show_opcodes() is used both for dumping kernel instructions and for dumping
user instructions. If userspace causes #PF by jumping to a kernel address,
show_opcodes() can be reached with regs->ip controlled by the user,
pointing to kernel code. Make sure that userspace can't trick us into
dumping kernel memory into dmesg.
Like d88b6d04: "cdrom: information leak in cdrom_ioctl_media_changed()"
There is another cast from unsigned long to int which causes
a bounds check to fail with specially crafted input. The value is
then used as an index in the slot array in cdrom_slot_status().
Signed-off-by: Scott Bauer <scott.bauer@intel.com> Signed-off-by: Scott Bauer <sbauer@plzdonthack.me> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some architectures need to use stop_machine() to patch functions for
ftrace, and the assumption is that the stopped CPUs do not make function
calls to traceable functions when they are in the stopped state.
Commit ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after
MULTI_STOP_PREPARE") added calls to the watchdog touch functions from
the stopped CPUs and those functions lack notrace annotations. This
leads to crashes when enabling/disabling ftrace on ARM kernels built
with the Thumb-2 instruction set.
Fix it by adding the necessary notrace annotations.
Fixes: ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after MULTI_STOP_PREPARE") Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: oleg@redhat.com Cc: tj@kernel.org Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180821152507.18313-1-vincent.whitchurch@axis.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We did have sporadic problems in the pinctrl framework during boot
where a pin group name unexpectedly became NULL leading to a NULL
dereference in strcmp.
Detailled analysis of the failing cases did reveal that there were
two devm allocated objects close to each other. The second one was
the affected group_desc in pinmux and the first one was the
psy_desc->properties buffer of the gab driver.
Review of the gab code showed that the address calculation for
one memcpy() is wrong. It does
properties + sizeof(type) * index
but C is defined to do the index multiplication already for
pointer + integer additions. Hence the factor was applied twice
and the memcpy() does write outside of the properties buffer.
Sometimes it happened to be the pinctrl and triggered the strcmp(NULL).
Anyways, it is overkill to use a memcpy() here instead of a simple
assignment, which is easier to read and has less risk for wrong
address calculations. So we change code to a simple assignment.
If we initialize the index to the first free location, we can even
remove the local variable 'properties'.
This bug seems to exist right from the beginning in 3.7-rc1 in
commit e60fea794e6e ("power: battery: Generic battery driver using IIO")
Signed-off-by: H. Nikolaus Schaller <hns@goldelico.com> Cc: stable@vger.kernel.org Fixes: e60fea794e6e ("power: battery: Generic battery driver using IIO") Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
"count" needs to be signed for the error handling to work. I made "i"
signed as well so they match.
Fixes: 02113ba93ea4 (PM / clk: Add support for obtaining clocks from device-tree) Cc: 4.6+ <stable@vger.kernel.org> # 4.6+ Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:
The BAM has 3 channels - tx, rx and command. command channel
is used for register read/writes, tx channel for data writes
and rx channel for data reads. Currently, the driver assumes the
transfer completion once it gets all the command descriptors
completed. Sometimes, there is race condition between data channel
(tx/rx) and command channel completion. In these cases,
the data present in buffer is not valid during small window
between command descriptor completion and data descriptor
completion.
This patch generates NAND transfer completion when both
(Data and Command) DMA channels have completed all its DMA
descriptors. It assigns completion callback in last
DMA descriptors of that channel and wait for completion.
This patch restores the suspend and resume hooks that the old driver used
to have. Apart from stopping and starting the clocks, the resume callback
also nullifies the selected_chip pointer, so the next command that is issued
will re-select the chip and thereby restore the timing registers.
Factor out some code from marvell_nfc_init() into a new function
marvell_nfc_reset() and also call it at resume time to reset some registers
that don't retain their contents during low-power mode.
Without this patch, a PXA3xx based system would cough up an error similar to
the one below after resume.
[ 44.660162] marvell-nfc 43100000.nand-controller: Timeout waiting for RB signal
[ 44.671492] ubi0 error: ubi_io_write: error -110 while writing 2048 bytes to PEB 102:38912, written 0 bytes
[ 44.682887] CPU: 0 PID: 1417 Comm: remote-control Not tainted 4.18.0-rc2+ #344
[ 44.691197] Hardware name: Marvell PXA3xx (Device Tree Support)
[ 44.697111] Backtrace:
[ 44.699593] [<c0106458>] (dump_backtrace) from [<c0106718>] (show_stack+0x18/0x1c)
[ 44.708931] r7:00000800 r6:00009800 r5:00000066 r4:c6139000
[ 44.715833] [<c0106700>] (show_stack) from [<c0678a60>] (dump_stack+0x20/0x28)
[ 44.724206] [<c0678a40>] (dump_stack) from [<c0456cbc>] (ubi_io_write+0x3d4/0x630)
[ 44.732925] [<c04568e8>] (ubi_io_write) from [<c0454428>] (ubi_eba_write_leb+0x690/0x6fc)
...
chip->read_buf is left unassigned since commit 4da712e70294 ("mtd: nand:
fsmc: use ->exec_op()"), leading to a NULL pointer dereference when it's
called from fsmc_read_page_hwecc(). Fix that by using the appropriate
helper to read data out of the NAND.
Modern NAND controller drivers implement ->exec_op() instead of
->cmdfunc(), make sure we don't end up with a NULL pointer dereference
when hynix_nand_reg_write_op() is called.
The problem is that iscsi_login_zero_tsih_s1 sets conn->sess early in
iscsi_login_set_conn_values. If the function fails later like when we
alloc the idr it does kfree(sess) and leaves the conn->sess pointer set.
iscsi_login_zero_tsih_s1 then returns -Exyz and we then call
iscsi_target_login_sess_out and access the freed memory.
This patch has iscsi_login_zero_tsih_s1 either completely setup the
session or completely tear it down, so later in
iscsi_target_login_sess_out we can just check for it being set to the
connection.
Cc: stable@vger.kernel.org Fixes: 0957627a9960 ("iscsi-target: Fix sess allocation leak in...") Signed-off-by: Mike Christie <mchristi@redhat.com> Acked-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Matthew Wilcox <willy@infradead.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A long time ago the unfortunate decision was taken to add a self-deletion
attribute to the sysfs SCSI device directory. That decision was unfortunate
because self-deletion is really tricky. We can't drop that attribute
because widely used user space software depends on it, namely the
rescan-scsi-bus.sh script. Hence this patch that avoids that writing into
that attribute triggers a deadlock. See also commit 7973cbd9fbd9 ("[PATCH]
add sysfs attributes to scan and delete scsi_devices").
