pm_runtime_get_sync will increment pm usage counter even it failed.
Forgetting to putting operation will result in reference leak here.
Fix it by replacing it with pm_runtime_resume_and_get to keep usage
counter balanced.
Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Wang Li <wangli74@huawei.com> Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Link: https://lore.kernel.org/r/20210409095458.29921-1-wangli74@huawei.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Call spi_master_get() holds the reference count to master device, thus
we need an additional spi_master_put() call to reduce the reference
count, otherwise we will leak a reference to master.
This commit fix it by removing the unnecessary spi_master_get().
Call spi_master_get() holds the reference count to master device, thus
we need an additional spi_master_put() call to reduce the reference
count, otherwise we will leak a reference to master.
This commit fix it by removing the unnecessary spi_master_get().
Some Chromebooks use hard-coded interrupts in their ACPI tables.
This is an excerpt as dumped on Relm:
...
Name (_HID, "ELAN0001") // _HID: Hardware ID
Name (_DDN, "Elan Touchscreen ") // _DDN: DOS Device Name
Name (_UID, 0x05) // _UID: Unique ID
Name (ISTP, Zero)
Method (_CRS, 0, NotSerialized) // _CRS: Current Resource Settings
{
Name (BUF0, ResourceTemplate ()
{
I2cSerialBusV2 (0x0010, ControllerInitiated, 0x00061A80,
AddressingMode7Bit, "\\_SB.I2C1",
0x00, ResourceConsumer, , Exclusive,
)
Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive, ,, )
{
0x000000B8,
}
})
Return (BUF0) /* \_SB_.I2C1.ETSA._CRS.BUF0 */
}
...
This interrupt is hard-coded to 0xB8 = 184 which is too high to be mapped
to IO-APIC, so no triggering information is propagated as acpi_register_gsi()
fails and irqresource_disabled() is issued, which leads to erasing triggering
and polarity information.
Do not overwrite flags as it leads to erasing triggering and polarity
information which might be useful in case of hard-coded interrupts.
This way the information can be read later on even though mapping to
APIC domain failed.
Signed-off-by: Angela Czubak <acz@semihalf.com>
[ rjw: Changelog rearrangement ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In some cases when firmware is busy or updating, some mailbox commands
still timeout on some newer CPUs. To fix this issue, change how we
process timeout.
With this change, replaced timeout from using simple count with real
timeout in micro-seconds using ktime. When the command response takes
more than average processing time, yield to other tasks. The worst case
timeout is extended upto 1 milli-second.
The current string size to print cpulist can accommodate upto 80
logical CPUs per package. But this limit is not enough. So increase
the string size. Also prevent buffer overflow, if the string size
reaches limit.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Having a button code and not a key code causes issues with libinput.
udev won't set ID_INPUT_KEY. If it is forced, then it causes a bug
within libinput.
Deinit the device on shutdown to halt MHI/PCI operation on device
side. This change fixes floating device state with some hosts that
do not fully shutdown PCIe device when rebooting.
If a channel was explicitly stopped but not reset and a driver
remove is issued, clean up the channel context such that it is
reflected on the device. This move is useful if a client driver
module is unloaded or a device crash occurs with the host having
placed the channel in a stopped state.
The same values are parsed several times from transfer and event
TRBs by different functions in the same call path, all while processing
one transfer event.
As the TRBs are in DMA memory and can be accessed by the xHC host we want
to avoid this to prevent double-fetch issues.
To resolve this pass the already parsed values to the different functions
in the path of parsing a transfer event
The Max Interrupters supported by the controller is given in a 10bit
wide bitfield, but the driver uses a fixed 128 size array to index these
interrupters.
Klockwork reports a possible array out of bounds case which in theory
is possible. In practice this hasn't been hit as a common number of Max
Interrupters for new controllers is 8, not even close to 128.
By virtue of using platform_irq_get_optional() under the covers,
platform_irq_count() needs the target interrupt controller to be
available and may return -EPROBE_DEFER if it isn't. Let's use
dev_err_probe() to avoid a spurious error log (and help debug any
deferral issues) in that case.
We sometimes see COMMAND_IGNORED responses during the clock stop
sequence. It turns out we already have information if devices are
present on a link, so we should only prepare those when they
are attached.
In addition, even when COMMAND_IGNORED are received, we should still
proceed with the clock stop. The device will not be prepared but
that's not a problem.
The only case where the clock stop will fail is if the Cadence IP
reports an error (including a timeout), or if the devices throw a
COMMAND_FAILED response.
When Secure World returns, it may have changed the size attribute of the
memory references passed as [in/out] parameters. The GlobalPlatform TEE
Internal Core API specification does not restrict the values that this
size can take. In particular, Secure World may increase the value to be
larger than the size of the input buffer to indicate that it needs more.
Therefore, the size check in optee_from_msg_param() is incorrect and
needs to be removed. This fixes a number of failed test cases in the
GlobalPlatform TEE Initial Configuratiom Test Suite v2_0_0_0-2017_06_09
when OP-TEE is compiled without dynamic shared memory support
(CFG_CORE_DYN_SHM=n).
Commit 99e71c029213 ("arm64: dts: imx8mq-librem5: Don't mark buck3 as always on")
removed always-on marking from GPU regulator, which is great for power
saving - however it introduces additional i2c0 traffic which can be deadly
for devices from the Dogwood batch.
To workaround the i2c0 shutdown issue on Dogwood, this commit marks
buck3 as always-on again - but only for Dogwood (r3).
Signed-off-by: Sebastian Krzyszkowiak <sebastian.krzyszkowiak@puri.sm> Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm> Signed-off-by: Shawn Guo <shawnguo@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The SW-initiated power gate toggling is dropped by PMC if there is
contention with a HW-initiated toggling, i.e. when one of CPU cores is
gated by cpuidle driver. Software should retry the toggling after 10
microseconds on Tegra20/30 SoCs, hence add the retrying. On Tegra114+ the
toggling method was changed in hardware, the TOGGLE_START bit indicates
whether PMC is busy or could accept the command to toggle, hence handle
that bit properly.
The problem pops up after enabling dynamic power gating of 3D hardware,
where 3D power domain fails to turn on/off "randomly".
