Sysbot has reported a warning where a kmalloc() attempt exceeds the
maximum limit. This has been identified as corruption of the xattr_ids
count when reading the xattr id lookup table.
This patch adds a number of additional sanity checks to detect this
corruption and others.
1. It checks for a corrupted xattr index read from the inode. This could
be because the metadata block is uncompressed, or because the
"compression" bit has been corrupted (turning a compressed block
into an uncompressed block). This would cause an out of bounds read.
2. It checks against corruption of the xattr_ids count. This can either
lead to the above kmalloc failure, or a smaller than expected
table to be read.
3. It checks the contents of the index table for corruption.
Sysbot has reported an "slab-out-of-bounds read" error which has been
identified as being caused by a corrupted "ino_num" value read from the
inode. This could be because the metadata block is uncompressed, or
because the "compression" bit has been corrupted (turning a compressed
block into an uncompressed block).
This patch adds additional sanity checks to detect this, and the
following corruption.
1. It checks against corruption of the inodes count. This can either
lead to a larger table to be read, or a smaller than expected
table to be read.
In the case of a too large inodes count, this would often have been
trapped by the existing sanity checks, but this patch introduces
a more exact check, which can identify too small values.
2. It checks the contents of the index table for corruption.
Sysbot has reported a number of "slab-out-of-bounds reads" and
"use-after-free read" errors which has been identified as being caused
by a corrupted index value read from the inode. This could be because
the metadata block is uncompressed, or because the "compression" bit has
been corrupted (turning a compressed block into an uncompressed block).
This patch adds additional sanity checks to detect this, and the
following corruption.
1. It checks against corruption of the ids count. This can either
lead to a larger table to be read, or a smaller than expected
table to be read.
In the case of a too large ids count, this would often have been
trapped by the existing sanity checks, but this patch introduces
a more exact check, which can identify too small values.
2. It checks the contents of the index table for corruption.
Patch series "Squashfs: fix BIO migration regression and add sanity checks".
Patch [1/4] fixes a regression introduced by the "migrate from
ll_rw_block usage to BIO" patch, which has produced a number of
Sysbot/Syzkaller reports.
Patches [2/4], [3/4], and [4/4] fix a number of filesystem corruption
issues which have produced Sysbot reports in the id, inode and xattr
lookup code.
Each patch has been tested against the Sysbot reproducers using the
given kernel configuration. They have the appropriate "Reported-by:"
lines added.
Additionally, all of the reproducer filesystems are indirectly fixed by
patch [4/4] due to the fact they all have xattr corruption which is now
detected there.
Additional testing with other configurations and architectures (32bit,
big endian), and normal filesystems has also been done to trap any
inadvertent regressions caused by the additional sanity checks.
This patch (of 4):
This is a regression introduced by the patch "migrate from ll_rw_block
usage to BIO".
Sysbot/Syskaller has reported a number of "out of bounds writes" and
"unable to handle kernel paging request in squashfs_decompress" errors
which have been identified as a regression introduced by the above
patch.
Specifically, the patch removed the following sanity check
1. It ensured any reads were not beyond the end of the filesystem
2. It ensured that the "length" field read from the filesystem
was within the expected maximum length. Without this any
corrupted values can over-run allocated buffers.
Link: https://lkml.kernel.org/r/20210204130249.4495-1-phillip@squashfs.org.uk Link: https://lkml.kernel.org/r/20210204130249.4495-2-phillip@squashfs.org.uk Fixes: 93e72b3c612adc ("squashfs: migrate from ll_rw_block usage to BIO") Reported-by: syzbot+6fba78f99b9afd4b5634@syzkaller.appspotmail.com Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk> Cc: Philippe Liard <pliard@google.com> 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>
This reverts commit 536d3bf261a2fc3b05b3e91e7eef7383443015cf, as it can
cause writers to memory.high to get stuck in the kernel forever,
performing page reclaim and consuming excessive amounts of CPU cycles.
Before the patch, a write to memory.high would first put the new limit
in place for the workload, and then reclaim the requested delta. After
the patch, the kernel tries to reclaim the delta before putting the new
limit into place, in order to not overwhelm the workload with a sudden,
large excess over the limit. However, if reclaim is actively racing
with new allocations from the uncurbed workload, it can keep the write()
working inside the kernel indefinitely.
This is causing problems in Facebook production. A privileged
system-level daemon that adjusts memory.high for various workloads
running on a host can get unexpectedly stuck in the kernel and
essentially turn into a sort of involuntary kswapd for one of the
workloads. We've observed that daemon busy-spin in a write() for
minutes at a time, neglecting its other duties on the system, and
expending privileged system resources on behalf of a workload.
To remedy this, we have first considered changing the reclaim logic to
break out after a couple of loops - whether the workload has converged
to the new limit or not - and bound the write() call this way. However,
the root cause that inspired the sequence change in the first place has
been fixed through other means, and so a revert back to the proven
limit-setting sequence, also used by memory.max, is preferable.
The sequence was changed to avoid extreme latencies in the workload when
the limit was lowered: the sudden, large excess created by the limit
lowering would erroneously trigger the penalty sleeping code that is
meant to throttle excessive growth from below. Allocating threads could
end up sleeping long after the write() had already reclaimed the delta
for which they were being punished.
However, erroneous throttling also caused problems in other scenarios at
around the same time. This resulted in commit b3ff92916af3 ("mm, memcg:
reclaim more aggressively before high allocator throttling"), included
in the same release as the offending commit. When allocating threads
now encounter large excess caused by a racing write() to memory.high,
instead of entering punitive sleeps, they will simply be tasked with
helping reclaim down the excess, and will be held no longer than it
takes to accomplish that. This is in line with regular limit
enforcement - i.e. if the workload allocates up against or over an
otherwise unchanged limit from below.
With the patch breaking userspace, and the root cause addressed by other
means already, revert it again.
Link: https://lkml.kernel.org/r/20210122184341.292461-1-hannes@cmpxchg.org Fixes: 536d3bf261a2 ("mm: memcontrol: avoid workload stalls when lowering memory.high") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Acked-by: Chris Down <chris@chrisdown.name> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Koutný <mkoutny@suse.com> Cc: <stable@vger.kernel.org> [5.8+] 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>
Since 5.10, splice() or sendfile() to NILFS2 return EINVAL. This was
caused by commit 36e2c7421f02 ("fs: don't allow splice read/write
without explicit ops").
This patch initializes the splice_write field in file_operations, like
most file systems do, to restore the functionality.
In thunderbolt mode the PHY is owned by the thunderbolt controller.
We are not supposed to touch it. So skip the vswing programming
as well (we already skipped the other steps not applicable to TBT).
Touching this stuff could supposedly interfere with the PHY
programming done by the thunderbolt controller.
The MH PHY vswing table does have all the entries these days. Get
rid of the old hacks in the code which claim otherwise.
This hack was totally bogus anyway. The correct way to handle the
lack of those two entries would have been to declare our max
vswing and pre-emph to both be level 2.