This patch avoids that self-removal triggers the following deadlock:
======================================================
WARNING: possible circular locking dependency detected
4.18.0-rc2-dbg+ #5 Not tainted
------------------------------------------------------
modprobe/6539 is trying to acquire lock: 000000008323c4cd (kn->count#202){++++}, at: kernfs_remove_by_name_ns+0x45/0x90
but task is already holding lock: 00000000a6ec2c69 (&shost->scan_mutex){+.+.}, at: scsi_remove_host+0x21/0x150 [scsi_mod]
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
This patch avoids that smatch complains about a double unlock on
ioc->transport_cmds.mutex.
Fixes: 651a01364994 ("scsi: scsi_transport_sas: switch to bsg-lib for SMP passthrough") Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Sathya Prakash <sathya.prakash@broadcom.com> Cc: Chaitra P B <chaitra.basappa@broadcom.com> Cc: Suganath Prabu Subramani <suganath-prabu.subramani@broadcom.com> Cc: stable@vger.kernel.org Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As a part of host reset operation, driver will flushout all IOs outstanding
at driver level with "DID_RESET" result. To find which are all commands
outstanding at the driver level, driver loops with smid starting from one
to HBA queue depth and calls mpt3sas_scsih_scsi_lookup_get() to get scmd as
shown below
for (smid = 1; smid <= ioc->scsiio_depth; smid++) {
scmd = mpt3sas_scsih_scsi_lookup_get(ioc, smid);
if (!scmd)
continue;
But in mpt3sas_scsih_scsi_lookup_get() function, driver returns some scsi
cmnds which are not outstanding at the driver level (possibly request is
constructed at block layer since QUEUE_FLAG_QUIESCED is not set. Even if
driver uses scsi_block_requests and scsi_unblock_requests, issue still
persists as they will be just blocking further IO from scsi layer and not
from block layer) and these commands are flushed with DID_RESET host bytes
thus resulting into above kernel BUG.
This issue got introduced by commit dbec4c9040ed ("scsi: mpt3sas: lockless
command submission").
To fix this issue, we have modified the mpt3sas_scsih_scsi_lookup_get() to
check for smid equals to zero (note: whenever any scsi cmnd is processing
at the driver level then smid for that scsi cmnd will be non-zero, always
it starts from one) before it returns the scmd pointer to the caller. If
smid is zero then this function returns scmd pointer as NULL and driver
won't flushout those scsi cmnds at driver level with DID_RESET host byte
thus this issue will not be observed.
[mkp: amended with updated fix from Sreekanth]
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Fixes: dbec4c9040ed ("scsi: mpt3sas: lockless command submission") Cc: stable@vger.kernel.org # v4.16+ Reviewed-by: Tomas Henzl <thenzl@redhat.com> Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix tpm ptt initialization error:
tpm tpm0: A TPM error (378) occurred get tpm pcr allocation.
We cannot use go_idle cmd_ready commands via runtime_pm handles
as with the introduction of localities this is no longer an optional
feature, while runtime pm can be not enabled.
Though cmd_ready/go_idle provides a power saving, it's also a part of
TPM2 protocol and should be called explicitly.
This patch exposes cmd_read/go_idle via tpm class ops and removes
runtime pm support as it is not used by any driver.
When calling from nested context always use both flags:
TPM_TRANSMIT_UNLOCKED and TPM_TRANSMIT_RAW. Both are needed to resolve
tpm spaces and locality request recursive calls to tpm_transmit().
TPM_TRANSMIT_RAW should never be used standalone as it will fail
on double locking. While TPM_TRANSMIT_UNLOCKED standalone should be
called from non-recursive locked contexts.
New wrappers are added tpm_cmd_ready() and tpm_go_idle() to
streamline tpm_try_transmit code.
tpm_crb no longer needs own power saving functions and can drop using
tpm_pm_suspend/resume.
This patch cannot be really separated from the locality fix. Fixes: 888d867df441 (tpm: cmd_ready command can be issued only after granting locality) Cc: stable@vger.kernel.org Fixes: 888d867df441 (tpm: cmd_ready command can be issued only after granting locality) Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Tested-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The userpace expects to read the number of bytes stated in the header.
Returning the size of the buffer instead would be unexpected.
Cc: stable@vger.kernel.org Fixes: 095531f891e6 ("tpm: return a TPM_RC_COMMAND_CODE response if command is not implemented") Signed-off-by: Ricardo Schwarzmeier <Ricardo.Schwarzmeier@infineon.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some versions of GCC suboptimally generate calls to the __multi3()
intrinsic for MIPS64r6 builds, resulting in link failures due to the
missing function:
LD vmlinux.o
MODPOST vmlinux.o
kernel/bpf/verifier.o: In function `kmalloc_array':
include/linux/slab.h:631: undefined reference to `__multi3'
fs/select.o: In function `kmalloc_array':
include/linux/slab.h:631: undefined reference to `__multi3'
...
We already have a workaround for this in which we provide the
instrinsic, but we do so selectively for GCC 7 only. Unfortunately the
issue occurs with older GCC versions too - it has been observed with
both GCC 5.4.0 & GCC 6.4.0.
MIPSr6 support was introduced in GCC 5, so all major GCC versions prior
to GCC 8 are affected and we extend our workaround accordingly to all
MIPS64r6 builds using GCC versions older than GCC 8.
Signed-off-by: Paul Burton <paul.burton@mips.com> Reported-by: Vladimir Kondratiev <vladimir.kondratiev@intel.com> Fixes: ebabcf17bcd7 ("MIPS: Implement __multi3 for GCC7 MIPS64r6 builds")
Patchwork: https://patchwork.linux-mips.org/patch/20297/ Cc: James Hogan <jhogan@kernel.org> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: linux-mips@linux-mips.org Cc: stable@vger.kernel.org # 4.15+ Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Linux expects that if a CPU modifies a memory location, then that
modification will eventually become visible to other CPUs in the system.
Loongson 3 CPUs include a Store Fill Buffer (SFB) which sits between a
core & its L1 data cache, queueing memory accesses & allowing for faster
forwarding of data from pending stores to younger loads from the core.
