4 years agox86/intel_rdt: Enable CMT and MBM on new Skylake stepping
Tony Luck [Fri, 8 Jun 2018 16:07:32 +0000 (09:07 -0700)]
x86/intel_rdt: Enable CMT and MBM on new Skylake stepping

New stepping of Skylake has fixes for cache occupancy and memory
bandwidth monitoring.

Update the code to enable these by default on newer steppings.

Signed-off-by: Tony Luck <>
Signed-off-by: Thomas Gleixner <>
Cc: Fenghua Yu <>
Cc: # v4.14
Cc: Vikas Shivappa <>
4 years agox86/apic/vector: Print APIC control bits in debugfs
Thomas Gleixner [Mon, 4 Jun 2018 15:34:00 +0000 (17:34 +0200)]
x86/apic/vector: Print APIC control bits in debugfs

Extend the debugability of the vector management by adding the state bits
to the debugfs output.

Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agogenirq/affinity: Defer affinity setting if irq chip is busy
Thomas Gleixner [Mon, 4 Jun 2018 15:33:59 +0000 (17:33 +0200)]
genirq/affinity: Defer affinity setting if irq chip is busy

The case that interrupt affinity setting fails with -EBUSY can be handled
in the kernel completely by using the already available generic pending

If a irq_chip::set_affinity() fails with -EBUSY, handle it like the
interrupts for which irq_chip::set_affinity() can only be invoked from
interrupt context. Copy the new affinity mask to irq_desc::pending_mask and
set the affinity pending bit. The next raised interrupt for the affected
irq will check the pending bit and try to set the new affinity from the
handler. This avoids that -EBUSY is returned when an affinity change is
requested from user space and the previous change has not been cleaned
up. The new affinity will take effect when the next interrupt is raised
from the device.

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agox86/platform/uv: Use apic_ack_irq()
Thomas Gleixner [Mon, 4 Jun 2018 15:33:58 +0000 (17:33 +0200)]
x86/platform/uv: Use apic_ack_irq()

To address the EBUSY fail of interrupt affinity settings in case that the
previous setting has not been cleaned up yet, use the new apic_ack_irq()
function instead of the special uv_ack_apic() implementation which is
merily a wrapper around ack_APIC_irq().

Preparatory change for the real fix

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Reported-by: Song Liu <>
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agox86/ioapic: Use apic_ack_irq()
Thomas Gleixner [Mon, 4 Jun 2018 15:33:57 +0000 (17:33 +0200)]
x86/ioapic: Use apic_ack_irq()

To address the EBUSY fail of interrupt affinity settings in case that the
previous setting has not been cleaned up yet, use the new apic_ack_irq()
function instead of directly invoking ack_APIC_irq().

Preparatory change for the real fix

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agoirq_remapping: Use apic_ack_irq()
Thomas Gleixner [Mon, 4 Jun 2018 15:33:56 +0000 (17:33 +0200)]
irq_remapping: Use apic_ack_irq()

To address the EBUSY fail of interrupt affinity settings in case that the
previous setting has not been cleaned up yet, use the new apic_ack_irq()
function instead of the special ir_ack_apic_edge() implementation which is
merily a wrapper around ack_APIC_irq().

Preparatory change for the real fix

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agox86/apic: Provide apic_ack_irq()
Thomas Gleixner [Mon, 4 Jun 2018 15:33:55 +0000 (17:33 +0200)]
x86/apic: Provide apic_ack_irq()

apic_ack_edge() is explicitely for handling interrupt affinity cleanup when
interrupt remapping is not available or disable.

Remapped interrupts and also some of the platform specific special
interrupts, e.g. UV, invoke ack_APIC_irq() directly.

To address the issue of failing an affinity update with -EBUSY the delayed
affinity mechanism can be reused, but ack_APIC_irq() does not handle
that. Adding this to ack_APIC_irq() is not possible, because that function
is also used for exceptions and directly handled interrupts like IPIs.

Create a new function, which just contains the conditional invocation of
irq_move_irq() and the final ack_APIC_irq().

Reuse the new function in apic_ack_edge().

Preparatory change for the real fix.