Cc: José Roberto de Souza <jose.souza@intel.com> Cc: Clinton Taylor <clinton.a.taylor@intel.com> Fixes: 9f7ffa297978 ("drm/i915/tc/icl: Update TC vswing tables") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20201207203512.1718-1-ville.syrjala@linux.intel.com Reviewed-by: Imre Deak <imre.deak@intel.com> Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
(cherry picked from commit 5ec346476e795089b7dac8ab9dcee30c8d80ad84) Signed-off-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix incorrect is_branch{32,64}_taken() analysis for the jsgt case. The return
code for both will tell the caller whether a given conditional jump is taken
or not, e.g. 1 means branch will be taken [for the involved registers] and the
goto target will be executed, 0 means branch will not be taken and instead we
fall-through to the next insn, and last but not least a -1 denotes that it is
not known at verification time whether a branch will be taken or not. Now while
the jsgt has the branch-taken case correct with reg->s32_min_value > sval, the
branch-not-taken case is off-by-one when testing for reg->s32_max_value < sval
since the branch will also be taken for reg->s32_max_value == sval. The jgt
branch analysis, for example, gets this right.
While reviewing a different fix, John and I noticed an oddity in one of the
BPF program dumps that stood out, for example:
# bpftool p d x i 13
0: (b7) r0 = 808464450
1: (b4) w4 = 808464432
2: (bc) w0 = w0
3: (15) if r0 == 0x0 goto pc+1
4: (9c) w4 %= w0
[...]
In line 2 we noticed that the mov32 would 32 bit truncate the original src
register for the div/mod operation. While for the two operations the dst
register is typically marked unknown e.g. from adjust_scalar_min_max_vals()
the src register is not, and thus verifier keeps tracking original bounds,
simplified:
Runtime result of r0 at exit is 0 instead of expected -1. Remove the
verifier mov32 src rewrite in div/mod and replace it with a jmp32 test
instead. After the fix, we result in the following code generation when
having dividend r1 and divisor r6:
x86 in particular can throw a 'divide error' exception for div
instruction not only for divisor being zero, but also for the case
when the quotient is too large for the designated register. For the
edx:eax and rdx:rax dividend pair it is not an issue in x86 BPF JIT
since we always zero edx (rdx). Hence really the only protection
needed is against divisor being zero.
Fixes: 68fda450a7df ("bpf: fix 32-bit divide by zero") Co-developed-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
from 6 to 7: safe
processed 15 insns (limit 1000000) max_states_per_insn 0 total_states 1 peak_states 1 mark_read 1
The underlying program was xlated as follows:
# bpftool p d x i 10
0: (b7) r0 = 808464450
1: (b4) w4 = 808464432
2: (bc) w0 = w0
3: (15) if r0 == 0x0 goto pc+1
4: (9c) w4 %= w0
5: (66) if w4 s> 0x30303030 goto pc+0
6: (7f) r0 >>= r0
7: (bc) w0 = w0
8: (15) if r0 == 0x0 goto pc+1
9: (9c) w4 %= w0
10: (66) if w0 s> 0x3030 goto pc+0
11: (d6) if w0 s<= 0x303030 goto pc+1
12: (05) goto pc-1
13: (95) exit
The verifier rewrote original instructions it recognized as dead code with
'goto pc-1', but reality differs from verifier simulation in that we are
actually able to trigger a hang due to hitting the 'goto pc-1' instructions.
Taking a closer look at the verifier analysis, the reason is that it misjudges
its pruning decision at the first 'from 6 to 7: safe' occasion. What happens
is that while both old/cur registers are marked as precise, they get misjudged
for the jmp32 case as range_within() yields true, meaning that the prior
verification path with a wider register bound could be verified successfully
and therefore the current path with a narrower register bound is deemed safe
as well whereas in reality it's not. R0 old/cur path's bounds compare as
follows:
The 64 bit bounds generally look okay and while the information that got
propagated from 32 to 64 bit looks correct as well, it's not precise enough
for judging a conditional jmp32. Given the latter only operates on subregisters
we also need to take these into account as well for a range_within() probe
in order to be able to prune paths. Extending the range_within() constraint
to both bounds will be able to tell us that the old signed 32 bit bounds are
not wider than the cur signed 32 bit bounds.
With the fix in place, the program will now verify the 'goto' branch case as
it should have been:
The bug is quite subtle in the sense that when verifier would determine that
a given branch is dead code, it would (here: wrongly) remove these instructions
from the program and hard-wire the taken branch for privileged programs instead
of the 'goto pc-1' rewrites which will cause hard to debug problems.
Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking") Reported-by: Anatoly Trosinenko <anatoly.trosinenko@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With commit eaa7995c529b54 (regulator: core: avoid
regulator_resolve_supply() race condition) we started holding the rdev
lock while resolving supplies, an operation that requires holding the
regulator_list_mutex. This results in lockdep warnings since in other
places we take the list mutex then the mutex on an individual rdev.
Since the goal is to make sure that we don't call set_supply() twice
rather than a concern about the cost of resolution pull the rdev lock
and check for duplicate resolution down to immediately before we do the
set_supply() and drop it again once the allocation is done.
Fixes: eaa7995c529b54 (regulator: core: avoid regulator_resolve_supply() race condition) Reported-by: Marek Szyprowski <m.szyprowski@samsung.com> Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20210122132042.10306-1-broonie@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
On !PREEMPT kernel, we can get below softlockup when doing stress
testing with creating and destroying block cgroup repeatly. The
reason is it may take a long time to acquire the queue's lock in
the loop of blkcg_destroy_blkgs(), or the system can accumulate a
huge number of blkgs in pathological cases. We can add a need_resched()
check on each loop and release locks and do cond_resched() if true
to avoid this issue, since the blkcg_destroy_blkgs() is not called
from atomic contexts.
Some i2c device driver indirectly uses I2C driver when it is now
being suspended. The i2c devices driver is suspended during the
NOIRQ phase and this cannot be changed due to other dependencies.
Therefore, we also need to move the suspend handling for the I2C
controller driver to the NOIRQ phase as well.
Signed-off-by: Qii Wang <qii.wang@mediatek.com> Signed-off-by: Wolfram Sang <wsa@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When handling an auth_gss downcall, it's possible to get 0-length
opaque object for the acceptor. In the case of a 0-length XDR
object, make sure simple_get_netobj() fills in dest->data = NULL,
and does not continue to kmemdup() which will set
dest->data = ZERO_SIZE_PTR for the acceptor.
The trace event code can handle NULL but not ZERO_SIZE_PTR for a
string, and so without this patch the rpcgss_context trace event
will crash the kernel as follows:
Remove duplicated helper functions to parse opaque XDR objects
and place inside new file net/sunrpc/auth_gss/auth_gss_internal.h.
In the new file carry the license and copyright from the source file
net/sunrpc/auth_gss/auth_gss.c. Finally, update the comment inside
include/linux/sunrpc/xdr.h since lockd is not the only user of
struct xdr_netobj.