Unfortunately the SFB prioritizes loads such that a continuous stream of
loads may cause a pending write to be buffered indefinitely. This is
problematic if we end up with 2 CPUs which each perform a store that the
other polls for - one or both CPUs may end up with their stores buffered
in the SFB, never reaching cache due to the continuous reads from the
poll loop. Such a deadlock condition has been observed whilst running
qspinlock code.
This patch changes the definition of cpu_relax() to smp_mb() for
Loongson-3, forcing a flush of the SFB on SMP systems which will cause
any pending writes to make it as far as the L1 caches where they will
become visible to other CPUs. If the kernel is not compiled for SMP
support, this will expand to a barrier() as before.
This workaround matches that currently implemented for ARM when
CONFIG_ARM_ERRATA_754327=y, which was introduced by commit 534be1d5a2da
("ARM: 6194/1: change definition of cpu_relax() for ARM11MPCore").
Although the workaround is only required when the Loongson 3 SFB
functionality is enabled, and we only began explicitly enabling that
functionality in v4.7 with commit 1e820da3c9af ("MIPS: Loongson-3:
Introduce CONFIG_LOONGSON3_ENHANCEMENT"), existing or future firmware
may enable the SFB which means we may need the workaround backported to
earlier kernels too.
[paul.burton@mips.com:
- Reword commit message & comment.
- Limit stable backport to v3.15+ where we support Loongson 3 CPUs.]
The VDSO Makefile filters CFLAGS to select a subset which it uses whilst
building the VDSO ELF. One of the flags it allows through is the -march=
flag that selects the architecture/ISA to target.
Unfortunately in cases where CONFIG_CPU_MIPS32_R{1,2}=y and the
toolchain defaults to building for MIPS64, the main MIPS Makefile ends
up using the short-form -<arch> flags in cflags-y. This is because the
calls to cc-option always fail to use the long-form -march=<arch> flag
due to the lack of an -mabi=<abi> flag in KBUILD_CFLAGS at the point
where the cc-option function is executed. The resulting GCC invocation
is something like:
$ mips64-linux-gcc -Werror -march=mips32r2 -c -x c /dev/null -o tmp
cc1: error: '-march=mips32r2' is not compatible with the selected ABI
These short-form -<arch> flags are dropped by the VDSO Makefile's
filtering, and so we attempt to build the VDSO without specifying any
architecture. This results in an attempt to build the VDSO using
whatever the compiler's default architecture is, regardless of whether
that is suitable for the kernel configuration.
One encountered build failure resulting from this mismatch is a
rejection of the sync instruction if the kernel is configured for a
MIPS32 or MIPS64 r1 or r2 target but the toolchain defaults to an older
architecture revision such as MIPS1 which did not include the sync
instruction:
CC arch/mips/vdso/gettimeofday.o
/tmp/ccGQKoOj.s: Assembler messages:
/tmp/ccGQKoOj.s:273: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:329: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:520: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:714: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1009: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1066: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1114: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1279: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1334: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1374: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1459: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1514: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:1814: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:2002: Error: opcode not supported on this processor: mips1 (mips1) `sync'
/tmp/ccGQKoOj.s:2066: Error: opcode not supported on this processor: mips1 (mips1) `sync'
make[2]: *** [scripts/Makefile.build:318: arch/mips/vdso/gettimeofday.o] Error 1
make[1]: *** [scripts/Makefile.build:558: arch/mips/vdso] Error 2
make[1]: *** Waiting for unfinished jobs....
This can be reproduced for example by attempting to build
pistachio_defconfig using Arnd's GCC 8.1.0 mips64 toolchain from
kernel.org:
Resolve this problem by using the long-form -march=<arch> in all cases,
which makes it through the arch/mips/vdso/Makefile's filtering & is thus
consistently used to build both the kernel proper & the VDSO.
The use of cc-option to prefer the long-form & fall back to the
short-form flags makes no sense since the short-form is just an
abbreviation for the also-supported long-form in all GCC versions that
we support building with. This means there is no case in which we have
to use the short-form -<arch> flags, so we can simply remove them.
The manual redefinition of _MIPS_ISA is removed naturally along with the
use of the short-form flags that it accompanied, and whilst here we
remove the separate assembler ISA selection. I suspect that both of
these were only required due to the mips32 vs mips2 mismatch that was
introduced by commit 59b3e8e9aac6 ("[MIPS] Makefile crapectomy.") and
fixed but not cleaned up by commit 9200c0b2a07c ("[MIPS] Fix Makefile
bugs for MIPS32/MIPS64 R1 and R2.").
I've marked this for backport as far as v4.4 where the MIPS VDSO was
introduced. In earlier kernels there should be no ill effect to using
the short-form flags.
The __clear_user function is defined to return the number of bytes that
could not be cleared. From the underlying memset / bzero implementation
this means setting register a2 to that number on return. Currently if a
page fault is triggered within the MIPSr6 version of setting of initial
unaligned bytes, the value loaded into a2 on return is meaningless.
During the MIPSr6 version of the initial unaligned bytes block, register
a2 contains the number of bytes to be set beyond the initial unaligned
bytes. The t0 register is initally set to the number of unaligned bytes
- STORSIZE, effectively a negative version of the number of unaligned
bytes. This is then incremented before each byte is saved.
The label .Lbyte_fixup\@ is jumped to on page fault. Currently the value
in a2 is incorrectly replaced by 0 - t0 + 1, effectively the number of
unaligned bytes remaining. This leads to the failures being reported by
the following test code:
static int __init test_clear_user(void)
{
int j, k;
Which reports:
[ 3.965439] Testing clear_user
[ 3.973169] clear_user (NULL 8) returned 6
[ 3.976782] clear_user (NULL 9) returned 6
[ 3.980390] clear_user (NULL 10) returned 6
[ 3.984052] clear_user (NULL 11) returned 6
[ 3.987524] clear_user (NULL 12) returned 6
Fix this by subtracting t0 from a2 (rather than $0), effectivey giving:
unset_bytes = (#bytes - (#unaligned bytes)) - (-#unaligned bytes remaining + 1) + 1
a2 = a2 - t0 + 1
This fixes the value returned from __clear user when the number of bytes
to set is > LONGSIZE and the address is invalid and unaligned.
Unfortunately, this breaks the fixup handling for unaligned bytes after
the final long, where register a2 still contains the number of bytes
remaining to be set and the t0 register is to 0 - the number of
unaligned bytes remaining.