The programming sequence and quirks are documented in TRMs, but PMC
driver obliviously re-used the Tegra20 logic for Tegra30+, which strikes
back now. The 10 microseconds and other timeouts aren't documented in TRM,
they are taken from downstream kernel.
When cross compiling x86 on an ARM machine with clang, there are several
errors along the lines of:
arch/x86/include/asm/page_64.h:52:7: error: invalid output constraint '=D' in asm
This happens because the x86 flags in the EFI stub are not derived from
KBUILD_CFLAGS like the other architectures are and the clang flags that
set the target architecture ('--target=') and the path to the GNU cross
tools ('--prefix=') are not present, meaning that the host architecture
is targeted.
These flags are available as $(CLANG_FLAGS) from the main Makefile so
add them to the cflags for x86 so that cross compiling works as expected.
When cross compiling x86 on an ARM machine with clang, there are several
errors along the lines of:
arch/x86/include/asm/string_64.h:27:10: error: invalid output constraint '=&c' in asm
This happens because the compressed boot Makefile reassigns KBUILD_CFLAGS
and drops the clang flags that set the target architecture ('--target=')
and the path to the GNU cross tools ('--prefix='), meaning that the host
architecture is targeted.
These flags are available as $(CLANG_FLAGS) from the main Makefile so
add them to the compressed boot folder's KBUILD_CFLAGS so that cross
compiling works as expected.
When cross-compiling with Clang, the `$(CLANG_FLAGS)' variable
contains additional flags needed to build C and assembly sources
for the target platform. Normally this variable is automatically
included in `$(KBUILD_CFLAGS)' via the top-level Makefile.
The x86 real-mode makefile builds `$(REALMODE_CFLAGS)' from a
plain assignment and therefore drops the Clang flags. This causes
Clang to not recognize x86-specific assembler directives:
Explicit propagation of `$(CLANG_FLAGS)' to `$(REALMODE_CFLAGS)',
which is inherited by real-mode make rules, fixes cross-compilation
with Clang for x86 targets.
Relevant flags:
* `--target' sets the target architecture when cross-compiling. This
flag must be set for both compilation and assembly (`KBUILD_AFLAGS')
to support architecture-specific assembler directives.
* `-no-integrated-as' tells clang to assemble with GNU Assembler
instead of its built-in LLVM assembler. This flag is set by default
unless `LLVM_IAS=1' is set, because the LLVM assembler can't yet
parse certain GNU extensions.
The TVK1281618 R3 sensors are different from the R2 board,
some incorrectness is fixed and some new sensors added, we
also rename the nodes appropriately with accelerometer@
etc.
This fixes warnings/errors like:
arch/arm/boot/dts/bcm4708-buffalo-wzr-1750dhp.dt.yaml: /: memory@0:reg:0: [0, 134217728, 2281701376, 402653184] is too long
From schema: /lib/python3.6/site-packages/dtschema/schemas/reg.yaml
To check whether the CPU and kernel support the MTE features we want
to test, we use an (emulated) CPU ID register read. However we only
check against a very particular feature version (0b0010), even though
the ARM ARM promises ID register features to be backwards compatible.
While this could be fixed by using ">=" instead of "==", we should
actually use the explicit HWCAP2_MTE hardware capability, exposed by the
kernel via the ELF auxiliary vectors.
That moves this responsibility to the kernel, and fixes running the
tests on machines with FEAT_MTE3 capability.
It should not be necessary to update the current_state field of
struct pci_dev in pci_enable_device_flags() before calling
do_pci_enable_device() for the device, because none of the
code between that point and the pci_set_power_state() call in
do_pci_enable_device() invoked later depends on it.
Moreover, doing that is actively harmful in some cases. For example,
if the given PCI device depends on an ACPI power resource whose _STA
method initially returns 0 ("off"), but the config space of the PCI
device is accessible and the power state retrieved from the
PCI_PM_CTRL register is D0, the current_state field in the struct
pci_dev representing that device will get out of sync with the
power.state of its ACPI companion object and that will lead to
power management issues going forward.
To avoid such issues it is better to leave the current_state value
as is until it is changed to PCI_D0 by do_pci_enable_device() as
appropriate. However, the power state of the device is not changed
to PCI_D0 if it is already enabled when pci_enable_device_flags()
gets called for it, so update its current_state in that case, but
use pci_update_current_state() covering platform PM too for that.
Link: https://lore.kernel.org/lkml/20210314000439.3138941-1-luzmaximilian@gmail.com/ Reported-by: Maximilian Luz <luzmaximilian@gmail.com> Tested-by: Maximilian Luz <luzmaximilian@gmail.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The mte selftest Makefile contains a check for GCC, to add the memtag
-march flag to the compiler options. This check fails if the compiler
is not explicitly specified, so reverts to the standard "cc", in which
case --version doesn't mention the "gcc" string we match against:
$ cc --version | head -n 1
cc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
This will not add the -march switch to the command line, so compilation
fails:
mte_helper.S: Assembler messages:
mte_helper.S:25: Error: selected processor does not support `irg x0,x0,xzr'
mte_helper.S:38: Error: selected processor does not support `gmi x1,x0,xzr'
...
Actually clang accepts the same -march option as well, so we can just
drop this check and add this unconditionally to the command line, to avoid
any future issues with this check altogether (gcc actually prints
basename(argv[0]) when called with --version).
Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Reviewed-by: Mark Brown <broone@kernel.org> Link: https://lore.kernel.org/r/20210319165334.29213-2-andre.przywara@arm.com Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Some hosts incorrectly use sub-minor version for minor version (i.e.
0x02 instead of 0x20 for bcdUSB 0x320 and 0x01 for bcdUSB 0x310).
Currently the xHCI driver works around this by just checking for minor
revision > 0x01 for USB 3.1 everywhere. With the addition of USB 3.2,
checking this gets a bit cumbersome. Since there is no USB release with
bcdUSB 0x301 to 0x309, we can assume that sub-minor version 01 to 09 is
incorrect. Let's try to fix this and use the minor revision that matches
with the USB/xHCI spec to help with the version checking within the
driver.