If we find an entry without an SKB, we currently continue, but
that will just result in an infinite loop since we won't increment
the read pointer, and will try the same thing over and over again.
Fix this.
If we get into a problem severe enough to attempt a reprobe,
we schedule a worker to do that. However, if the problem gets
more severe and the device is actually destroyed before this
worker has a chance to run, we use a free device. Bump up the
reference count of the device until the worker runs to avoid
this situation.
Until now we have been relying on matching the PCI ID and subsystem
device ID in order to recognize Qu devices with Hr2. Add rules to
match these devices, so that we don't have to add a new rule for every
new ID we get.
Having sta_id not set for aux_sta and snif_sta can potentially lead to a
hard to debug issue in case remove station is called without an add. In
this case sta_id 0, an unrelated regular station, will be removed.
In fact, we do have a FW assert that occures rarely and from the debug
data analysis it looks like sta_id 0 is removed by mistake, though it's
hard to pinpoint the exact flow. The WARN_ON in this patch should help
to find it.
In the new CSA flow, we remain associated during CSA, but
still do a unbind-bind to the vif. However, sending the power
command right after when vif is unbound but still associated
causes FW to assert (0x3400) since it cannot tell the LMAC id.
Just skip this command, we will send it again in a bit, when
assigning the new context.
If a layoutget ends up being reordered w.r.t. a layoutreturn, e.g. due
to a layoutget-on-open not knowing a priori which file to lock, then we
must assume the layout is no longer being considered valid state by the
server.
Incrementally improve our ability to reject such states by using the
cached old stateid in conjunction with the plh_barrier to try to
identify them.
If the server returns a new stateid that does not match the one in our
cache, then try to return the one we hold instead of just invalidating
it on the client side. This ensures that both client and server will
agree that the stateid is invalid.
Upon receiving CSA with 160MHz extended NSS BW from associated AP,
STA should set the HT operation_mode based on new_center_freq_seg1
because it is later used as ccfs2 in ieee80211_chandef_vht_oper().
Signed-off-by: Aviad Brikman <aviad.brikman@celeno.com> Signed-off-by: Shay Bar <shay.bar@celeno.com> Link: https://lore.kernel.org/r/20201222064714.24888-1-shay.bar@celeno.com Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When switching between firmware types, the wrong control
can be selected when requesting control in kernel API.
Use the currently selected DSP firwmare type to select
the proper mixer control.
The final step in regulator_register() is to call
regulator_resolve_supply() for each registered regulator
(including the one in the process of being registered). The
regulator_resolve_supply() function first checks if rdev->supply
is NULL, then it performs various steps to try to find the supply.
If successful, rdev->supply is set inside of set_supply().
This procedure can encounter a race condition if two concurrent
tasks call regulator_register() near to each other on separate CPUs
and one of the regulators has rdev->supply_name specified. There
is currently nothing guaranteeing atomicity between the rdev->supply
check and set steps. Thus, both tasks can observe rdev->supply==NULL
in their regulator_resolve_supply() calls. This then results in
both creating a struct regulator for the supply. One ends up
actually stored in rdev->supply and the other is lost (though still
present in the supply's consumer_list).
Here is a kernel log snippet showing the issue:
[ 12.421768] gpu_cc_gx_gdsc: supplied by pm8350_s5_level
[ 12.425854] gpu_cc_gx_gdsc: supplied by pm8350_s5_level
[ 12.429064] debugfs: Directory 'regulator.4-SUPPLY' with parent
'17a00000.rsc:rpmh-regulator-gfxlvl-pm8350_s5_level'
already present!
Avoid this race condition by holding the rdev->mutex lock inside
of regulator_resolve_supply() while checking and setting
rdev->supply.
xfrm_probe_algs() probes kernel crypto modules and changes the
availability of struct xfrm_algo_desc. But there is a small window
where ealg->available and aalg->available get changed between
count_ah_combs()/count_esp_combs() and dump_ah_combs()/dump_esp_combs(),
in this case we may allocate a smaller skb but later put a larger
amount of data and trigger the panic in skb_put().
Fix this by relaxing the checks when counting the size, that is,
skipping the test of ->available. We may waste some memory for a few
of sizeof(struct sadb_comb), but it is still much better than a panic.
Commit 0138ba5783ae ("powerpc/64/signal: Balance return predictor
stack in signal trampoline") changed __kernel_sigtramp_rt64() VDSO and
trampoline code, and introduced a regression in the way glibc's
backtrace()[1] detects the signal-handler stack frame. Apart from the
practical implications, __kernel_sigtramp_rt64() was a VDSO function
with the semantics that it is a function you can call from userspace
to end a signal handling. Now this semantics are no longer valid.
I believe the aforementioned change affects all releases since 5.9.
This patch tries to fix both the semantics and practical aspect of
__kernel_sigtramp_rt64() returning it to the previous code, whilst
keeping the intended behaviour of 0138ba5783ae by adding a new symbol
to serve as the jump target from the kernel to the trampoline. Now the
trampoline has two parts, a new entry point and the old return point.
When set_config changes a line from input to output debounce is
implicitly disabled, as debounce makes no sense for outputs, but the
debounce period is not being cleared and is still reported in the
line info.
So clear the debounce period when the debouncer is stopped in
edge_detector_stop().
What 84965ff8a84f0 ("io_uring: if we see flush on exit, cancel related tasks")
really wants is to cancel all relevant REQ_F_INFLIGHT requests reliably.
That can be achieved by io_uring_cancel_files(), but we'll miss it
calling io_uring_cancel_task_requests(files=NULL) from io_uring_flush(),
because it will go through __io_uring_cancel_task_requests().
Just always call io_uring_cancel_files() during cancel, it's good enough
for now.
Now io_uring_cancel_task_requests() can be called not through file
notes but directly, remove a WARN_ONCE() there that give us false
positives. That check is not very important and we catch it in other
places.
Fixes: 84965ff8a84f0 ("io_uring: if we see flush on exit, cancel related tasks") Cc: stable@vger.kernel.org # 5.9+ Reported-by: syzbot+3e3d9bd0c6ce9efbc3ef@syzkaller.appspotmail.com Signed-off-by: Pavel Begunkov <asml.silence@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is caused by calling io_cqring_overflow_flush() which may sleep
after calling prepare_to_wait_exclusive() which set task state to
TASK_INTERRUPTIBLE
Rewrite io_uring_cancel_files() to mimic __io_uring_task_cancel()'s
counting scheme, so it does all the heavy work before setting
TASK_UNINTERRUPTIBLE.
As tasks now cancel only theirs requests, and inflight_wait is awaited
only in io_uring_cancel_files(), which should be called with ->in_idle
set, instead of keeping a separate inflight_wait use tctx->wait.
That will add some spurious wakeups but actually is safer from point of
not hanging the task.
e.g.
task1 | IRQ
| *start* io_complete_rw_common(link)
| link: req1 -> req2 -> req3(with files)
*cancel_files() |
io_wq_cancel(), etc. |
| put_req(link), adds to io-wq req2
schedule() |
So, task1 will never try to cancel req2 or req3. If req2 is
long-standing (e.g. read(empty_pipe)), this may hang.