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agogenirq/migration: Avoid out of line call if pending is not set
Thomas Gleixner [Wed, 6 Jun 2018 12:46:59 +0000 (14:46 +0200)]
genirq/migration: Avoid out of line call if pending is not set

The upcoming fix for the -EBUSY return from affinity settings requires to
use the irq_move_irq() functionality even on irq remapped interrupts. To
avoid the out of line call, move the check for the pending bit into an
inline helper.

Preparatory change for the real fix. No functional change.

Fixes: dccfe3147b42 ("x86/vector: Simplify vector move cleanup")
Signed-off-by: Thomas Gleixner <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
Cc: Dou Liyang <>
4 years agogenirq/generic_pending: Do not lose pending affinity update
Thomas Gleixner [Mon, 4 Jun 2018 15:33:54 +0000 (17:33 +0200)]
genirq/generic_pending: Do not lose pending affinity update

The generic pending interrupt mechanism moves interrupts from the interrupt
handler on the original target CPU to the new destination CPU. This is
required for x86 and ia64 due to the way the interrupt delivery and
acknowledge works if the interrupts are not remapped.

However that update can fail for various reasons. Some of them are valid
reasons to discard the pending update, but the case, when the previous move
has not been fully cleaned up is not a legit reason to fail.

Check the return value of irq_do_set_affinity() for -EBUSY, which indicates
a pending cleanup, and rearm the pending move in the irq dexcriptor so it's
tried again when the next interrupt arrives.

Fixes: 996c591227d9 ("x86/irq: Plug vector cleanup race")
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Song Liu <>
Cc: Dmitry Safonov <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
Cc: Tariq Toukan <>
4 years agox86/apic/vector: Prevent hlist corruption and leaks
Thomas Gleixner [Mon, 4 Jun 2018 15:33:53 +0000 (17:33 +0200)]
x86/apic/vector: Prevent hlist corruption and leaks

Several people observed the WARN_ON() in irq_matrix_free() which triggers
when the caller tries to free an vector which is not in the allocation
range. Song provided the trace information which allowed to decode the root

The rework of the vector allocation mechanism failed to preserve a sanity
check, which prevents setting a new target vector/CPU when the previous
affinity change has not fully completed.

As a result a half finished affinity change can be overwritten, which can
cause the leak of a irq descriptor pointer on the previous target CPU and
double enqueue of the hlist head into the cleanup lists of two or more
CPUs. After one CPU cleaned up its vector the next CPU will invoke the
cleanup handler with vector 0, which triggers the out of range warning in
the matrix allocator.

Prevent this by checking the apic_data of the interrupt whether the
move_in_progress flag is false and the hlist node is not hashed. Return
-EBUSY if not.

This prevents the damage and restores the behaviour before the vector
allocation rework, but due to other changes in that area it also widens the
chance that user space can observe -EBUSY. In theory this should be fine,
but actually not all user space tools handle -EBUSY correctly. Addressing
that is not part of this fix, but will be addressed in follow up patches.

Fixes: 69cde0004a4b ("x86/vector: Use matrix allocator for vector assignment")
Reported-by: Dmitry Safonov <>
Reported-by: Tariq Toukan <>
Reported-by: Song Liu <>
Signed-off-by: Thomas Gleixner <>
Tested-by: Song Liu <>
Cc: Joerg Roedel <>
Cc: Peter Zijlstra <>
Cc: Mike Travis <>
Cc: Borislav Petkov <>
4 years agox86/vector: Fix the args of vector_alloc tracepoint
Dou Liyang [Fri, 1 Jun 2018 06:50:31 +0000 (14:50 +0800)]
x86/vector: Fix the args of vector_alloc tracepoint

The vector_alloc tracepont reversed the reserved and ret aggs, that made
the trace print wrong. Exchange them.

Fixes: 8d1e3dca7de6 ("x86/vector: Add tracepoints for vector management")
Signed-off-by: Dou Liyang <>
Signed-off-by: Thomas Gleixner <>
4 years agox86/idt: Simplify the idt_setup_apic_and_irq_gates()
Dou Liyang [Wed, 23 May 2018 02:35:55 +0000 (10:35 +0800)]
x86/idt: Simplify the idt_setup_apic_and_irq_gates()

The idt_setup_apic_and_irq_gates() sets the gates from
FIRST_EXTERNAL_VECTOR up to FIRST_SYSTEM_VECTOR first. then secondly, from
FIRST_SYSTEM_VECTOR to NR_VECTORS, it takes both APIC=y and APIC=n into

But for APIC=n, the FIRST_SYSTEM_VECTOR is equal to NR_VECTORS, all
vectors has been set at the first step.