The current dwc3_gadget_reset_interrupt() will stop any active
transfers, but only addresses blocking of EP queuing for while we are
coming from a disconnected scenario, i.e. after receiving the disconnect
event. If the host decides to issue a bus reset on the device, the
connected parameter will still be set to true, allowing for EP queuing
to continue while we are disabling the functions. To avoid this, set the
connected flag to false until the stop active transfers is complete.
Currently user can configure UAC1 function with
parameters that violate UAC1 spec or are not supported
by UAC1 gadget implementation.
This can lead to incorrect behavior if such gadget
is connected to the host - like enumeration failure
or other issues depending on host's UAC1 driver
implementation, bringing user to a long hours
of debugging the issue.
Instead of silently accept these parameters, throw
an error if they are not valid.
Currently user can configure UAC2 function with
parameters that violate UAC2 spec or are not supported
by UAC2 gadget implementation.
This can lead to incorrect behavior if such gadget
is connected to the host - like enumeration failure
or other issues depending on host's UAC2 driver
implementation, bringing user to a long hours
of debugging the issue.
Instead of silently accept these parameters, throw
an error if they are not valid.
When irq_matrix_free() is called for an unallocated vector the
managed_allocated and total_allocated counters get out of sync with the
real state of the matrix. Later, when the last interrupt is freed, these
counters will underflow resulting in UINTMAX because the counters are
unsigned.
While this is certainly a problem of the calling code, this can be catched
in the allocator by checking the allocation bit for the to be freed vector
which simplifies debugging.
An example of the problem described above:
https://lore.kernel.org/lkml/20210318192819.636943062@linutronix.de/
Add the missing sanity check and emit a warning when it triggers.
Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/r/20210319111823.1105248-1-vkuznets@redhat.com Signed-off-by: Sasha Levin <sashal@kernel.org>
A malicious hypervisor could disable the CPUID intercept for an SEV or
SEV-ES guest and trick it into the no-SEV boot path, where it could
potentially reveal secrets. This is not an issue for SEV-SNP guests,
as the CPUID intercept can't be disabled for those.
Remove the Hypervisor CPUID bit check from the SEV detection code to
protect against this kind of attack and add a Hypervisor bit equals zero
check to the SME detection path to prevent non-encrypted guests from
trying to enable SME.
This handles the following cases:
1) SEV(-ES) guest where CPUID intercept is disabled. The guest
will still see leaf 0x8000001f and the SEV bit. It can
retrieve the C-bit and boot normally.
2) Non-encrypted guests with intercepted CPUID will check
the SEV_STATUS MSR and find it 0 and will try to enable SME.
This will fail when the guest finds MSR_K8_SYSCFG to be zero,
as it is emulated by KVM. But we can't rely on that, as there
might be other hypervisors which return this MSR with bit
23 set. The Hypervisor bit check will prevent that the guest
tries to enable SME in this case.
3) Non-encrypted guests on SEV capable hosts with CPUID intercept
disabled (by a malicious hypervisor) will try to boot into
the SME path. This will fail, but it is also not considered
a problem because non-encrypted guests have no protection
against the hypervisor anyway.
According with USB Device Class Definition for Video Device the
Processing Unit Descriptor bLength should be 12 (10 + bmControlSize),
but it has 11.
Invalid length caused that Processing Unit Descriptor Test Video form
CV tool failed. To fix this issue patch adds bmVideoStandards into
uvc_processing_unit_descriptor structure.
The bmVideoStandards field was added in UVC 1.1 and it wasn't part of
UVC 1.0a.
Patch adds extra checking for bInterval passed by configfs.
The 5.6.4 chapter of USB Specification (rev. 2.0) say:
"A high-bandwidth endpoint must specify a period of 1x125 µs
(i.e., a bInterval value of 1)."
The issue was observed during testing UVC class on CV.
I treat this change as improvement because we can control
bInterval by configfs.
Given that crypto_alloc_tfm() may return ERR pointers, and to avoid
crashes on obscure error paths where such pointers are presented to
crypto_destroy_tfm() (such as [0]), add an ERR_PTR check there
before dereferencing the second argument as a struct crypto_tfm
pointer.
In current design, whenever the BHI interrupt is fired, the
execution environment is updated. This can cause race conditions
and impede ongoing power up/down processing. For example, if a
power down is in progress, MHI host updates to a local "disabled"
execution environment. If a BHI interrupt fires later, that value
gets replaced with one from the BHI EE register. This impacts the
controller as it does not expect multiple RDDM execution
environment change status callbacks as an example. Another issue
would be that the device can enter mission mode and the execution
environment is updated, while device creation for SBL channels is
still going on due to slower PM state worker thread run, leading
to multiple attempts at opening the same channel.
Ensure that EE changes are handled only from appropriate places
and occur one after another and handle only PBL modes or RDDM EE
changes as critical events directly from the interrupt handler.
Simplify handling by waiting for SYS ERROR before handling RDDM.
This also makes sure that we use the correct execution environment
to notify the controller driver when the device resets to one of
the PBL execution environments.
Currently, client devices are created in SBL or AMSS (mission
mode) and only destroyed after power down or SYS ERROR. When
moving between certain execution environments, such as from SBL
to AMSS, no clean-up is required. This presents an issue where
SBL-specific channels are left open and client drivers now run in
an execution environment where they cannot operate. Fix this by
expanding the mhi_destroy_device() to do an execution environment
specific clean-up if one is requested. Close the gap and destroy
devices in such scenarios that allow SBL client drivers to clean
up once device enters mission mode.
The wake_db register presence is highly speculative and can fuze MHI
devices. Indeed, currently the wake_db register address is defined at
entry 127 of the 'Channel doorbell array', thus writing to this address
is equivalent to ringing the doorbell for channel 127, causing trouble
with some devics (e.g. SDX24 based modems) that get an unexpected
channel 127 doorbell interrupt.
This change fixes that issue by setting wake get/put as no-op for
pci_generic devices. The wake device sideband mechanism seems really
specific to each device, and is AFAIK not defined by the MHI spec.
It also removes zeroing initialization of wake_db register during MMIO
initialization, the register being set via wake_get/put accessors few
cycles later during M0 transition.