If the tctx inflight number haven't changed because of cancellation,
__io_uring_task_cancel() will continue leaving the task in
TASK_UNINTERRUPTIBLE state, that's not expected by
__io_uring_files_cancel(). Ensure we always call finish_wait() before
retrying.
We need to actively cancel anything that introduces a potential circular
loop, where io_uring holds a reference to itself. If the file in question
is an io_uring file, then add the request to the inflight list.
1. There can't be a request in the inflight list without
IO_WQ_WORK_FILES, kill this check to keep the whole condition simpler.
2. Also, don't call the function for files==NULL to not do such a check,
all that staff is already handled well by its counter part,
__io_uring_cancel_task_requests().
With that just flip the task check.
Also, it iowq-cancels all request of current task there, don't forget to
set right ->files into struct io_task_cancel.
Fixes: c1973b38bf639 ("io_uring: cancel only requests of current task") Reported-by: syzbot+c0d52d0b3c0c3ffb9525@syzkaller.appspotmail.com Signed-off-by: Pavel Begunkov <asml.silence@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Instead of iterating over each request and cancelling it individually in
io_uring_cancel_files(), try to cancel all matching requests and use
->inflight_list only to check if there anything left.
In many cases it should be faster, and we can reuse a lot of code from
task cancellation.
io_uring_cancel_files() guarantees to cancel all matching requests,
that's not necessary to do that in a loop. Move it up in the callchain
into io_uring_cancel_task_requests().
Commit e5f0e8f8e456 ("net: sched: introduce and use qdisc tree flush/purge helpers")
introduced qdisc tree flush/purge helpers, but erroneously used flush helper
instead of purge helper in qdisc_replace function.
This issue was found in our CI, that tests various qdisc setups by configuring
qdisc and sending data through it. Call of invalid helper sporadically leads
to corruption of vt_tree/cf_tree of hfsc_class that causes kernel oops:
Having multiple destination ports for a unicast address does not make
sense.
Make port_db_load_purge override existent unicast portvec instead of
adding a new port bit.
Fixes: 884729399260 ("net: dsa: mv88e6xxx: handle multiple ports in ATU") Signed-off-by: DENG Qingfang <dqfext@gmail.com> Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Link: https://lore.kernel.org/r/20210130134334.10243-1-dqfext@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
UDP/IP header of UDP GROed frag_skbs are not updated even after NAT
forwarding. Only the header of head_skb from ip_finish_output_gso ->
skb_gso_segment is updated but following frag_skbs are not updated.
A call path skb_mac_gso_segment -> inet_gso_segment ->
udp4_ufo_fragment -> __udp_gso_segment -> __udp_gso_segment_list
does not try to update UDP/IP header of the segment list but copy
only the MAC header.
Update port, addr and check of each skb of the segment list in
__udp_gso_segment_list. It covers both SNAT and DNAT.
Fixes: 9fd1ff5d2ac7 (udp: Support UDP fraglist GRO/GSO.) Signed-off-by: Dongseok Yi <dseok.yi@samsung.com> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Link: https://lore.kernel.org/r/1611962007-80092-1-git-send-email-dseok.yi@samsung.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dev->hard_header_len for tunnel interface is set only when header_ops
are set too and already contains full overhead of any tunnel encapsulation.
That's why there is not need to use this overhead twice in mtu calc.
Fixes: fdafed459998 ("ip_gre: set dev->hard_header_len and dev->needed_headroom properly") Reported-by: Slava Bacherikov <mail@slava.cc> Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru> Link: https://lore.kernel.org/r/1611959267-20536-1-git-send-email-vfedorenko@novek.ru Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Following race condition was detected:
<CPU A, t0> - neigh_flush_dev() is under execution and calls
neigh_mark_dead(n) marking the neighbour entry 'n' as dead.
<CPU B, t1> - Executing: __netif_receive_skb() ->
__netif_receive_skb_core() -> arp_rcv() -> arp_process().arp_process()
calls __neigh_lookup() which takes a reference on neighbour entry 'n'.
<CPU A, t2> - Moves further along neigh_flush_dev() and calls
neigh_cleanup_and_release(n), but since reference count increased in t2,
'n' couldn't be destroyed.
<CPU B, t3> - Moves further along, arp_process() and calls
neigh_update()-> __neigh_update() -> neigh_update_gc_list(), which adds
the neighbour entry back in gc_list(neigh_mark_dead(), removed it
earlier in t0 from gc_list)
This leads to 'n' still being part of gc_list, but the actual
neighbour structure has been freed.
The situation can be prevented from happening if we disallow a dead
entry to have any possibility of updating gc_list. This is what the
patch intends to achieve.
Fixes: 9c29a2f55ec0 ("neighbor: Fix locking order for gc_list changes") Signed-off-by: Chinmay Agarwal <chinagar@codeaurora.org> Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://lore.kernel.org/r/20210127165453.GA20514@chinagar-linux.qualcomm.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Similar to commit 165ae7a8feb5 ("igb: Report speed and duplex as unknown
when device is runtime suspended"), if we try to read speed and duplex
sysfs while the device is runtime suspended, igc will complain and
stops working:
The more generic approach will be wrap get_link_ksettings() with begin()
and complete() callbacks, and calls runtime resume and runtime suspend
routine respectively. However, igc is like igb, runtime resume routine
uses rtnl_lock() which upper ethtool layer also uses.
So to prevent a deadlock on rtnl, take a different approach, use
pm_runtime_suspended() to avoid reading register while device is runtime
suspended.
One customer reports a crash problem which causes by flush request. It
triggers a warning before crash.
/* new request after previous flush is completed */
if (ktime_after(req_start, mddev->prev_flush_start)) {
WARN_ON(mddev->flush_bio);
mddev->flush_bio = bio;
bio = NULL;
}
The WARN_ON is triggered. We use spin lock to protect prev_flush_start and
flush_bio in md_flush_request. But there is no lock protection in
md_submit_flush_data. It can set flush_bio to NULL first because of
compiler reordering write instructions.
For example, flush bio1 sets flush bio to NULL first in
md_submit_flush_data. An interrupt or vmware causing an extended stall
happen between updating flush_bio and prev_flush_start. Because flush_bio
is NULL, flush bio2 can get the lock and submit to underlayer disks. Then
flush bio1 updates prev_flush_start after the interrupt or extended stall.
Then flush bio3 enters in md_flush_request. The start time req_start is
behind prev_flush_start. The flush_bio is not NULL(flush bio2 hasn't
finished). So it can trigger the WARN_ON now. Then it calls INIT_WORK
again. INIT_WORK() will re-initialize the list pointers in the
work_struct, which then can result in a corrupted work list and the
work_struct queued a second time. With the work list corrupted, it can
lead in invalid work items being used and cause a crash in
process_one_work.
We need to make sure only one flush bio can be handled at one same time.
