RSB fill sequence does not have any protection for miss-prediction of
conditional branch at the end of the sequence. CPU can speculatively
execute code immediately after the sequence, while RSB filling hasn't
completed yet.
tl;dr: The Enhanced IBRS mitigation for Spectre v2 does not work as
documented for RET instructions after VM exits. Mitigate it with a new
one-entry RSB stuffing mechanism and a new LFENCE.
== Background ==
Indirect Branch Restricted Speculation (IBRS) was designed to help
mitigate Branch Target Injection and Speculative Store Bypass, i.e.
Spectre, attacks. IBRS prevents software run in less privileged modes
from affecting branch prediction in more privileged modes. IBRS requires
the MSR to be written on every privilege level change.
To overcome some of the performance issues of IBRS, Enhanced IBRS was
introduced. eIBRS is an "always on" IBRS, in other words, just turn
it on once instead of writing the MSR on every privilege level change.
When eIBRS is enabled, more privileged modes should be protected from
less privileged modes, including protecting VMMs from guests.
== Problem ==
Here's a simplification of how guests are run on Linux' KVM:
void run_kvm_guest(void)
{
// Prepare to run guest
VMRESUME();
// Clean up after guest runs
}
The execution flow for that would look something like this to the
processor:
1. Host-side: call run_kvm_guest()
2. Host-side: VMRESUME
3. Guest runs, does "CALL guest_function"
4. VM exit, host runs again
5. Host might make some "cleanup" function calls
6. Host-side: RET from run_kvm_guest()
Now, when back on the host, there are a couple of possible scenarios of
post-guest activity the host needs to do before executing host code:
* on pre-eIBRS hardware (legacy IBRS, or nothing at all), the RSB is not
touched and Linux has to do a 32-entry stuffing.
* on eIBRS hardware, VM exit with IBRS enabled, or restoring the host
IBRS=1 shortly after VM exit, has a documented side effect of flushing
the RSB except in this PBRSB situation where the software needs to stuff
the last RSB entry "by hand".
IOW, with eIBRS supported, host RET instructions should no longer be
influenced by guest behavior after the host retires a single CALL
instruction.
However, if the RET instructions are "unbalanced" with CALLs after a VM
exit as is the RET in #6, it might speculatively use the address for the
instruction after the CALL in #3 as an RSB prediction. This is a problem
since the (untrusted) guest controls this address.
Balanced CALL/RET instruction pairs such as in step #5 are not affected.
== Solution ==
The PBRSB issue affects a wide variety of Intel processors which
support eIBRS. But not all of them need mitigation. Today,
X86_FEATURE_RSB_VMEXIT triggers an RSB filling sequence that mitigates
PBRSB. Systems setting RSB_VMEXIT need no further mitigation - i.e.,
eIBRS systems which enable legacy IBRS explicitly.
However, such systems (X86_FEATURE_IBRS_ENHANCED) do not set RSB_VMEXIT
and most of them need a new mitigation.
Therefore, introduce a new feature flag X86_FEATURE_RSB_VMEXIT_LITE
which triggers a lighter-weight PBRSB mitigation versus RSB_VMEXIT.
The lighter-weight mitigation performs a CALL instruction which is
immediately followed by a speculative execution barrier (INT3). This
steers speculative execution to the barrier -- just like a retpoline
-- which ensures that speculation can never reach an unbalanced RET.
Then, ensure this CALL is retired before continuing execution with an
LFENCE.
In other words, the window of exposure is opened at VM exit where RET
behavior is troublesome. While the window is open, force RSB predictions
sampling for RET targets to a dead end at the INT3. Close the window
with the LFENCE.
There is a subset of eIBRS systems which are not vulnerable to PBRSB.
Add these systems to the cpu_vuln_whitelist[] as NO_EIBRS_PBRSB.
Future systems that aren't vulnerable will set ARCH_CAP_PBRSB_NO.
[ bp: Massage, incorporate review comments from Andy Cooper. ]
In do_adb_query() function of drivers/macintosh/adb.c, req->data is copied
form userland. The parameter "req->data[2]" is missing check, the array
size of adb_handler[] is 16, so adb_handler[req->data[2]].original_address and
adb_handler[req->data[2]].handler_id will lead to oob read.
Cc: stable <stable@kernel.org> Signed-off-by: Ning Qiang <sohu0106@126.com> Reviewed-by: Kees Cook <keescook@chromium.org> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220713153734.2248-1-sohu0106@126.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The BCM4349B1, aka CYW/BCM89359, is a WiFi+BT chip and its Bluetooth
portion can be controlled over serial.
Two subversions are added for the chip, because ROM firmware reports
002.002.013 (at least for the chips I have here), while depending on
patchram firmware revision, either 002.002.013 or 002.002.014 is
reported.
The selftests, when built with newer versions of clang, is found
to have over optimized guests' ucall() function, and eliminating
the stores for uc.cmd (perhaps due to no immediate readers). This
resulted in the userspace side always reading a value of '0', and
causing multiple test failures.
As a result, prevent the compiler from optimizing the stores in
ucall() with WRITE_ONCE().
Suggested-by: Ricardo Koller <ricarkol@google.com> Suggested-by: Reiji Watanabe <reijiw@google.com> Signed-off-by: Raghavendra Rao Ananta <rananta@google.com>
Message-Id: <20220615185706.1099208-1-rananta@google.com> Reviewed-by: Andrew Jones <drjones@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Instead of printing an error message, kvm_stat script fails when we
restrict statistics to a guest by its name and there are multiple guests
with such name:
# kvm_stat -g my_vm
Traceback (most recent call last):
File "/usr/bin/kvm_stat", line 1819, in <module>
main()
File "/usr/bin/kvm_stat", line 1779, in main
options = get_options()
File "/usr/bin/kvm_stat", line 1718, in get_options
options = argparser.parse_args()
File "/usr/lib64/python3.10/argparse.py", line 1825, in parse_args
args, argv = self.parse_known_args(args, namespace)
File "/usr/lib64/python3.10/argparse.py", line 1858, in parse_known_args
namespace, args = self._parse_known_args(args, namespace)
File "/usr/lib64/python3.10/argparse.py", line 2067, in _parse_known_args
start_index = consume_optional(start_index)
File "/usr/lib64/python3.10/argparse.py", line 2007, in consume_optional
take_action(action, args, option_string)
File "/usr/lib64/python3.10/argparse.py", line 1935, in take_action
action(self, namespace, argument_values, option_string)
File "/usr/bin/kvm_stat", line 1649, in __call__
' to specify the desired pid'.format(" ".join(pids)))
TypeError: sequence item 0: expected str instance, int found
To avoid this, it's needed to convert pids int values to strings before
pass them to join().
The root cause of the bug sits in neon_poly1305_blocks. The logic
neon_poly1305_blocks() performed is that if it was called with both s[]
and r[] uninitialized, it will first try to initialize them with the
data from the first "block" that it believed to be 32 bytes in length.