Simplify the second step, make it just work for APIC=y.

Signed-off-by: Dou Liyang <>
Signed-off-by: Thomas Gleixner <>
4 years agox86/platform/uv: Remove extra parentheses
Varsha Rao [Sun, 20 May 2018 08:00:12 +0000 (13:30 +0530)]
x86/platform/uv: Remove extra parentheses

Remove extra parentheses to fix the extraneous parentheses clang

Suggested-by: Lukas Bulwahn <>
Signed-off-by: Varsha Rao <>
Signed-off-by: Thomas Gleixner <>
Cc: Nicholas Mc Guire <>
Cc: "H. Peter Anvin" <>
4 years agox86/mm: Decouple dynamic __PHYSICAL_MASK from AMD SME
Kirill A. Shutemov [Fri, 18 May 2018 11:30:28 +0000 (14:30 +0300)]
x86/mm: Decouple dynamic __PHYSICAL_MASK from AMD SME

AMD SME claims one bit from physical address to indicate whether the
page is encrypted or not. To achieve that we clear out the bit from

The capability to adjust __PHYSICAL_MASK is required beyond AMD SME.
For instance for upcoming Intel Multi-Key Total Memory Encryption.

Factor it out into a separate feature with own Kconfig handle.

It also helps with overhead of AMD SME. It saves more than 3k in .text
on defconfig + AMD_MEM_ENCRYPT:

add/remove: 3/2 grow/shrink: 5/110 up/down: 189/-3753 (-3564)

We would need to return to this once we have infrastructure to patch
constants in code. That's good candidate for it.

Signed-off-by: Kirill A. Shutemov <>
Signed-off-by: Thomas Gleixner <>
Reviewed-by: Tom Lendacky <>
Cc: "H. Peter Anvin" <>
4 years agox86: Mark native_set_p4d() as __always_inline
Arnd Bergmann [Tue, 5 Jun 2018 11:35:15 +0000 (13:35 +0200)]
x86: Mark native_set_p4d() as __always_inline

When CONFIG_OPTIMIZE_INLINING is enabled, the function native_set_p4d()
may not be fully inlined into the caller, resulting in a false-positive
warning about an access to the __pgtable_l5_enabled variable from a
non-__init function, despite the original caller being an __init function:

WARNING: vmlinux.o(.text.unlikely+0x1429): Section mismatch in reference from the function native_set_p4d() to the variable
WARNING: vmlinux.o(.text.unlikely+0x1429): Section mismatch in reference from the function native_p4d_clear() to the variable

The function native_set_p4d() references the variable __initdata
__pgtable_l5_enabled.  This is often because native_set_p4d lacks a
__initdata annotation or the annotation of __pgtable_l5_enabled is wrong.

Marking the native_set_p4d function and its caller native_p4d_clear()
avoids this problem.

I did not bisect the original cause, but I assume this is related to the
recent rework that turned pgtable_l5_enabled() into an inline function,
which in turn caused the compiler to make different inlining decisions.

Fixes: ad3fe525b950 ("x86/mm: Unify pgtable_l5_enabled usage in early boot code")
Signed-off-by: Arnd Bergmann <>
Signed-off-by: Thomas Gleixner <>
Acked-by: Kirill A. Shutemov <>
Cc: Greg Kroah-Hartman <>
Cc: Dave Hansen <>
Cc: "H. Peter Anvin" <>
Cc: Andrew Morton <>
Cc: Zi Yan <>
Cc: Naoya Horiguchi <>
4 years agoMerge branches 'x86/dma', 'x86/microcode', 'x86/mm' and 'x86/vdso' into x86/urgent
Ingo Molnar [Mon, 4 Jun 2018 16:50:32 +0000 (18:50 +0200)]
Merge branches 'x86/dma', 'x86/microcode', 'x86/mm' and 'x86/vdso' into x86/urgent

Merge these small and simple 1-2 commit branches into the urgent branch.

Signed-off-by: Ingo Molnar <>
4 years agoLinux 4.17 v4.17
Linus Torvalds [Sun, 3 Jun 2018 21:15:21 +0000 (14:15 -0700)]
Linux 4.17

4 years agoMerge branch 'fixes' of git://
Linus Torvalds [Sun, 3 Jun 2018 18:01:28 +0000 (11:01 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/viro/vfs

Pull vfs fixes from Al Viro.