So add spin lock in md_submit_flush_data to protect prev_flush_start and
flush_bio in an atomic way.
Reviewed-by: David Jeffery <djeffery@redhat.com> Signed-off-by: Xiao Ni <xni@redhat.com> Signed-off-by: Song Liu <songliubraving@fb.com> Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ILI251x seems to report pressure information in the 5th byte of
each per-finger touch data element. On the available hardware, this
information has the values ranging from 0x0 to 0xa, which is also
matching the downstream example code. Report pressure information
on the ILI251x.
Jan Kiszka reported that the x2apic_wrmsr_fence() function uses a plain
MFENCE while the Intel SDM (10.12.3 MSR Access in x2APIC Mode) calls for
MFENCE; LFENCE.
Short summary: we have special MSRs that have weaker ordering than all
the rest. Add fencing consistent with current SDM recommendations.
This is not known to cause any issues in practice, only in theory.
Longer story below:
The reason the kernel uses a different semantic is that the SDM changed
(roughly in late 2017). The SDM changed because folks at Intel were
auditing all of the recommended fences in the SDM and realized that the
x2apic fences were insufficient.
Why was the pain MFENCE judged insufficient?
WRMSR itself is normally a serializing instruction. No fences are needed
because the instruction itself serializes everything.
But, there are explicit exceptions for this serializing behavior written
into the WRMSR instruction documentation for two classes of MSRs:
IA32_TSC_DEADLINE and the X2APIC MSRs.
Back to x2apic: WRMSR is *not* serializing in this specific case.
But why is MFENCE insufficient? MFENCE makes writes visible, but
only affects load/store instructions. WRMSR is unfortunately not a
load/store instruction and is unaffected by MFENCE. This means that a
non-serializing WRMSR could be reordered by the CPU to execute before
the writes made visible by the MFENCE have even occurred in the first
place.
This means that an x2apic IPI could theoretically be triggered before
there is any (visible) data to process.
Does this affect anything in practice? I honestly don't know. It seems
quite possible that by the time an interrupt gets to consume the (not
yet) MFENCE'd data, it has become visible, mostly by accident.
To be safe, add the SDM-recommended fences for all x2apic WRMSRs.
This also leaves open the question of the _other_ weakly-ordered WRMSR:
MSR_IA32_TSC_DEADLINE. While it has the same ordering architecture as
the x2APIC MSRs, it seems substantially less likely to be a problem in
practice. While writes to the in-memory Local Vector Table (LVT) might
theoretically be reordered with respect to a weakly-ordered WRMSR like
TSC_DEADLINE, the SDM has this to say:
In x2APIC mode, the WRMSR instruction is used to write to the LVT
entry. The processor ensures the ordering of this write and any
subsequent WRMSR to the deadline; no fencing is required.
But, that might still leave xAPIC exposed. The safest thing to do for
now is to add the extra, recommended LFENCE.
[ bp: Massage commit message, fix typos, drop accidentally added
newline to tools/arch/x86/include/asm/barrier.h. ]
Reported-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com> Signed-off-by: Borislav Petkov <bp@suse.de> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: <stable@vger.kernel.org> Link: https://lkml.kernel.org/r/20200305174708.F77040DD@viggo.jf.intel.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
local_db_save() is called at the start of exc_debug_kernel(), reads DR7 and
disables breakpoints to prevent recursion.
When running in a guest (X86_FEATURE_HYPERVISOR), local_db_save() reads the
per-cpu variable cpu_dr7 to check whether a breakpoint is active or not
before it accesses DR7.
A data breakpoint on cpu_dr7 therefore results in infinite #DB recursion.
Disallow data breakpoints on cpu_dr7 to prevent that.
Fixes: 84b6a3491567a("x86/entry: Optimize local_db_save() for virt") Signed-off-by: Lai Jiangshan <laijs@linux.alibaba.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20210204152708.21308-2-jiangshanlai@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When FSGSBASE is enabled, paranoid_entry() fetches the per-CPU GSBASE value
via __per_cpu_offset or pcpu_unit_offsets.
When a data breakpoint is set on __per_cpu_offset[cpu] (read-write
operation), the specific CPU will be stuck in an infinite #DB loop.
RCU will try to send an NMI to the specific CPU, but it is not working
either since NMI also relies on paranoid_entry(). Which means it's
undebuggable.
Fixes: eaad981291ee3("x86/entry/64: Introduce the FIND_PERCPU_BASE macro") Signed-off-by: Lai Jiangshan <laijs@linux.alibaba.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20210204152708.21308-1-jiangshanlai@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tom reported that one of the GDB test-cases failed, and Boris bisected
it to commit:
d53d9bc0cf78 ("x86/debug: Change thread.debugreg6 to thread.virtual_dr6")
The debugging session led us to commit:
6c0aca288e72 ("x86: Ignore trap bits on single step exceptions")
It turns out that TF and data breakpoints are both traps and will be
merged, while instruction breakpoints are faults and will not be merged.
This means 6c0aca288e72 is wrong, only TF and instruction breakpoints
need to be excluded while TF and data breakpoints can be merged.
With retpolines disabled, some configurations of GCC, and specifically
the GCC versions 9 and 10 in Ubuntu will add Intel CET instrumentation
to the kernel by default. That breaks certain tracing scenarios by
adding a superfluous ENDBR64 instruction before the fentry call, for
functions which can be called indirectly.
CET instrumentation isn't currently necessary in the kernel, as CET is
only supported in user space. Disable it unconditionally and move it
into the x86's Makefile as CET/CFI... enablement should be a per-arch
decision anyway.
[ bp: Massage and extend commit message. ]
Fixes: 29be86d7f9cb ("kbuild: add -fcf-protection=none when using retpoline flags") Reported-by: Nikolay Borisov <nborisov@suse.com> Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com> Signed-off-by: Borislav Petkov <bp@suse.de> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Nikolay Borisov <nborisov@suse.com> Cc: <stable@vger.kernel.org> Cc: Seth Forshee <seth.forshee@canonical.com> Cc: Masahiro Yamada <yamada.masahiro@socionext.com> Link: https://lkml.kernel.org/r/20210128215219.6kct3h2eiustncws@treble Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 3fea5a499d57 ("mm: memcontrol: convert page cache to a new
mem_cgroup_charge() API") introduced a bug in __add_to_page_cache_locked()
causing the following splat:
page dumped because: VM_BUG_ON_PAGE(page_memcg(page))
pages's memcg:ffff8889a4116000
------------[ cut here ]------------
kernel BUG at mm/memcontrol.c:2924!
invalid opcode: 0000 [#1] SMP KASAN PTI
CPU: 35 PID: 12345 Comm: cat Tainted: G S W I 5.11.0-rc4-debug+ #1
Hardware name: HP HP Z8 G4 Workstation/81C7, BIOS P60 v01.25 12/06/2017
RIP: commit_charge+0xf4/0x130
Call Trace:
mem_cgroup_charge+0x175/0x770
__add_to_page_cache_locked+0x712/0xad0
add_to_page_cache_lru+0xc5/0x1f0
cachefiles_read_or_alloc_pages+0x895/0x2e10 [cachefiles]
__fscache_read_or_alloc_pages+0x6c0/0xa00 [fscache]
__nfs_readpages_from_fscache+0x16d/0x630 [nfs]
nfs_readpages+0x24e/0x540 [nfs]
read_pages+0x5b1/0xc40
page_cache_ra_unbounded+0x460/0x750
generic_file_buffered_read_get_pages+0x290/0x1710
generic_file_buffered_read+0x2a9/0xc30
nfs_file_read+0x13f/0x230 [nfs]
new_sync_read+0x3af/0x610
vfs_read+0x339/0x4b0
ksys_read+0xf1/0x1c0
do_syscall_64+0x33/0x40
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Before that commit, there was a try_charge() and commit_charge() in
__add_to_page_cache_locked(). These two separated charge functions were
replaced by a single mem_cgroup_charge(). However, it forgot to add a
matching mem_cgroup_uncharge() when the xarray insertion failed with the
page released back to the pool.