First 16 bytes are used as the key and the next 16 bytes for s[]. This
would lead to the aforementioned read out-of-bound. However, after
calling poly1305_init_arch(), only 16 bytes were deducted from the input
and s[] is initialized yet again with the following 16 bytes. The second
initialization of s[] is certainly redundent which indicates that the
first initialization should be for r[] only.
This patch fixes the issue by calling poly1305_init_arm64() instead of
poly1305_init_arch(). This is also the implementation for the same
algorithm on arm platform.
The fix in commit 3f8dec116210 ("ACPI/APEI: Limit printable size of BERT
table data") does not work as intended on systems where the BIOS has a
fixed size block of memory for the BERT table, relying on s/w to quit
when it finds a record with estatus->block_status == 0. On these systems
all errors are suppressed because the check:
if (region_len < ACPI_BERT_PRINT_MAX_LEN)
always fails.
New scheme skips individual CPER records that are too large, and also
limits the total number of records that will be printed to 5.
Fixes: 3f8dec116210 ("ACPI/APEI: Limit printable size of BERT table data") Cc: All applicable <stable@vger.kernel.org> Signed-off-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Taking a recent change in the i8042 quirklist to this one: Clevo
board_names are somewhat unique, and if not: The generic Board_-/Sys_Vendor
string "Notebook" doesn't help much anyway. So identifying the devices just
by the board_name helps keeping the list significantly shorter and might
even hit more devices requiring the fix.
Signed-off-by: Werner Sembach <wse@tuxedocomputers.com> Fixes: c844d22fe0c0 ("ACPI: video: Force backlight native for Clevo NL5xRU and NL5xNU") Cc: All applicable <stable@vger.kernel.org> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The TongFang PF5PU1G, PF4NU1F, PF5NU1G, and PF5LUXG/TUXEDO BA15 Gen10,
Pulse 14/15 Gen1, and Pulse 15 Gen2 have the same problem as the Clevo
NL5xRU and NL5xNU/TUXEDO Aura 15 Gen1 and Gen2:
They have a working native and video interface. However the default
detection mechanism first registers the video interface before
unregistering it again and switching to the native interface during boot.
This results in a dangling SBIOS request for backlight change for some
reason, causing the backlight to switch to ~2% once per boot on the first
power cord connect or disconnect event. Setting the native interface
explicitly circumvents this buggy behaviour by avoiding the unregistering
process.
Signed-off-by: Werner Sembach <wse@tuxedocomputers.com> Cc: All applicable <stable@vger.kernel.org> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Avoid double free in tun_free_netdev() by moving the
dev->tstats and tun->security allocs to a new ndo_init routine
(tun_net_init()) that will be called by register_netdevice().
ndo_init is paired with the desctructor (tun_free_netdev()),
so if there's an error in register_netdevice() the destructor
will handle the frees.
BUG: KASAN: double-free or invalid-free in selinux_tun_dev_free_security+0x1a/0x20 security/selinux/hooks.c:5605
cgroup_skb/egress programs which sock_fields test installs process packets
flying in both directions, from the client to the server, and in reverse
direction.
Recently added dst_port check relies on the fact that destination
port (remote peer port) of the socket which sends the packet is known ahead
of time. This holds true only for the client socket, which connects to the
known server port.
Filter out any traffic that is not egressing from the client socket in the
BPF program that tests reading the dst_port.
Fixes: 8f50f16ff39d ("selftests/bpf: Extend verifier and bpf_sock tests for dst_port loads") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> Link: https://lore.kernel.org/bpf/20220317113920.1068535-3-jakub@cloudflare.com Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add coverage to the verifier tests and tests for reading bpf_sock fields to
ensure that 32-bit, 16-bit, and 8-bit loads from dst_port field are allowed
only at intended offsets and produce expected values.
While 16-bit and 8-bit access to dst_port field is straight-forward, 32-bit
wide loads need be allowed and produce a zero-padded 16-bit value for
backward compatibility.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/r/20220130115518.213259-3-jakub@cloudflare.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
[OP: backport to 5.10: adjusted context in sock_fields.c] Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
syzbot is reporting lockdep warning at ath9k_wmi_event_tasklet() followed
by kernel panic at get_htc_epid_queue() from ath9k_htc_tx_get_packet() from
ath9k_htc_txstatus() [1], for ath9k_wmi_event_tasklet(WMI_TXSTATUS_EVENTID)
depends on spin_lock_init() from ath9k_init_priv() being already completed.
Since ath9k_wmi_event_tasklet() is set by ath9k_init_wmi() from
ath9k_htc_probe_device(), it is possible that ath9k_wmi_event_tasklet() is
called via tasklet interrupt before spin_lock_init() from ath9k_init_priv()
from ath9k_init_device() from ath9k_htc_probe_device() is called.
Let's hold ath9k_wmi_event_tasklet(WMI_TXSTATUS_EVENTID) no-op until
ath9k_tx_init() completes.
syzbot is reporting lockdep warning followed by kernel panic at
ath9k_htc_rxep() [1], for ath9k_htc_rxep() depends on ath9k_rx_init()
being already completed.
Since ath9k_htc_rxep() is set by ath9k_htc_connect_svc(WMI_BEACON_SVC)
from ath9k_init_htc_services(), it is possible that ath9k_htc_rxep() is
called via timer interrupt before ath9k_rx_init() from ath9k_init_device()
is called.
Since we can't call ath9k_init_device() before ath9k_init_htc_services(),
let's hold ath9k_htc_rxep() no-op until ath9k_rx_init() completes.
The mitigations for RETBleed are currently ineffective on x86_32 since
entry_32.S does not use the required macros. However, for an x86_32
target, the kconfig symbols for them are still enabled by default and
/sys/devices/system/cpu/vulnerabilities/retbleed will wrongly report
that mitigations are in place.
Make all of these symbols depend on X86_64, and only enable RETHUNK by
default on X86_64.
Fixes: f43b9876e857 ("x86/retbleed: Add fine grained Kconfig knobs") Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Borislav Petkov <bp@suse.de> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/YtwSR3NNsWp1ohfV@decadent.org.uk
[bwh: Backported to 5.10/5.15/5.18: adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sk_lookup doesn't allow setting data_in for bpf_prog_run. This doesn't
play well with the verifier tests, since they always set a 64 byte
input buffer. Allow not running verifier tests by setting
bpf_test.runs to a negative value and don't run the ctx access case
for sk_lookup. We have dedicated ctx access tests so skipping here
doesn't reduce coverage.
Allow to pass sk_lookup programs to PROG_TEST_RUN. User space
provides the full bpf_sk_lookup struct as context. Since the
context includes a socket pointer that can't be exposed
to user space we define that PROG_TEST_RUN returns the cookie
of the selected socket or zero in place of the socket pointer.
