openwrt/staging/blogic.git
6 years agoxprtrdma: Fix calculation of ri_max_send_sges
Chuck Lever [Wed, 31 Jan 2018 17:34:05 +0000 (12:34 -0500)]
xprtrdma: Fix calculation of ri_max_send_sges

Commit 16f906d66cd7 ("xprtrdma: Reduce required number of send
SGEs") introduced the rpcrdma_ia::ri_max_send_sges field. This fixes
a problem where xprtrdma would not work if the device's max_sge
capability was small (low single digits).

At least RPCRDMA_MIN_SEND_SGES are needed for the inline parts of
each RPC. ri_max_send_sges is set to this value:

  ia->ri_max_send_sges = max_sge - RPCRDMA_MIN_SEND_SGES;

Then when marshaling each RPC, rpcrdma_args_inline uses that value
to determine whether the device has enough Send SGEs to convey an
NFS WRITE payload inline, or whether instead a Read chunk is
required.

More recently, commit ae72950abf99 ("xprtrdma: Add data structure to
manage RDMA Send arguments") used the ri_max_send_sges value to
calculate the size of an array, but that commit erroneously assumed
ri_max_send_sges contains a value similar to the device's max_sge,
and not one that was reduced by the minimum SGE count.

This assumption results in the calculated size of the sendctx's
Send SGE array to be too small. When the array is used to marshal
an RPC, the code can write Send SGEs into the following sendctx
element in that array, corrupting it. When the device's max_sge is
large, this issue is entirely harmless; but it results in an oops
in the provider's post_send method, if dev.attrs.max_sge is small.

So let's straighten this out: ri_max_send_sges will now contain a
value with the same meaning as dev.attrs.max_sge, which makes
the code easier to understand, and enables rpcrdma_sendctx_create
to calculate the size of the SGE array correctly.

Reported-by: Michal Kalderon <Michal.Kalderon@cavium.com>
Fixes: 16f906d66cd7 ("xprtrdma: Reduce required number of send SGEs")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Tested-by: Michal Kalderon <Michal.Kalderon@cavium.com>
Cc: stable@vger.kernel.org # v4.10+
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoSUNRPC: Micro-optimize __rpc_execute
Chuck Lever [Wed, 3 Jan 2018 20:38:49 +0000 (15:38 -0500)]
SUNRPC: Micro-optimize __rpc_execute

The common case: There are 13 to 14 actions per RPC, and tk_callback
is non-NULL in only one of them. There's no need to store a NULL in
the tk_callback field during each FSM step.

This slightly improves throughput results in dbench and other multi-
threaded benchmarks on my two-socket client on 56Gb InfiniBand, but
will probably be inconsequential on slower systems.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoSUNRPC: task_run_action should display tk_callback
Chuck Lever [Wed, 3 Jan 2018 20:38:41 +0000 (15:38 -0500)]
SUNRPC: task_run_action should display tk_callback

This shows up in every RPC:

     kworker/4:1-19772 [004]  3467.373443: rpc_task_run_action:  task:4711@2 flags=0e81 state=0005 status=0 action=call_status
     kworker/4:1-19772 [004]  3467.373444: rpc_task_run_action:  task:4711@2 flags=0e81 state=0005 status=0 action=call_status

What's actually going on is that the first iteration of the RPC
scheduler is invoking the function in tk_callback (in this case,
xprt_timer), then invoking call_status on the next iteration.

Feeding do_action, rather than tk_action, to the "task_run_action"
trace point will now always display the correct FSM step.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agosunrpc: Format RPC events consistently for display
Chuck Lever [Wed, 3 Jan 2018 20:38:33 +0000 (15:38 -0500)]
sunrpc: Format RPC events consistently for display

Clean up: Make it easier to use text search when browsing a trace
report. Other events use "status=%d".

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoSUNRPC: Trace xprt_timer events
Chuck Lever [Wed, 3 Jan 2018 20:38:25 +0000 (15:38 -0500)]
SUNRPC: Trace xprt_timer events

Track RPC timeouts: report the XID and the server address to match
the content of network capture.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Correct some documenting comments
Chuck Lever [Wed, 3 Jan 2018 20:38:17 +0000 (15:38 -0500)]
xprtrdma: Correct some documenting comments

Fix kernel-doc warnings in net/sunrpc/xprtrdma/ .

net/sunrpc/xprtrdma/verbs.c:1575: warning: No description found for parameter 'count'
net/sunrpc/xprtrdma/verbs.c:1575: warning: Excess function parameter 'min_reqs' description in 'rpcrdma_ep_post_extra_recv'

net/sunrpc/xprtrdma/backchannel.c:288: warning: No description found for parameter 'r_xprt'
net/sunrpc/xprtrdma/backchannel.c:288: warning: Excess function parameter 'xprt' description in 'rpcrdma_bc_receive_call'

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Fix "bytes registered" accounting
Chuck Lever [Wed, 3 Jan 2018 20:38:09 +0000 (15:38 -0500)]
xprtrdma: Fix "bytes registered" accounting

The contents of seg->mr_len changed when ->ro_map stopped returning
the full chunk length in the first segment. Count the full length of
each Write chunk, not the length of the first segment (which now can
only be as large as a page).

Fixes: 9d6b04097882 ("xprtrdma: Place registered MWs on a ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Instrument allocation/release of rpcrdma_req/rep objects
Chuck Lever [Wed, 20 Dec 2017 21:31:53 +0000 (16:31 -0500)]
xprtrdma: Instrument allocation/release of rpcrdma_req/rep objects

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points to instrument QP and CQ access upcalls
Chuck Lever [Wed, 20 Dec 2017 21:31:45 +0000 (16:31 -0500)]
xprtrdma: Add trace points to instrument QP and CQ access upcalls

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points in the client-side backchannel code paths
Chuck Lever [Wed, 20 Dec 2017 21:31:37 +0000 (16:31 -0500)]
xprtrdma: Add trace points in the client-side backchannel code paths

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points for connect events
Chuck Lever [Wed, 20 Dec 2017 21:31:29 +0000 (16:31 -0500)]
xprtrdma: Add trace points for connect events

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points to instrument MR allocation and recovery
Chuck Lever [Wed, 20 Dec 2017 21:31:21 +0000 (16:31 -0500)]
xprtrdma: Add trace points to instrument MR allocation and recovery

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points to instrument memory invalidation
Chuck Lever [Wed, 20 Dec 2017 21:31:12 +0000 (16:31 -0500)]
xprtrdma: Add trace points to instrument memory invalidation

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points in reply decoder path
Chuck Lever [Wed, 20 Dec 2017 21:31:04 +0000 (16:31 -0500)]
xprtrdma: Add trace points in reply decoder path

This includes decoding Write and Reply chunks, and fixing up inline
payloads.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points to instrument memory registration
Chuck Lever [Wed, 20 Dec 2017 21:30:56 +0000 (16:30 -0500)]
xprtrdma: Add trace points to instrument memory registration

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points in the RPC Reply handler paths
Chuck Lever [Wed, 20 Dec 2017 21:30:48 +0000 (16:30 -0500)]
xprtrdma: Add trace points in the RPC Reply handler paths

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Add trace points in RPC Call transmit paths
Chuck Lever [Wed, 20 Dec 2017 21:30:40 +0000 (16:30 -0500)]
xprtrdma: Add trace points in RPC Call transmit paths

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agorpcrdma: infrastructure for static trace points in rpcrdma.ko
Chuck Lever [Sat, 20 Jan 2018 16:16:34 +0000 (11:16 -0500)]
rpcrdma: infrastructure for static trace points in rpcrdma.ko

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agordma/ib: Add trace point macros to display human-readable values
Chuck Lever [Sat, 20 Jan 2018 16:16:25 +0000 (11:16 -0500)]
rdma/ib: Add trace point macros to display human-readable values

These can be shared with all kernel ULPs, and more can easily be
added as needed.

Note: checkpatch.pl has some heartburn with the TRACE_DEFINE_ENUM
macros and the LIST macros. These follow the same style as other
header files under include/tracing/events , thus should be
considered acceptable exceptions.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Acked-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Introduce rpcrdma_mw_unmap_and_put
Chuck Lever [Fri, 15 Dec 2017 01:58:04 +0000 (20:58 -0500)]
xprtrdma: Introduce rpcrdma_mw_unmap_and_put

Clean up: Code review suggested that a common bit of code can be
placed into a helper function, and this gives us fewer places to
stick an "I DMA unmapped something" trace point.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Remove usage of "mw"
Chuck Lever [Fri, 15 Dec 2017 01:57:55 +0000 (20:57 -0500)]
xprtrdma: Remove usage of "mw"

Clean up: struct rpcrdma_mw was named after Memory Windows, but
xprtrdma no longer supports a Memory Window registration mode.
Rename rpcrdma_mw and its fields to reduce confusion and make
the code more sensible to read.