Because t0 is now is now subtracted from a2 rather than 0, the number of
bytes unset is reported incorrectly:
static int __init test_clear_user(void)
{
char *test;
int j, k;
pr_info("\n\n\nTesting clear_user\n");
test = vmalloc(PAGE_SIZE);
Use the `unsigned long' rather than `__u32' type for DSP accumulator
registers, like with the regular MIPS multiply/divide accumulator and
general-purpose registers, as all are 64-bit in 64-bit implementations
and using a 32-bit data type leads to contents truncation on context
saving.
Update `arch_ptrace' and `compat_arch_ptrace' accordingly, removing
casts that are similarly not used with multiply/divide accumulator or
general-purpose register accesses.
Signed-off-by: Maciej W. Rozycki <macro@mips.com> Signed-off-by: Paul Burton <paul.burton@mips.com> Fixes: e50c0a8fa60d ("Support the MIPS32 / MIPS64 DSP ASE.")
Patchwork: https://patchwork.linux-mips.org/patch/19329/ Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: James Hogan <jhogan@kernel.org> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: linux-fsdevel@vger.kernel.org Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org # 2.6.15+ Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Show kprobes blacklist addresses under same condition of
showing kallsyms addresses.
Since there are several name conflict for local symbols,
kprobe blacklist needs to show each addresses so that
user can identify where is on blacklist by comparing
with kallsyms.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: David Howells <dhowells@redhat.com> Cc: David S . Miller <davem@davemloft.net> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Jon Medhurst <tixy@linaro.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Thomas Richter <tmricht@linux.ibm.com> Cc: Tobin C . Harding <me@tobin.cc> Cc: Will Deacon <will.deacon@arm.com> Cc: acme@kernel.org Cc: akpm@linux-foundation.org Cc: brueckner@linux.vnet.ibm.com Cc: linux-arch@vger.kernel.org Cc: rostedt@goodmis.org Cc: schwidefsky@de.ibm.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/lkml/152491893217.9916.14760965896164273464.stgit@devbox Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Without FORCE make does not detect changes only made to the command line
options. So object files might not be re-built even when they should be.
Fix this by adding FORCE where it is missing.
When the kernel is built with CONFIG_EXPOLINE=y and a compiler with
indirect branch mitigation enabled the purgatory crashes. The reason for
that is that the macros defined for expoline are used in mem.S. These
macros define new sections (.text.__s390x_indirect_*) which are marked
executable. Due to the missing linker script those sections are linked to
address 0, just as the .text section. In combination with the entry point
also being at address 0 this causes the purgatory load code
(kernel/kexec_file.c: kexec_purgatory_setup_sechdrs) to update the entry
point twice. Thus the old kernel jumps to some 'random' address causing the
crash.
To fix this turn off expolines for the purgatory. There is no problem with
this in this case due to the fact that the purgatory only runs once and the
tlb is purged (diag 308) in the end.
During interrupt setup we allocate interrupt vectors, walk the list of msi
descriptors, and fill in the message data. Requesting more interrupts than
supported on s390 can lead to an out of bounds access.
When we restrict the number of interrupts we should also stop walking the
msi list after all supported interrupts are handled.
Cc: stable@vger.kernel.org Signed-off-by: Sebastian Ott <sebott@linux.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The numa_init_early initcall sets the node_to_cpumask_map[0] to the
full cpu_possible_mask. Unfortunately this early_initcall is too late,
the NUMA setup for numa=emu is done even earlier. The order of calls
is numa_setup() -> emu_update_cpu_topology(), then the early_initcalls(),
followed by sched_init_domains().
Starting with git commit 051f3ca02e46432c0965e8948f00c07d8a2f09c0
"sched/topology: Introduce NUMA identity node sched domain"
the incorrect node_to_cpumask_map[0] really screws up the domain
setup and the kernel panics with the follow oops:
When allocating a new AOB fails, handle_outbound() is still capable of
transmitting the selected buffer (just without async completion).
But if a previous transfer on this queue slot used async completion, its
sbal_state flags field is still set to QDIO_OUTBUF_STATE_FLAG_PENDING.
So when the upper layer driver sees this stale flag, it expects an async
completion that never happens.
Fix this by unconditionally clearing the flags field.
Fixes: 104ea556ee7f ("qdio: support asynchronous delivery of storage blocks") Cc: <stable@vger.kernel.org> #v3.2+ Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For machines without the exrl instruction the BFP jit generates
code that uses an "br %r1" instruction located in the lowcore page.
Unfortunately there is a cut & paste error that puts an additional
"larl %r1,.+14" instruction in the code that clobbers the branch
target address in %r1. Remove the larl instruction.
Cc: <stable@vger.kernel.org> # v4.17+ Fixes: de5cb6eb51 ("s390: use expoline thunks in the BPF JIT") Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The memove, memset, memcpy, __memset16, __memset32 and __memset64
function have an additional indirect return branch in form of a
"bzr" instruction. These need to use expolines as well.
Cc: <stable@vger.kernel.org> # v4.17+ Fixes: 97489e0663 ("s390/lib: use expoline for indirect branches") Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit c9b5ad546e7d "s390/mm: tag normal pages vs pages used in page tables"
accidentally changed the logic in arch_set_page_states(), which is used by
the suspend/resume code. set_page_stable(page, order) was changed to
set_page_stable_dat(page, 0). After this, only the first page of higher order
pages will be set to stable, and a write to one of the unstable pages will
result in an addressing exception.
Fix this by using "order" again, instead of "0".
Fixes: c9b5ad546e7d ("s390/mm: tag normal pages vs pages used in page tables") Cc: stable@vger.kernel.org # 4.14+ Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When bootstrapping an architecture, it's usual to generate the kernel's
user-space headers (make headers_install) before building a compiler. Move
the compiler check (for asm goto support) to the archprepare target so that
it is only done when building code for the target.
Fixes: e501ce957a78 ("x86: Force asm-goto") Reported-by: Helmut Grohne <helmutg@debian.org> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180829194317.GA4765@decadent.org.uk Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reset the KASAN shadow state of the task stack before rewinding RSP.
Without this, a kernel oops will leave parts of the stack poisoned, and
code running under do_exit() can trip over such poisoned regions and cause
nonsensical false-positive KASAN reports about stack-out-of-bounds bugs.