 - fix io_destroy()/aio_complete() race

 - the vfs_open() change to get rid of open_check_o_direct() boilerplate
   was nice, but buggy. Al has a patch avoiding a revert, but that's
   definitely not a last-day fodder, so for now revert it is...

* 'fixes' of git://
  Revert "fs: fold open_check_o_direct into do_dentry_open"
  fix io_destroy()/aio_complete() race

4 years agoRevert "fs: fold open_check_o_direct into do_dentry_open"
Al Viro [Sat, 2 Jun 2018 05:31:02 +0000 (01:31 -0400)]
Revert "fs: fold open_check_o_direct into do_dentry_open"

This reverts commit cab64df194667dc5d9d786f0a895f647f5501c0d.

Having vfs_open() in some cases drop the reference to
struct file combined with

error = vfs_open(path, f, cred);
if (error) {
return ERR_PTR(error);
return f;

is flat-out wrong.  It used to be

error = vfs_open(path, f, cred);
if (!error) {
/* from now on we need fput() to dispose of f */
error = open_check_o_direct(f);
if (error) {
f = ERR_PTR(error);
} else {
f = ERR_PTR(error);

and sure, having that open_check_o_direct() boilerplate gotten rid of is
nice, but not that way...

Worse, another call chain (via finish_open()) is FUBAR now wrt
FILE_OPENED handling - in that case we get error returned, with file
already hit by fput() *AND* FILE_OPENED not set.  Guess what happens in
path_openat(), when it hits

if (!(opened & FILE_OPENED)) {

The root cause of all that crap is that the callers of do_dentry_open()
have no way to tell which way did it fail; while that could be fixed up
(by passing something like int *opened to do_dentry_open() and have it
marked if we'd called ->open()), it's probably much too late in the
cycle to do so right now.

Signed-off-by: Al Viro <>
Signed-off-by: Linus Torvalds <>
4 years agoMerge branch 'sched-urgent-for-linus' of git://
Linus Torvalds [Sun, 3 Jun 2018 16:01:41 +0000 (09:01 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Thomas Gleixner:

 - two patches addressing the problem that the scheduler allows under
   certain conditions user space tasks to be scheduled on CPUs which are
   not yet fully booted which causes a few subtle and hard to debug

 - add a missing runqueue clock update in the deadline scheduler which
   triggers a warning under certain circumstances

 - fix a silly typo in the scheduler header file

* 'sched-urgent-for-linus' of git://
  sched/headers: Fix typo
  sched/deadline: Fix missing clock update
  sched/core: Require cpu_active() in select_task_rq(), for user tasks
  sched/core: Fix rules for running on online && !active CPUs

4 years agoMerge branch 'perf-urgent-for-linus' of git://
Linus Torvalds [Sun, 3 Jun 2018 15:58:59 +0000 (08:58 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf tooling fixes from Thomas Gleixner:

 - fix 'perf test Session topology' segfault on s390 (Thomas Richter)

 - fix NULL return handling in bpf__prepare_load() (YueHaibing)

 - fix indexing on Coresight ETM packet queue decoder (Mathieu Poirier)

 - fix format description of NRCPUS header (Arnaldo Carvalho
   de Melo)

 - update documentation section on cpu topology

 - handle uncore event aliases in small groups properly (Kan Liang)

 - add missing perf_sample.addr into python sample dictionary (Leo Yan)

* 'perf-urgent-for-linus' of git://
  perf tools: Fix format description of NRCPUS header
  perf script python: Add addr into perf sample dict
  perf data: Update documentation section on cpu topology
  perf cs-etm: Fix indexing for decoder packet queue
  perf bpf: Fix NULL return handling in bpf__prepare_load()
  perf test: "Session topology" dumps core on s390
  perf parse-events: Handle uncore event aliases in small groups properly

4 years agoMerge git://
Linus Torvalds [Sun, 3 Jun 2018 00:35:53 +0000 (17:35 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Infinite loop in _decode_session6(), from Eric Dumazet.

 2) Pass correct argument to nla_strlcpy() in netfilter, also from Eric

 3) Out of bounds memory access in ipv6 srh code, from Mathieu Xhonneux.