Renaming "mw" was suggested by Tom Talpey, the author of the
original xprtrdma implementation. It's a good idea, but I haven't
done this until now because it's a huge diffstat for no benefit
other than code readability.

However, I'm about to introduce static trace points that expose
a few of xprtrdma's internal data structures. They should make sense
in the trace report, and it's reasonable to treat trace points as a
kernel API contract which might be difficult to change later.

While I'm churning things up, two additional changes:
- rename variables unhelpfully called "r" to "mr", to improve code
  clarity, and
- rename the MR-related helper functions using the form
  "rpcrdma_mr_<verb>", to be consistent with other areas of the
  code.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Replace all usage of "frmr" with "frwr"
Chuck Lever [Fri, 15 Dec 2017 01:57:47 +0000 (20:57 -0500)]
xprtrdma: Replace all usage of "frmr" with "frwr"

Clean up: Over time, the industry has adopted the term "frwr"
instead of "frmr". The term "frwr" is now more widely recognized.

For the past couple of years I've attempted to add new code using
"frwr" , but there still remains plenty of older code that still
uses "frmr". Replace all usage of "frmr" to avoid confusion.

While we're churning code, rename variables unhelpfully called "f"
to "frwr", to improve code clarity.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Don't clear RPC_BC_PA_IN_USE on pre-allocated rpc_rqst's
Chuck Lever [Fri, 15 Dec 2017 01:57:39 +0000 (20:57 -0500)]
xprtrdma: Don't clear RPC_BC_PA_IN_USE on pre-allocated rpc_rqst's

No need for the overhead of atomically setting and clearing this bit
flag for every use of a pre-allocated backchannel rpc_rqst. These
are a distinct pool of rpc_rqsts that are used only for callback
operations, so it is safe to simply leave the bit set.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Split xprt_rdma_send_request
Chuck Lever [Fri, 15 Dec 2017 01:57:31 +0000 (20:57 -0500)]
xprtrdma: Split xprt_rdma_send_request

Clean up. @rqst is set up differently for backchannel Replies. For
example, rqst->rq_task and task->tk_client are both NULL. So it is
easier to understand and maintain this code path if it is separated.

Also, we can get rid of the confusing rl_connect_cookie hack in
rpcrdma_bc_receive_call.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: buf_free not called for CB replies
Chuck Lever [Fri, 15 Dec 2017 01:57:23 +0000 (20:57 -0500)]
xprtrdma: buf_free not called for CB replies

Since commit 5a6d1db45569 ("SUNRPC: Add a transport-specific private
field in rpc_rqst"), the rpc_rqst's for RPC-over-RDMA backchannel
operations leave rq_buffer set to NULL.

xprt_release does not invoke ->op->buf_free when rq_buffer is NULL.
The RPCRDMA_REQ_F_BACKCHANNEL check in xprt_rdma_free is therefore
redundant because xprt_rdma_free is not invoked for backchannel
requests.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Move unmap-safe logic to rpcrdma_marshal_req
Chuck Lever [Fri, 15 Dec 2017 01:57:14 +0000 (20:57 -0500)]
xprtrdma: Move unmap-safe logic to rpcrdma_marshal_req

Clean up. This logic is related to marshaling the request, and I'd
like to keep everything that touches req->rl_registered close
together, for CPU cache efficiency.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Support IPv6 in xprt_rdma_set_port
Chuck Lever [Fri, 15 Dec 2017 01:57:06 +0000 (20:57 -0500)]
xprtrdma: Support IPv6 in xprt_rdma_set_port

Clean up a harmless oversight. xprtrdma's ->set_port method has
never properly supported IPv6.

This issue has never been a problem because NFS/RDMA mounts have
always required "port=20049", thus so far, rpcbind is not invoked
for these mounts.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Remove another sockaddr_storage field (cdata::addr)
Chuck Lever [Fri, 15 Dec 2017 01:56:58 +0000 (20:56 -0500)]
xprtrdma: Remove another sockaddr_storage field (cdata::addr)

Save more space in struct rpcrdma_xprt by removing the redundant
"addr" field from struct rpcrdma_create_data_internal. Wherever
we have rpcrdma_xprt, we also have the rpc_xprt, which has a
sockaddr_storage field with the same content.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Initialize the xprt address string array earlier
Chuck Lever [Fri, 15 Dec 2017 01:56:50 +0000 (20:56 -0500)]
xprtrdma: Initialize the xprt address string array earlier

This makes the address strings available for debugging messages in
earlier stages of transport set up.

The first benefit is to get rid of the single-use rep_remote_addr
field, saving 128+ bytes in struct rpcrdma_ep.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Remove unused padding variables
Chuck Lever [Fri, 15 Dec 2017 01:56:42 +0000 (20:56 -0500)]
xprtrdma: Remove unused padding variables

Clean up. Remove fields that should have been removed by
commit b3221d6a53c4 ("xprtrdma: Remove logic that constructs
RDMA_MSGP type calls").

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Remove ri_reminv_expected
Chuck Lever [Fri, 15 Dec 2017 01:56:34 +0000 (20:56 -0500)]
xprtrdma: Remove ri_reminv_expected

Clean up.

Commit b5f0afbea4f2 ("xprtrdma: Per-connection pad optimization")
should have removed this.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Per-mode handling for Remote Invalidation
Chuck Lever [Fri, 15 Dec 2017 01:56:26 +0000 (20:56 -0500)]
xprtrdma: Per-mode handling for Remote Invalidation

Refactoring change: Remote Invalidation is particular to the memory
registration mode that is use. Use a callout instead of a generic
function to handle Remote Invalidation.

This gets rid of the 8-byte flags field in struct rpcrdma_mw, of
which only a single bit flag has been allocated.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Eliminate unnecessary lock cycle in xprt_rdma_send_request
Chuck Lever [Fri, 15 Dec 2017 01:56:18 +0000 (20:56 -0500)]
xprtrdma: Eliminate unnecessary lock cycle in xprt_rdma_send_request

The rpcrdma_req is not shared yet, and its associated Send hasn't
been posted, thus RMW should be safe. There's no need for the
expense of a lock cycle here.

Fixes: 0ba6f37012db ("xprtrdma: Refactor rpcrdma_deferred_completion")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Fix backchannel allocation of extra rpcrdma_reps
Chuck Lever [Fri, 15 Dec 2017 01:56:09 +0000 (20:56 -0500)]
xprtrdma: Fix backchannel allocation of extra rpcrdma_reps

The backchannel code uses rpcrdma_recv_buffer_put to add new reps
to the free rep list. This also decrements rb_recv_count, which
spoofs the receive overrun logic in rpcrdma_buffer_get_rep.

Commit 9b06688bc3b9 ("xprtrdma: Fix additional uses of
spin_lock_irqsave(rb_lock)") replaced the original open-coded
list_add with a call to rpcrdma_recv_buffer_put(), but then a year
later, commit 05c974669ece ("xprtrdma: Fix receive buffer
accounting") added rep accounting to rpcrdma_recv_buffer_put.
It was an oversight to let the backchannel continue to use this
function.

The fix this, let's combine the "add to free list" logic with
rpcrdma_create_rep.

Also, do not allocate RPCRDMA_MAX_BC_REQUESTS rpcrdma_reps in
rpcrdma_buffer_create and then allocate additional rpcrdma_reps in
rpcrdma_bc_setup_reps. Allocating the extra reps during backchannel
set-up is sufficient.

Fixes: 05c974669ece ("xprtrdma: Fix receive buffer accounting")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
6 years agoxprtrdma: Fix buffer leak after transport set up failure
Chuck Lever [Fri, 15 Dec 2017 01:56:01 +0000 (20:56 -0500)]
xprtrdma: Fix buffer leak after transport set up failure

This leak has been around forever, and is exceptionally rare.

EINVAL causes mount to fail with "an incorrect mount option was
specified" although it's not likely that one of the mount
options is incorrect. Instead, return ENODEV in this case, as this
appears to be an issue with system or device configuration rather
than a specific mount option.

Some obsolete comments are also removed.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
7 years agoLinux 4.15-rc4
Linus Torvalds [Mon, 18 Dec 2017 02:59:59 +0000 (18:59 -0800)]
Linux 4.15-rc4

7 years agoRevert "exec: avoid RLIMIT_STACK races with prlimit()"
Kees Cook [Tue, 12 Dec 2017 19:28:38 +0000 (11:28 -0800)]
Revert "exec: avoid RLIMIT_STACK races with prlimit()"

This reverts commit 04e35f4495dd560db30c25efca4eecae8ec8c375.

