Skip to content

K
Kernel

openeuler / Kernel
1 年多 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 03 8月, 2020 1 次提交
  3. 02 8月, 2020 1 次提交
  5. 31 7月, 2020 2 次提交
  7. 30 7月, 2020 2 次提交
    • L
      random32: remove net_rand_state from the latent entropy gcc plugin · 83bdc727
      由 Linus Torvalds 提交于 
      It turns out that the plugin right now ends up being really unhappy
about the change from 'static' to 'extern' storage that happened in
commit f227e3ec ("random32: update the net random state on interrupt
and activity").

This is probably a trivial fix for the latent_entropy plugin, but for
now, just remove net_rand_state from the list of things the plugin
worries about.
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Cc: Emese Revfy <re.emese@gmail.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      83bdc727
    • W
      random32: update the net random state on interrupt and activity · f227e3ec
      由 Willy Tarreau 提交于 
      This modifies the first 32 bits out of the 128 bits of a random CPU's
net_rand_state on interrupt or CPU activity to complicate remote
observations that could lead to guessing the network RNG's internal
state.

Note that depending on some network devices' interrupt rate moderation
or binding, this re-seeding might happen on every packet or even almost
never.

In addition, with NOHZ some CPUs might not even get timer interrupts,
leaving their local state rarely updated, while they are running
networked processes making use of the random state.  For this reason, we
also perform this update in update_process_times() in order to at least
update the state when there is user or system activity, since it's the
only case we care about.
Reported-by: NAmit Klein <aksecurity@gmail.com>
Suggested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Dumazet <edumazet@google.com>
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      f227e3ec
  9. 29 7月, 2020 4 次提交
  11. 28 7月, 2020 8 次提交
  13. 27 7月, 2020 1 次提交
    • T
      genirq/affinity: Make affinity setting if activated opt-in · f0c7baca
      由 Thomas Gleixner 提交于 
      John reported that on a RK3288 system the perf per CPU interrupts are all
affine to CPU0 and provided the analysis:

 "It looks like what happens is that because the interrupts are not per-CPU
  in the hardware, armpmu_request_irq() calls irq_force_affinity() while
  the interrupt is deactivated and then request_irq() with IRQF_PERCPU |
  IRQF_NOBALANCING.  

  Now when irq_startup() runs with IRQ_STARTUP_NORMAL, it calls
  irq_setup_affinity() which returns early because IRQF_PERCPU and
  IRQF_NOBALANCING are set, leaving the interrupt on its original CPU."

This was broken by the recent commit which blocked interrupt affinity
setting in hardware before activation of the interrupt. While this works in
general, it does not work for this particular case. As contrary to the
initial analysis not all interrupt chip drivers implement an activate
callback, the safe cure is to make the deferred interrupt affinity setting
at activation time opt-in.

Implement the necessary core logic and make the two irqchip implementations
for which this is required opt-in. In hindsight this would have been the
right thing to do, but ...

Fixes: baedb87d ("genirq/affinity: Handle affinity setting on inactive interrupts correctly")
Reported-by: NJohn Keeping <john@metanate.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Tested-by: NMarc Zyngier <maz@kernel.org>
Acked-by: NMarc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/87blk4tzgm.fsf@nanos.tec.linutronix.de
      f0c7baca
  15. 25 7月, 2020 3 次提交
  17. 24 7月, 2020 4 次提交
    • J
      tpm: Unify the mismatching TPM space buffer sizes · 6c4e79d9
      由 Jarkko Sakkinen 提交于 
      The size of the buffers for storing context's and sessions can vary from
arch to arch as PAGE_SIZE can be anything between 4 kB and 256 kB (the
maximum for PPC64). Define a fixed buffer size set to 16 kB. This should be
enough for most use with three handles (that is how many we allow at the
moment). Parametrize the buffer size while doing this, so that it is easier
to revisit this later on if required.

Cc: stable@vger.kernel.org
Reported-by: NStefan Berger <stefanb@linux.ibm.com>
Fixes: 745b361e ("tpm: infrastructure for TPM spaces")
Reviewed-by: NJerry Snitselaar <jsnitsel@redhat.com>
Tested-by: NStefan Berger <stefanb@linux.ibm.com>
Signed-off-by: NJarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
      6c4e79d9
    • T
      tpm: Require that all digests are present in TCG_PCR_EVENT2 structures · 7f3d176f
      由 Tyler Hicks 提交于 
      Require that the TCG_PCR_EVENT2.digests.count value strictly matches the
value of TCG_EfiSpecIdEvent.numberOfAlgorithms in the event field of the
TCG_PCClientPCREvent event log header. Also require that
TCG_EfiSpecIdEvent.numberOfAlgorithms is non-zero.