 4) NULL deref in XDP_REDIRECT handling of tun driver, from Toshiaki

 5) Incorrect idr release in cls_flower, from Paul Blakey.

 6) Probe error handling fix in davinci_emac, from Dan Carpenter.

 7) Memory leak in XPS configuration, from Alexander Duyck.

 8) Use after free with cloned sockets in kcm, from Kirill Tkhai.

 9) MTU handling fixes fo ip_tunnel and ip6_tunnel, from Nicolas

10) Fix UAPI hole in bpf data structure for 32-bit compat applications,
    from Daniel Borkmann.

* git:// (33 commits)
  bpf: fix uapi hole for 32 bit compat applications
  net: usb: cdc_mbim: add flag FLAG_SEND_ZLP
  ip6_tunnel: remove magic mtu value 0xFFF8
  ip_tunnel: restore binding to ifaces with a large mtu
  net: dsa: b53: Add BCM5389 support
  kcm: Fix use-after-free caused by clonned sockets
  net-sysfs: Fix memory leak in XPS configuration
  ixgbe: fix parsing of TC actions for HW offload
  net: ethernet: davinci_emac: fix error handling in probe()
  net/ncsi: Fix array size in dumpit handler
  cls_flower: Fix incorrect idr release when failing to modify rule
  net/sonic: Use dma_mapping_error()
  xfrm Fix potential error pointer dereference in xfrm_bundle_create.
  vhost_net: flush batched heads before trying to busy polling
  tun: Fix NULL pointer dereference in XDP redirect
  be2net: Fix error detection logic for BE3
  net: qmi_wwan: Add Netgear Aircard 779S
  mlxsw: spectrum: Forbid creation of VLAN 1 over port/LAG
  atm: zatm: fix memcmp casting
  iwlwifi: pcie: compare with number of IRQs requested for, not number of CPUs

4 years agoMerge tag 'scsi-fixes' of git://
Linus Torvalds [Sat, 2 Jun 2018 22:54:49 +0000 (15:54 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fix from James Bottomley:
 "Eve of merge window fix: The original code was so bogus as to be
  casting the wrong generic device to an rport and proceeding to take
  actions based on the bogus values it found.

  Fortunately it seems the location that is dereferenced always exists,
  so the code hasn't oopsed yet, but it certainly annoys the memory

* tag 'scsi-fixes' of git://
  scsi: scsi_transport_srp: Fix shost to rport translation

4 years agoMerge tag 'drm-fixes-for-v4.17-rc8' of git://
Linus Torvalds [Sat, 2 Jun 2018 22:24:45 +0000 (15:24 -0700)]
Merge tag 'drm-fixes-for-v4.17-rc8' of git://

Pull drm fixes from Dave Airlie:
 "A few final fixes:

   - fix for potential Spectre vector in the new query uAPI
   - fix NULL pointer deref (FDO #106559)
   - DMI fix to hide LVDS for Radiant P845 (FDO #105468)

   - suspend/resume DC regression fix
   - underscan flicker fix on fiji
   - gamma setting fix after dpms

   - fix oops regression

   - fix PSR timing

   - fix oops regression"

* tag 'drm-fixes-for-v4.17-rc8' of git://
  drm/amd/display: Update color props when modeset is required
  drm/amd/display: Make atomic-check validate underscan changes
  drm/bridge/synopsys: dw-hdmi: fix dw_hdmi_setup_rx_sense
  drm/amd/display: Fix BUG_ON during CRTC atomic check update
  drm/i915/query: nospec expects no more than an unsigned long
  drm/i915/query: Protect tainted function pointer lookup
  drm/i915/lvds: Move acpi lid notification registration to registration phase
  drm/i915: Disable LVDS on Radiant P845
  drm/omap: fix NULL deref crash with SDI displays
  drm/psr: Fix missed entry in PSR setup time table.

4 years agoMerge branch 'drm-fixes-4.17' of git:// into drm...
Dave Airlie [Sat, 2 Jun 2018 20:13:57 +0000 (06:13 +1000)]
Merge branch 'drm-fixes-4.17' of git:// into drm-fixes

Two last minute DC fixes for 4.17.  A fix for underscan on fiji and
a fix for gamma settings getting after dpms.