This does not wipe the exception stacks; if an oops happens on an exception
stack, it might result in random KASAN false-positives from other tasks
afterwards. This is probably relatively uninteresting, since if the kernel
oopses on an exception stack, there are most likely bigger things to worry
about. It'd be more interesting if vmapped stacks and KASAN were
compatible, since then handle_stack_overflow() would oops from exception
stack context.
Fixes: 2deb4be28077 ("x86/dumpstack: When OOPSing, rewind the stack before do_exit()") Signed-off-by: Jann Horn <jannh@google.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Dmitry Vyukov <dvyukov@google.com> Cc: Alexander Potapenko <glider@google.com> Cc: Kees Cook <keescook@chromium.org> Cc: kasan-dev@googlegroups.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180828184033.93712-1-jannh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix this by sanitizing val before using it to index data->temp_src
Notice that given that speculation windows are large, the policy is
to kill the speculation on the first load and not worry if it can be
completed with a dependent load/store [1].
On Nehalem and newer core CPUs the CPU cache internally uses 44 bits
physical address space. The L1TF workaround is limited by this internal
cache address width, and needs to have one bit free there for the
mitigation to work.
Older client systems report only 36bit physical address space so the range
check decides that L1TF is not mitigated for a 36bit phys/32GB system with
some memory holes.
But since these actually have the larger internal cache width this warning
is bogus because it would only really be needed if the system had more than
43bits of memory.
Add a new internal x86_cache_bits field. Normally it is the same as the
physical bits field reported by CPUID, but for Nehalem and newerforce it to
be at least 44bits.
Change the L1TF memory size warning to use the new cache_bits field to
avoid bogus warnings and remove the bogus comment about memory size.
Fixes: 17dbca119312 ("x86/speculation/l1tf: Add sysfs reporting for l1tf") Reported-by: George Anchev <studio@anchev.net> Reported-by: Christopher Snowhill <kode54@gmail.com> Signed-off-by: Andi Kleen <ak@linux.intel.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: x86@kernel.org Cc: linux-kernel@vger.kernel.org Cc: Michael Hocko <mhocko@suse.com> Cc: vbabka@suse.cz Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180824170351.34874-1-andi@firstfloor.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A NMI can hit in the middle of context switching or in the middle of
switch_mm_irqs_off(). In either case, CR3 might not match current->mm,
which could cause copy_from_user_nmi() and friends to read the wrong
memory.
Fix it by adding a new nmi_uaccess_okay() helper and checking it in
copy_from_user_nmi() and in __copy_from_user_nmi()'s callers.
Signed-off-by: Andy Lutomirski <luto@kernel.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Rik van Riel <riel@surriel.com> Cc: Nadav Amit <nadav.amit@gmail.com> Cc: Borislav Petkov <bp@alien8.de> Cc: Jann Horn <jannh@google.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/dd956eba16646fd0b15c3c0741269dfd84452dac.1535557289.git.luto@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the __getcpu function, lsl is using the wrong target and destination
registers. Luckily, the compiler tends to choose %eax for both variables,
so it has been working so far.
Fixes: a582c540ac1b ("x86/vdso: Use RDPID in preference to LSL when available") Signed-off-by: Samuel Neves <sneves@dei.uc.pt> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Andy Lutomirski <luto@kernel.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180901201452.27828-1-sneves@dei.uc.pt Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch sets and clears FCF_ASYNC_{SENT|ACTIVE} flags to prevent
stalling of relogin attempt. Once flag are correctly set/cleared, relogin
timer can retry relogin attempt for driver to continue login.
Fixes: fa83e65885b9 ("scsi: qla2xxx: ensure async flags are reset correctly") Cc: stable@vger.kernel.org #4.17 Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The info->groups[] array is allocated in imx1_pinctrl_parse_dt(). It
has info->ngroups elements. Thus the > here should be >= to prevent
reading one element beyond the end of the array.
Make sure to set the mem device release callback before calling
put_device() in a couple of probe error paths so that the containing
object also gets freed.
The offset of the DSP core needs to be taken into account for the DSP
preloader control get and put. Currently the dsp->preloaded variable
will only ever be read/updated on the first DSP, whilst this doesn't
affect the operation of the control the readback will be incorrect.
Signed-off-by: Ajit Pandey <ajit.pandey@cirrus.com> Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com> Signed-off-by: Mark Brown <broonie@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a potential execution path in which function
platform_get_resource() returns NULL. If this happens,
we will end up having a NULL pointer dereference.
Fix this by replacing devm_ioremap with devm_ioremap_resource,
which has the NULL check and the memory region request.
This code was detected with the help of Coccinelle.
Cc: stable@vger.kernel.org Fixes: 2bd8d1d5cf89 ("ASoC: sirf: Add audio usp interface driver") Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com> Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When merging codec formats, dpcm_runtime_base_format() should skip
the codecs which are not supporting the current stream direction.
At the moment, if a BE link has more than one codec, and only one
of these codecs has no capture DAI, it becomes impossible to start
a capture stream because the merged format would be 0.
Skipping invalid codec DAI solves the problem.
Fixes: b073ed4e2126 ("ASoC: soc-pcm: DPCM cares BE format") Signed-off-by: Jerome Brunet <jbrunet@baylibre.com> Signed-off-by: Mark Brown <broonie@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
strncpy might not NUL-terminate the string, if the name equals the buffer size.
Use strlcpy instead.
Signed-off-by: Michael Buesch <m@bues.ch> Cc: stable@vger.kernel.org Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
strncpy might not NUL-terminate the string, if the name equals the buffer size.
Use strlcpy instead.
Signed-off-by: Michael Buesch <m@bues.ch> Cc: stable@vger.kernel.org Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Division is slow, so it shouldn't be done by the pixel generating code.
The driver supports only 2 or 4 bytes per pixel, so we can replace
division with a shift.
Allocations larger than PAGE_ALLOC_COSTLY_ORDER are unreliable and they
may fail anytime. This patch fixes the udl kms driver so that when a large
alloactions fails, it tries to do multiple smaller allocations.
If we leave urbs around, it causes not only leak, but also memory
corruption. This patch fixes the function udl_free_urb_list, so that it
always waits for all urbs that are in progress.
syzbot is hitting NULL pointer dereference at process_init_reply().
This is because deactivate_locked_super() is called before response for
initial request is processed.
Fix this by aborting and waiting for all requests (including FUSE_INIT)
before resetting fc->sb.
Original patch by Tetsuo Handa <penguin-kernel@I-love.SKAURA.ne.jp>.
fuse_abort_conn() does not guarantee that all async requests have actually
finished aborting (i.e. their ->end() function is called). This could
actually result in still used inodes after umount.