The TCG PC Client Platform Firmware Profile Specification section 9.1
(Family "2.0", Level 00 Revision 1.04) states:

 For each Hash algorithm enumerated in the TCG_PCClientPCREvent entry,
 there SHALL be a corresponding digest in all TCG_PCR_EVENT2 structures.
 Note: This includes EV_NO_ACTION events which do not extend the PCR.

Section 9.4.5.1 provides this description of
TCG_EfiSpecIdEvent.numberOfAlgorithms:

 The number of Hash algorithms in the digestSizes field. This field MUST
 be set to a value of 0x01 or greater.

Enforce these restrictions, as required by the above specification, in
order to better identify and ignore invalid sequences of bytes at the
end of an otherwise valid TPM2 event log. Firmware doesn't always have
the means necessary to inform the kernel of the actual event log size so
the kernel's event log parsing code should be stringent when parsing the
event log for resiliency against firmware bugs. This is true, for
example, when firmware passes the event log to the kernel via a reserved
memory region described in device tree.

POWER and some ARM systems use the "linux,sml-base" and "linux,sml-size"
device tree properties to describe the memory region used to pass the
event log from firmware to the kernel. Unfortunately, the
"linux,sml-size" property describes the size of the entire reserved
memory region rather than the size of the event long within the memory
region and the event log format does not include information describing
the size of the event log.

tpm_read_log_of(), in drivers/char/tpm/eventlog/of.c, is where the
"linux,sml-size" property is used. At the end of that function,
log->bios_event_log_end is pointing at the end of the reserved memory
region. That's typically 0x10000 bytes offset from "linux,sml-base",
depending on what's defined in the device tree source.

The firmware event log only fills a portion of those 0x10000 bytes and
the rest of the memory region should be zeroed out by firmware. Even in
the case of a properly zeroed bytes in the remainder of the memory
region, the only thing allowing the kernel's event log parser to detect
the end of the event log is the following conditional in
__calc_tpm2_event_size():

        if (event_type == 0 && event_field->event_size == 0)
                size = 0;

If that wasn't there, __calc_tpm2_event_size() would think that a 16
byte sequence of zeroes, following an otherwise valid event log, was
a valid event.

However, problems can occur if a single bit is set in the offset
corresponding to either the TCG_PCR_EVENT2.eventType or
TCG_PCR_EVENT2.eventSize fields, after the last valid event log entry.
This could confuse the parser into thinking that an additional entry is
present in the event log and exposing this invalid entry to userspace in
the /sys/kernel/security/tpm0/binary_bios_measurements file. Such
problems have been seen if firmware does not fully zero the memory
region upon a warm reboot.

This patch significantly raises the bar on how difficult it is for
stale/invalid memory to confuse the kernel's event log parser but
there's still, ultimately, a reliance on firmware to properly initialize
the remainder of the memory region reserved for the event log as the
parser cannot be expected to detect a stale but otherwise properly
formatted firmware event log entry.

Fixes: fd5c7869 ("tpm: fix handling of the TPM 2.0 event logs")
Signed-off-by: NTyler Hicks <tyhicks@linux.microsoft.com>
Reviewed-by: NJarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Signed-off-by: NJarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
      7f3d176f
    • Y
      tcp: allow at most one TLP probe per flight · 76be93fc
      由 Yuchung Cheng 提交于 
      Previously TLP may send multiple probes of new data in one
flight. This happens when the sender is cwnd limited. After the
initial TLP containing new data is sent, the sender receives another
ACK that acks partial inflight.  It may re-arm another TLP timer
to send more, if no further ACK returns before the next TLP timeout
(PTO) expires. The sender may send in theory a large amount of TLP
until send queue is depleted. This only happens if the sender sees
such irregular uncommon ACK pattern. But it is generally undesirable
behavior during congestion especially.

The original TLP design restrict only one TLP probe per inflight as
published in "Reducing Web Latency: the Virtue of Gentle Aggression",
SIGCOMM 2013. This patch changes TLP to send at most one probe
per inflight.

Note that if the sender is app-limited, TLP retransmits old data
and did not have this issue.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      76be93fc
    • M
      dm integrity: fix integrity recalculation that is improperly skipped · 5df96f2b
      由 Mikulas Patocka 提交于 
      Commit adc0daad ("dm: report suspended
device during destroy") broke integrity recalculation.