* 'drm-fixes-4.17' of git://
  drm/amd/display: Update color props when modeset is required
  drm/amd/display: Make atomic-check validate underscan changes

4 years agoMerge tag 'mips_fixes_4.17_3' of git://
Linus Torvalds [Sat, 2 Jun 2018 17:12:23 +0000 (10:12 -0700)]
Merge tag 'mips_fixes_4.17_3' of git://git./linux/kernel/git/mips/linux

Pull MIPS fixes from James Hogan:
 "A final few MIPS fixes for 4.17:

   - drop Lantiq gphy reboot/remove reset (4.14)

   - prctl(PR_SET_FP_MODE): Disallow PRE without FR (4.0)

   - ptrace(PTRACE_PEEKUSR): Fix 64-bit FGRs (3.15)"

* tag 'mips_fixes_4.17_3' of git://
  MIPS: ptrace: Fix PTRACE_PEEKUSR requests for 64-bit FGRs
  MIPS: prctl: Disallow FRE without FR with PR_SET_FP_MODE requests
  MIPS: lantiq: gphy: Drop reboot/remove reset asserts

4 years agoMerge tag 'vfio-v4.17' of git://
Linus Torvalds [Sat, 2 Jun 2018 17:08:45 +0000 (10:08 -0700)]
Merge tag 'vfio-v4.17' of git://

Pull VFIO fix from Alex Williamson:
 "Revert a pfn page mapping optimization identified as introducing a bad
  page state regression (Alex Williamson)"

* tag 'vfio-v4.17' of git://
  Revert "vfio/type1: Improve memory pinning process for raw PFN mapping"

4 years agoMerge tag 'char-misc-4.17-rc8' of git://
Linus Torvalds [Sat, 2 Jun 2018 17:05:45 +0000 (10:05 -0700)]
Merge tag 'char-misc-4.17-rc8' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes from Greg KH:
 "Here are four small bugfixes for some char/misc drivers. Well, really
  three fixes and one fix for one of those fixes due to problems found
  by 0-day.

  This resolves some reported issues with the hwtracing drivers, and a
  reported regression for the thunderbolt subsystem. All of these have
  been in linux-next for a while now with no reported problems"

* tag 'char-misc-4.17-rc8' of git://
  hwtracing: stm: fix build error on some arches
  intel_th: Use correct device when freeing buffers
  stm class: Use vmalloc for the master map
  thunderbolt: Handle NULL boot ACL entries properly

4 years agoMerge tag 'staging-4.17-rc8' of git://
Linus Torvalds [Sat, 2 Jun 2018 17:02:14 +0000 (10:02 -0700)]
Merge tag 'staging-4.17-rc8' of git://git./linux/kernel/git/gregkh/staging

Pull IIO driver fixes from Greg KH:
 "Here are some old IIO driver fixes that were sitting in my tree for a
  few weeks. Sorry about not getting them to you sooner. They fix a
  number of small IIO driver issues that have been reported.

  All of these have been in linux-next for a while with no reported

* tag 'staging-4.17-rc8' of git://
  iio: adc: select buffer for at91-sama5d2_adc
  iio: hid-sensor-trigger: Fix sometimes not powering up the sensor after resume
  iio: adc: at91-sama5d2_adc: fix channel configuration for differential channels
  iio:kfifo_buf: check for uint overflow
  iio:buffer: make length types match kfifo types
  iio: adc: stm32-dfsdm: fix sample rate for div2 spi clock
  iio: adc: stm32-dfsdm: fix successive oversampling settings
  iio: ad7793: implement IIO_CHAN_INFO_SAMP_FREQ

4 years agoMerge tag 'for-linus' of git://
Linus Torvalds [Sat, 2 Jun 2018 16:55:44 +0000 (09:55 -0700)]
Merge tag 'for-linus' of git://git./linux/kernel/git/rdma/rdma

Pull rdma fixes from Jason Gunthorpe:
 "Just three small last minute regressions that were found in the last
  week. The Broadcom fix is a bit big for rc7, but since it is fixing
  driver crash regressions that were merged via netdev into rc1, I am
  sending it.

   - bnxt netdev changes merged this cycle caused the bnxt RDMA driver
     to crash under certain situations

   - Arnd found (several, unfortunately) kconfig problems with the
     patches adding INFINIBAND_ADDR_TRANS. Reverting this last part,
     will fix it more fully outside -rc.