Add a helper to wait until all requests are fully done. This is done by
looking at the "num_waiting" counter. When this counter drops to zero, we
can be sure that no more requests are outstanding.
fuse_dev_release() assumes that it's the only one referencing the
fpq->processing list, but that's not true, since fuse_abort_conn() can be
doing the same without any serialization between the two.
Fixes: c3696046beb3 ("fuse: separate pqueue for clones") Cc: <stable@vger.kernel.org> # v4.2 Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fuse_dev_splice_write() reads pipe->buffers to determine the size of
'bufs' array before taking the pipe_lock(). This is not safe as
another thread might change the 'pipe->buffers' between the allocation
and taking the pipe_lock(). So we end up with too small 'bufs' array.
Move the bufs allocations inside pipe_lock()/pipe_unlock() to fix this.
WARNING: possible irq lock inversion dependency detected
CPU 0/KVM/10284 just changed the state of lock: 000000000d538a88 (&st->lock){+...}, at:
speculative_store_bypass_update+0x10b/0x170
but this lock was taken by another, HARDIRQ-safe lock
in the past:
(&(&sighand->siglock)->rlock){-.-.}
and interrupts could create inverse lock ordering between them.
In svm_vcpu_run() speculative_store_bypass_update() is called with
interupts enabled via x86_virt_spec_ctrl_set_guest/host().
This is actually a false positive, because GIF=0 so interrupts are
disabled even if IF=1; however, we can easily move the invocations of
x86_virt_spec_ctrl_set_guest/host() into the interrupt disabled region to
cure it, and it's a good idea to keep the GIF=0/IF=1 area as small
and self-contained as possible.
Fixes: 1f50ddb4f418 ("x86/speculation: Handle HT correctly on AMD") Reported-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com> Cc: Joerg Roedel <joro@8bytes.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krčmář <rkrcmar@redhat.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Borislav Petkov <bp@suse.de> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Tom Lendacky <thomas.lendacky@amd.com> Cc: kvm@vger.kernel.org Cc: x86@kernel.org Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some of the MSRs returned by GET_MSR_INDEX_LIST currently cannot be sent back
to KVM_GET_MSR and/or KVM_SET_MSR; either they can never be sent back, or you
they are only accepted under special conditions. This makes the API a pain to
use.
To avoid this pain, this patch makes it so that the result of the get-list
ioctl can always be used for host-initiated get and set. Since we don't have
a separate way to check for read-only MSRs, this means some Hyper-V MSRs are
ignored when written. Arguably they should not even be in the result of
GET_MSR_INDEX_LIST, but I am leaving there in case userspace is using the
outcome of GET_MSR_INDEX_LIST to derive the support for the corresponding
Hyper-V feature.
Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Two users have reported [1] that they have an "extremely unlikely" system
with more than MAX_PA/2 memory and L1TF mitigation is not effective.
Make the warning more helpful by suggesting the proper mem=X kernel boot
parameter to make it effective and a link to the L1TF document to help
decide if the mitigation is worth the unusable RAM.
Two users have reported [1] that they have an "extremely unlikely" system
with more than MAX_PA/2 memory and L1TF mitigation is not effective. In
fact it's a CPU with 36bits phys limit (64GB) and 32GB memory, but due to
holes in the e820 map, the main region is almost 500MB over the 32GB limit:
Suggestions to use 'mem=32G' to enable the L1TF mitigation while losing the
500MB revealed, that there's an off-by-one error in the check in
l1tf_select_mitigation().
l1tf_pfn_limit() returns the last usable pfn (inclusive) and the range
check in the mitigation path does not take this into account.
Instead of amending the range check, make l1tf_pfn_limit() return the first
PFN which is over the limit which is less error prone. Adjust the other
users accordingly.
On 32bit PAE kernels on 64bit hardware with enough physical bits,
l1tf_pfn_limit() will overflow unsigned long. This in turn affects
max_swapfile_size() and can lead to swapon returning -EINVAL. This has been
observed in a 32bit guest with 42 bits physical address size, where
max_swapfile_size() overflows exactly to 1 << 32, thus zero, and produces
the following warning to dmesg:
[ 6.396845] Truncating oversized swap area, only using 0k out of 2047996k
Fix this by using unsigned long long instead.
Fixes: 17dbca119312 ("x86/speculation/l1tf: Add sysfs reporting for l1tf") Fixes: 377eeaa8e11f ("x86/speculation/l1tf: Limit swap file size to MAX_PA/2") Reported-by: Dominique Leuenberger <dimstar@suse.de> Reported-by: Adrian Schroeter <adrian@suse.de> Signed-off-by: Vlastimil Babka <vbabka@suse.cz> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Andi Kleen <ak@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: "H . Peter Anvin" <hpa@zytor.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180820095835.5298-1-vbabka@suse.cz Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jann reported that x86 was missing required TLB invalidates when he
hit the !*batch slow path in tlb_remove_table().
This is indeed the case; RCU_TABLE_FREE does not provide TLB (cache)
invalidates, the PowerPC-hash where this code originated and the
Sparc-hash where this was subsequently used did not need that. ARM
which later used this put an explicit TLB invalidate in their
__p*_free_tlb() functions, and PowerPC-radix followed that example.
But when we hooked up x86 we failed to consider this. Fix this by
(optionally) hooking tlb_remove_table() into the TLB invalidate code.
NOTE: s390 was also needing something like this and might now
be able to use the generic code again.
[ Modified to be on top of Nick's cleanups, which simplified this patch
now that tlb_flush_mmu_tlbonly() really only flushes the TLB - Linus ]
Fixes: 9e52fc2b50de ("x86/mm: Enable RCU based page table freeing (CONFIG_HAVE_RCU_TABLE_FREE=y)") Reported-by: Jann Horn <jannh@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Rik van Riel <riel@surriel.com> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: David Miller <davem@davemloft.net> Cc: Will Deacon <will.deacon@arm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no need to call this from tlb_flush_mmu_tlbonly, it logically
belongs with tlb_flush_mmu_free. This makes future fixes simpler.
[ This was originally done to allow code consolidation for the
mmu_notifier fix, but it also ends up helping simplify the
HAVE_RCU_TABLE_INVALIDATE fix. - Linus ]
Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Acked-by: Will Deacon <will.deacon@arm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 5d9f40b56630 ("platform/x86: ideapad-laptop: Add
Y520-15IKBN to no_hw_rfkill") added the entry for Y20-15IKBN, and it
turned out that another variant, Y20-15IKBM, also requires the
no_hw_rfkill.