This removes the assignment of setup and cleanup functions for the ath79
target. Assigning the setup-method will lead to 'setup_transfer' not
being assigned in spi_bitbang_init. Because of this, performing any
TX/RX operation will lead to a kernel oops.
Also drop the redundant cleanup assignment, as it's also assigned in
spi_bitbang_init.
spi-bitbang has to call the chipselect function on the ath79 SPI driver
in order to communicate with the SPI slave device, as the ath79 SPI
driver has three dedicated chipselect lines but can also be used with
GPIOs for the CS lines.
Fixes commit 4a07b8bcd503 ("spi: bitbang: Make chipselect callback optional")
We want to probe l4_wkup and l4_cfg interconnect devices first to avoid
issues with missing resources. Otherwise we attempt to probe l4_per
devices first causing pointless deferred probe and also annoyingh
renumbering of the MMC devices for example.
Signed-off-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Trusted Foundation firmware doesn't implement the do_idle call and in
this case suspending should fall back to the common suspend path. In order
to fix this issue we will unconditionally set the NOFLUSH_L2 mode via
firmware call, which is a NO-OP on Tegra30/124, and then proceed to the
C7 idling, like it was done by the older Tegra114 cpuidle driver.
Fixes: 14e086baca50 ("cpuidle: tegra: Squash Tegra114 driver into the common driver") Cc: stable@vger.kernel.org # 5.7+ Reported-by: Anton Bambura <jenneron@protonmail.com> # TF701 T114 Tested-by: Anton Bambura <jenneron@protonmail.com> # TF701 T114 Tested-by: Matt Merhar <mattmerhar@protonmail.com> # Ouya T30 Tested-by: Peter Geis <pgwipeout@gmail.com> # Ouya T30 Signed-off-by: Dmitry Osipenko <digetx@gmail.com> Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> Link: https://lore.kernel.org/r/20210302095405.28453-1-digetx@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use kzalloc() rather than kmalloc() for the dynamically allocated parts
of the colormap in fb_alloc_cmap_gfp, to prevent a leak of random kernel
data to userspace under certain circumstances.
The return value on success (>= 0) is overwritten by the return value of
put_old_timex32(). That works correct in the fault case, but is wrong for
the success case where put_old_timex32() returns 0.
Just check the return value of put_old_timex32() and return -EFAULT in case
it is not zero.
[ tglx: Massage changelog ]
Fixes: 3a4d44b61625 ("ntp: Move adjtimex related compat syscalls to native counterparts") Signed-off-by: Chen Jun <chenjun102@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Richard Cochran <richardcochran@gmail.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20210414030449.90692-1-chenjun102@huawei.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For zoned btrfs, zone append is mandatory to write to a sequential write
only zone, otherwise parallel writes to the same zone could result in
unaligned write errors.
If a zoned block device does not support zone append (e.g. a dm-crypt
zoned device using a non-NULL IV cypher), fail to mount.
CC: stable@vger.kernel.org # 5.12 Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.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>
When doing a device replace on a zoned filesystem, if we find a block
group with ->to_copy == 0, we jump to the label 'done', which will result
in later calling btrfs_unfreeze_block_group(), even though at this point
we never called btrfs_freeze_block_group().
Since at this point we have neither turned the block group to RO mode nor
made any progress, we don't need to jump to the label 'done'. So fix this
by jumping instead to the label 'skip' and dropping our reference on the
block group before the jump.
Fixes: 78ce9fc269af6e ("btrfs: zoned: mark block groups to copy for device-replace") CC: stable@vger.kernel.org # 5.12 Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Filipe Manana <fdmanana@suse.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>
There is a race between a task aborting a transaction during a commit,
a task doing an fsync and the transaction kthread, which leads to an
use-after-free of the log root tree. When this happens, it results in a
stack trace like the following:
The steps that lead to this crash are the following:
1) We are at transaction N;
2) We have two tasks with a transaction handle attached to transaction N.
Task A and Task B. Task B is doing an fsync;
3) Task B is at btrfs_sync_log(), and has saved fs_info->log_root_tree
into a local variable named 'log_root_tree' at the top of
btrfs_sync_log(). Task B is about to call write_all_supers(), but
before that...
4) Task A calls btrfs_commit_transaction(), and after it sets the
transaction state to TRANS_STATE_COMMIT_START, an error happens before
it waits for the transaction's 'num_writers' counter to reach a value
of 1 (no one else attached to the transaction), so it jumps to the
label "cleanup_transaction";
5) Task A then calls cleanup_transaction(), where it aborts the
transaction, setting BTRFS_FS_STATE_TRANS_ABORTED on fs_info->fs_state,
setting the ->aborted field of the transaction and the handle to an
errno value and also setting BTRFS_FS_STATE_ERROR on fs_info->fs_state.
After that, at cleanup_transaction(), it deletes the transaction from
the list of transactions (fs_info->trans_list), sets the transaction
to the state TRANS_STATE_COMMIT_DOING and then waits for the number
of writers to go down to 1, as it's currently 2 (1 for task A and 1
for task B);
6) The transaction kthread is running and sees that BTRFS_FS_STATE_ERROR
is set in fs_info->fs_state, so it calls btrfs_cleanup_transaction().
There it sees the list fs_info->trans_list is empty, and then proceeds
into calling btrfs_drop_all_logs(), which frees the log root tree with
a call to btrfs_free_log_root_tree();
7) Task B calls write_all_supers() and, shortly after, under the label
'out_wake_log_root', it deferences the pointer stored in
'log_root_tree', which was already freed in the previous step by the
transaction kthread. This results in a use-after-free leading to a
crash.
Fix this by deleting the transaction from the list of transactions at
cleanup_transaction() only after setting the transaction state to
TRANS_STATE_COMMIT_DOING and waiting for all existing tasks that are
attached to the transaction to release their transaction handles.
This makes the transaction kthread wait for all the tasks attached to
the transaction to be done with the transaction before dropping the
log roots and doing other cleanups.