We don't support testing programs that select a reuseport socket,
since this would mean running another (unrelated) BPF program
from the sk_lookup test handler.
Share the timing / signal interruption logic between different
implementations of PROG_TEST_RUN. There is a change in behaviour
as well. We check the loop exit condition before checking for
pending signals. This resolves an edge case where a signal
arrives during the last iteration. Instead of aborting with
EINTR we return the successful result to user space.
Some cloud hypervisors do not provide IBPB on very recent CPU processors,
including AMD processors affected by Retbleed.
Using IBPB before firmware calls on such systems would cause a GPF at boot
like the one below. Do not enable such calls when IBPB support is not
present.
>From the department of "WTAF? How did we miss that!?"...
When we are recovering a buffer, the first thing we do is check the
buffer magic number and extract the LSN from the buffer. If the LSN
is older than the current LSN, we replay the modification to it. If
the metadata on disk is newer than the transaction in the log, we
skip it. This is a fundamental v5 filesystem metadata recovery
behaviour.
generic/482 failed with an attribute writeback failure during log
recovery. The write verifier caught the corruption before it got
written to disk, and the attr buffer dump looked like:
The highlighted bytes are the LSN that was replayed into the
buffer: 0x100000538. This is cycle 1, block 0x538. Prior to replay,
that block on disk looks like this:
Note the LSN stamped into the buffer on disk: 1/0x900. The version
on disk is much newer than the log transaction that was being
replayed. That's a bug, and should -never- happen.
So I immediately went to look at xlog_recover_get_buf_lsn() to check
that we handled the LSN correctly. I was wondering if there was a
similar "two commits with the same start LSN skips the second
replay" problem with buffers. I didn't get that far, because I found
a much more basic, rudimentary bug: xlog_recover_get_buf_lsn()
doesn't recognise buffers with XFS_ATTR3_LEAF_MAGIC set in them!!!
IOWs, attr3 leaf buffers fall through the magic number checks
unrecognised, so trigger the "recover immediately" behaviour instead
of undergoing an LSN check. IOWs, we incorrectly replay ATTR3 leaf
buffers and that causes silent on disk corruption of inode attribute
forks and potentially other things....
Git history shows this is *another* zero day bug, this time
introduced in commit 50d5c8d8e938 ("xfs: check LSN ordering for v5
superblocks during recovery") which failed to handle the attr3 leaf
buffers in recovery. And we've failed to handle them ever since...
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
to format the log inode. It writes the LSN from the inode item into
the log inode, and if recovery decides the inode item needs to be
replayed, it recovers the log inode LSN field and writes it into the
on disk inode LSN field.
Now this might seem like a reasonable thing to do, but it is wrong
on multiple levels. Firstly, if the item is not yet in the AIL,
item->li_lsn is zero. i.e. the first time the inode it is logged and
formatted, the LSN we write into the log inode will be zero. If we
only log it once, recovery will run and can write this zero LSN into
the inode.
This means that the next time the inode is logged and log recovery
runs, it will *always* replay changes to the inode regardless of
whether the inode is newer on disk than the version in the log and
that violates the entire purpose of recording the LSN in the inode
at writeback time (i.e. to stop it going backwards in time on disk
during recovery).
Secondly, if we commit the CIL to the journal so the inode item
moves to the AIL, and then relog the inode, the LSN that gets
stamped into the log inode will be the LSN of the inode's current
location in the AIL, not it's age on disk. And it's not the LSN that
will be associated with the current change. That means when log
recovery replays this inode item, the LSN that ends up on disk is
the LSN for the previous changes in the log, not the current
changes being replayed. IOWs, after recovery the LSN on disk is not
in sync with the LSN of the modifications that were replayed into
the inode. This, again, violates the recovery ordering semantics
that on-disk writeback LSNs provide.
Hence the inode LSN in the log dinode is -always- invalid.
Thirdly, recovery actually has the LSN of the log transaction it is
replaying right at hand - it uses it to determine if it should
replay the inode by comparing it to the on-disk inode's LSN. But it
doesn't use that LSN to stamp the LSN into the inode which will be
written back when the transaction is fully replayed. It uses the one
in the log dinode, which we know is always going to be incorrect.
Looking back at the change history, the inode logging was broken by
commit 93f958f9c41f ("xfs: cull unnecessary icdinode fields") way
back in 2016 by a stupid idiot who thought he knew how this code
worked. i.e. me. That commit replaced an in memory di_lsn field that
was updated only at inode writeback time from the inode item.li_lsn
value - and hence always contained the same LSN that appeared in the
on-disk inode - with a read of the inode item LSN at inode format
time. CLearly these are not the same thing.
Before 93f958f9c41f, the log recovery behaviour was irrelevant,
because the LSN in the log inode always matched the on-disk LSN at
the time the inode was logged, hence recovery of the transaction
would never make the on-disk LSN in the inode go backwards or get
out of sync.
A symptom of the problem is this, caught from a failure of
generic/482. Before log recovery, the inode has been allocated but
never used:
You can see that the LSN of the on-disk inode is 0, even though it
clearly has been written to disk. I point out this inode, because
the generic/482 failure occurred because several adjacent inodes in
this specific inode cluster were not replayed correctly and still
appeared to be zero on disk when all the other metadata (inobt,
finobt, directories, etc) indicated they should be allocated and
written back.
The fix for this is two-fold. The first is that we need to either
revert the LSN changes in 93f958f9c41f or stop logging the inode LSN
altogether. If we do the former, log recovery does not need to
change but we add 8 bytes of memory per inode to store what is
largely a write-only inode field. If we do the latter, log recovery
needs to stamp the on-disk inode in the same manner that inode
writeback does.
I prefer the latter, because we shouldn't really be trying to log
and replay changes to the on disk LSN as the on-disk value is the
canonical source of the on-disk version of the inode. It also
matches the way we recover buffer items - we create a buf_log_item
that carries the current recovery transaction LSN that gets stamped
into the buffer by the write verifier when it gets written back
when the transaction is fully recovered.
However, this might break log recovery on older kernels even more,
so I'm going to simply ignore the logged value in recovery and stamp
the on-disk inode with the LSN of the transaction being recovered
that will trigger writeback on transaction recovery completion. This
will ensure that the on-disk inode LSN always reflects the LSN of
the last change that was written to disk, regardless of whether it
comes from log recovery or runtime writeback.
Fixes: 93f958f9c41f ("xfs: cull unnecessary icdinode fields") Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This code goes back to a time when transaction commits wrote
directly to iclogs. The associated log items were pinned, written to
the log, and then "uncommitted" if some part of the log write had
failed. This uncommit sequence called an ->iop_unpin_remove()
handler that was eventually folded into ->iop_unpin() via the remove
parameter. The log subsystem has since changed significantly in that
transactions commit to the CIL instead of direct to iclogs, though
log items must still be aborted in the event of an eventual log I/O
error. However, the context for a log item abort is now asynchronous
from transaction commit, which means the committing transaction has
been freed by this point in time and the transaction uncommit
sequence of events is no longer relevant.