Fix this by adding a mem_cgroup_uncharge() call when insertion error
happens.
Link: https://lkml.kernel.org/r/20210125042441.20030-1-longman@redhat.com Fixes: 3fea5a499d57 ("mm: memcontrol: convert page cache to a new mem_cgroup_charge() API") Signed-off-by: Waiman Long <longman@redhat.com> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Matthew Wilcox <willy@infradead.org> Cc: Miaohe Lin <linmiaohe@huawei.com> Cc: Muchun Song <smuchun@gmail.com> Cc: Michal Hocko <mhocko@suse.com> 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>
Sergey reported deadlock between kswapd correctly doing its usual
lock_page(page) followed by down_read(page->mapping->i_mmap_rwsem), and
madvise(MADV_REMOVE) on an madvise(MADV_HUGEPAGE) area doing
down_write(page->mapping->i_mmap_rwsem) followed by lock_page(page).
This happened when shmem_fallocate(punch hole)'s unmap_mapping_range()
reaches zap_pmd_range()'s call to __split_huge_pmd(). The same deadlock
could occur when partially truncating a mapped huge tmpfs file, or using
fallocate(FALLOC_FL_PUNCH_HOLE) on it.
__split_huge_pmd()'s page lock was added in 5.8, to make sure that any
concurrent use of reuse_swap_page() (holding page lock) could not catch
the anon THP's mapcounts and swapcounts while they were being split.
Fortunately, reuse_swap_page() is never applied to a shmem or file THP
(not even by khugepaged, which checks PageSwapCache before calling), and
anonymous THPs are never created in shmem or file areas: so that
__split_huge_pmd()'s page lock can only be necessary for anonymous THPs,
on which there is no risk of deadlock with i_mmap_rwsem.
Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2101161409470.2022@eggly.anvils Fixes: c444eb564fb1 ("mm: thp: make the THP mapcount atomic against __split_huge_pmd_locked()") Signed-off-by: Hugh Dickins <hughd@google.com> Reported-by: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> Reviewed-by: Andrea Arcangeli <aarcange@redhat.com> 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 VM_MAP_PUT_PAGES was added, it was defined with the same value as
VM_FLUSH_RESET_PERMS. This doesn't seem like it will cause any big
functional problems other than some excess flushing for VM_MAP_PUT_PAGES
allocations.
Redefine VM_MAP_PUT_PAGES to have its own value. Also, rearrange things
so flags are less likely to be missed in the future.
Link: https://lkml.kernel.org/r/20210122233706.9304-1-rick.p.edgecombe@intel.com Fixes: b944afc9d64d ("mm: add a VM_MAP_PUT_PAGES flag for vmap") Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com> Suggested-by: Matthew Wilcox <willy@infradead.org> Cc: Miaohe Lin <linmiaohe@huawei.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Daniel Axtens <dja@axtens.net> 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>
In fast_isolate_freepages, high_pfn will be used if a prefered one (ie
PFN >= low_fn) not found.
But the high_pfn is not reset before searching an free area, so when it
was used as freepage, it may from another free area searched before. As
a result move_freelist_head(freelist, freepage) will have unexpected
behavior (eg corrupt the MOVABLE freelist)
Unable to handle kernel paging request at virtual address dead000000000200
Mem abort info:
ESR = 0x96000044
Exception class = DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
Data abort info:
ISV = 0, ISS = 0x00000044
CM = 0, WnR = 1
[dead000000000200] address between user and kernel address ranges
The issue was reported on an smart phone product with 6GB ram and 3GB
zram as swap device.
This patch fixes the issue by reset high_pfn before searching each free
area, which ensure freepage and freelist match when call
move_freelist_head in fast_isolate_freepages().
Link: http://lkml.kernel.org/r/20190118175136.31341-12-mgorman@techsingularity.net Link: https://lkml.kernel.org/r/20210112094720.1238444-1-wu-yan@tcl.com Fixes: 5a811889de10f1eb ("mm, compaction: use free lists to quickly locate a migration target") Signed-off-by: Rokudo Yan <wu-yan@tcl.com> Acked-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Vlastimil Babka <vbabka@suse.cz> 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>
The page_huge_active() can be called from scan_movable_pages() which do
not hold a reference count to the HugeTLB page. So when we call
page_huge_active() from scan_movable_pages(), the HugeTLB page can be
freed parallel. Then we will trigger a BUG_ON which is in the
page_huge_active() when CONFIG_DEBUG_VM is enabled. Just remove the
VM_BUG_ON_PAGE.
Link: https://lkml.kernel.org/r/20210115124942.46403-6-songmuchun@bytedance.com Fixes: 7e1f049efb86 ("mm: hugetlb: cleanup using paeg_huge_active()") Signed-off-by: Muchun Song <songmuchun@bytedance.com> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Cc: David Hildenbrand <david@redhat.com> Cc: Yang Shi <shy828301@gmail.com> 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 we isolate a HugeTLB page on CPU0. Meanwhile, we free it to the
buddy allocator on CPU1. Then, we can trigger a BUG_ON on CPU0, because
it is already freed to the buddy allocator.
Link: https://lkml.kernel.org/r/20210115124942.46403-5-songmuchun@bytedance.com Fixes: c8721bbbdd36 ("mm: memory-hotplug: enable memory hotplug to handle hugepage") Signed-off-by: Muchun Song <songmuchun@bytedance.com> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Cc: David Hildenbrand <david@redhat.com> Cc: Yang Shi <shy828301@gmail.com> 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>
There is a race condition between __free_huge_page()
and dissolve_free_huge_page().
CPU0: CPU1:
// page_count(page) == 1
put_page(page)
__free_huge_page(page)
dissolve_free_huge_page(page)
spin_lock(&hugetlb_lock)
// PageHuge(page) && !page_count(page)
update_and_free_page(page)
// page is freed to the buddy
spin_unlock(&hugetlb_lock)
spin_lock(&hugetlb_lock)
clear_page_huge_active(page)
enqueue_huge_page(page)
// It is wrong, the page is already freed
spin_unlock(&hugetlb_lock)
The race window is between put_page() and dissolve_free_huge_page().