Fixes: ef67963dac255b ("btrfs: drop logs when we've aborted a transaction") CC: stable@vger.kernel.org # 5.10+ Reviewed-by: Josef Bacik <josef@toxicpanda.com> 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>
When creating a subvolume we allocate an extent buffer for its root node
after starting a transaction. We setup a root item for the subvolume that
points to that extent buffer and then attempt to insert the root item into
the root tree - however if that fails, due to ENOMEM for example, we do
not free the extent buffer previously allocated and we do not abort the
transaction (as at that point we did nothing that can not be undone).
This means that we effectively do not return the metadata extent back to
the free space cache/tree and we leave a delayed reference for it which
causes a metadata extent item to be added to the extent tree, in the next
transaction commit, without having backreferences. When this happens
'btrfs check' reports the following:
$ btrfs check /dev/sdi
Opening filesystem to check...
Checking filesystem on /dev/sdi
UUID: dce2cb9d-025f-4b05-a4bf-cee0ad3785eb
[1/7] checking root items
[2/7] checking extents
ref mismatch on [30425088 16384] extent item 1, found 0
backref 30425088 root 256 not referenced back 0x564a91c23d70
incorrect global backref count on 30425088 found 1 wanted 0
backpointer mismatch on [30425088 16384]
owner ref check failed [30425088 16384]
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 212992 bytes used, error(s) found
total csum bytes: 0
total tree bytes: 131072
total fs tree bytes: 32768
total extent tree bytes: 16384
btree space waste bytes: 124669
file data blocks allocated: 65536
referenced 65536
So fix this by freeing the metadata extent if btrfs_insert_root() returns
an error.
CC: stable@vger.kernel.org # 4.4+ Signed-off-by: Filipe Manana <fdmanana@suse.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>
Fix a regression caused by making the 486SX separately selectable in
Kconfig, for which the HIGHMEM64G setting has not been updated and
therefore has become exposed as a user-selectable option for the M486SX
configuration setting unlike with original M486 and all the other
settings that choose non-PAE-enabled processors:
High Memory Support
> 1. off (NOHIGHMEM)
2. 4GB (HIGHMEM4G)
3. 64GB (HIGHMEM64G)
choice[1-3?]:
With the fix in place the setting is now correctly removed:
High Memory Support
> 1. off (NOHIGHMEM)
2. 4GB (HIGHMEM4G)
choice[1-2?]:
[ bp: Massage commit message. ]
Fixes: 87d6021b8143 ("x86/math-emu: Limit MATH_EMULATION to 486SX compatibles") Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk> Signed-off-by: Borislav Petkov <bp@suse.de> Cc: stable@vger.kernel.org # v5.5+ Link: https://lkml.kernel.org/r/alpine.DEB.2.21.2104141221340.44318@angie.orcam.me.uk Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is already after the patch "btrfs: inode: fix NULL pointer
dereference if inode doesn't need compression."
[CAUSE]
@pages is firstly created by kcalloc() in compress_file_extent():
pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
Then passed to btrfs_compress_pages() to be utilized there:
ret = btrfs_compress_pages(...
pages,
&nr_pages,
...);
btrfs_compress_pages() will initialize each page as output, in
zlib_compress_pages() we have:
pages[nr_pages] = out_page;
nr_pages++;
Normally this is completely fine, but there is a special case which
is in btrfs_compress_pages() itself:
switch (type) {
default:
return -E2BIG;
}
In this case, we didn't modify @pages nor @out_pages, leaving them
untouched, then when we cleanup pages, the we can hit NULL pointer
dereference again:
if (pages) {
for (i = 0; i < nr_pages; i++) {
WARN_ON(pages[i]->mapping);
put_page(pages[i]);
}
...
}
Since pages[i] are all initialized to zero, and btrfs_compress_pages()
doesn't change them at all, accessing pages[i]->mapping would lead to
NULL pointer dereference.
This is not possible for current kernel, as we check
inode_need_compress() before doing pages allocation.
But if we're going to remove that inode_need_compress() in
compress_file_extent(), then it's going to be a problem.
[FIX]
When btrfs_compress_pages() hits its default case, modify @out_pages to
0 to prevent such problem from happening.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=212331 CC: stable@vger.kernel.org # 5.10+ Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Qu Wenruo <wqu@suse.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>
* rqst[1,2,3] is allocated in vars
* each rqst->rq_iov is also allocated in vars or using pooled memory
SMB2_open_free, SMB2_ioctl_free, SMB2_query_info_free are iterating on
each rqst after vars has been freed (use-after-free), and they are
freeing the kvec a second time (double-free).
==================================================================
BUG: KASAN: use-after-free in SMB2_open_free+0x1c/0xa0
Read of size 8 at addr ffff888007b10c00 by task python3/1200
Memory state around the buggy address: ffff888007b10b00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff888007b10b80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff888007b10c00: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
^ ffff888007b10c80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb ffff888007b10d00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Signed-off-by: Aurelien Aptel <aaptel@suse.com> CC: <stable@vger.kernel.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 315db9a05b7a ("cifs: fix leak in cifs_smb3_do_mount() ctx")
revealed an existing bug when mounting shares that contain a prefix
path or DFS links.
cifs_setup_volume_info() requires the @devname to contain the full
path (UNC + prefix) to update the fs context with the new UNC and
prepath values, however we were passing only the UNC
path (old_ctx->UNC) in @device thus discarding any prefix paths.
Instead of concatenating both old_ctx->{UNC,prepath} and pass it in
@devname, just keep the dup'ed values of UNC and prepath in
cifs_sb->ctx after calling smb3_fs_context_dup(), and fix
smb3_parse_devname() to correctly parse and not leak the new UNC and
prefix paths.
Cc: <stable@vger.kernel.org> # v5.11+ Fixes: 315db9a05b7a ("cifs: fix leak in cifs_smb3_do_mount() ctx") Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Acked-by: David Disseldorp <ddiss@suse.de> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We can detect server unresponsiveness only if echoes are enabled.
Echoes can be disabled under two scenarios:
1. The connection is low on credits, so we've disabled echoes/oplocks.
2. The connection has not seen any request till now (other than
negotiate/sess-setup), which is when we enable these two, based on
the credits available.
So this fix will check for dead connection, only when echo is enabled.
Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> CC: <stable@vger.kernel.org> # v5.8+ Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cifs_smb3_do_mount() calls smb3_fs_context_dup() and then
cifs_setup_volume_info(). The latter's subsequent smb3_parse_devname()
call overwrites the cifs_sb->ctx->UNC string already dup'ed by
smb3_fs_context_dup(), resulting in a leak. E.g.
This change is a bandaid until the cifs_setup_volume_info() TODO and
error handling issues are resolved.
Signed-off-by: David Disseldorp <ddiss@suse.de> Acked-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> CC: <stable@vger.kernel.org> # v5.11+ Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If smb3_notify() is called at mount point of CIFS, build_path_from_dentry()
returns the pointer to kmalloc-ed memory with terminating zero (this is
empty FileName to be passed to SMB2 CREATE request). This pointer is assigned
to the `path` variable.
Then `path + 1` (to skip first backslash symbol) is passed to
cifs_convert_path_to_utf16(). This is incorrect for empty path and causes
out-of-bound memory access.
Get rid of this "increase by one". cifs_convert_path_to_utf16() already
contains the check for leading backslash in the path.
BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=212693 CC: <stable@vger.kernel.org> # v5.6+ Signed-off-by: Eugene Korenevsky <ekorenevsky@astralinux.ru> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[440700.376476] CIFS VFS: \\otters.example.com crypt_message: Could not get encryption key
[440700.386947] ------------[ cut here ]------------
[440700.386948] err = 1
[440700.386977] WARNING: CPU: 11 PID: 2733 at /build/linux-hwe-5.4-p6lk6L/linux-hwe-5.4-5.4.0/lib/errseq.c:74 errseq_set+0x5c/0x70
...
[440700.397304] CPU: 11 PID: 2733 Comm: tar Tainted: G OE 5.4.0-70-generic #78~18.04.1-Ubuntu
...
[440700.397334] Call Trace:
[440700.397346] __filemap_set_wb_err+0x1a/0x70
[440700.397419] cifs_writepages+0x9c7/0xb30 [cifs]
[440700.397426] do_writepages+0x4b/0xe0
[440700.397444] __filemap_fdatawrite_range+0xcb/0x100
[440700.397455] filemap_write_and_wait+0x42/0xa0
[440700.397486] cifs_setattr+0x68b/0xf30 [cifs]
[440700.397493] notify_change+0x358/0x4a0
[440700.397500] utimes_common+0xe9/0x1c0
[440700.397510] do_utimes+0xc5/0x150
[440700.397520] __x64_sys_utimensat+0x88/0xd0
Fixes: 61cfac6f267d ("CIFS: Fix possible use after free in demultiplex thread") Signed-off-by: Paul Aurich <paul@darkrain42.org> CC: stable@vger.kernel.org Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
which is cause by a 'BUG_ON(in_nmi())' in nmi_enter().
From the call trace, we can find three interrupts (noted A, B, C above):
interrupt (A) is preempted by (B), which is further interrupted by (C).
Subsequent investigations show that (B) results in nmi_enter() being
called, but that it actually is a spurious interrupt. Furthermore,
interrupts are reenabled in the context of (B), and (C) fires with
NMI priority. We end-up with a nested NMI situation, something
we definitely do not want to (and cannot) handle.
The bug here is that spurious interrupts should never result in any
state change, and we should just return to the interrupted context.
Moving the handling of spurious interrupts as early as possible in
the GICv3 handler fixes this issue.
Fixes: 3f1f3234bc2d ("irqchip/gic-v3: Switch to PMR masking before calling IRQ handler") Acked-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: He Ying <heying24@huawei.com>
[maz: rewrote commit message, corrected Fixes: tag] Signed-off-by: Marc Zyngier <maz@kernel.org> Link: https://lore.kernel.org/r/20210423083516.170111-1-heying24@huawei.com Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The mmc core uses a PM notifier to temporarily during system suspend, turn
off the card detection mechanism for removal/insertion of (e)MMC/SD/SDIO
cards. Additionally, the notifier may be used to remove an SDIO card
entirely, if a corresponding SDIO functional driver don't have the system
suspend/resume callbacks assigned. This behaviour has been around for a
very long time.
However, a recent bug report tells us there are problems with this
approach. More precisely, when receiving the PM_SUSPEND_PREPARE
notification, we may end up hanging on I/O to be completed, thus also
preventing the system from getting suspended.
In the end what happens, is that the cancel_delayed_work_sync() in
mmc_pm_notify() ends up waiting for mmc_rescan() to complete - and since
mmc_rescan() wants to claim the host, it needs to wait for the I/O to be
completed first.
Typically, this problem is triggered in Android, if there is ongoing I/O
while the user decides to suspend, resume and then suspend the system
again. This due to that after the resume, an mmc_rescan() work gets punted
to the workqueue, which job is to verify that the card remains inserted
after the system has resumed.
To fix this problem, userspace needs to become frozen to suspend the I/O,
prior to turning off the card detection mechanism. Therefore, let's drop
the PM notifiers for mmc subsystem altogether and rely on the card
detection to be turned off/on as a part of the system_freezable_wq, that we
are already using.
Moreover, to allow and SDIO card to be removed during system suspend, let's
manage this from a ->prepare() callback, assigned at the mmc_host_class
level. In this way, we can use the parent device (the mmc_host_class
device), to remove the card device that is the child, in the
device_prepare() phase.
Some of SD cards sets permanent write protection bit in their CSD register,
due to lifespan or internal problem. To avoid unnecessary I/O write
operations, let's parse the bits in the CSD during initialization and mark
the card as read only for this case.
A CMD11 is sent to the SD/SDIO card to start the voltage switch procedure
into 1.8V I/O. According to the SD spec a power cycle is needed of the
card, if it turns out that the CMD11 fails. Let's fix this, to allow a
retry of the initialization without the voltage switch, to succeed.
Note that, whether it makes sense to also retry with the voltage switch
after the power cycle is a bit more difficult to know. At this point, we
treat it like the CMD11 isn't supported and therefore we skip it when
retrying.
In command queueing mode, the cache isn't flushed via the mmc_flush_cache()
function, but instead by issuing a CMDQ_TASK_MGMT (CMD48) with a
FLUSH_CACHE opcode. In this path, we need to check if cache has been
enabled, before deciding to flush the cache, along the lines of what's
being done in mmc_flush_cache().