Further, since stale buffers remain locked at transaction commit
through unpin, we can be certain that the buffer is not associated
with any transaction when the unpin callback executes. Remove this
unused hunk of code and replace it with an assertion that the buffer
is disassociated from transaction context.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The special processing used to simulate a buffer I/O failure on fs
shutdown has a difficult to reproduce race that can result in a use
after free of the associated buffer. Consider a buffer that has been
committed to the on-disk log and thus is AIL resident. The buffer
lands on the writeback delwri queue, but is subsequently locked,
committed and pinned by another transaction before submitted for
I/O. At this point, the buffer is stuck on the delwri queue as it
cannot be submitted for I/O until it is unpinned. A log checkpoint
I/O failure occurs sometime later, which aborts the bli. The unpin
handler is called with the aborted log item, drops the bli reference
count, the pin count, and falls into the I/O failure simulation
path.
The potential problem here is that once the pin count falls to zero
in ->iop_unpin(), xfsaild is free to retry delwri submission of the
buffer at any time, before the unpin handler even completes. If
delwri queue submission wins the race to the buffer lock, it
observes the shutdown state and simulates the I/O failure itself.
This releases both the bli and delwri queue holds and frees the
buffer while xfs_buf_item_unpin() sits on xfs_buf_lock() waiting to
run through the same failure sequence. This problem is rare and
requires many iterations of fstest generic/019 (which simulates disk
I/O failures) to reproduce.
To avoid this problem, grab a hold on the buffer before the log item
is unpinned if the associated item has been aborted and will require
a simulated I/O failure. The hold is already required for the
simulated I/O failure, so the ordering simply guarantees the unpin
handler access to the buffer before it is unpinned and thus
processed by the AIL. This particular ordering is required so long
as the AIL does not acquire a reference on the bli, which is the
long term solution to this problem.
Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If any part of log intent item recovery fails, we should shut down the
log immediately to stop the log from writing a clean unmount record to
disk, because the metadata is not consistent. The inability to cancel a
dirty transaction catches most of these cases, but there are a few
things that have slipped through the cracks, such as ENOSPC from a
transaction allocation, or runtime errors that result in cancellation of
a non-dirty transaction.
This solves some weird behaviors reported by customers where a system
goes down, the first mount fails, the second succeeds, but then the fs
goes down later because of inconsistent metadata.
Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
During regular operation, the xfs_inactive operations create
transactions with zero block reservation because in general we're
freeing space, not asking for more. The per-AG space reservations
created at mount time enable us to handle expansions of the refcount
btree without needing to reserve blocks to the transaction.
Unfortunately, log recovery doesn't create the per-AG space reservations
when intent items are being recovered. This isn't an issue for intent
item recovery itself because they explicitly request blocks, but any
inode inactivation that can happen during log recovery uses the same
xfs_inactive paths as regular runtime. If a refcount btree expansion
happens, the transaction will fail due to blk_res_used > blk_res, and we
shut down the filesystem unnecessarily.
Fix this problem by making per-AG reservations temporarily so that we
can handle the inactivations, and releasing them at the end. This
brings the recovery environment closer to the runtime environment.
Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While I was running with KASAN and lockdep enabled, I stumbled upon an
KASAN report about a UAF to a freed CIL checkpoint. Looking at the
comment for xfs_log_item_in_current_chkpt, it seems pretty obvious to me
that the original patch to xfs_defer_finish_noroll should have done
something to lock the CIL to prevent it from switching the CIL contexts
while the predicate runs.
For upper level code that needs to know if a given log item is new
enough not to need relogging, add a new wrapper that takes the CIL
context lock long enough to sample the current CIL context. This is
kind of racy in that the CIL can switch the contexts immediately after
sampling, but that's ok because the consequence is that the defer ops
code is a little slow to relog items.
==================================================================
BUG: KASAN: use-after-free in xfs_log_item_in_current_chkpt+0x139/0x160 [xfs]
Read of size 8 at addr ffff88804ea5f608 by task fsstress/527999
Memory state around the buggy address: ffff88804ea5f500: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff88804ea5f580: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff88804ea5f600: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
^ ffff88804ea5f680: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb ffff88804ea5f700: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================
Fixes: 4e919af7827a ("xfs: periodically relog deferred intent items") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[backported from CIL scalability series for dependency]
In doing an investigation into AIL push stalls, I was looking at the
log force code to see if an async CIL push could be done instead.
This lead me to xfs_log_force_lsn() and looking at how it works.
xfs_log_force_lsn() is only called from inode synchronisation
contexts such as fsync(), and it takes the ip->i_itemp->ili_last_lsn
value as the LSN to sync the log to. This gets passed to
xlog_cil_force_lsn() via xfs_log_force_lsn() to flush the CIL to the
journal, and then used by xfs_log_force_lsn() to flush the iclogs to
the journal.
The problem is that ip->i_itemp->ili_last_lsn does not store a
log sequence number. What it stores is passed to it from the
->iop_committing method, which is called by xfs_log_commit_cil().
The value this passes to the iop_committing method is the CIL
context sequence number that the item was committed to.
As it turns out, xlog_cil_force_lsn() converts the sequence to an
actual commit LSN for the related context and returns that to
xfs_log_force_lsn(). xfs_log_force_lsn() overwrites it's "lsn"
variable that contained a sequence with an actual LSN and then uses
that to sync the iclogs.
This caused me some confusion for a while, even though I originally
wrote all this code a decade ago. ->iop_committing is only used by
a couple of log item types, and only inode items use the sequence
number it is passed.
Let's clean up the API, CIL structures and inode log item to call it
a sequence number, and make it clear that the high level code is
using CIL sequence numbers and not on-disk LSNs for integrity
synchronisation purposes.
Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Allison Henderson <allison.henderson@oracle.com> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Factor out the log syncing logic into two helpers to make the code easier
to read and more maintainable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Acked-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Handle 0x0020, DMI type 17, 84 bytes
Memory Device
Array Handle: 0x0013
Error Information Handle: Not Provided
Total Width: 72 bits
Data Width: 64 bits
Size: 32 GB
Form Factor: DIMM
Set: None
Locator: PROC 1 DIMM 1 <===== device
Bank Locator: Not Specified <===== bank
This results in a buffer overflow because ghes_edac_register() calls
strlen() on an uninitialized label, which had non-zero values left over
from krealloc_array():
The label contains garbage because the commit in Fixes reallocs the
DIMMs array while scanning the system but doesn't clear the newly
allocated memory.
Change dimm_setup_label() to always initialize the label to fix the
issue. Set it to the empty string in case BIOS does not provide both
bank and device so that ghes_edac_register() can keep the default label
given by edac_mc_alloc_dimms().
