Skip to content

R
riscv-linux

OpenXiangShan / riscv-linux
12 个月 前同步成功

通知 6
Star 2
Fork 2
R
riscv-linux

体验新版 GitCode，发现更多精彩内容 >>

切换分支/标签
  1. 21 8月, 2018 2 次提交
    • W
      base: fix order of OF initialization · 5ca42fb9
      由 Wesley W. Terpstra 提交于 
      This fixes: [    0.010000] cpu cpu0: Error -2 creating of_node link
... which you get for every CPU on all architectures that use
CONFIG_GENERIC_CPU_DEVICES.

In that case, driver_init() calls cpu_dev_init() before calling
of_core_init(). Then we get the callchain:

  cpu_dev_init()
    -> cpu_dev_register_generic()
    -> register_cpu(cpu, i)
    -> device_register(&cpu->dev)
    -> device_add(dev)
    -> device_add_class_symlinks(dev)

... in device_add_class_symlinks, we we dev->of_node, and call
sysfs_create_link(), which fails because we haven't called
of_core_init() to register the sysfs devicetree directory yet.
Signed-off-by: NWesley W. Terpstra <wesley@sifive.com>
[hch: updated the changelog based on review feedback]
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Acked-by: NFrank Rowand <frowand.list@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      5ca42fb9
    • P
      RISC-V Updates for the 4.19 Merge Window · 33e7d009
      由 Palmer Dabbelt 提交于 
      This tag contains some major improvements to the RISC-V port, including
the necessary interrupt controller and timer support to actually make it
to userspace.  Support for three devices has been added:

* Support for the ISA-mandated timers on RISC-V systems.
* Support for the ISA-mandated first-level interrupt controller on
  RISC-V systems, which is handled as part of our core arch code because
  it's very small and tightly tied to the ISA.
* Support for SiFive's platform-level interrupt controller, which talks
  to the actual devices.

In addition to these new devices, there are a handful of cleanups all
over the RISC-V tree:

* Build fixes for various configurations
    * A fix to the vDSO build's makefile so it respects CFLAGS.
    * The addition of __lshrti3, a libgcc derived function necessary for
      some 32-bit configurations.
    * !SMP && PERF_EVENTS
* Cleanups to the arch code to remove the remnants of old versions of
  the drivers that were just properly submitted.
    * Some dead code from the timer driver, most of which wasn't ever
      even compiled.
    * Cleanups of some interrupt #defines, which are now local to the
      interrupt handling code.
* Fixes to ptrace(), which while not being sufficient to fully make GDB
  work are at least sufficient to get simple GDB tasks to work.
* Early printk support via RISC-V's architecturally mandated SBI console
  device.
* A fix to our early debug trap handler to ensure it's always aligned.

These patches have all been through a fairly extensive review process,
but as this enables a whole pile of functionality (ie, userspace) I'm
confident we'll need to submit a few more patches.  The only concrete
issues I know about are the sys_riscv_flush_icache patches, but as I
managed to screw those up on Friday I figured it'd be best to let them
bake another week.

This tag boots a Fedora root filesystem on QEMU's master branch for me,
and before this morning's rebase (from 4.18-rc8 to 4.18) it booted on
the HiFive Unleashed.

Thanks to Christoph Hellwig and the other guys at WD for getting the new
drivers in shape!
      33e7d009
  3. 14 8月, 2018 2 次提交
  5. 13 8月, 2018 17 次提交
  7. 12 8月, 2018 5 次提交
  9. 11 8月, 2018 5 次提交
  11. 10 8月, 2018 9 次提交
    • D
      Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf · e91e2189
      由 David S. Miller 提交于 
      Daniel Borkmann says:

====================
pull-request: bpf 2018-08-10

The following pull-request contains BPF updates for your *net* tree.

The main changes are:

1) Fix cpumap and devmap on teardown as they're under RCU context
   and won't have same assumption as running under NAPI protection,
   from Jesper.

2) Fix various sockmap bugs in bpf_tcp_sendmsg() code, e.g. we had
   a bug where socket error was not propagated correctly, from Daniel.

