded78bb ("ext4: use journal inode to determine journal overhead") Reported-by: Harish Sriram <harish@linux.ibm.com> Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20200316093038.25485-1-riteshh@linux.ibm.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The qed_chain data structure was modified in
commit 1a4a69751f4d ("qed: Chain support for external PBL") to support
receiving an external pbl (due to iWARP FW requirements).
The pages pointed to by the pbl are allocated in qed_chain_alloc
and their virtual address are stored in an virtual addresses array to
enable accessing and freeing the data. The physical addresses however
weren't stored and were accessed directly from the external-pbl
during free.
Destroy-qp flow, leads to freeing the external pbl before the chain is
freed, when the chain is freed it tries accessing the already freed
external pbl, leading to a use-after-free. Therefore we need to store
the physical addresses in additional to the virtual addresses in a
new data structure.
Fixes: 1a4a69751f4d ("qed: Chain support for external PBL") Signed-off-by: Michal Kalderon <mkalderon@marvell.com> Signed-off-by: Yuval Bason <ybason@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current JIT clobbers the destination register for BPF_JSET BPF_X
and BPF_K by using "and" and "or" instructions. This is fine when the
destination register is a temporary loaded from a register stored on
the stack but not otherwise.
This patch fixes the problem (for both BPF_K and BPF_X) by always loading
the destination register into temporaries since BPF_JSET should not
modify the destination register.
This bug may not be currently triggerable as BPF_REG_AX is the only
register not stored on the stack and the verifier uses it in a limited
way.
Fixes: 03f5781be2c7b ("bpf, x86_32: add eBPF JIT compiler for ia32") Signed-off-by: Xi Wang <xi.wang@gmail.com> Signed-off-by: Luke Nelson <luke.r.nels@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Wang YanQing <udknight@gmail.com> Link: https://lore.kernel.org/bpf/20200422173630.8351-2-luke.r.nels@gmail.com Signed-off-by: Wang YanQing <udknight@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The jc42 driver passes I2C client's name as hwmon device name. In case
of device tree probed devices this ends up being part of the compatible
string, "jc-42.4-temp". This name contains hyphens and the hwmon core
doesn't like this:
jc42 2-0018: hwmon: 'jc-42.4-temp' is not a valid name attribute, please fix
This changes the name to "jc42" which doesn't have any illegal
characters.
If the in-core buddy bitmap gets corrupted (or out of sync with the
block bitmap), issue a WARN_ON and try to recover. In most cases this
involves skipping trying to allocate out of a particular block group.
We can end up declaring the file system corrupted, which is fair,
since the file system probably should be checked before we proceed any
further.
Current wait times have proven to be too short to protect against inode
reuses that lead to metadata inconsistencies.
Now that we will retry the inode allocation if we can't find any
recently deleted inodes, it's a lot safer to increase the recently
deleted time from 5 seconds to a minute.
Run generic/388 with journal data mode sometimes may trigger the warning
in ext4_invalidatepage. Actually, we should use the matching invalidatepage
in ext4_writepage.
In assembly, many instances of __emit_inst(x) expand to a directive. In
a few places __emit_inst(x) is used as an assembler macro argument. For
example, in arch/arm64/kvm/hyp/entry.S
Both comma and space are separators, with an exception that content
inside a pair of parentheses/quotes is not split, so the clang
integrated assembler splits the arguments to:
GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend
(along with many other non-x86 backends) sees:
alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1
# .inst(...) is parsed as one argument
while its x86 backend sees:
alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1
# The extra space before '(' makes the whole .inst (...) parsed as two arguments
The non-x86 backend's behavior is considered unintentional
(https://sourceware.org/bugzilla/show_bug.cgi?id=25750).
So drop the space separator inside `.inst (...)` to make the clang
integrated assembler work.
Before the pci_driver->probe() is called, the pci subsystem calls
runtime_forbid() and runtime_get_sync() on this pci dev, so only call
runtime_put_autosuspend() is not enough to enable the runtime_pm on
this device.
For controllers with vgaswitcheroo feature, the pci/quirks.c will call
runtime_allow() for this dev, then the controllers could enter
rt_idle/suspend/resume, but for non-vgaswitcheroo controllers like
Intel hda controllers, the runtime_pm is not enabled because the
runtime_allow() is not called.
Since it is no harm calling runtime_allow() twice, here let hda
driver call runtime_allow() for all controllers. Then the runtime_pm
is enabled on all controllers after the put_autosuspend() is called.
xenbus_map_ring_valloc() maps a ring page and returns the status of the
used grant (0 meaning success).
There are Xen hypervisors which might return the value 1 for the status
of a failed grant mapping due to a bug. Some callers of
xenbus_map_ring_valloc() test for errors by testing the returned status
to be less than zero, resulting in no error detected and crashing later
due to a not available ring page.
Set the return value of xenbus_map_ring_valloc() to GNTST_general_error
in case the grant status reported by Xen is greater than zero.