The problem is dm_suspended() returns true not only during suspend,
but also during resume. So this race condition could occur:
1. dm_integrity_resume calls queue_work(ic->recalc_wq, &ic->recalc_work)
2. integrity_recalc (&ic->recalc_work) preempts the current thread
3. integrity_recalc calls if (unlikely(dm_suspended(ic->ti))) goto unlock_ret;
4. integrity_recalc exits and no recalculating is done.

To fix this race condition, add a function dm_post_suspending that is
only true during the postsuspend phase and use it instead of
dm_suspended().

Signed-off-by: Mikulas Patocka <mpatocka redhat com>
Fixes: adc0daad ("dm: report suspended device during destroy")
Cc: stable vger kernel org # v4.18+
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
      5df96f2b
  19. 23 7月, 2020 5 次提交
  21. 22 7月, 2020 3 次提交
    • R
      i2c: drop duplicated word in the header file · aca7ed09
      由 Randy Dunlap 提交于 
      Drop the doubled word "be" in a comment.
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NWolfram Sang <wsa@kernel.org>
      aca7ed09
    • E
      fs-verity: use smp_load_acquire() for ->i_verity_info · f3db0bed
      由 Eric Biggers 提交于 
      Normally smp_store_release() or cmpxchg_release() is paired with
smp_load_acquire().  Sometimes smp_load_acquire() can be replaced with
the more lightweight READ_ONCE().  However, for this to be safe, all the
published memory must only be accessed in a way that involves the
pointer itself.  This may not be the case if allocating the object also
involves initializing a static or global variable, for example.

fsverity_info::tree_params.hash_alg->tfm is a crypto_ahash object that's
internal to and is allocated by the crypto subsystem.  So by using
READ_ONCE() for ->i_verity_info, we're relying on internal
implementation details of the crypto subsystem.

Remove this fragile assumption by using smp_load_acquire() instead.

Also fix the cmpxchg logic to correctly execute an ACQUIRE barrier when
losing the cmpxchg race, since cmpxchg doesn't guarantee a memory
barrier on failure.

(Note: I haven't seen any real-world problems here.  This change is just
fixing the code to be guaranteed correct and less fragile.)

Fixes: fd2d1acf ("fs-verity: add the hook for file ->open()")
Link: https://lore.kernel.org/r/20200721225920.114347-6-ebiggers@kernel.orgSigned-off-by: NEric Biggers <ebiggers@google.com>
      f3db0bed
    • E
      fscrypt: use smp_load_acquire() for ->i_crypt_info · ab673b98
      由 Eric Biggers 提交于 
      Normally smp_store_release() or cmpxchg_release() is paired with
smp_load_acquire().  Sometimes smp_load_acquire() can be replaced with
the more lightweight READ_ONCE().  However, for this to be safe, all the
published memory must only be accessed in a way that involves the
pointer itself.  This may not be the case if allocating the object also
involves initializing a static or global variable, for example.

fscrypt_info includes various sub-objects which are internal to and are
allocated by other kernel subsystems such as keyrings and crypto.  So by
using READ_ONCE() for ->i_crypt_info, we're relying on internal
implementation details of these other kernel subsystems.

Remove this fragile assumption by using smp_load_acquire() instead.

(Note: I haven't seen any real-world problems here.  This change is just
fixing the code to be guaranteed correct and less fragile.)

Fixes: e37a784d ("fscrypt: use READ_ONCE() to access ->i_crypt_info")
Link: https://lore.kernel.org/r/20200721225920.114347-5-ebiggers@kernel.orgSigned-off-by: NEric Biggers <ebiggers@google.com>
      ab673b98
  23. 21 7月, 2020 4 次提交
  25. 20 7月, 2020 1 次提交
  27. 18 7月, 2020 1 次提交
    • B
      blk-cgroup: show global disk stats in root cgroup io.stat · ef45fe47
      由 Boris Burkov 提交于 
      In order to improve consistency and usability in cgroup stat accounting,
we would like to support the root cgroup's io.stat.

Since the root cgroup has processes doing io even if the system has no
explicitly created cgroups, we need to be careful to avoid overhead in
that case.  For that reason, the rstat algorithms don't handle the root
cgroup, so just turning the file on wouldn't give correct statistics.

To get around this, we simulate flushing the iostat struct by filling it
out directly from global disk stats. The result is a root cgroup io.stat
file consistent with both /proc/diskstats and io.stat.

Note that in order to collect the disk stats, we needed to iterate over
devices. To facilitate that, we had to change the linkage of a disk_type
to external so that it can be used from blk-cgroup.c to iterate over
disks.
Suggested-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NBoris Burkov <boris@bur.io>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>
      ef45fe47