[ bp: Rewrite commit message. ]
Fixes: b9cae27728d1f ("EDAC/ghes: Scan the system once on driver init") Co-developed-by: Robert Richter <rric@kernel.org> Signed-off-by: Robert Richter <rric@kernel.org> Signed-off-by: Toshi Kani <toshi.kani@hpe.com> Signed-off-by: Borislav Petkov <bp@suse.de> Tested-by: Robert Elliott <elliott@hpe.com> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20220719220124.760359-1-toshi.kani@hpe.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 26f09e9b3a06 ("mm/memblock: add memblock memory allocation apis")
added a check to determine whether arm_dma_zone_size is exceeding the
amount of kernel virtual address space available between the upper 4GB
virtual address limit and PAGE_OFFSET in order to provide a suitable
definition of MAX_DMA_ADDRESS that should fit within the 32-bit virtual
address space. The quantity used for comparison was off by a missing
trailing 0, leading to MAX_DMA_ADDRESS to be overflowing a 32-bit
quantity.
This was caught thanks to CONFIG_DEBUG_VIRTUAL on the bcm2711 platform
where we define a dma_zone_size of 1GB and we have a PAGE_OFFSET value
of 0xc000_0000 (CONFIG_VMSPLIT_3G) leading to MAX_DMA_ADDRESS being
0x1_0000_0000 which overflows the unsigned long type used throughout
__pa() and then __virt_addr_valid(). Because the virtual address passed
to __virt_addr_valid() would now be 0, the function would loudly warn
and flood the kernel log, thus making the platform unable to boot
properly.
There was a report that a task is waiting at the
throttle_direct_reclaim. The pgscan_direct_throttle in vmstat was
increasing.
This is a bug where zone_watermark_fast returns true even when the free
is very low. The commit f27ce0e14088 ("page_alloc: consider highatomic
reserve in watermark fast") changed the watermark fast to consider
highatomic reserve. But it did not handle a negative value case which
can be happened when reserved_highatomic pageblock is bigger than the
actual free.
If watermark is considered as ok for the negative value, allocating
contexts for order-0 will consume all free pages without direct reclaim,
and finally free page may become depleted except highatomic free.
Then allocating contexts may fall into throttle_direct_reclaim. This
symptom may easily happen in a system where wmark min is low and other
reclaimers like kswapd does not make free pages quickly.
Handle the negative case by using MIN.
Link: https://lkml.kernel.org/r/20220725095212.25388-1-jaewon31.kim@samsung.com Fixes: f27ce0e14088 ("page_alloc: consider highatomic reserve in watermark fast") Signed-off-by: Jaewon Kim <jaewon31.kim@samsung.com> Reported-by: GyeongHwan Hong <gh21.hong@samsung.com> Acked-by: Mel Gorman <mgorman@techsingularity.net> Cc: Minchan Kim <minchan@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Yong-Taek Lee <ytk.lee@samsung.com> Cc: <stable@vger.kerenl.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This happens when calling sctp_sendmsg without connecting to server first.
In this case, a data chunk already queues up in send queue of client side
when processing the INIT_ACK from server in sctp_process_init() where it
calls sctp_stream_init() to alloc stream_in. If it fails to alloc stream_in
all stream_out will be freed in sctp_stream_init's err path. Then in the
asoc freeing it will crash when dequeuing this data chunk as stream_out
is missing.
As we can't free stream out before dequeuing all data from send queue, and
this patch is to fix it by moving the err path stream_out/in freeing in
sctp_stream_init() to sctp_stream_free() which is eventually called when
freeing the asoc in sctp_association_free(). This fix also makes the code
in sctp_process_init() more clear.
Note that in sctp_association_init() when it fails in sctp_stream_init(),
sctp_association_free() will not be called, and in that case it should
go to 'stream_free' err path to free stream instead of 'fail_init'.
Sending a PTP packet can imply to use the normal TX driver datapath but
invoked from the driver's ptp worker. The kernel generic TX code
disables softirqs and preemption before calling specific driver TX code,
but the ptp worker does not. Although current ptp driver functionality
does not require it, there are several reasons for doing so:
1) The invoked code is always executed with softirqs disabled for non
PTP packets.
2) Better if a ptp packet transmission is not interrupted by softirq
handling which could lead to high latencies.
3) netdev_xmit_more used by the TX code requires preemption to be
disabled.
Indeed a solution for dealing with kernel preemption state based on static
kernel configuration is not possible since the introduction of dynamic
preemption level configuration at boot time using the static calls
functionality.
When using 'perf mem' and 'perf c2c', an issue is observed that tool
reports the wrong offset for global data symbols. This is a common
issue on both x86 and Arm64 platforms.
Let's see an example, for a test program, below is the disassembly for
its .bss section which is dumped with objdump:
First we used 'perf mem record' to run the test program and then used
'perf --debug verbose=4 mem report' to observe what's the symbol info
for 'buf1' and 'buf2' structures.
The perf tool relies on libelf to parse symbols, in executable and
shared object files, 'st_value' holds a virtual address; 'sh_addr' is
the address at which section's first byte should reside in memory, and
'sh_offset' is the byte offset from the beginning of the file to the
first byte in the section. The perf tool uses below formula to convert
a symbol's memory address to a file address:
We can see the final adjusted address ranges for buf1 and buf2 are
[0x30a8-0x30e8) and [0x3068-0x30a8) respectively, apparently this is
incorrect, in the code, the structure for 'buf1' and 'buf2' specifies
compiler attribute with 64-byte alignment.
The problem happens for 'sh_offset', libelf returns it as 0x3028 which
is not 64-byte aligned, combining with disassembly, it's likely libelf
doesn't respect the alignment for .bss section, therefore, it doesn't
return the aligned value for 'sh_offset'.
Suggested by Fangrui Song, ELF file contains program header which
contains PT_LOAD segments, the fields p_vaddr and p_offset in PT_LOAD
segments contain the execution info. A better choice for converting
memory address to file address is using the formula:
file_address = st_value - p_vaddr + p_offset
This patch introduces elf_read_program_header() which returns the
program header based on the passed 'st_value', then it uses the formula
above to calculate the symbol file address; and the debugging log is
updated respectively.
We try using cancel_delayed_work_sync() to prevent the work from
enabling NAPI. This is insufficient since we don't disable the source
of the refill work scheduling. This means an NAPI poll callback after
cancel_delayed_work_sync() can schedule the refill work then can
re-enable the NAPI that leads to use-after-free [1].
Since the work can enable NAPI, we can't simply disable NAPI before
calling cancel_delayed_work_sync(). So fix this by introducing a
dedicated boolean to control whether or not the work could be
scheduled from NAPI.