CONFIG_UBSAN_TRAP causes GCC to emit a UD2 whenever it encounters an
unreachable code path. This includes __builtin_unreachable(). Because
the BUG() macro uses __builtin_unreachable() after it emits its own UD2,
this results in a double UD2. In this case objtool rightfully detects
that the second UD2 is unreachable:
In case command ring buffer becomes inconsistent, tcmu sets device flag
TCMU_DEV_BIT_BROKEN. If the bit is set, tcmu rejects new commands from LIO
core with TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE, and no longer processes
completions from the ring. The reset_ring attribute can be used to
completely clean up the command ring, so after reset_ring the ring no
longer is inconsistent.
Therefore reset_ring also should reset bit TCMU_DEV_BIT_BROKEN to allow
normal processing.
Creation of the response to READ FULL STATUS fails for FC based
reservations. Reason is the too high loop limit (< 24) in
fc_get_pr_transport_id(). The string representation of FC WWPN is 23 chars
long only ("11:22:33:44:55:66:77:88"). So when i is 23, the loop body is
executed a last time for the ending '\0' of the string and thus hex2bin()
reports an error.
Link: https://lore.kernel.org/r/20200408132610.14623-3-bstroesser@ts.fujitsu.com Signed-off-by: Bodo Stroesser <bstroesser@ts.fujitsu.com> Reviewed-by: Mike Christie <mchristi@redhat.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently, when the HD-audio controller driver doesn't detect any
codecs, it tries to abort the probe. But this abort happens at the
delayed probe, i.e. the primary probe call already returned success,
hence the driver is never unbound until user does so explicitly.
As a result, it may leave the HD-audio device in the running state
without the runtime PM. More badly, if the device is a HD-audio bus
that is tied with a GPU, GPU cannot reach to the full power down and
consumes unnecessarily much power.
This patch changes the logic after no-codec situation; it continues
probing without the further codec initialization but keep the
controller driver running normally.
In the reflink extent remap function, it turns out that uirec (the block
mapping corresponding only to the part of the passed-in mapping that got
unmapped) was not fully initialized. Specifically, br_state was not
being copied from the passed-in struct to the uirec. This could lead to
unpredictable results such as the reflinked mapping being marked
unwritten in the destination file.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
On some SoCs, such as the i.MX6, it is necessary to set a bit
in the SoC level GPR register before suspending for wake on lan
to work.
The fec platform callback sleep_mode_enable was intended to allow this
but the platform implementation was NAK'd back in 2015 [1]
This means that, currently, wake on lan is broken on mainline for
the i.MX6 at least.
So implement the required bit setting in the fec driver by itself
by adding a new optional DT property indicating the GPR register
and adding the offset and bit information to the driver.
Signed-off-by: Martin Fuzzey <martin.fuzzey@flowbird.group> Signed-off-by: Fugang Duan <fugang.duan@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch fixes an encoding bug in emit_stx for BPF_B when the source
register is BPF_REG_FP.
The current implementation for BPF_STX BPF_B in emit_stx saves one REX
byte when the operands can be encoded using Mod-R/M alone. The lower 8
bits of registers %rax, %rbx, %rcx, and %rdx can be accessed without using
a REX prefix via %al, %bl, %cl, and %dl, respectively. Other registers,
(e.g., %rsi, %rdi, %rbp, %rsp) require a REX prefix to use their 8-bit
equivalents (%sil, %dil, %bpl, %spl).
The current code checks if the source for BPF_STX BPF_B is BPF_REG_1
or BPF_REG_2 (which map to %rdi and %rsi), in which case it emits the
required REX prefix. However, it misses the case when the source is
BPF_REG_FP (mapped to %rbp).
The result is that BPF_STX BPF_B with BPF_REG_FP as the source operand
will read from register %ch instead of the correct %bpl. This patch fixes
the problem by fixing and refactoring the check on which registers need
the extra REX byte. Since no BPF registers map to %rsp, there is no need
to handle %spl.
Fixes: 622582786c9e0 ("net: filter: x86: internal BPF JIT") Signed-off-by: Xi Wang <xi.wang@gmail.com> Signed-off-by: Luke Nelson <luke.r.nels@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Link: https://lore.kernel.org/bpf/20200418232655.23870-1-luke.r.nels@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
Leaving PF_MEMALLOC set when exiting a kthread causes it to remain set
during do_exit(). That can confuse things. In particular, if BSD
process accounting is enabled, then do_exit() writes data to an
accounting file. If that file has FS_SYNC_FL set, then this write
occurs synchronously and can misbehave if PF_MEMALLOC is set.
For example, if the accounting file is located on an XFS filesystem,
then a WARN_ON_ONCE() in iomap_do_writepage() is triggered and the data
doesn't get written when it should. Or if the accounting file is
located on an ext4 filesystem without a journal, then a WARN_ON_ONCE()
in ext4_write_inode() is triggered and the inode doesn't get written.
Fix this in xfsaild() by using the helper functions to save and restore
PF_MEMALLOC.