SELinux runs with secureexec for all non-"noatsecure" domain transitions,
which means lots of processes end up hitting the stack hard-limit change
that was introduced in order to fix a race with prlimit(). That race fix
will need to be redesigned.

Reported-by: Laura Abbott <labbott@redhat.com>
Reported-by: Tomáš Trnka <trnka@scm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge branch 'WIP.x86-pti.base-for-linus' of git://git.kernel.org/pub/scm/linux/kerne...
Linus Torvalds [Sun, 17 Dec 2017 21:57:08 +0000 (13:57 -0800)]
Merge branch 'WIP.x86-pti.base-for-linus' of git://git./linux/kernel/git/tip/tip

Pull Page Table Isolation (PTI) v4.14 backporting base tree from Ingo Molnar:
 "This tree contains the v4.14 PTI backport preparatory tree, which
  consists of four merges of upstream trees and 7 cherry-picked commits,
  which the upcoming PTI work depends on"

NOTE! The resulting tree is exactly the same as the original base tree
(ie the diff between this commit and its immediate first parent is
empty).

The only reason for this merge is literally to have a common point for
the actual PTI changes so that the commits can be shared in both the
4.15 and 4.14 trees.

* 'WIP.x86-pti.base-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mm/kasan: Don't use vmemmap_populate() to initialize shadow
  locking/barriers: Convert users of lockless_dereference() to READ_ONCE()
  locking/barriers: Add implicit smp_read_barrier_depends() to READ_ONCE()
  bpf: fix build issues on um due to mising bpf_perf_event.h
  perf/x86: Enable free running PEBS for REGS_USER/INTR
  x86: Make X86_BUG_FXSAVE_LEAK detectable in CPUID on AMD
  x86/cpufeature: Add User-Mode Instruction Prevention definitions

7 years agoMerge branch 'WIP.x86-pti.base.prep-for-linus' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sun, 17 Dec 2017 21:54:31 +0000 (13:54 -0800)]
Merge branch 'WIP.x86-pti.base.prep-for-linus' of git://git./linux/kernel/git/tip/tip

Pull Page Table Isolation (PTI) preparatory tree from Ingo Molnar:
 "This does a rename to free up linux/pti.h to be used by the upcoming
  page table isolation feature"

* 'WIP.x86-pti.base.prep-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  drivers/misc/intel/pti: Rename the header file to free up the namespace

7 years agoMerge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Dec 2017 21:48:50 +0000 (13:48 -0800)]
Merge branch 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull timer fix from Thomas Gleixner:
 "A single bugfix which prevents arbitrary sigev_notify values in
  posix-timers"

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  posix-timer: Properly check sigevent->sigev_notify

7 years agoMerge tag 'dmaengine-fix-4.15-rc4' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Sun, 17 Dec 2017 21:28:49 +0000 (13:28 -0800)]
Merge tag 'dmaengine-fix-4.15-rc4' of git://git.infradead.org/users/vkoul/slave-dma

Pull dmaengine fixes from Vinod Koul:
 "This time consisting of fixes in a bunch of drivers and the dmatest
  module:

   - Fix for disable clk on error path in fsl-edma driver
   - Disable clk fail fix in jz4740 driver
   - Fix long pending bug in dmatest driver for dangling pointer
   - Fix potential NULL pointer dereference in at_hdmac driver
   - Error handling path in ioat driver"

* tag 'dmaengine-fix-4.15-rc4' of git://git.infradead.org/users/vkoul/slave-dma:
  dmaengine: fsl-edma: disable clks on all error paths
  dmaengine: jz4740: disable/unprepare clk if probe fails
  dmaengine: dmatest: move callback wait queue to thread context
  dmaengine: at_hdmac: fix potential NULL pointer dereference in atc_prep_dma_interleaved
  dmaengine: ioat: Fix error handling path

7 years agocramfs: fix MTD dependency
Arnd Bergmann [Fri, 10 Nov 2017 14:57:21 +0000 (15:57 +0100)]
cramfs: fix MTD dependency

With CONFIG_MTD=m and CONFIG_CRAMFS=y, we now get a link failure:

  fs/cramfs/inode.o: In function `cramfs_mount': inode.c:(.text+0x220): undefined reference to `mount_mtd'
  fs/cramfs/inode.o: In function `cramfs_mtd_fill_super':
  inode.c:(.text+0x6d8): undefined reference to `mtd_point'
  inode.c:(.text+0xae4): undefined reference to `mtd_unpoint'

This adds a more specific Kconfig dependency to avoid the broken
configuration.

Alternatively we could make CRAMFS itself depend on "MTD || !MTD" with a
similar result.

Fixes: 99c18ce580c6 ("cramfs: direct memory access support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sun, 17 Dec 2017 20:18:35 +0000 (12:18 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull vfs fixes from Al Viro:
 "The alloc_super() one is a regression in this merge window, lazytime
  thing is older..."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  VFS: Handle lazytime in do_mount()
  alloc_super(): do ->s_umount initialization earlier

7 years agoMerge tag 'ext4_for_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Sun, 17 Dec 2017 20:14:33 +0000 (12:14 -0800)]
Merge tag 'ext4_for_stable' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 fixes from Ted Ts'o:
 "Fix a regression which caused us to fail to interpret symlinks in very
  ancient ext3 file system images.

  Also fix two xfstests failures, one of which could cause an OOPS, plus
  an additional bug fix caught by fuzz testing"

* tag 'ext4_for_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix crash when a directory's i_size is too small
  ext4: add missing error check in __ext4_new_inode()
  ext4: fix fdatasync(2) after fallocate(2) operation
  ext4: support fast symlinks from ext3 file systems

7 years agox86/mm/kasan: Don't use vmemmap_populate() to initialize shadow
Andrey Ryabinin [Thu, 16 Nov 2017 01:36:35 +0000 (17:36 -0800)]
x86/mm/kasan: Don't use vmemmap_populate() to initialize shadow

[ Note, this is a Git cherry-pick of the following commit:

    d17a1d97dc20: ("x86/mm/kasan: don't use vmemmap_populate() to initialize shadow")

  ... for easier x86 PTI code testing and back-porting. ]

The KASAN shadow is currently mapped using vmemmap_populate() since that
provides a semi-convenient way to map pages into init_top_pgt.  However,
since that no longer zeroes the mapped pages, it is not suitable for
KASAN, which requires zeroed shadow memory.

Add kasan_populate_shadow() interface and use it instead of
vmemmap_populate().  Besides, this allows us to take advantage of
gigantic pages and use them to populate the shadow, which should save us
some memory wasted on page tables and reduce TLB pressure.

Link: http://lkml.kernel.org/r/20171103185147.2688-2-pasha.tatashin@oracle.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Steven Sistare <steven.sistare@oracle.com>
Cc: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Bob Picco <bob.picco@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agolocking/barriers: Convert users of lockless_dereference() to READ_ONCE()
Will Deacon [Tue, 24 Oct 2017 10:22:48 +0000 (11:22 +0100)]
locking/barriers: Convert users of lockless_dereference() to READ_ONCE()

[ Note, this is a Git cherry-pick of the following commit:

    506458efaf15 ("locking/barriers: Convert users of lockless_dereference() to READ_ONCE()")

  ... for easier x86 PTI code testing and back-porting. ]

READ_ONCE() now has an implicit smp_read_barrier_depends() call, so it
can be used instead of lockless_dereference() without any change in
semantics.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1508840570-22169-4-git-send-email-will.deacon@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agolocking/barriers: Add implicit smp_read_barrier_depends() to READ_ONCE()
Will Deacon [Tue, 24 Oct 2017 10:22:47 +0000 (11:22 +0100)]
locking/barriers: Add implicit smp_read_barrier_depends() to READ_ONCE()

[ Note, this is a Git cherry-pick of the following commit:

    76ebbe78f739 ("locking/barriers: Add implicit smp_read_barrier_depends() to READ_ONCE()")

  ... for easier x86 PTI code testing and back-porting. ]

In preparation for the removal of lockless_dereference(), which is the
same as READ_ONCE() on all architectures other than Alpha, add an
implicit smp_read_barrier_depends() to READ_ONCE() so that it can be
used to head dependency chains on all architectures.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1508840570-22169-3-git-send-email-will.deacon@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agobpf: fix build issues on um due to mising bpf_perf_event.h
Daniel Borkmann [Tue, 12 Dec 2017 01:25:31 +0000 (02:25 +0100)]
bpf: fix build issues on um due to mising bpf_perf_event.h

[ Note, this is a Git cherry-pick of the following commit:

    a23f06f06dbe ("bpf: fix build issues on um due to mising bpf_perf_event.h")

  ... for easier x86 PTI code testing and back-porting. ]

Since c895f6f703ad ("bpf: correct broken uapi for
BPF_PROG_TYPE_PERF_EVENT program type") um (uml) won't build
on i386 or x86_64:

  [...]
    CC      init/main.o
  In file included from ../include/linux/perf_event.h:18:0,
                   from ../include/linux/trace_events.h:10,
                   from ../include/trace/syscall.h:7,
                   from ../include/linux/syscalls.h:82,
                   from ../init/main.c:20:
  ../include/uapi/linux/bpf_perf_event.h:11:32: fatal error:
  asm/bpf_perf_event.h: No such file or directory #include
  <asm/bpf_perf_event.h>
  [...]