3) Fix incompatible libbpf header license for BTF code and match it
   before it gets officially released with the rest of libbpf which
   is LGPL-2.1, from Martin.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      e91e2189
    • A
      make sure that __dentry_kill() always invalidates d_seq, unhashed or not · 4c0d7cd5
      由 Al Viro 提交于 
      RCU pathwalk relies upon the assumption that anything that changes
->d_inode of a dentry will invalidate its ->d_seq.  That's almost
true - the one exception is that the final dput() of already unhashed
dentry does *not* touch ->d_seq at all.  Unhashing does, though,
so for anything we'd found by RCU dcache lookup we are fine.
Unfortunately, we can *start* with an unhashed dentry or jump into
it.

We could try and be careful in the (few) places where that could
happen.  Or we could just make the final dput() invalidate the damn
thing, unhashed or not.  The latter is much simpler and easier to
backport, so let's do it that way.
Reported-by: N"Dae R. Jeong" <threeearcat@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
      4c0d7cd5
    • A
      fix __legitimize_mnt()/mntput() race · 119e1ef8
      由 Al Viro 提交于 
      __legitimize_mnt() has two problems - one is that in case of success
the check of mount_lock is not ordered wrt preceding increment of
refcount, making it possible to have successful __legitimize_mnt()
on one CPU just before the otherwise final mntpu() on another,
with __legitimize_mnt() not seeing mntput() taking the lock and
mntput() not seeing the increment done by __legitimize_mnt().
Solved by a pair of barriers.

Another is that failure of __legitimize_mnt() on the second
read_seqretry() leaves us with reference that'll need to be
dropped by caller; however, if that races with final mntput()
we can end up with caller dropping rcu_read_lock() and doing
mntput() to release that reference - with the first mntput()
having freed the damn thing just as rcu_read_lock() had been
dropped.  Solution: in "do mntput() yourself" failure case
grab mount_lock, check if MNT_DOOMED has been set by racing
final mntput() that has missed our increment and if it has -
undo the increment and treat that as "failure, caller doesn't
need to drop anything" case.

It's not easy to hit - the final mntput() has to come right
after the first read_seqretry() in __legitimize_mnt() *and*
manage to miss the increment done by __legitimize_mnt() before
the second read_seqretry() in there.  The things that are almost
impossible to hit on bare hardware are not impossible on SMP
KVM, though...
Reported-by: NOleg Nesterov <oleg@redhat.com>
Fixes: 48a066e7 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
      119e1ef8
    • A
      fix mntput/mntput race · 9ea0a46c
      由 Al Viro 提交于 
      mntput_no_expire() does the calculation of total refcount under mount_lock;
unfortunately, the decrement (as well as all increments) are done outside
of it, leading to false positives in the "are we dropping the last reference"
test.  Consider the following situation:
	* mnt is a lazy-umounted mount, kept alive by two opened files.  One
of those files gets closed.  Total refcount of mnt is 2.  On CPU 42
mntput(mnt) (called from __fput()) drops one reference, decrementing component
	* After it has looked at component #0, the process on CPU 0 does
mntget(), incrementing component #0, gets preempted and gets to run again -
on CPU 69.  There it does mntput(), which drops the reference (component #69)
and proceeds to spin on mount_lock.
	* On CPU 42 our first mntput() finishes counting.  It observes the
decrement of component #69, but not the increment of component #0.  As the
result, the total it gets is not 1 as it should've been - it's 0.  At which
point we decide that vfsmount needs to be killed and proceed to free it and
shut the filesystem down.  However, there's still another opened file
on that filesystem, with reference to (now freed) vfsmount, etc. and we are
screwed.

It's not a wide race, but it can be reproduced with artificial slowdown of
the mnt_get_count() loop, and it should be easier to hit on SMP KVM setups.