To fix this problem, let's add a new bus ops callback ->cache_enabled() and
implement it for the mmc bus type. In this way, the mmc block device driver
can call it to know whether cache flushing should be done.
Fixes: 1e8e55b67030 (mmc: block: Add CQE support) Cc: stable@vger.kernel.org Reported-by: Brendan Peter <bpeter@lytx.com> Signed-off-by: Avri Altman <avri.altman@wdc.com> Tested-by: Brendan Peter <bpeter@lytx.com> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Link: https://lore.kernel.org/r/20210425060207.2591-2-avri.altman@wdc.com Link: https://lore.kernel.org/r/20210425060207.2591-3-avri.altman@wdc.com
[Ulf: Squashed the two patches and made some minor updates] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The cache function can be turned ON and OFF by writing to the CACHE_CTRL
byte (EXT_CSD byte [33]). However, card->ext_csd.cache_ctrl is only
set on init if cache size > 0.
Fix that by explicitly setting ext_csd.cache_ctrl on ext-csd write.
CMD8 is not supported with Command Queue Enabled. Add required callback
to clear CQE_EN and CQE_INTR fields in the host controller register
before sending CMD8. Add corresponding callback in the CQHCI resume path
to re-enable CQE_EN and CQE_INTR fields.
Reported-by: Kamal Mostafa <kamal@canonical.com> Tested-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Aniruddha Tvs Rao <anrao@nvidia.com> Signed-off-by: Jon Hunter <jonathanh@nvidia.com> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Thierry Reding <treding@nvidia.com> Link: https://lore.kernel.org/r/20210407094617.770495-1-jonathanh@nvidia.com Cc: stable@vger.kernel.org # v5.10+ Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Bus power may control card power, but the full reset done by SDHCI at
initialization still may not reset the power, whereas a direct write to
SDHCI_POWER_CONTROL can. That might be needed to initialize correctly, if
the card was left powered on previously.
For data read commands, SDHC may initiate data transfers even before it
completely process the command response. In case command itself fails,
driver un-maps the memory associated with data transfer but this memory
can still be accessed by SDHC for the already initiated data transfer.
This scenario can lead to un-mapped memory access error.
To avoid this scenario, reset SDHC (when command fails) prior to
un-mapping memory. Resetting SDHC ensures that all in-flight data
transfers are either aborted or completed. So we don't run into this
scenario.
Swap the reset, un-map steps sequence in sdhci_request_done().
Suggested-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org> Signed-off-by: Pradeep P V K <pragalla@codeaurora.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Link: https://lore.kernel.org/r/1614760331-43499-1-git-send-email-pragalla@qti.qualcomm.com Cc: stable@vger.kernel.org # v4.9+ Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A 'tmio_mmc_host_free()' call is missing in the remove function, in order
to balance a 'tmio_mmc_host_alloc()' call in the probe.
This is done in the error handling path of the probe, but not in the remove
function.
A 'uniphier_sd_clk_enable()' call should be balanced by a corresponding
'uniphier_sd_clk_disable()' call.
This is done in the remove function, but not in the error handling path of
the probe.
While diag reset is in progress there is short duration where all access to
controller's PCI config space from the host needs to be blocked. This is
due to a hardware limitation of the IOC controllers.
Block all access to controller's config space from userland applications by
calling pci_cfg_access_lock() while diag reset is in progress and unlocking
it again after the controller comes back to ready state.
Link: https://lore.kernel.org/r/20210330105137.20728-1-sreekanth.reddy@broadcom.com Cc: stable@vger.kernel.org #v5.4.108+ Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Whenever the driver is adding a vSES to virtual-phys list it is
reinitializing the list head. Hence those vSES devices which were added
previously are lost.
Stop reinitializing the list every time a new vSES device is added.
Link: https://lore.kernel.org/r/20210330105004.20413-1-sreekanth.reddy@broadcom.com Cc: stable@vger.kernel.org #v5.11.10+ Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rmmod on SLI-4 adapters is sometimes hitting a bad ptr dereference in
lpfc_els_free_iocb().
A prior patch refactored the lpfc_sli_abort_iocb() routine. One of the
changes was to convert from building/sending an abort within the routine to
using a common routine. The reworked routine passes, without modification,
the pring ptr to the new common routine. The older routine had logic to
check SLI-3 vs SLI-4 and adapt the pring ptr if necessary as callers were
passing SLI-3 pointers even when not on an SLI-4 adapter. The new routine
is missing this check and adapt, so the SLI-3 ring pointers are being used
in SLI-4 paths.
Fix by cleaning up the calling routines. In review, there is no need to
pass the ring ptr argument to abort_iocb at all. The routine can look at
the adapter type itself and reference the proper ring.
Link: https://lore.kernel.org/r/20210412013127.2387-2-jsmart2021@gmail.com Fixes: db7531d2b377 ("scsi: lpfc: Convert abort handling to SLI-3 and SLI-4 handlers") Cc: <stable@vger.kernel.org> # v5.11+ Co-developed-by: Justin Tee <justin.tee@broadcom.com> Signed-off-by: Justin Tee <justin.tee@broadcom.com> Signed-off-by: James Smart <jsmart2021@gmail.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit a6dcfe08487e ("scsi: qla2xxx: Limit interrupt vectors to number of
CPUs") lowers the number of allocated MSI-X vectors to the number of CPUs.
That breaks vector allocation assumptions in qla83xx_iospace_config(),
qla24xx_enable_msix() and qla2x00_iospace_config(). Either of the functions
computes maximum number of qpairs as:
ha->max_qpairs = ha->msix_count - 1 (MB interrupt) - 1 (default
response queue) - 1 (ATIO, in dual or pure target mode)
max_qpairs is set to zero in case of two CPUs and initiator mode. The
number is then used to allocate ha->queue_pair_map inside
qla2x00_alloc_queues(). No allocation happens and ha->queue_pair_map is
left NULL but the driver thinks there are queue pairs available.
qla2xxx_queuecommand() tries to find a qpair in the map and crashes:
Normally, an unused OSD id/slot is represented by an empty addrvec.