Lets add missing bpf_perf_event.h also to um arch. This seems
to be the only one still missing.

Fixes: c895f6f703ad ("bpf: correct broken uapi for BPF_PROG_TYPE_PERF_EVENT program type")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Suggested-by: Richard Weinberger <richard@sigma-star.at>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Cc: Richard Weinberger <richard@sigma-star.at>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoperf/x86: Enable free running PEBS for REGS_USER/INTR
Andi Kleen [Thu, 31 Aug 2017 21:46:30 +0000 (14:46 -0700)]
perf/x86: Enable free running PEBS for REGS_USER/INTR

[ Note, this is a Git cherry-pick of the following commit:

    a47ba4d77e12 ("perf/x86: Enable free running PEBS for REGS_USER/INTR")

  ... for easier x86 PTI code testing and back-porting. ]

Currently free running PEBS is disabled when user or interrupt
registers are requested. Most of the registers are actually
available in the PEBS record and can be supported.

So we just need to check for the supported registers and then
allow it: it is all except for the segment register.

For user registers this only works when the counter is limited
to ring 3 only, so this also needs to be checked.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170831214630.21892-1-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agox86: Make X86_BUG_FXSAVE_LEAK detectable in CPUID on AMD
Rudolf Marek [Tue, 28 Nov 2017 21:01:06 +0000 (22:01 +0100)]
x86: Make X86_BUG_FXSAVE_LEAK detectable in CPUID on AMD

[ Note, this is a Git cherry-pick of the following commit:

    2b67799bdf25 ("x86: Make X86_BUG_FXSAVE_LEAK detectable in CPUID on AMD")

  ... for easier x86 PTI code testing and back-porting. ]

The latest AMD AMD64 Architecture Programmer's Manual
adds a CPUID feature XSaveErPtr (CPUID_Fn80000008_EBX[2]).

If this feature is set, the FXSAVE, XSAVE, FXSAVEOPT, XSAVEC, XSAVES
/ FXRSTOR, XRSTOR, XRSTORS always save/restore error pointers,
thus making the X86_BUG_FXSAVE_LEAK workaround obsolete on such CPUs.

Signed-Off-By: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Tested-by: Borislav Petkov <bp@suse.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Link: https://lkml.kernel.org/r/bdcebe90-62c5-1f05-083c-eba7f08b2540@assembler.cz
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agox86/cpufeature: Add User-Mode Instruction Prevention definitions
Ricardo Neri [Mon, 6 Nov 2017 02:27:51 +0000 (18:27 -0800)]
x86/cpufeature: Add User-Mode Instruction Prevention definitions

[ Note, this is a Git cherry-pick of the following commit: (limited to the cpufeatures.h file)

    3522c2a6a4f3 ("x86/cpufeature: Add User-Mode Instruction Prevention definitions")

  ... for easier x86 PTI code testing and back-porting. ]

User-Mode Instruction Prevention is a security feature present in new
Intel processors that, when set, prevents the execution of a subset of
instructions if such instructions are executed in user mode (CPL > 0).
Attempting to execute such instructions causes a general protection
exception.

The subset of instructions comprises:

 * SGDT - Store Global Descriptor Table
 * SIDT - Store Interrupt Descriptor Table
 * SLDT - Store Local Descriptor Table
 * SMSW - Store Machine Status Word
 * STR  - Store Task Register

This feature is also added to the list of disabled-features to allow
a cleaner handling of build-time configuration.

Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Chen Yucong <slaoub@gmail.com>
Cc: Chris Metcalf <cmetcalf@mellanox.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Huang Rui <ray.huang@amd.com>
Cc: Jiri Slaby <jslaby@suse.cz>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi V. Shankar <ravi.v.shankar@intel.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: ricardo.neri@intel.com
Link: http://lkml.kernel.org/r/1509935277-22138-7-git-send-email-ricardo.neri-calderon@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoMerge commit 'upstream-x86-virt' into WIP.x86/mm
Ingo Molnar [Fri, 1 Dec 2017 09:34:04 +0000 (10:34 +0100)]
Merge commit 'upstream-x86-virt' into WIP.x86/mm

Merge a minimal set of virt cleanups, for a base for the MM isolation patches.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoMerge branch 'upstream-acpi-fixes' into WIP.x86/pti.base
Ingo Molnar [Sun, 17 Dec 2017 12:09:31 +0000 (13:09 +0100)]
Merge branch 'upstream-acpi-fixes' into WIP.x86/pti.base

Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoMerge branch 'upstream-x86-selftests' into WIP.x86/pti.base
Ingo Molnar [Sun, 17 Dec 2017 12:04:28 +0000 (13:04 +0100)]
Merge branch 'upstream-x86-selftests' into WIP.x86/pti.base

Conflicts:
arch/x86/kernel/cpu/Makefile

Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoMerge commit 'upstream-x86-entry' into WIP.x86/mm
Ingo Molnar [Fri, 1 Dec 2017 09:32:48 +0000 (10:32 +0100)]
Merge commit 'upstream-x86-entry' into WIP.x86/mm

Pull in a minimal set of v4.15 entry code changes, for a base for the MM isolation patches.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agodrivers/misc/intel/pti: Rename the header file to free up the namespace
Ingo Molnar [Tue, 5 Dec 2017 13:14:47 +0000 (14:14 +0100)]
drivers/misc/intel/pti: Rename the header file to free up the namespace

We'd like to use the 'PTI' acronym for 'Page Table Isolation' - free up the
namespace by renaming the <linux/pti.h> driver header to <linux/intel-pti.h>.

(Also standardize the header guard name while at it.)

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: J Freyensee <james_p_freyensee@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
7 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Linus Torvalds [Sat, 16 Dec 2017 21:43:08 +0000 (13:43 -0800)]
Merge tag 'for-linus' of git://git./linux/kernel/git/rdma/rdma

Pull rdma fixes from Jason Gunthorpe:
 "More fixes from testing done on the rc kernel, including more SELinux
  testing. Looking forward, lockdep found regression today in ipoib
  which is still being fixed.

  Summary:

   - Fix for SELinux on the umad SMI path. Some old hardware does not
     fill the PKey properly exposing another bug in the newer SELinux
     code.

   - Check the input port as we can exceed array bounds from this user
     supplied value

   - Users are unable to use the hash field support as they want due to
     incorrect checks on the field restrictions, correct that so the
     feature works as intended

   - User triggerable oops in the NETLINK_RDMA handler

   - cxgb4 driver fix for a bad interaction with CQ flushing in iser
     caused by patches in this merge window, and bad CQ flushing during
     normal close.

   - Unbalanced memalloc_noio in ipoib in an error path"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  IB/ipoib: Restore MM behavior in case of tx_ring allocation failure
  iw_cxgb4: only insert drain cqes if wq is flushed
  iw_cxgb4: only clear the ARMED bit if a notification is needed
  RDMA/netlink: Fix general protection fault
  IB/mlx4: Fix RSS hash fields restrictions
  IB/core: Don't enforce PKey security on SMI MADs
  IB/core: Bound check alternate path port number

7 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 16 Dec 2017 21:34:38 +0000 (13:34 -0800)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Two bugfixes for the AT24 I2C eeprom driver and some minor corrections
  for I2C bus drivers"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: piix4: Fix port number check on release
  i2c: stm32: Fix copyrights
  i2c-cht-wc: constify platform_device_id
  eeprom: at24: change nvmem stride to 1
  eeprom: at24: fix I2C device selection for runtime PM

7 years agoMerge tag 'nfs-for-4.15-3' of git://git.linux-nfs.org/projects/anna/linux-nfs
Linus Torvalds [Sat, 16 Dec 2017 21:12:53 +0000 (13:12 -0800)]
Merge tag 'nfs-for-4.15-3' of git://git.linux-nfs.org/projects/anna/linux-nfs

Pull NFS client fixes from Anna Schumaker:
 "This has two stable bugfixes, one to fix a BUG_ON() when
  nfs_commit_inode() is called with no outstanding commit requests and
  another to fix a race in the SUNRPC receive codepath.

  Additionally, there are also fixes for an NFS client deadlock and an
  xprtrdma performance regression.