[1]
==================================================================
BUG: KASAN: use-after-free in refill_work+0x43/0xd4
Read of size 2 at addr ffff88810562c92e by task kworker/2:1/42
Fixes: b2baed69e605c ("virtio_net: set/cancel work on ndo_open/ndo_stop") Signed-off-by: Jason Wang <jasowang@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are sleep in atomic context bugs in timer handlers of sctp
such as sctp_generate_t3_rtx_event(), sctp_generate_probe_event(),
sctp_generate_t1_init_event(), sctp_generate_timeout_event(),
sctp_generate_t3_rtx_event() and so on.
The root cause is sctp_sched_prio_init_sid() with GFP_KERNEL parameter
that may sleep could be called by different timer handlers which is in
interrupt context.
One of the call paths that could trigger bug is shown below:
This patch changes gfp_t parameter of init_sid in sctp_sched_set_sched()
from GFP_KERNEL to GFP_ATOMIC in order to prevent sleep in atomic
context bugs.
Fix the inability to bring an interface up on a setup with
only MSI interrupts enabled (no MSI-X).
Solution is to add a default number of QPs = 1. This is enough,
since without MSI-X support driver enables only a basic feature set.
Fixes: bc6d33c8d93f ("i40e: Fix the number of queues available to be mapped for use") Signed-off-by: Dawid Lukwinski <dawid.lukwinski@intel.com> Signed-off-by: Michal Maloszewski <michal.maloszewski@intel.com> Tested-by: Dave Switzer <david.switzer@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Link: https://lore.kernel.org/r/20220722175401.112572-1-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
While reading sysctl_tcp_reflect_tos, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its readers.
Fixes: ac8f1710c12b ("tcp: reflect tos value received in SYN to the socket") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Acked-by: Wei Wang <weiwan@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
init_rx_sa() allocates relevant resource for rx_sa->stats and rx_sa->
key.tfm with alloc_percpu() and macsec_alloc_tfm(). When some error
occurs after init_rx_sa() is called in macsec_add_rxsa(), the function
released rx_sa with kfree() without releasing rx_sa->stats and rx_sa->
key.tfm, which will lead to a resource leak.
We should call macsec_rxsa_put() instead of kfree() to decrease the ref
count of rx_sa and release the relevant resource if the refcount is 0.
The same bug exists in macsec_add_txsa() for tx_sa as well. This patch
fixes the above two bugs.
Fixes: 3cf3227a21d1 ("net: macsec: hardware offloading infrastructure") Signed-off-by: Jianglei Nie <niejianglei2021@163.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently, MACSEC_SA_ATTR_PN is handled inconsistently, sometimes as a
u32, sometimes forced into a u64 without checking the actual length of
the attribute. Instead, we can use nla_get_u64 everywhere, which will
read up to 64 bits into a u64, capped by the actual length of the
attribute coming from userspace.
This fixes several issues:
- the check in validate_add_rxsa doesn't work with 32-bit attributes
- the checks in validate_add_txsa and validate_upd_sa incorrectly
reject X << 32 (with X != 0)
Fixes: 48ef50fa866a ("macsec: Netlink support of XPN cipher suites (IEEE 802.1AEbw)") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
IEEE 802.1AEbw-2013 (section 10.7.8) specifies that the maximum value
of the replay window is 2^30-1, to help with recovery of the upper
bits of the PN.
To avoid leaving the existing macsec device in an inconsistent state
if this test fails during changelink, reuse the cleanup mechanism
introduced for HW offload. This wasn't needed until now because
macsec_changelink_common could not fail during changelink, as
modifying the cipher suite was not allowed.
Finally, this must happen after handling IFLA_MACSEC_CIPHER_SUITE so
that secy->xpn is set.
Fixes: 48ef50fa866a ("macsec: Netlink support of XPN cipher suites (IEEE 802.1AEbw)") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The expected length is MACSEC_SALT_LEN, not MACSEC_SA_ATTR_SALT.
Fixes: 48ef50fa866a ("macsec: Netlink support of XPN cipher suites (IEEE 802.1AEbw)") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 48ef50fa866a added a test on tb_sa[MACSEC_SA_ATTR_PN], but
nothing guarantees that it's not NULL at this point. The same code was
added to macsec_add_txsa, but there it's not a problem because
validate_add_txsa checks that the MACSEC_SA_ATTR_PN attribute is
present.
Note: it's not possible to reproduce with iproute, because iproute
doesn't allow creating an SA without specifying the PN.
Fixes: 48ef50fa866a ("macsec: Netlink support of XPN cipher suites (IEEE 802.1AEbw)") Link: https://bugzilla.kernel.org/show_bug.cgi?id=208315 Reported-by: Frantisek Sumsal <fsumsal@redhat.com> Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since commit 1033990ac5b2 ("sctp: implement memory accounting on tx path"),
SCTP has supported memory accounting on tx path where 'sctp_wmem' is used
by sk_wmem_schedule(). So we should fix the description for this option in
ip-sysctl.rst accordingly.
v1->v2:
- Improve the description as Marcelo suggested.
Fixes: 1033990ac5b2 ("sctp: implement memory accounting on tx path") Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
While reading sysctl_tcp_invalid_ratelimit, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its reader.
Fixes: 032ee4236954 ("tcp: helpers to mitigate ACK loops by rate-limiting out-of-window dupacks") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
While reading sysctl_tcp_autocorking, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: f54b311142a9 ("tcp: auto corking") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
While reading sysctl_tcp_min_rtt_wlen, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: f672258391b4 ("tcp: track min RTT using windowed min-filter") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
tls_device_down takes a reference on all contexts it's going to move to
the degraded state (software fallback). If sk_destruct runs afterwards,
it can reduce the reference counter back to 1 and return early without
destroying the context. Then tls_device_down will release the reference
it took and call tls_device_free_ctx. However, the context will still
stay in tls_device_down_list forever. The list will contain an item,
memory for which is released, making a memory corruption possible.
Fix the above bug by properly removing the context from all lists before
any call to tls_device_free_ctx.
Fixes: 3740651bf7e2 ("tls: Fix context leak on tls_device_down") Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change net device's MTU to smaller than IPV6_MIN_MTU or unregister
device while matching route. That may trigger null-ptr-deref bug
for ip6_ptr probability as following.
=========================================================
BUG: KASAN: null-ptr-deref in find_match.part.0+0x70/0x134
Read of size 4 at addr 0000000000000308 by task ping6/263
Reproducer as following:
Firstly, prepare conditions:
$ip netns add ns1
$ip netns add ns2
$ip link add veth1 type veth peer name veth2
$ip link set veth1 netns ns1
$ip link set veth2 netns ns2
$ip netns exec ns1 ip -6 addr add 2001:0db8:0:f101::1/64 dev veth1
$ip netns exec ns2 ip -6 addr add 2001:0db8:0:f101::2/64 dev veth2
$ip netns exec ns1 ifconfig veth1 up
$ip netns exec ns2 ifconfig veth2 up
$ip netns exec ns1 ip -6 route add 2000::/64 dev veth1 metric 1
$ip netns exec ns2 ip -6 route add 2001::/64 dev veth2 metric 1
Secondly, execute the following two commands in two ssh windows
respectively:
$ip netns exec ns1 sh
$while true; do ip -6 addr add 2001:0db8:0:f101::1/64 dev veth1; ip -6 route add 2000::/64 dev veth1 metric 1; ping6 2000::2; done
$ip netns exec ns1 sh
$while true; do ip link set veth1 mtu 1000; ip link set veth1 mtu 1500; sleep 5; done
It is because ip6_ptr has been assigned to NULL in addrconf_ifdown() firstly,
then ip6_ignore_linkdown() accesses ip6_ptr directly without NULL check.