We should make sure that the page is already on the free list when it is
dissolved.
As a result __free_huge_page would corrupt page(s) already in the buddy
allocator.
Link: https://lkml.kernel.org/r/20210115124942.46403-4-songmuchun@bytedance.com Fixes: c8721bbbdd36 ("mm: memory-hotplug: enable memory hotplug to handle hugepage") Signed-off-by: Muchun Song <songmuchun@bytedance.com> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Hildenbrand <david@redhat.com> Cc: Yang Shi <shy828301@gmail.com> 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>
If a new hugetlb page is allocated during fallocate it will not be
marked as active (set_page_huge_active) which will result in a later
isolate_huge_page failure when the page migration code would like to
move that page. Such a failure would be unexpected and wrong.
Only export set_page_huge_active, just leave clear_page_huge_active as
static. Because there are no external users.
Link: https://lkml.kernel.org/r/20210115124942.46403-3-songmuchun@bytedance.com Fixes: 70c3547e36f5 (hugetlbfs: add hugetlbfs_fallocate()) Signed-off-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Reviewed-by: Oscar Salvador <osalvador@suse.de> Cc: David Hildenbrand <david@redhat.com> Cc: Yang Shi <shy828301@gmail.com> 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>
The tegra_uart_config of the DEBUG_LL code is now placed right at the
start of the .text section after commit which enabled debug output in the
decompressor. Tegra devices are not booting anymore if DEBUG_LL is enabled
since tegra_uart_config data is executes as a code. Fix the misplaced
tegra_uart_config storage by embedding it into the code.
Cc: stable@vger.kernel.org Fixes: 2596a72d3384 ("ARM: 9009/1: uncompress: Enable debug in head.S") Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Dmitry Osipenko <digetx@gmail.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Building with gcc 4.9.2 reveals a latent bug in the PCI accessors
for Footbridge platforms, which causes a fatal alignment fault
while accessing IO memory. Fix this by making the assembly volatile.
Cc: stable@vger.kernel.org Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2fee9583198eb9 ("spi: dt-bindings: clarify CS behavior for spi-cs-high and gpio descriptors")
it was clarified what the proper state for cs-gpios should be, even if the
flag is ignored. The driver code is doing the right thing since
766c6b63aa04 ("spi: fix client driver breakages when using GPIO descriptors")
The chip-select of the td028ttec1 panel is active-low, so we must omit spi-cs-high;
attribute (already removed by separate patch) and should now use GPIO_ACTIVE_LOW for
the client device description to be fully consistent.
Fixes: 766c6b63aa04 ("spi: fix client driver breakages when using GPIO descriptors") CC: stable@vger.kernel.org Signed-off-by: H. Nikolaus Schaller <hns@goldelico.com> Signed-off-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
broke the GTA04 display. This contradicted the data sheet but was the only
way to get it as an spi client operational again.
The panel data sheet defines the chip-select to be active low.
Now, with the arrival of
commit 766c6b63aa04 ("spi: fix client driver breakages when using GPIO descriptors")
the logic of interaction between spi-cs-high and the gpio descriptor flags
has been changed a second time, making the display broken again. So we have
to remove the original fix which in retrospect was a workaround of a bug in
the spi subsystem and not a feature of the panel or bug in the device tree.
With this fix the device tree is back in sync with the data sheet and
spi subsystem code.
Fixes: 766c6b63aa04 ("spi: fix client driver breakages when using GPIO descriptors") CC: stable@vger.kernel.org Signed-off-by: H. Nikolaus Schaller <hns@goldelico.com> Signed-off-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Set cr3_lm_rsvd_bits, which is effectively an invalid GPA mask, at vCPU
reset. The reserved bits check needs to be done even if userspace never
configures the guest's CPUID model.
Cc: stable@vger.kernel.org Fixes: 0107973a80ad ("KVM: x86: Introduce cr3_lm_rsvd_bits in kvm_vcpu_arch") Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20210204000117.3303214-2-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Set the emulator context to PROT64 if SYSENTER transitions from 32-bit
userspace (compat mode) to a 64-bit kernel, otherwise the RIP update at
the end of x86_emulate_insn() will incorrectly truncate the new RIP.
Note, this bug is mostly limited to running an Intel virtual CPU model on
an AMD physical CPU, as other combinations of virtual and physical CPUs
do not trigger full emulation. On Intel CPUs, SYSENTER in compatibility
mode is legal, and unconditionally transitions to 64-bit mode. On AMD
CPUs, SYSENTER is illegal in compatibility mode and #UDs. If the vCPU is
AMD, KVM injects a #UD on SYSENTER in compat mode. If the pCPU is Intel,
SYSENTER will execute natively and not trigger #UD->VM-Exit (ignoring
guest TLB shenanigans).
Fixes: fede8076aab4 ("KVM: x86: handle wrap around 32-bit address space") Cc: stable@vger.kernel.org Signed-off-by: Jonny Barker <jonny@jonnybarker.com>
[sean: wrote changelog] Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20210202165546.2390296-1-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Recent commit 255cbecfe0 modified struct kvm_vcpu_arch to make
'cpuid_entries' a pointer to an array of kvm_cpuid_entry2 entries
rather than embedding the array in the struct. KVM_SET_CPUID and
KVM_SET_CPUID2 were updated accordingly, but KVM_GET_CPUID2 was missed.
As a result, KVM_GET_CPUID2 currently returns random fields from struct
kvm_vcpu_arch to userspace rather than the expected CPUID values. Fix
this by treating 'cpuid_entries' as a pointer when copying its
contents to userspace buffer.
Userspace that does not know about KVM_GET_MSR_FEATURE_INDEX_LIST
will generally use the default value for MSR_IA32_ARCH_CAPABILITIES.
When this happens and the host has tsx=on, it is possible to end up with
virtual machines that have HLE and RTM disabled, but TSX_CTRL available.
If the fleet is then switched to tsx=off, kvm_get_arch_capabilities()
will clear the ARCH_CAP_TSX_CTRL_MSR bit and it will not be possible to
use the tsx=off hosts as migration destinations, even though the guests
do not have TSX enabled.
To allow this migration, allow guests to write to their TSX_CTRL MSR,
while keeping the host MSR unchanged for the entire life of the guests.
This ensures that TSX remains disabled and also saves MSR reads and
writes, and it's okay to do because with tsx=off we know that guests will
not have the HLE and RTM features in their CPUID. (If userspace sets
bogus CPUID data, we do not expect HLE and RTM to work in guests anyway).
Cc: stable@vger.kernel.org Fixes: cbbaa2727aa3 ("KVM: x86: fix presentation of TSX feature in ARCH_CAPABILITIES") Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a bug in the TDP MMU function to zap SPTEs which could be
replaced with a larger mapping which prevents the function from doing
anything. Fix this by correctly zapping the last level SPTEs.