  Summary:

  Stable bugfixes:
   - NFS: Avoid a BUG_ON() in nfs_commit_inode() by not waiting for a
     commit in the case that there were no commit requests.
   - SUNRPC: Fix a race in the receive code path

  Other fixes:
   - NFS: Fix a deadlock in nfs client initialization
   - xprtrdma: Fix a performance regression for small IOs"

* tag 'nfs-for-4.15-3' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  SUNRPC: Fix a race in the receive code path
  nfs: don't wait on commit in nfs_commit_inode() if there were no commit requests
  xprtrdma: Spread reply processing over more CPUs
  nfs: fix a deadlock in nfs client initialization

7 years agoRevert "mm: replace p??_write with pte_access_permitted in fault + gup paths"
Linus Torvalds [Sat, 16 Dec 2017 02:53:22 +0000 (18:53 -0800)]
Revert "mm: replace p??_write with pte_access_permitted in fault + gup paths"

This reverts commits 5c9d2d5c269cc7da82b894e9, and e7fe7b5cae90.

We'll probably need to revisit this, but basically we should not
complicate the get_user_pages_fast() case, and checking the actual page
table protection key bits will require more care anyway, since the
protection keys depend on the exact state of the VM in question.

Particularly when doing a "remote" page lookup (ie in somebody elses VM,
not your own), you need to be much more careful than this was.  Dave
Hansen says:

 "So, the underlying bug here is that we now a get_user_pages_remote()
  and then go ahead and do the p*_access_permitted() checks against the
  current PKRU. This was introduced recently with the addition of the
  new p??_access_permitted() calls.

  We have checks in the VMA path for the "remote" gups and we avoid
  consulting PKRU for them. This got missed in the pkeys selftests
  because I did a ptrace read, but not a *write*. I also didn't
  explicitly test it against something where a COW needed to be done"

It's also not entirely clear that it makes sense to check the protection
key bits at this level at all.  But one possible eventual solution is to
make the get_user_pages_fast() case just abort if it sees protection key
bits set, which makes us fall back to the regular get_user_pages() case,
which then has a vma and can do the check there if we want to.

We'll see.

Somewhat related to this all: what we _do_ want to do some day is to
check the PAGE_USER bit - it should obviously always be set for user
pages, but it would be a good check to have back.  Because we have no
generic way to test for it, we lost it as part of moving over from the
architecture-specific x86 GUP implementation to the generic one in
commit e585513b76f7 ("x86/mm/gup: Switch GUP to the generic
get_user_page_fast() implementation").

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 15 Dec 2017 21:08:37 +0000 (13:08 -0800)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Clamp timeouts to INT_MAX in conntrack, from Jay Elliot.

 2) Fix broken UAPI for BPF_PROG_TYPE_PERF_EVENT, from Hendrik
    Brueckner.

 3) Fix locking in ieee80211_sta_tear_down_BA_sessions, from Johannes
    Berg.

 4) Add missing barriers to ptr_ring, from Michael S. Tsirkin.

 5) Don't advertise gigabit in sh_eth when not available, from Thomas
    Petazzoni.

 6) Check network namespace when delivering to netlink taps, from Kevin
    Cernekee.

 7) Kill a race in raw_sendmsg(), from Mohamed Ghannam.

 8) Use correct address in TCP md5 lookups when replying to an incoming
    segment, from Christoph Paasch.

 9) Add schedule points to BPF map alloc/free, from Eric Dumazet.

10) Don't allow silly mtu values to be used in ipv4/ipv6 multicast, also
    from Eric Dumazet.

11) Fix SKB leak in tipc, from Jon Maloy.

12) Disable MAC learning on OVS ports of mlxsw, from Yuval Mintz.

13) SKB leak fix in skB_complete_tx_timestamp(), from Willem de Bruijn.

14) Add some new qmi_wwan device IDs, from Daniele Palmas.

15) Fix static key imbalance in ingress qdisc, from Jiri Pirko.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (76 commits)
  net: qcom/emac: Reduce timeout for mdio read/write
  net: sched: fix static key imbalance in case of ingress/clsact_init error
  net: sched: fix clsact init error path
  ip_gre: fix wrong return value of erspan_rcv
  net: usb: qmi_wwan: add Telit ME910 PID 0x1101 support
  pkt_sched: Remove TC_RED_OFFLOADED from uapi
  net: sched: Move to new offload indication in RED
  net: sched: Add TCA_HW_OFFLOAD
  net: aquantia: Increment driver version
  net: aquantia: Fix typo in ethtool statistics names
  net: aquantia: Update hw counters on hw init
  net: aquantia: Improve link state and statistics check interval callback
  net: aquantia: Fill in multicast counter in ndev stats from hardware
  net: aquantia: Fill ndev stat couters from hardware
  net: aquantia: Extend stat counters to 64bit values
  net: aquantia: Fix hardware DMA stream overload on large MRRS
  net: aquantia: Fix actual speed capabilities reporting
  sock: free skb in skb_complete_tx_timestamp on error
  s390/qeth: update takeover IPs after configuration change
  s390/qeth: lock IP table while applying takeover changes
  ...

7 years agoMerge tag 'usb-4.15-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 15 Dec 2017 21:03:25 +0000 (13:03 -0800)]
Merge tag 'usb-4.15-rc4' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some USB fixes for 4.15-rc4.

  There is the usual handful gadget/dwc2/dwc3 fixes as always, for
  reported issues. But the most important things in here is the core fix
  from Alan Stern to resolve a nasty security bug (my first attempt is
  reverted, Alan's was much cleaner), as well as a number of usbip fixes
  from Shuah Khan to resolve those reported security issues.

  All of these have been in linux-next with no reported issues"

* tag 'usb-4.15-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  USB: core: prevent malicious bNumInterfaces overflow
  Revert "USB: core: only clean up what we allocated"
  USB: core: only clean up what we allocated
  Revert "usb: gadget: allow to enable legacy drivers without USB_ETH"
  usb: gadget: webcam: fix V4L2 Kconfig dependency
  usb: dwc2: Fix TxFIFOn sizes and total TxFIFO size issues
  usb: dwc3: gadget: Fix PCM1 for ISOC EP with ep->mult less than 3
  usb: dwc3: of-simple: set dev_pm_ops
  usb: dwc3: of-simple: fix missing clk_disable_unprepare
  usb: dwc3: gadget: Wait longer for controller to end command processing
  usb: xhci: fix TDS for MTK xHCI1.1
  xhci: Don't add a virt_dev to the devs array before it's fully allocated
  usbip: fix stub_send_ret_submit() vulnerability to null transfer_buffer
  usbip: prevent vhci_hcd driver from leaking a socket pointer address
  usbip: fix stub_rx: harden CMD_SUBMIT path to handle malicious input
  usbip: fix stub_rx: get_pipe() to validate endpoint number
  tools/usbip: fixes potential (minor) "buffer overflow" (detected on recent gcc with -Werror)
  USB: uas and storage: Add US_FL_BROKEN_FUA for another JMicron JMS567 ID
  usb: musb: da8xx: fix babble condition handling

7 years agoMerge tag 'staging-4.15-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Fri, 15 Dec 2017 20:59:48 +0000 (12:59 -0800)]
Merge tag 'staging-4.15-rc4' of git://git./linux/kernel/git/gregkh/staging

Pull staging fixes from Greg KH:
 "Here are some small staging driver fixes for 4.15-rc4.

  One patch for the ccree driver to prevent an unitialized value from
  being returned to a caller, and the other fixes a logic error in the
  pi433 driver"

* tag 'staging-4.15-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: pi433: Fixes issue with bit shift in rf69_get_modulation
  staging: ccree: Uninitialized return in ssi_ahash_import()

7 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Fri, 15 Dec 2017 20:56:23 +0000 (12:56 -0800)]
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio regression fixes from Michael Tsirkin:
 "Fixes two issues in the latest kernel"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  virtio_mmio: fix devm cleanup
  ptr_ring: fix up after recent ptr_ring changes

7 years agoMerge tag 'for-4.15/dm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device...
Linus Torvalds [Fri, 15 Dec 2017 20:53:37 +0000 (12:53 -0800)]
Merge tag 'for-4.15/dm-fixes' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:

 - fix a particularly nasty DM core bug in a 4.15 refcount_t conversion.

 - fix various targets to dm_register_target after module __init
   resources created; otherwise racing lvm2 commands could result in a
   NULL pointer during initialization of associated DM kernel module.

 - fix regression in bio-based DM multipath queue_if_no_path handling.

 - fix DM bufio's shrinker to reclaim more than one buffer per scan.