When we close ping6 sockets, some resources are left unfreed because
pingv6_prot is missing sk->sk_prot->destroy(). As reported by
syzbot [0], just three syscalls leak 96 bytes and easily cause OOM.
struct ipv6_sr_hdr *hdr;
char data[24] = {0};
int fd;
To fix memory leaks, let's add a destroy function.
Note the socket() syscall checks if the GID is within the range of
net.ipv4.ping_group_range. The default value is [1, 0] so that no
GID meets the condition (1 <= GID <= 0). Thus, the local DoS does
not succeed until we change the default value. However, at least
Ubuntu/Fedora/RHEL loosen it.
While reading sysctl_tcp_limit_output_bytes, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its reader.
Fixes: 46d3ceabd8d9 ("tcp: TCP Small Queues") Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This to-be-reverted commit was meant to apply a stricter rule for the
stack to enter pingpong mode. However, the condition used to check for
interactive session "before(tp->lsndtime, icsk->icsk_ack.lrcvtime)" is
jiffy based and might be too coarse, which delays the stack entering
pingpong mode.
We revert this patch so that we no longer use the above condition to
determine interactive session, and also reduce pingpong threshold to 1.
Fixes: 4a41f453bedf ("tcp: change pingpong threshold to 3") Reported-by: LemmyHuang <hlm3280@163.com> Suggested-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: Wei Wang <weiwan@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://lore.kernel.org/r/20220721204404.388396-1-weiwan@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tx side sets EOP and RS bits on descriptors to indicate that a
particular descriptor is the last one and needs to generate an irq when
it was sent. These bits should not be checked on completion path
regardless whether it's the Tx or the Rx. DD bit serves this purpose and
it indicates that a particular descriptor is either for Rx or was
successfully Txed. EOF is also set as loopback test does not xmit
fragmented frames.
Look at (DD | EOF) bits setting in ice_lbtest_receive_frames() instead
of EOP and RS pair.
Fixes: 0e674aeb0b77 ("ice: Add handler for ethtool selftest") Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a watch is being added to a queue, it needs to guard against
interference from addition of a new watch, manual removal of a watch and
removal of a watch due to some other queue being destroyed.
KEYCTL_WATCH_KEY guards against this for the same {key,queue} pair by
holding the key->sem writelocked and by holding refs on both the key and
the queue - but that doesn't prevent interaction from other {key,queue}
pairs.
While add_watch_to_object() does take the spinlock on the event queue,
it doesn't take the lock on the source's watch list. The assumption was
that the caller would prevent that (say by taking key->sem) - but that
doesn't prevent interference from the destruction of another queue.
Fix this by locking the watcher list in add_watch_to_object().
Since __post_watch_notification() walks wlist->watchers with only the
RCU read lock held, we need to use RCU methods to add to the list (we
already use RCU methods to remove from the list).
Fix add_watch_to_object() to use hlist_add_head_rcu() instead of
hlist_add_head() for that list.
Fixes: c73be61cede5 ("pipe: Add general notification queue support") Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Users may request that pages from an OpenCL SVM allocation be migrated
to the GPU with clEnqueueSVMMigrateMem(). In Nouveau this will call into
nouveau_dmem_migrate_vma() to do the migration. If the total range to be
migrated exceeds SG_MAX_SINGLE_ALLOC the pages will be migrated in
chunks of size SG_MAX_SINGLE_ALLOC. However a typo in updating the
starting address means that only the first chunk will get migrated.
Fix the calculation so that the entire range will get migrated if
possible.
Signed-off-by: Alistair Popple <apopple@nvidia.com> Fixes: e3d8b0890469 ("drm/nouveau/svm: map pages after migration") Reviewed-by: Ralph Campbell <rcampbell@nvidia.com> Reviewed-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220720062745.960701-1-apopple@nvidia.com Cc: <stable@vger.kernel.org> # v5.8+ Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch slightly reworks the s390 arch_get_random_seed_{int,long}
implementation: Make sure the CPACF trng instruction is never
called in any interrupt context. This is done by adding an
additional condition in_task().
Justification:
There are some constrains to satisfy for the invocation of the
arch_get_random_seed_{int,long}() functions:
- They should provide good random data during kernel initialization.
- They should not be called in interrupt context as the TRNG
instruction is relatively heavy weight and may for example
make some network loads cause to timeout and buck.
However, it was not clear what kind of interrupt context is exactly
encountered during kernel init or network traffic eventually calling
arch_get_random_seed_long().
After some days of investigations it is clear that the s390
start_kernel function is not running in any interrupt context and
so the trng is called:
which confirms that the call is in softirq context. So in_task() covers exactly
the cases where we want to have CPACF trng called: not in nmi, not in hard irq,
not in soft irq but in normal task context and during kernel init.
Signed-off-by: Harald Freudenberger <freude@linux.ibm.com> Acked-by: Jason A. Donenfeld <Jason@zx2c4.com> Reviewed-by: Juergen Christ <jchrist@linux.ibm.com> Link: https://lore.kernel.org/r/20220713131721.257907-1-freude@linux.ibm.com Fixes: e4f74400308c ("s390/archrandom: simplify back to earlier design and initialize earlier")
[agordeev@linux.ibm.com changed desc, added Fixes and Link, removed -stable] Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
==================================================================
BUG: KASAN: use-after-free in ntfs_ucsncmp+0x123/0x130
Read of size 2 at addr ffff8880751acee8 by task a.out/879
This commit introduced a regression that can cause mount hung. The
changes in __ocfs2_find_empty_slot causes that any node with none-zero
node number can grab the slot that was already taken by node 0, so node 1
will access the same journal with node 0, when it try to grab journal
cluster lock, it will hung because it was already acquired by node 0.
It's very easy to reproduce this, in one cluster, mount node 0 first, then
node 1, you will see the following call trace from node 1.
To fix it, we can just fix __ocfs2_find_empty_slot. But original commit
introduced the feature to mount ocfs2 locally even it is cluster based,
that is a very dangerous, it can easily cause serious data corruption,
there is no way to stop other nodes mounting the fs and corrupting it.
Setup ha or other cluster-aware stack is just the cost that we have to
take for avoiding corruption, otherwise we have to do it in kernel.