Cc: stable@vger.kernel.org Fixes: 14881998566d ("kvm: x86/mmu: Support disabling dirty logging for the tdp MMU") Signed-off-by: Ben Gardon <bgardon@google.com>
Message-Id: <20210202185734.1680553-11-bgardon@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Don't let KVM load when running as an SEV guest, regardless of what
CPUID says. Memory is encrypted with a key that is not accessible to
the host (L0), thus it's impossible for L0 to emulate SVM, e.g. it'll
see garbage when reading the VMCB.
Technically, KVM could decrypt all memory that needs to be accessible to
the L0 and use shadow paging so that L0 does not need to shadow NPT, but
exposing such information to L0 largely defeats the purpose of running as
an SEV guest. This can always be revisited if someone comes up with a
use case for running VMs inside SEV guests.
Note, VMLOAD, VMRUN, etc... will also #GP on GPAs with C-bit set, i.e. KVM
is doomed even if the SEV guest is debuggable and the hypervisor is willing
to decrypt the VMCB. This may or may not be fixed on CPUs that have the
SVME_ADDR_CHK fix.
Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20210202212017.2486595-1-seanjc@google.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some Kingston A2000 NVMe SSDs sooner or later get confused and stop
working when they use the deepest APST sleep while running Linux. The
system then crashes and one has to cold boot it to get the SSD working
again.
Kingston seems to known about this since at least mid-September 2020:
https://bbs.archlinux.org/viewtopic.php?pid=1926994#p1926994
Someone working for a German company representing Kingston to the German
press confirmed to me Kingston engineering is aware of the issue and
investigating; the person stated that to their current knowledge only
the deepest APST sleep state causes trouble. Therefore, make Linux avoid
it for now by applying the NVME_QUIRK_NO_DEEPEST_PS to this SSD.
I have two such SSDs, but it seems the problem doesn't occur with them.
I hence couldn't verify if this patch really fixes the problem, but all
the data in front of me suggests it should.
This patch can easily be reverted or improved upon if a better solution
surfaces.
FWIW, there are many reports about the issue scattered around the web;
most of the users disabled APST completely to make things work, some
just made Linux avoid the deepest sleep state:
I guess case below can trigger above panic: there're two threads which
operates different io_uring ctxs and share same sqthread identity, and
later one thread exits, io_uring_cancel_task_requests() will clear
task->io_uring->identity->files to be NULL in sqpoll mode, then another
ctx that uses same identity will panic.
Indeed we don't need to clear task->io_uring->identity->files here,
io_grab_identity() should handle identity->files changes well, if
task->io_uring->identity->files is not equal to current->files,
io_cow_identity() should handle this changes well.
Prevent the ICL HDR plane pipeline from performing YUV color range
correction twice when the input is in limited range. This is done by
removing the limited-range code from icl_program_input_csc().
Before this patch the following could happen: user space gives us a YUV
buffer in limited range; per the pipeline in [1], the plane would first
go through a "YUV Range correct" stage that expands the range; the plane
would then go through the "Input CSC" stage which would also expand the
range because icl_program_input_csc() would use a matrix and an offset
that assume limited-range input; this would ultimately cause dark and
light colors to appear darker and lighter than they should respectively.
This is an issue because if a buffer switches between being scanned out
and being composited with the GPU, the user will see a color difference.
If this switching happens quickly and frequently, the user will perceive
this as a flickering.
If we enable_breadcrumbs for a request while that request is being
removed from HW; we may see that the request is active as we take the
ce->signal_lock and proceed to attach the request to ce->signals.
However, during unsubmission after marking the request as inactive, we
see that the request has not yet been added to ce->signals and so skip
the removal. Pull the check during cancel_breadcrumbs under the same
spinlock as enabling so that we the two tests are consistent in
enable/cancel.
Otherwise, we may insert a request onto ce->signals that we expect should
not be there:
While updating, we can note that we are always called with
irqs-disabled, due to the engine->active.lock being held at the single
caller, and so remove the irqsave/restore making it symmetric to
enable_breadcrumbs.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2931 Fixes: c18636f76344 ("drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Andi Shyti <andi.shyti@intel.com> Cc: <stable@vger.kernel.org> # v5.10+ Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210119162057.31097-1-chris@chris-wilson.co.uk
(cherry picked from commit e7004ea4f5f528f5a5018f0b70cab36d25315498) Signed-off-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Simplify the frontbuffer unpin by removing the lock requirement. The LRU
bumping was primarily to protect the GTT from being evicted and from
frontbuffers being eagerly shrunk. Now we protect frontbuffers from the
shrinker, and we avoid accidentally evicting from the GTT, so the
benefit from bumping LRU is no more, and we can save more time by not.
Reported-and-tested-by: Matti Hämäläinen <ccr@tnsp.org> Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2905 Fixes: c1793ba86a41 ("drm/i915: Add ww locking to pin_to_display_plane, v2.") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-6-chris@chris-wilson.co.uk
(cherry picked from commit 14ca83eece9565a2d2177291ceb122982dc38420) Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: <stable@vger.kernel.org> # v5.10+ Signed-off-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Atm the driver will calculate a wrong MST timeslots/MTP (aka time unit)
value for MST streams if the link parameters (link rate or lane count)
are limited in a way independent of the sink capabilities (reported by
DPCD).
One example of such a limitation is when a MUX between the sink and
source connects only a limited number of lanes to the display and
connects the rest of the lanes to other peripherals (USB).
Another issue is that atm MST core calculates the divider based on the
backwards compatible DPCD (at address 0x0000) vs. the extended
capability info (at address 0x2200). This can result in leaving some
part of the MST BW unused (For instance in case of the WD19TB dock).
Fix the above two issues by calculating the PBN divider value based on
the rate and lane count link parameters that the driver uses for all
other computation.
Bugzilla: https://gitlab.freedesktop.org/drm/intel/-/issues/2977 Cc: Lyude Paul <lyude@redhat.com> Cc: Ville Syrjala <ville.syrjala@intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Ville Syrjala <ville.syrjala@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210125173636.1733812-2-imre.deak@intel.com
(cherry picked from commit b59c27cab257cfbff939615a87b72bce83925710) Signed-off-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Grab kvm->lock before pinning memory when registering an encrypted
region; sev_pin_memory() relies on kvm->lock being held to ensure
correctness when checking and updating the number of pinned pages.
Add a lockdep assertion to help prevent future regressions.
Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Joerg Roedel <joro@8bytes.org> Cc: Tom Lendacky <thomas.lendacky@amd.com> Cc: Brijesh Singh <brijesh.singh@amd.com> Cc: Sean Christopherson <seanjc@google.com> Cc: x86@kernel.org Cc: kvm@vger.kernel.org Cc: stable@vger.kernel.org Cc: linux-kernel@vger.kernel.org Fixes: 1e80fdc09d12 ("KVM: SVM: Pin guest memory when SEV is active") Signed-off-by: Peter Gonda <pgonda@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
V2
- Fix up patch description
- Correct file paths svm.c -> sev.c
- Add unlock of kvm->lock on sev_pin_memory error