* tag 'for-4.15/dm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm bufio: fix shrinker scans when (nr_to_scan < retain_target)
  dm mpath: fix bio-based multipath queue_if_no_path handling
  dm: fix various targets to dm_register_target after module __init resources created
  dm table: fix regression from improper dm_dev_internal.count refcount_t conversion

7 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Fri, 15 Dec 2017 20:51:42 +0000 (12:51 -0800)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "The most important one is the bfa fix because it's easy to oops the
  kernel with this driver (this includes the commit that corrects the
  compiler warning in the original), a regression in the new timespec
  conversion in aacraid and a regression in the Fibre Channel ELS
  handling patch.

  The other three are a theoretical problem with termination in the
  vendor/host matching code and a use after free in lpfc.

  The additional patches are a fix for an I/O hang in the mq code under
  certain circumstances and a rare oops in some debugging code"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: core: Fix a scsi_show_rq() NULL pointer dereference
  scsi: MAINTAINERS: change FCoE list to linux-scsi
  scsi: libsas: fix length error in sas_smp_handler()
  scsi: bfa: fix type conversion warning
  scsi: core: run queue if SCSI device queue isn't ready and queue is idle
  scsi: scsi_devinfo: cleanly zero-pad devinfo strings
  scsi: scsi_devinfo: handle non-terminated strings
  scsi: bfa: fix access to bfad_im_port_s
  scsi: aacraid: address UBSAN warning regression
  scsi: libfc: fix ELS request handling
  scsi: lpfc: Use after free in lpfc_rq_buf_free()

7 years agoMerge tag 'mmc-v4.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Fri, 15 Dec 2017 20:49:54 +0000 (12:49 -0800)]
Merge tag 'mmc-v4.15-rc2' of git://git./linux/kernel/git/ulfh/mmc

Pull MMC fixes from Ulf Hansson:
 "A couple of MMC fixes:

   - fix use of uninitialized drv_typ variable

   - apply NO_CMD23 quirk to some specific SD cards to make them work"

* tag 'mmc-v4.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: core: apply NO_CMD23 quirk to some specific cards
  mmc: core: properly init drv_type

7 years agoMerge tag 'ceph-for-4.15-rc4' of git://github.com/ceph/ceph-client
Linus Torvalds [Fri, 15 Dec 2017 20:48:27 +0000 (12:48 -0800)]
Merge tag 'ceph-for-4.15-rc4' of git://github.com/ceph/ceph-client

Pull ceph fix from Ilya Dryomov:
 "CephFS inode trimming fix from Zheng, marked for stable"

* tag 'ceph-for-4.15-rc4' of git://github.com/ceph/ceph-client:
  ceph: drop negative child dentries before try pruning inode's alias

7 years agoMerge branch 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszer...
Linus Torvalds [Fri, 15 Dec 2017 20:46:48 +0000 (12:46 -0800)]
Merge branch 'overlayfs-linus' of git://git./linux/kernel/git/mszeredi/vfs

Pull overlayfs fixes from Miklos Szeredi:

 - fix incomplete syncing of filesystem

 - fix regression in readdir on ovl over 9p

 - only follow redirects when needed

 - misc fixes and cleanups

* 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
  ovl: fix overlay: warning prefix
  ovl: Use PTR_ERR_OR_ZERO()
  ovl: Sync upper dirty data when syncing overlayfs
  ovl: update ctx->pos on impure dir iteration
  ovl: Pass ovl_get_nlink() parameters in right order
  ovl: don't follow redirects if redirect_dir=off

7 years agonet: qcom/emac: Reduce timeout for mdio read/write
Hemanth Puranik [Fri, 15 Dec 2017 14:35:58 +0000 (20:05 +0530)]
net: qcom/emac: Reduce timeout for mdio read/write

Currently mdio read/write takes around ~115us as the timeout
between status check is set to 100us.
By reducing the timeout to 1us mdio read/write takes ~15us to
complete. This improves the link up event response.

Signed-off-by: Hemanth Puranik <hpuranik@codeaurora.org>
Acked-by: Timur Tabi <timur@codeaurora.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 15 Dec 2017 20:44:49 +0000 (12:44 -0800)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Will Deacon:
 "There are some significant fixes in here for FP state corruption,
  hardware access/dirty PTE corruption and an erratum workaround for the
  Falkor CPU.

  I'm hoping that things finally settle down now, but never say never...

  Summary:

   - Fix FPSIMD context switch regression introduced in -rc2

   - Fix ABI break with SVE CPUID register reporting

   - Fix use of uninitialised variable

   - Fixes to hardware access/dirty management and sanity checking

   - CPU erratum workaround for Falkor CPUs

   - Fix reporting of writeable+executable mappings

   - Fix signal reporting for RAS errors"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: fpsimd: Fix copying of FP state from signal frame into task struct
  arm64/sve: Report SVE to userspace via CPUID only if supported
  arm64: fix CONFIG_DEBUG_WX address reporting
  arm64: fault: avoid send SIGBUS two times
  arm64: hw_breakpoint: Use linux/uaccess.h instead of asm/uaccess.h
  arm64: Add software workaround for Falkor erratum 1041
  arm64: Define cputype macros for Falkor CPU
  arm64: mm: Fix false positives in set_pte_at access/dirty race detection
  arm64: mm: Fix pte_mkclean, pte_mkdirty semantics
  arm64: Initialise high_memory global variable earlier

7 years agonet: sched: fix static key imbalance in case of ingress/clsact_init error
Jiri Pirko [Fri, 15 Dec 2017 11:40:13 +0000 (12:40 +0100)]
net: sched: fix static key imbalance in case of ingress/clsact_init error

Move static key increments to the beginning of the init function
so they pair 1:1 with decrements in ingress/clsact_destroy,
which is called in case ingress/clsact_init fails.

Fixes: 6529eaba33f0 ("net: sched: introduce tcf block infractructure")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: sched: fix clsact init error path
Jiri Pirko [Fri, 15 Dec 2017 11:40:12 +0000 (12:40 +0100)]
net: sched: fix clsact init error path

Since in qdisc_create, the destroy op is called when init fails, we
don't do cleanup in init and leave it up to destroy.
This fixes use-after-free when trying to put already freed block.

Fixes: 6e40cf2d4dee ("net: sched: use extended variants of block_get/put in ingress and clsact qdiscs")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Dec 2017 20:14:33 +0000 (12:14 -0800)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "Misc fixes:

   - fix the s2ram regression related to confusion around segment
     register restoration, plus related cleanups that make the code more
     robust

   - a guess-unwinder Kconfig dependency fix

   - an isoimage build target fix for certain tool chain combinations

   - instruction decoder opcode map fixes+updates, and the syncing of
     the kernel decoder headers to the objtool headers

   - a kmmio tracing fix

   - two 5-level paging related fixes

   - a topology enumeration fix on certain SMP systems"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  objtool: Resync objtool's instruction decoder source code copy with the kernel's latest version
  x86/decoder: Fix and update the opcodes map
  x86/power: Make restore_processor_context() sane
  x86/power/32: Move SYSENTER MSR restoration to fix_processor_context()
  x86/power/64: Use struct desc_ptr for the IDT in struct saved_context
  x86/unwinder/guess: Prevent using CONFIG_UNWINDER_GUESS=y with CONFIG_STACKDEPOT=y
  x86/build: Don't verify mtools configuration file for isoimage
  x86/mm/kmmio: Fix mmiotrace for page unaligned addresses
  x86/boot/compressed/64: Print error if 5-level paging is not supported
  x86/boot/compressed/64: Detect and handle 5-level paging at boot-time
  x86/smpboot: Do not use smp_num_siblings in __max_logical_packages calculation

7 years agoMerge branch 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Dec 2017 19:44:59 +0000 (11:44 -0800)]
Merge branch 'locking-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull locking fixes from Ingo Molnar:
 "Misc fixes:

   - Fix a S390 boot hang that was caused by the lock-break logic.
     Remove lock-break to begin with, as review suggested it was
     unreasonably fragile and our confidence in its continued good
     health is lower than our confidence in its removal.

   - Remove the lockdep cross-release checking code for now, because of
     unresolved false positive warnings. This should make lockdep work
     well everywhere again.

   - Get rid of the final (and single) ACCESS_ONCE() straggler and
     remove the API from v4.15.