This can be reproduced as follows in the kvm-xfstests test appliance
modified to add the 'acct' Debian package, and with kvm-xfstests's
recommended kconfig modified to add CONFIG_BSD_PROCESS_ACCT=y:
mkfs.xfs -f /dev/vdb
mount /vdb
touch /vdb/file
chattr +S /vdb/file
accton /vdb/file
mkfs.xfs -f /dev/vdc
mount /vdc
umount /vdc
This bug was originally reported by syzbot at
https://lore.kernel.org/r/0000000000000e7156059f751d7b@google.com.
Reported-by: syzbot+1f9dc49e8de2582d90c2@syzkaller.appspotmail.com Signed-off-by: Eric Biggers <ebiggers@google.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
WARNING: possible irq lock inversion dependency detected
5.6.0-rc7-syzkaller #0 Not tainted
--------------------------------------------------------
syz-executor.0/10317 just changed the state of lock: ffff888021d16568 (&(&info->lock)->rlock){+.+.}, at: spin_lock include/linux/spinlock.h:338 [inline] ffff888021d16568 (&(&info->lock)->rlock){+.+.}, at: shmem_mfill_atomic_pte+0x1012/0x21c0 mm/shmem.c:2407
but this lock was taken by another, SOFTIRQ-safe lock in the past:
(&(&xa->xa_lock)->rlock#5){..-.}
and interrupts could create inverse lock ordering between them.
other info that might help us debug this:
Possible interrupt unsafe locking scenario:
It is because CPU 0 held info->lock with IRQ enabled in userfaultfd_copy
path, then CPU 1 is splitting a THP which held xa_lock and info->lock in
IRQ disabled context at the same time. If softirq comes in to acquire
xa_lock, the deadlock would be triggered.
The fix is to acquire/release info->lock with *_irq version instead of
plain spin_{lock,unlock} to make it softirq safe.
Fixes: 4c27fe4c4c84 ("userfaultfd: shmem: add shmem_mcopy_atomic_pte for userfaultfd support") Reported-by: syzbot+e27980339d305f2dbfd9@syzkaller.appspotmail.com Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Tested-by: syzbot+e27980339d305f2dbfd9@syzkaller.appspotmail.com Acked-by: Hugh Dickins <hughd@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Link: http://lkml.kernel.org/r/1587061357-122619-1-git-send-email-yang.shi@linux.alibaba.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current JIT uses the following sequence to zero-extend into the
upper 32 bits of the destination register for BPF_LDX BPF_{B,H,W},
when the destination register is not on the stack:
EMIT3(0xC7, add_1reg(0xC0, dst_hi), 0);
The problem is that C7 /0 encodes a MOV instruction that requires a 4-byte
immediate; the current code emits only 1 byte of the immediate. This
means that the first 3 bytes of the next instruction will be treated as
the rest of the immediate, breaking the stream of instructions.
This patch fixes the problem by instead emitting "xor dst_hi,dst_hi"
to clear the upper 32 bits. This fixes the problem and is more efficient
than using MOV to load a zero immediate.
This bug may not be currently triggerable as BPF_REG_AX is the only
register not stored on the stack and the verifier uses it in a limited
way, and the verifier implements a zero-extension optimization. But the
JIT should avoid emitting incorrect encodings regardless.
Fixes: 03f5781be2c7b ("bpf, x86_32: add eBPF JIT compiler for ia32") Signed-off-by: Xi Wang <xi.wang@gmail.com> Signed-off-by: Luke Nelson <luke.r.nels@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Reviewed-by: H. Peter Anvin (Intel) <hpa@zytor.com> Acked-by: Wang YanQing <udknight@gmail.com> Link: https://lore.kernel.org/bpf/20200422173630.8351-1-luke.r.nels@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On s390 FORCE_MAX_ZONEORDER is 9 instead of 11, thus a larger kzalloc()
allocation as done for the firmware tracer will always fail.
Looking at mlx5_fw_tracer_save_trace(), it is actually the driver itself
that copies the debug data into the trace array and there is no need for
the allocation to be contiguous in physical memory. We can therefor use
kvzalloc() instead of kzalloc() and get rid of the large contiguous
allcoation.
When the kernel is built with CONFIG_DEBUG_PER_CPU_MAPS, the cpumap code
can trigger a spurious warning if CONFIG_CPUMASK_OFFSTACK is also set. This
happens because in this configuration, NR_CPUS can be larger than
nr_cpumask_bits, so the initial check in cpu_map_alloc() is not sufficient
to guard against hitting the warning in cpumask_check().
Fix this by explicitly checking the supplied key against the
nr_cpumask_bits variable before calling cpu_possible().
Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Tested-by: Xiumei Mu <xmu@redhat.com> Acked-by: Jesper Dangaard Brouer <brouer@redhat.com> Acked-by: Song Liu <songliubraving@fb.com> Link: https://lore.kernel.org/bpf/20200416083120.453718-1-toke@redhat.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / linux-stable / log