Trim the last letter from the string so that it matches to both
Y20-15IKBN and Y20-15IKBM models.
The probe handler_data was being allocated with __get_free_pages()
for no reason I could find. The error path was using kfree(). Since
other things are happily using kmalloc() in the probe path, switch to
kmalloc() entirely. This fixes the error path mismatch and will avoid
issues with CONFIG_HARDENED_USERCOPY_PAGESPAN=y.
Reported-by: Mihai Donțu <mihai.dontu@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Mario Limonciello <Mario.limonciello@dell.com> Cc: stable@vger.kernel.org Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config
read functions, but I concluded this wouldn't be trivial, so it's not
worth doing it for a single user that can't have any racing problems
while reading the register in two separate operations.
v2:
- Scrub the development (non-public) changelog (Joonas).
- Remove the i915.ko bits so this can be easily backported in order
to properly avoid stolen memory even on machines without i915.ko
(Joonas).
- CC stable for the reasons above.
In many architectures loads may be reordered with older stores to
different locations. In the nvme driver the following two operations
could be reordered:
- Write shadow doorbell (dbbuf_db) into memory.
- Read EventIdx (dbbuf_ei) from memory.
This can result in a potential race condition between driver and VM host
processing requests (if given virtual NVMe controller has a support for
shadow doorbell). If that occurs, then the NVMe controller may decide to
wait for MMIO doorbell from guest operating system, and guest driver may
decide not to issue MMIO doorbell on any of subsequent commands.
This issue is purely timing-dependent one, so there is no easy way to
reproduce it. Currently the easiest known approach is to run "Oracle IO
Numbers" (orion) that is shipped with Oracle DB:
Where nvme_test is a .lun file that contains a list of NVMe block
devices to run test against. Limiting number of vCPUs assigned to given
VM instance seems to increase chances for this bug to occur. On test
environment with VM that got 4 NVMe drives and 1 vCPU assigned the
virtual NVMe controller hang could be observed within 10-20 minutes.
That correspond to about 400-500k IO operations processed (or about
100GB of IO read/writes).
Orion tool was used as a validation and set to run in a loop for 36
hours (equivalent of pushing 550M IO operations). No issues were
observed. That suggest that the patch fixes the issue.
Fixes: f9f38e33389c ("nvme: improve performance for virtual NVMe devices") Signed-off-by: Michal Wnukowski <wnukowski@google.com> Reviewed-by: Keith Busch <keith.busch@intel.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
[hch: updated changelog and comment a bit] Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Whenever we hit block or inode bitmap corruptions we set
bit and then reduce this block group free inode/clusters
counter to expose right available space.
However some of ext4_mark_group_bitmap_corrupted() is called
inside group spinlock, some are not, this could make it happen
that we double reduce one block group free counters from system.
Always hold group spinlock for it could fix it, but it looks
a little heavy, we could use test_and_set_bit() to fix race
problems here.
When ext4_find_entry() falls back to "searching the old fashioned
way" due to a corrupt dx dir, it needs to reset the error code
to NULL so that the nonstandard ERR_BAD_DX_DIR code isn't returned
to userspace.
While working on extended rand for last_error/first_error timestamps,
I noticed that the endianess is wrong; we access the little-endian
fields in struct ext4_super_block as native-endian when we print them.
This adds a special case in ext4_attr_show() and ext4_attr_store()
to byteswap the superblock fields if needed.
In older kernels, this code was part of super.c, it got moved to
sysfs.c in linux-4.4.
Extended attribute names are defined to be NUL-terminated, so the name
must not contain a NUL character. This is important because there are
places when remove extended attribute, the code uses strlen to
determine the length of the entry. That should probably be fixed at
some point, but code is currently really messy, so the simplest fix
for now is to simply validate that the extended attributes are sane.
When cpu_stop_queue_work() releases the lock for the stopper
thread that was queued into its wake queue, preemption is
enabled, which leads to the following deadlock:
-grabs lock for migration/0-
-spins with preemption disabled,
waiting for migration/0's lock to be
released-
-adds work items for migration/0
and queues migration/0 to its
wake_q-
-releases lock for migration/0
and preemption is enabled-
-current thread is preempted,
and __set_cpus_allowed_ptr
has changed the thread's
cpu allowed mask to CPU1 only-
-acquires migration/0 and migration/1's
locks-
-adds work for migration/0 but does not
add migration/0 to wake_q, since it is
already in a wake_q-
-adds work for migration/1 and adds
migration/1 to its wake_q-
-releases migration/0 and migration/1's
locks, wakes migration/1, and enables
preemption-
-since migration/1 is requested to run,
migration/1 begins to run and waits on
migration/0, but migration/0 will never
be able to run, since the thread that
can wake it is affine to CPU1-
Disable preemption in cpu_stop_queue_work() before queueing works for
stopper threads, and queueing the stopper thread in the wake queue, to
ensure that the operation of queueing the works and waking the stopper
threads is atomic.
Fixes: 0b26351b910f ("stop_machine, sched: Fix migrate_swap() vs. active_balance() deadlock") Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org> Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: peterz@infradead.org Cc: matt@codeblueprint.co.uk Cc: bigeasy@linutronix.de Cc: gregkh@linuxfoundation.org Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/1533329766-4856-1-git-send-email-isaacm@codeaurora.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Co-Developed-by: Isaac J. Manjarres <isaacm@codeaurora.org>
The code flow in cpu_stop_queue_two_works() is a little arcane; fix this by
lifting the preempt_disable() to the top to create more natural nesting wrt
the spinlocks and make the wake_up_q() and preempt_enable() unconditional
at the end.
Furthermore, enable preemption in the -EDEADLK case, such that we spin-wait
with preemption enabled.
Suggested-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: isaacm@codeaurora.org Cc: matt@codeblueprint.co.uk Cc: psodagud@codeaurora.org Cc: gregkh@linuxfoundation.org Cc: pkondeti@codeaurora.org Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180730112140.GH2494@hirez.programming.kicks-ass.net Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With commit eca0fa28cd0d ("perf record: Provide detailed information on
s390 CPU") s390 platform provides detailed type/model/capacity
information in the CPU identifier string instead of just "IBM/S390".