   - Fix a liblockdep build warning"

* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tools/lib/lockdep: Add missing declaration of 'pr_cont()'
  checkpatch: Remove ACCESS_ONCE() warning
  compiler.h: Remove ACCESS_ONCE()
  tools/include: Remove ACCESS_ONCE()
  tools/perf: Convert ACCESS_ONCE() to READ_ONCE()
  locking/lockdep: Remove the cross-release locking checks
  locking/core: Remove break_lock field when CONFIG_GENERIC_LOCKBREAK=y
  locking/core: Fix deadlock during boot on systems with GENERIC_LOCKBREAK

7 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Dec 2017 19:40:24 +0000 (11:40 -0800)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "Two fixes: a crash fix for an ARM SoC platform, and kernel-doc
  warnings fixes"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/rt: Do not pull from current CPU if only one CPU to pull
  sched/core: Fix kernel-doc warnings after code movement

7 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Dec 2017 19:36:20 +0000 (11:36 -0800)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf tooling fix from Ingo Molnar:
 "Synchronize kernel <-> tooling headers to resolve two build warnings
  in the perf build"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tools/headers: Synchronize kernel <-> tooling headers

7 years agoMerge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Dec 2017 19:34:29 +0000 (11:34 -0800)]
Merge branch 'core-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull early_ioremap fix from Ingo Molnar:
 "A boot hang fix when the EFI earlyprintk driver is enabled"

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  mm/early_ioremap: Fix boot hang with earlyprintk=efi,keep

7 years agoMerge tag 'for-linus-4.15-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 15 Dec 2017 19:32:09 +0000 (11:32 -0800)]
Merge tag 'for-linus-4.15-rc4-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fixes from Juergen Gross:
 "Two minor fixes for running as Xen dom0:

   - when built as 32 bit kernel on large machines the Xen LAPIC
     emulation should report a rather modern LAPIC in order to support
     enough APIC-Ids

   - The Xen LAPIC emulation is needed for dom0 only, so build it only
     for kernels supporting to run as Xen dom0"

* tag 'for-linus-4.15-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen: XEN_ACPI_PROCESSOR is Dom0-only
  x86/Xen: don't report ancient LAPIC version

7 years agoSUNRPC: Fix a race in the receive code path
Trond Myklebust [Fri, 15 Dec 2017 02:24:08 +0000 (21:24 -0500)]
SUNRPC: Fix a race in the receive code path

We must ensure that the call to rpc_sleep_on() in xprt_transmit() cannot
race with the call to xprt_complete_rqst().

Reported-by: Chuck Lever <chuck.lever@oracle.com>
Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=317
Fixes: ce7c252a8c74 ("SUNRPC: Add a separate spinlock to protect..")
Cc: stable@vger.kernel.org # 4.14+
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
7 years agonfs: don't wait on commit in nfs_commit_inode() if there were no commit requests
Scott Mayhew [Fri, 8 Dec 2017 21:00:12 +0000 (16:00 -0500)]
nfs: don't wait on commit in nfs_commit_inode() if there were no commit requests

If there were no commit requests, then nfs_commit_inode() should not
wait on the commit or mark the inode dirty, otherwise the following
BUG_ON can be triggered:

[ 1917.130762] kernel BUG at fs/inode.c:578!
[ 1917.130766] Oops: Exception in kernel mode, sig: 5 [#1]
[ 1917.130768] SMP NR_CPUS=2048 NUMA pSeries
[ 1917.130772] Modules linked in: iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi blocklayoutdriver rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc sg nx_crypto pseries_rng ip_tables xfs libcrc32c sd_mod crc_t10dif crct10dif_generic crct10dif_common ibmvscsi scsi_transport_srp ibmveth scsi_tgt dm_mirror dm_region_hash dm_log dm_mod
[ 1917.130805] CPU: 2 PID: 14923 Comm: umount.nfs4 Tainted: G               ------------ T 3.10.0-768.el7.ppc64 #1
[ 1917.130810] task: c0000005ecd88040 ti: c00000004cea0000 task.ti: c00000004cea0000
[ 1917.130813] NIP: c000000000354178 LR: c000000000354160 CTR: c00000000012db80
[ 1917.130816] REGS: c00000004cea3720 TRAP: 0700   Tainted: G               ------------ T  (3.10.0-768.el7.ppc64)
[ 1917.130820] MSR: 8000000100029032 <SF,EE,ME,IR,DR,RI>  CR: 22002822  XER: 20000000
[ 1917.130828] CFAR: c00000000011f594 SOFTE: 1
GPR00: c000000000354160 c00000004cea39a0 c0000000014c4700 c0000000018cc750
GPR04: 000000000000c750 80c0000000000000 0600000000000000 04eeb76bea749a03
GPR08: 0000000000000034 c0000000018cc758 0000000000000001 d000000005e619e8
GPR12: c00000000012db80 c000000007b31200 0000000000000000 0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR20: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR24: 0000000000000000 c000000000dfc3ec 0000000000000000 c0000005eefc02c0
GPR28: d0000000079dbd50 c0000005b94a02c0 c0000005b94a0250 c0000005b94a01c8
[ 1917.130867] NIP [c000000000354178] .evict+0x1c8/0x350
[ 1917.130871] LR [c000000000354160] .evict+0x1b0/0x350
[ 1917.130873] Call Trace:
[ 1917.130876] [c00000004cea39a0] [c000000000354160] .evict+0x1b0/0x350 (unreliable)
[ 1917.130880] [c00000004cea3a30] [c0000000003558cc] .evict_inodes+0x13c/0x270
[ 1917.130884] [c00000004cea3af0] [c000000000327d20] .kill_anon_super+0x70/0x1e0
[ 1917.130896] [c00000004cea3b80] [d000000005e43e30] .nfs_kill_super+0x20/0x60 [nfs]
[ 1917.130900] [c00000004cea3c00] [c000000000328a20] .deactivate_locked_super+0xa0/0x1b0
[ 1917.130903] [c00000004cea3c80] [c00000000035ba54] .cleanup_mnt+0xd4/0x180
[ 1917.130907] [c00000004cea3d10] [c000000000119034] .task_work_run+0x114/0x150
[ 1917.130912] [c00000004cea3db0] [c00000000001ba6c] .do_notify_resume+0xcc/0x100
[ 1917.130916] [c00000004cea3e30] [c00000000000a7b0] .ret_from_except_lite+0x5c/0x60
[ 1917.130919] Instruction dump:
[ 1917.130921] 7fc3f378 486734b5 60000000 387f00a0 38800003 4bdcb365 60000000 e95f00a0
[ 1917.130927] 694a0060 7d4a0074 794ad182 694a0001 <0b0a0000892d02a4 2f890000 40de0134

Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Cc: stable@vger.kernel.org # 4.5+
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
7 years agoxprtrdma: Spread reply processing over more CPUs
Chuck Lever [Mon, 4 Dec 2017 19:04:04 +0000 (14:04 -0500)]
xprtrdma: Spread reply processing over more CPUs

Commit d8f532d20ee4 ("xprtrdma: Invoke rpcrdma_reply_handler
directly from RECV completion") introduced a performance regression
for NFS I/O small enough to not need memory registration. In multi-
threaded benchmarks that generate primarily small I/O requests,
IOPS throughput is reduced by nearly a third. This patch restores
the previous level of throughput.

Because workqueues are typically BOUND (in particular ib_comp_wq,
nfsiod_workqueue, and rpciod_workqueue), NFS/RDMA workloads tend
to aggregate on the CPU that is handling Receive completions.

The usual approach to addressing this problem is to create a QP
and CQ for each CPU, and then schedule transactions on the QP
for the CPU where you want the transaction to complete. The
transaction then does not require an extra context switch during
completion to end up on the same CPU where the transaction was
started.

This approach doesn't work for the Linux NFS/RDMA client because
currently the Linux NFS client does not support multiple connections
per client-server pair, and the RDMA core API does not make it
straightforward for ULPs to determine which CPU is responsible for
handling Receive completions for a CQ.

So for the moment, record the CPU number in the rpcrdma_req before
the transport sends each RPC Call. Then during Receive completion,
queue the RPC completion on that same CPU.

Additionally, move all RPC completion processing to the deferred
handler so that even RPCs with simple small replies complete on
the CPU that sent the corresponding RPC Call.

Fixes: d8f532d20ee4 ("xprtrdma: Invoke rpcrdma_reply_handler ...")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
7 years agonfs: fix a deadlock in nfs client initialization
Scott Mayhew [Tue, 5 Dec 2017 18:55:44 +0000 (13:55 -0500)]
nfs: fix a deadlock in nfs client initialization

The following deadlock can occur between a process waiting for a client
to initialize in while walking the client list during nfsv4 server trunking
detection and another process waiting for the nfs_clid_init_mutex so it
can initialize that client:

Process 1                               Process 2
---------                               ---------
spin_lock(&nn->nfs_client_lock);
list_add_tail(&CLIENTA->cl_share_link,
        &nn->nfs_client_list);
spin_unlock(&nn->nfs_client_lock);
                                        spin_lock(&nn->nfs_client_lock);
                                        list_add_tail(&CLIENTB->cl_share_link,
                                                &nn->nfs_client_list);
                                        spin_unlock(&nn->nfs_client_lock);
                                        mutex_lock(&nfs_clid_init_mutex);
                                        nfs41_walk_client_list(clp, result, cred);
                                        nfs_wait_client_init_complete(CLIENTA);
(waiting for nfs_clid_init_mutex)

Make sure nfs_match_client() only evaluates clients that have completed
initialization in order to prevent that deadlock.