Link: https://lkml.kernel.org/r/20220603222801.42488-1-junxiao.bi@oracle.com Fixes: 912f655d78c5("ocfs2: mount shared volume without ha stack") Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com> Acked-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Changwei Ge <gechangwei@live.cn> Cc: Gang He <ghe@suse.com> Cc: Jun Piao <piaojun@huawei.com> Cc: <heming.zhao@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This fixes the following trace which is caused by hci_rx_work starting up
*after* the final channel reference has been put() during sock_close() but
*before* the references to the channel have been destroyed, so instead
the code now rely on kref_get_unless_zero/l2cap_chan_hold_unless_zero to
prevent referencing a channel that is about to be destroyed.
refcount_t: increment on 0; use-after-free.
BUG: KASAN: use-after-free in refcount_dec_and_test+0x20/0xd0
Read of size 4 at addr ffffffc114f5bf18 by task kworker/u17:14/705
CPU: 4 PID: 705 Comm: kworker/u17:14 Tainted: G S W 4.14.234-00003-g1fb6d0bd49a4-dirty #28
Hardware name: Qualcomm Technologies, Inc. SM8150 V2 PM8150
Google Inc. MSM sm8150 Flame DVT (DT)
Workqueue: hci0 hci_rx_work
Call trace:
dump_backtrace+0x0/0x378
show_stack+0x20/0x2c
dump_stack+0x124/0x148
print_address_description+0x80/0x2e8
__kasan_report+0x168/0x188
kasan_report+0x10/0x18
__asan_load4+0x84/0x8c
refcount_dec_and_test+0x20/0xd0
l2cap_chan_put+0x48/0x12c
l2cap_recv_frame+0x4770/0x6550
l2cap_recv_acldata+0x44c/0x7a4
hci_acldata_packet+0x100/0x188
hci_rx_work+0x178/0x23c
process_one_work+0x35c/0x95c
worker_thread+0x4cc/0x960
kthread+0x1a8/0x1c4
ret_from_fork+0x10/0x18
Cc: stable@kernel.org Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Tested-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Sedat Dilek noticed that I had an extraneous semicolon at the end of a
line in the previous patch.
It's harmless, but unintentional, and while compilers just treat it as
an extra empty statement, for all I know some other tooling might warn
about it. So clean it up before other people notice too ;)
Problems observed:
======================================================================
1) Using ssh/sshfs. The remote sshd daemon can abort with the message:
"message authentication code incorrect"
This happens because the tcp message sent is corrupted during the
USB "Bulk out". The device calculate the tcp checksum and send a
valid tcp message to the remote sshd. Then the encryption detects
the error and aborts.
2) NETDEV WATCHDOG: ... (ax88179_178a): transmit queue 0 timed out
3) Stop normal work without any log message.
The "Bulk in" continue receiving packets normally.
The host sends "Bulk out" and the device responds with -ECONNRESET.
(The netusb.c code tx_complete ignore -ECONNRESET)
Under normal conditions these errors take days to happen and in
intense usage take hours.
A test with ping gives packet loss, showing that something is wrong:
ping -4 -s 462 {destination} # 462 = 512 - 42 - 8
Not all packets fail.
My guess is that the device tries to find another packet starting
at the extra byte and will fail or not depending on the next
bytes (old buffer content).
======================================================================
Signed-off-by: Jose Alonso <joalonsof@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a race in pty_write(). pty_write() can be called in parallel
with e.g. ioctl(TIOCSTI) or ioctl(TCXONC) which also inserts chars to
the buffer. Provided, tty_flip_buffer_push() in pty_write() is called
outside the lock, it can commit inconsistent tail. This can lead to out
of bounds writes and other issues. See the Link below.
To fix this, we have to introduce a new helper called
tty_insert_flip_string_and_push_buffer(). It does both
tty_insert_flip_string() and tty_flip_buffer_commit() under the port
lock. It also calls queue_work(), but outside the lock. See 71a174b39f10 (pty: do tty_flip_buffer_push without port->lock in
pty_write) for the reasons.
Keep the helper internal-only (in drivers' tty.h). It is not intended to
be used widely.
Link: https://seclists.org/oss-sec/2022/q2/155 Fixes: 71a174b39f10 (pty: do tty_flip_buffer_push without port->lock in pty_write) Cc: 一只狗 <chennbnbnb@gmail.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Suggested-by: Hillf Danton <hdanton@sina.com> Signed-off-by: Jiri Slaby <jslaby@suse.cz> Link: https://lore.kernel.org/r/20220707082558.9250-2-jslaby@suse.cz Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commit a9c3f68f3cd8d (tty: Fix low_latency BUG) in 2014,
tty_flip_buffer_push() is only a wrapper to tty_schedule_flip(). All
users were converted in the previous patches, so remove
tty_schedule_flip() completely while inlining its body into
tty_flip_buffer_push().
Since commit a9c3f68f3cd8d (tty: Fix low_latency BUG) in 2014,
tty_flip_buffer_push() is only a wrapper to tty_schedule_flip(). We are
going to remove the latter (as it is used less), so call the former in
the rest of the users.
Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Matt Turner <mattst88@gmail.com> Cc: William Hubbs <w.d.hubbs@gmail.com> Cc: Chris Brannon <chris@the-brannons.com> Cc: Kirk Reiser <kirk@reisers.ca> Cc: Samuel Thibault <samuel.thibault@ens-lyon.org> Cc: Heiko Carstens <hca@linux.ibm.com> Cc: Vasily Gorbik <gor@linux.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Alexander Gordeev <agordeev@linux.ibm.com> Reviewed-by: Johan Hovold <johan@kernel.org> Signed-off-by: Jiri Slaby <jslaby@suse.cz> Link: https://lore.kernel.org/r/20211122111648.30379-3-jslaby@suse.cz Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commit a9c3f68f3cd8d (tty: Fix low_latency BUG) in 2014,
tty_flip_buffer_push() is only a wrapper to tty_schedule_flip(). We are
going to remove the latter (as it is used less), so call the former in
drivers/tty/.
Cc: Vladimir Zapolskiy <vz@mleia.com> Reviewed-by: Johan Hovold <johan@kernel.org> Signed-off-by: Jiri Slaby <jslaby@suse.cz> Link: https://lore.kernel.org/r/20211122111648.30379-2-jslaby@suse.cz Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the pipe is closed, we mark the associated watchqueue defunct by
calling watch_queue_clear(). However, while that is protected by the
watchqueue lock, new watchqueue entries aren't actually added under that
lock at all: they use the pipe->rd_wait.lock instead, and looking up
that pipe happens without any locking.
The watchqueue code uses the RCU read-side section to make sure that the
wqueue entry itself hasn't disappeared, but that does not protect the
pipe_info in any way.
So make sure to actually hold the wqueue lock when posting watch events,
properly serializing against the pipe being torn down.
projects / linux-stable / log