However, it also appears to be possible to generate an osdmap where
an unused OSD id/slot has an addrvec with a single blank address of
type NONE. Allow such addrvecs and make the end result be exactly
the same as for the empty addrvec case -- leave addr intact.
A dummy v3 encoding (exactly the same as v2) was introduced so that
the monitors can distinguish broken clients that may not include their
auth ticket in CEPHX_GET_AUTH_SESSION_KEY request on reconnects, thus
failing to prove previous possession of their global_id (one part of
CVE-2021-20288).
The kernel client has always included its auth ticket, so it is
compatible with enforcing mode as is. However we want to bump the
encoding version to avoid having to authenticate twice on the initial
connect -- all legacy (CephXAuthenticate < v3) are now forced do so in
order to expose insecure global_id reclaim.
Marking for stable since at least for 5.11 and 5.12 it is trivial
(v2 -> v3).
pm_runtime usage_count counter is not well managed.
pm_runtime_put_autosuspend callback drops the usage_counter but this
one has never been increased. Add pm_runtime_get_sync callback to bump up
the usage counter. It is also needed to use pm_runtime_force_suspend and
pm_runtime_force_resume APIs to handle properly the clock.
Cast &data to (char *) in order to avoid unintentionally accessing
the stack.
Notice that data is of type u32, so any increment to &data
will be in the order of 4-byte chunks, and this piece of code
is actually intended to be a byte offset.
The module misses MODULE_DEVICE_TABLE() for both SPI and OF ID tables
and thus never autoloads on ID matches.
Add the missing declarations.
Present since day-0 of spinand framework introduction.
Fixes: 7529df465248 ("mtd: nand: Add core infrastructure to support SPI NANDs") Cc: stable@vger.kernel.org # 4.19+ Signed-off-by: Alexander Lobakin <alobakin@pm.me> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/linux-mtd/20210323173714.317884-1-alobakin@pm.me Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mx25l51245g and mx66l51235l have the same flash ID. The flash
detection returns the first entry in the flash_info array that
matches the flash ID that was read, thus for the 0xc2201a ID,
mx25l51245g was always hit, introducing a regression for
mx66l51235l.
If one wants to differentiate the flash names, a better fix would be
to differentiate between the two at run-time, depending on SFDP,
and choose the correct name from a list of flash names, depending on
the SFDP differentiator.
If rmmod the driver during read or write, the driver will release the
resources which are used during read or write, so it is possible to
refer to NULL pointer.
Use the testcase "mtd_debug read /dev/mtd0 0xc00000 0x400000 dest_file &
sleep 0.5;rmmod spi_hisi_sfc_v3xx.ko", the issue can be reproduced in
hisi_sfc_v3xx driver.
To avoid the issue, fill the interface _get_device and _put_device of
mtd_info to grab the reference to the spi controller driver module, so
the request of rmmod the driver is rejected before read/write is finished.
Fixes: b199489d37b2 ("mtd: spi-nor: add the framework for SPI NOR") Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com> Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Tested-by: Michael Walle <michael@walle.cc> Tested-by: Tudor Ambarus <tudor.ambarus@microchip.com> Reviewed-by: Michael Walle <michael@walle.cc> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/1617262486-4223-1-git-send-email-yangyicong@hisilicon.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 339ddb53d373 ("fs/epoll: remove unnecessary wakeups of nested
epoll") changed the userspace visible behavior of exclusive waiters
blocked on a common epoll descriptor upon a single event becoming ready.
Previously, all tasks doing epoll_wait would awake, and now only one is
awoken, potentially causing missed wakeups on applications that rely on
this behavior, such as Apache Qpid.
While the aforementioned commit aims at having only a wakeup single path
in ep_poll_callback (with the exceptions of epoll_ctl cases), we need to
restore the wakeup in what was the old ep_scan_ready_list() such that
the next thread can be awoken, in a cascading style, after the waker's
corresponding ep_send_events().
Link: https://lkml.kernel.org/r/20210405231025.33829-3-dave@stgolabs.net Fixes: 339ddb53d373 ("fs/epoll: remove unnecessary wakeups of nested epoll") Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Jason Baron <jbaron@akamai.com> Cc: Roman Penyaev <rpenyaev@suse.de> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When mounting eCryptfs, a null "dev_name" argument to ecryptfs_mount()
causes a kernel panic if the parsed options are valid. The easiest way to
reproduce this is to call mount() from userspace with an existing
eCryptfs mount's options and a "source" argument of 0.
Error out if "dev_name" is null in ecryptfs_mount()
The LLVM ld.lld linker uses a different symbol type for __bss_start,
resulting in the calculation of KBSS_SZ to be thrown off. Up until now,
this has gone unnoticed as it only affects the appended DTB case, but
pending changes for ARM in the way the decompressed kernel is cleaned
from the caches has uncovered this problem.
On a ld.lld build:
$ nm vmlinux |grep bss_ c1c22034 D __bss_start c1c86e98 B __bss_stop
which is obviously incorrect, and may cause the cache clean to access
unmapped memory, or cause the size calculation to wrap, resulting in no
cache clean to be performed at all.
Fix this by updating the sed regex to take D type symbols into account.
The reason is that the parsing in the write function only processes
commands if it finished parsing (there is white space written after the
command). That's to handle:
cases, where the command is broken over multiple writes.
The problem is if the file descriptor is closed, then the write call is
not processed, and the command needs to be processed in the release code.
The release code can handle matching of functions, but does not handle
commands.
Cc: stable@vger.kernel.org Fixes: eda1e32855656 ("tracing: handle broken names in ftrace filter") Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In cm_write(), if the 'buf' is allocated memory but not fully consumed,
it is possible to reallocate the buffer without freeing it by passing
'*ppos' as 0 on a subsequent call.
Add an explicit kfree() before kzalloc() to prevent the possible memory
leak.
Fixes: 526b4af47f44 ("ACPI: Split out custom_method functionality into an own driver") Signed-off-by: Mark Langsdorf <mlangsdo@redhat.com> Cc: 5.4+ <stable@vger.kernel.org> # 5.4+ Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / thirdparty / kernel / stable.git / log