This patch also fixes v4.0 trunking behavior by not marking the client
NFS_CS_READY until the clientid has been confirmed.

Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
7 years agoip_gre: fix wrong return value of erspan_rcv
Haishuang Yan [Fri, 15 Dec 2017 02:46:16 +0000 (10:46 +0800)]
ip_gre: fix wrong return value of erspan_rcv

If pskb_may_pull return failed, return PACKET_REJECT instead of -ENOMEM.

Fixes: 84e54fe0a5ea ("gre: introduce native tunnel support for ERSPAN")
Cc: William Tu <u9012063@gmail.com>
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Acked-by: William Tu <u9012063@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: usb: qmi_wwan: add Telit ME910 PID 0x1101 support
Daniele Palmas [Thu, 14 Dec 2017 15:56:14 +0000 (16:56 +0100)]
net: usb: qmi_wwan: add Telit ME910 PID 0x1101 support

This patch adds support for Telit ME910 PID 0x1101.

Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Acked-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoMerge branch 'net-sched-Make-qdisc-offload-uapi-uniform'
David S. Miller [Fri, 15 Dec 2017 18:35:37 +0000 (13:35 -0500)]
Merge branch 'net-sched-Make-qdisc-offload-uapi-uniform'

Yuval Mintz says:

====================
net: sched: Make qdisc offload uapi uniform

Several qdiscs can already be offloaded to hardware, but there's an
inconsistecy in regard to the uapi through which they indicate such
an offload is taking place - indication is passed to the user via
TCA_OPTIONS where each qdisc retains private logic for setting it.

The recent addition of offloading to RED in
602f3baf2218 ("net_sch: red: Add offload ability to RED qdisc") caused
the addition of yet another uapi field for this purpose -
TC_RED_OFFLOADED.

For clarity and prevention of bloat in the uapi we want to eliminate
said added uapi, replacing it with a common mechanism that can be used
to reflect offload status of the various qdiscs.

The first patch introduces TCA_HW_OFFLOAD as the generic message meant
for this purpose. The second changes the current RED implementation into
setting the internal bits necessary for passing it, and the third removes
TC_RED_OFFLOADED as its no longer needed.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agopkt_sched: Remove TC_RED_OFFLOADED from uapi
Yuval Mintz [Thu, 14 Dec 2017 13:54:31 +0000 (15:54 +0200)]
pkt_sched: Remove TC_RED_OFFLOADED from uapi

Following the previous patch, RED is now using the new uniform uapi
for indicating it's offloaded. As a result, TC_RED_OFFLOADED is no
longer utilized by kernel and can be removed [as it's still not
part of any stable release].

Fixes: 602f3baf2218 ("net_sch: red: Add offload ability to RED qdisc")
Signed-off-by: Yuval Mintz <yuvalm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: sched: Move to new offload indication in RED
Yuval Mintz [Thu, 14 Dec 2017 13:54:30 +0000 (15:54 +0200)]
net: sched: Move to new offload indication in RED

Let RED utilize the new internal flag, TCQ_F_OFFLOADED,
to mark a given qdisc as offloaded instead of using a dedicated
indication.

Also, change internal logic into looking at said flag when possible.

Fixes: 602f3baf2218 ("net_sch: red: Add offload ability to RED qdisc")
Signed-off-by: Yuval Mintz <yuvalm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: sched: Add TCA_HW_OFFLOAD
Yuval Mintz [Thu, 14 Dec 2017 13:54:29 +0000 (15:54 +0200)]
net: sched: Add TCA_HW_OFFLOAD

Qdiscs can be offloaded to HW, but current implementation isn't uniform.
Instead, qdiscs either pass information about offload status via their
TCA_OPTIONS or omit it altogether.

Introduce a new attribute - TCA_HW_OFFLOAD that would form a uniform
uAPI for the offloading status of qdiscs.

Signed-off-by: Yuval Mintz <yuvalm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoMerge branch 'aquantia-fixes'
David S. Miller [Fri, 15 Dec 2017 17:46:43 +0000 (12:46 -0500)]
Merge branch 'aquantia-fixes'

Igor Russkikh says:

====================
net: aquantia: Atlantic driver 12/2017 updates

The patchset contains important hardware fix for machines with large MRRS
and couple of improvement in stats and capabilities reporting

patch v3:
 - Fixed patch #7 after Andrew's finding. NIC level stats actually
   have to be cleaned only on hw struct creation (and this is done
   in kzalloc). On each hwinit we only have to reset link state
   to make sure hw stats update will not increment nic stats during init.

patch v2:
 - split into more detailed commits

Comment from David on wrong defines case will be submitted separately later
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Increment driver version
Igor Russkikh [Thu, 14 Dec 2017 09:34:48 +0000 (12:34 +0300)]
net: aquantia: Increment driver version

Add a suffix to distinguish kernel mainline version and aquantia releases

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Fix typo in ethtool statistics names
Igor Russkikh [Thu, 14 Dec 2017 09:34:47 +0000 (12:34 +0300)]
net: aquantia: Fix typo in ethtool statistics names

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Update hw counters on hw init
Igor Russkikh [Thu, 14 Dec 2017 09:34:46 +0000 (12:34 +0300)]
net: aquantia: Update hw counters on hw init

On very first start we should read out current HW counter values
to make diff based calculations later.
This also should be done each time NIC gets down/up or wakes up
after sleep state. We reset link state explicitly to prevent diffs
from being summed this first time.

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Improve link state and statistics check interval callback
Igor Russkikh [Thu, 14 Dec 2017 09:34:45 +0000 (12:34 +0300)]
net: aquantia: Improve link state and statistics check interval callback

Reduce timeout from 2 secs to 1 sec. If link is down,
reduce it to 500msec. This speeds up link detection.

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Fill in multicast counter in ndev stats from hardware
Igor Russkikh [Thu, 14 Dec 2017 09:34:44 +0000 (12:34 +0300)]
net: aquantia: Fill in multicast counter in ndev stats from hardware

This metric comes from HW and is also diff-calculated, like other counters

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Fill ndev stat couters from hardware
Igor Russkikh [Thu, 14 Dec 2017 09:34:43 +0000 (12:34 +0300)]
net: aquantia: Fill ndev stat couters from hardware

Originally they were filled from ring sw counters.
These sometimes incorrectly calculate byte and packet amounts
when using LRO/LSO and jumboframes. Filling ndev counters from
hardware makes them precise.

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Extend stat counters to 64bit values
Igor Russkikh [Thu, 14 Dec 2017 09:34:42 +0000 (12:34 +0300)]
net: aquantia: Extend stat counters to 64bit values

Device hardware provides only 32bit counters. Using these directly
causes byte counters to overflow soon. A separate nic level structure
with 64 bit counters is now used to collect incrementally all the stats
and report these counters to ethtool stats and ndev stats.

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Fix hardware DMA stream overload on large MRRS
Igor Russkikh [Thu, 14 Dec 2017 09:34:41 +0000 (12:34 +0300)]
net: aquantia: Fix hardware DMA stream overload on large MRRS

Systems with large MRRS on device (2K, 4K) with high data rates and/or
large MTU, atlantic observes DMA packet buffer overflow. On some systems
that causes PCIe transaction errors, hardware NMIs or datapath freeze.
This patch
1) Limits MRRS from device side to 2K (thats maximum our hardware supports)
2) Limit maximum size of outstanding TX DMA data read requests. This makes
hardware buffers running fine.

Signed-off-by: Pavel Belous <pavel.belous@aquantia.com>
Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agonet: aquantia: Fix actual speed capabilities reporting
Igor Russkikh [Thu, 14 Dec 2017 09:34:40 +0000 (12:34 +0300)]
net: aquantia: Fix actual speed capabilities reporting

Different hardware device Ids correspond to different maximum speed
available. Extra checks were added for devices D108 and D109 to
remove unsupported speeds from these device capabilities list.

Signed-off-by: Igor Russkikh <igor.russkikh@aquantia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agosock: free skb in skb_complete_tx_timestamp on error
Willem de Bruijn [Wed, 13 Dec 2017 19:41:06 +0000 (14:41 -0500)]
sock: free skb in skb_complete_tx_timestamp on error

skb_complete_tx_timestamp must ingest the skb it is passed. Call
kfree_skb if the skb cannot be enqueued.

Fixes: b245be1f4db1 ("net-timestamp: no-payload only sysctl")
Fixes: 9ac25fc06375 ("net: fix socket refcounting in skb_complete_tx_timestamp()")
Reported-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>