This breaks 'perf kvm' support which uses hard coded string IBM/S390 to
compare with the CPU identifier string. Fix this by changing the
comparison.
Reported-by: Stefan Raspl <raspl@linux.ibm.com> Signed-off-by: Thomas Richter <tmricht@linux.ibm.com> Reviewed-by: Hendrik Brueckner <brueckner@linux.ibm.com> Tested-by: Stefan Raspl <raspl@linux.ibm.com> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: stable@vger.kernel.org Fixes: eca0fa28cd0d ("perf record: Provide detailed information on s390 CPU") Link: http://lkml.kernel.org/r/20180712070936.67547-1-tmricht@linux.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the oom killer kills a userspace process in the page fault handler
while in guest context, the fault handler fails to release the mm_sem
if the FAULT_FLAG_RETRY_NOWAIT option is set. This leads to a deadlock
when tearing down the mm when the process terminates. This bug can only
happen when pfault is enabled, so only KVM clients are affected.
The problem arises in the rare cases in which handle_mm_fault does not
release the mm_sem. This patch fixes the issue by manually releasing
the mm_sem when needed.
When there is contention on faulting in a particular page table entry
at stage 2, the break-before-make requirement of the architecture can
lead to additional refaulting due to TLB invalidation.
Avoid this by skipping a page table update if the new value of the PTE
matches the previous value.
Contention on updating a PMD entry by a large number of vcpus can lead
to duplicate work when handling stage 2 page faults. As the page table
update follows the break-before-make requirement of the architecture,
it can lead to repeated refaults due to clearing the entry and
flushing the tlbs.
This problem is more likely when -
* there are large number of vcpus
* the mapping is large block mapping
such as when using PMD hugepages (512MB) with 64k pages.
Fix this by skipping the page table update if there is no change in
the entry being updated.
When the VCPU is blocked (for example from WFI) we don't inject the
physical timer interrupt if it should fire while the CPU is blocked, but
instead we just wake up the VCPU and expect kvm_timer_vcpu_load to take
care of injecting the interrupt.
Unfortunately, kvm_timer_vcpu_load() doesn't actually do that, it only
has support to schedule a soft timer if the emulated phys timer is
expected to fire in the future.
Follow the same pattern as kvm_timer_update_state() and update the irq
state after potentially scheduling a soft timer.
Reported-by: Andre Przywara <andre.przywara@arm.com> Cc: Stable <stable@vger.kernel.org> # 4.15+ Fixes: bbdd52cfcba29 ("KVM: arm/arm64: Avoid phys timer emulation in vcpu entry/exit") Signed-off-by: Christoffer Dall <christoffer.dall@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
kvm_timer_update_state() is called when changing the phys timer
configuration registers, either via vcpu reset, as a result of a trap
from the guest, or when userspace programs the registers.
phys_timer_emulate() is in turn called by kvm_timer_update_state() to
either cancel an existing software timer, or program a new software
timer, to emulate the behavior of a real phys timer, based on the change
in configuration registers.
Unfortunately, the interaction between these two functions left a small
race; if the conceptual emulated phys timer should actually fire, but
the soft timer hasn't executed its callback yet, we cancel the timer in
phys_timer_emulate without injecting an irq. This only happens if the
check in kvm_timer_update_state is called before the timer should fire,
which is relatively unlikely, but possible.
The solution is to update the state of the phys timer after calling
phys_timer_emulate, which will pick up the pending timer state and
update the interrupt value.
Note that this leaves the opportunity of raising the interrupt twice,
once in the just-programmed soft timer, and once in
kvm_timer_update_state. Since this always happens synchronously with
the VCPU execution, there is no harm in this, and the guest ever only
sees a single timer interrupt.
Cc: Stable <stable@vger.kernel.org> # 4.15+ Signed-off-by: Christoffer Dall <christoffer.dall@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ARM64's pfn_valid() shifts away the upper PAGE_SHIFT bits of the input
before seeing if the PFN is valid. This leads to false positives when
some of the upper bits are set, but the lower bits match a valid PFN.
For example, the following userspace code looks up a bogus entry in
/proc/kpageflags:
int pagemap = open("/proc/self/pagemap", O_RDONLY);
int pageflags = open("/proc/kpageflags", O_RDONLY);
uint64_t pfn, val;
On ARM64 this causes the userspace process to crash with SIGSEGV rather
than reading (1 << KPF_NOPAGE). kpageflags_read() treats the offset as
valid, and stable_page_flags() will try to access an address between the
user and kernel address ranges.
If there is a mismatch in the I/D min line size, we must
always use the system wide safe value both in applications
and in the kernel, while performing cache operations. However,
we have been checking more bits than just the min line sizes,
which triggers false negatives. We may need to trap the user
accesses in such cases, but not necessarily patch the kernel.
This patch fixes the check to do the right thing as advertised.
A new capability will be added to check mismatches in other
fields and ensure we trap the CTR accesses.
Fixes: be68a8aaf925 ("arm64: cpufeature: Fix CTR_EL0 field definitions") Cc: <stable@vger.kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Reported-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 719f6a7040f1bdaf96 ("printk: Use the main logbuf in NMI
when logbuf_lock is available") brought back the possible deadlocks
in printk() and NMI.
The check of logbuf_lock is done only in printk_nmi_enter() to prevent
mixed output. But another CPU might take the lock later, enter NMI, and:
+ Both NMIs might be serialized by yet another lock, for example,
the one in nmi_cpu_backtrace().
+ The other CPU might get stopped in NMI, see smp_send_stop()
in panic().
The only safe solution is to use trylock when storing the message
into the main log-buffer. It might cause reordering when some lines
go to the main lock buffer directly and others are delayed via
the per-CPU buffer. It means that it is not useful in general.
This patch replaces the problematic NMI deferred context with NMI
direct context. It can be used to mark a code that might produce
many messages in NMI and the risk of losing them is more critical
than problems with eventual reordering.
The context is then used when dumping trace buffers on oops. It was
the primary motivation for the original fix. Also the reordering is
even smaller issue there because some traces have their own time stamps.
Finally, nmi_cpu_backtrace() need not longer be serialized because
it will always us the per-CPU buffers again.
Fixes: 719f6a7040f1bdaf96 ("printk: Use the main logbuf in NMI when logbuf_lock is available") Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/20180627142028.11259-1-pmladek@suse.com
To: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> Cc: linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by: Petr Mladek <pmladek@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>