Fix consists of moving the refcount decrement under mount_lock; the tricky
part is that we want (and can) keep the fast case (i.e. mount that still
has non-NULL ->mnt_ns) entirely out of mount_lock.  All places that zero
mnt->mnt_ns are dropping some reference to mnt and they call synchronize_rcu()
before that mntput().  IOW, if mntput() observes (under rcu_read_lock())
a non-NULL ->mnt_ns, it is guaranteed that there is another reference yet to
be dropped.
Reported-by: NJann Horn <jannh@google.com>
Tested-by: NJann Horn <jannh@google.com>
Fixes: 48a066e7 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
      9ea0a46c
    • D
      Merge branch 'bpf-fix-cpu-and-devmap-teardown' · 9c954201
      由 Daniel Borkmann 提交于 
      Jesper Dangaard Brouer says:

====================
Removing entries from cpumap and devmap, goes through a number of
syncronization steps to make sure no new xdp_frames can be enqueued.
But there is a small chance, that xdp_frames remains which have not
been flushed/processed yet.  Flushing these during teardown, happens
from RCU context and not as usual under RX NAPI context.

The optimization introduced in commt 389ab7f0 ("xdp: introduce
xdp_return_frame_rx_napi"), missed that the flush operation can also
be called from RCU context.  Thus, we cannot always use the
xdp_return_frame_rx_napi call, which take advantage of the protection
provided by XDP RX running under NAPI protection.

The samples/bpf xdp_redirect_cpu have a --stress-mode, that is
adjusted to easier reproduce (verified by Red Hat QA).
====================
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
      9c954201
    • J
      xdp: fix bug in devmap teardown code path · 1bf9116d
      由 Jesper Dangaard Brouer 提交于 
      Like cpumap teardown, the devmap teardown code also flush remaining
xdp_frames, via bq_xmit_all() in case map entry is removed.  The code
can call xdp_return_frame_rx_napi, from the the wrong context, in-case
ndo_xdp_xmit() fails.

Fixes: 389ab7f0 ("xdp: introduce xdp_return_frame_rx_napi")
Fixes: 735fc405 ("xdp: change ndo_xdp_xmit API to support bulking")
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
      1bf9116d
    • J
      samples/bpf: xdp_redirect_cpu adjustment to reproduce teardown race easier · 37d7ff25
      由 Jesper Dangaard Brouer 提交于 
      The teardown race in cpumap is really hard to reproduce.  These changes
makes it easier to reproduce, for QA.

The --stress-mode now have a case of a very small queue size of 8, that helps
to trigger teardown flush to encounter a full queue, which results in calling
xdp_return_frame API, in a non-NAPI protect context.

Also increase MAX_CPUS, as my QA department have larger machines than me.
Tested-by: NJean-Tsung Hsiao <jhsiao@redhat.com>
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
      37d7ff25
    • J
      xdp: fix bug in cpumap teardown code path · ad0ab027
      由 Jesper Dangaard Brouer 提交于 
      When removing a cpumap entry, a number of syncronization steps happen.
Eventually the teardown code __cpu_map_entry_free is invoked from/via
call_rcu.

The teardown code __cpu_map_entry_free() flushes remaining xdp_frames,
by invoking bq_flush_to_queue, which calls xdp_return_frame_rx_napi().
The issues is that the teardown code is not running in the RX NAPI
code path.  Thus, it is not allowed to invoke the NAPI variant of
xdp_return_frame.

This bug was found and triggered by using the --stress-mode option to
the samples/bpf program xdp_redirect_cpu.  It is hard to trigger,
because the ptr_ring have to be full and cpumap bulk queue max
contains 8 packets, and a remote CPU is racing to empty the ptr_ring
queue.

Fixes: 389ab7f0 ("xdp: introduce xdp_return_frame_rx_napi")
Tested-by: NJean-Tsung Hsiao <jhsiao@redhat.com>
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
      ad0ab027
    • L
      Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6 · 112cbae2
      由 Linus Torvalds 提交于 
      Pull crypto fix from Herbert Xu:
 "This fixes a performance regression in arm64 NEON crypto as well as a
  crash in x86 aegis/morus on unsupported CPUs"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: x86/aegis,morus - Fix and simplify CPUID checks
  crypto: arm64 - revert NEON yield for fast AEAD implementations
      112cbae2