   - Subtle change in error code for a uapi function caused breakage in
     userspace. This was bug was subtly introduced cycle"

* tag 'for-linus' of git://
  IB/core: Fix error code for invalid GID entry
  IB: Revert "remove redundant INFINIBAND kconfig dependencies"
  RDMA/bnxt_re: Fix broken RoCE driver due to recent L2 driver changes

4 years agoMerge branch 'i2c/for-current' of git://
Linus Torvalds [Sat, 2 Jun 2018 16:52:22 +0000 (09:52 -0700)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "A documentation bugfix and a MAINTAINERS addition"

* 'i2c/for-current' of git://
  i2c: ocores: update HDL sources URL
  i2c: xlp9xx: Add MAINTAINERS entry

4 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Sat, 2 Jun 2018 16:44:15 +0000 (09:44 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge two fixes from Andrew Morton.

* emailed patches from Andrew Morton <>:
  mm: fix the NULL mapping case in __isolate_lru_page()
  mm/huge_memory.c: __split_huge_page() use atomic ClearPageDirty()

4 years agomm: fix the NULL mapping case in __isolate_lru_page()
Hugh Dickins [Fri, 1 Jun 2018 23:50:50 +0000 (16:50 -0700)]
mm: fix the NULL mapping case in __isolate_lru_page()

George Boole would have noticed a slight error in 4.16 commit
69d763fc6d3a ("mm: pin address_space before dereferencing it while
isolating an LRU page").  Fix it, to match both the comment above it,
and the original behaviour.

Although anonymous pages are not marked PageDirty at first, we have an
old habit of calling SetPageDirty when a page is removed from swap
cache: so there's a category of ex-swap pages that are easily
migratable, but were inadvertently excluded from compaction's async
migration in 4.16.

Fixes: 69d763fc6d3a ("mm: pin address_space before dereferencing it while isolating an LRU page")
Signed-off-by: Hugh Dickins <>
Acked-by: Minchan Kim <>
Acked-by: Mel Gorman <>
Reported-by: Ivan Kalvachev <>
Cc: "Huang, Ying" <>
Cc: Jan Kara <>
Signed-off-by: Andrew Morton <>
Signed-off-by: Linus Torvalds <>
4 years agomm/huge_memory.c: __split_huge_page() use atomic ClearPageDirty()
Hugh Dickins [Fri, 1 Jun 2018 23:50:45 +0000 (16:50 -0700)]
mm/huge_memory.c: __split_huge_page() use atomic ClearPageDirty()

Swapping load on huge=always tmpfs (with khugepaged tuned up to be very
eager, but I'm not sure that is relevant) soon hung uninterruptibly,
waiting for page lock in shmem_getpage_gfp()'s find_lock_entry(), most
often when "cp -a" was trying to write to a smallish file.  Debug showed
that the page in question was not locked, and page->mapping NULL by now,
but page->index consistent with having been in a huge page before.

Reproduced in minutes on a 4.15 kernel, even with 4.17's 605ca5ede764
("mm/huge_memory.c: reorder operations in __split_huge_page_tail()") added
in; but took hours to reproduce on a 4.17 kernel (no idea why).

The culprit proved to be the __ClearPageDirty() on tails beyond i_size in
__split_huge_page(): the non-atomic __bitoperation may have been safe when
4.8's baa355fd3314 ("thp: file pages support for split_huge_page()")
introduced it, but liable to erase PageWaiters after 4.10's 62906027091f
("mm: add PageWaiters indicating tasks are waiting for a page bit").

Fixes: 62906027091f ("mm: add PageWaiters indicating tasks are waiting for a page bit")
Signed-off-by: Hugh Dickins <>
Acked-by: Kirill A. Shutemov <>
Cc: Konstantin Khlebnikov <>
Cc: Nicholas Piggin <>
Cc: <>
Signed-off-by: Andrew Morton <>
Signed-off-by: Linus Torvalds <>
4 years agoRevert "vfio/type1: Improve memory pinning process for raw PFN mapping"
Alex Williamson [Sat, 2 Jun 2018 14:41:44 +0000 (08:41 -0600)]
Revert "vfio/type1: Improve memory pinning process for raw PFN mapping"

Bisection by Amadeusz Sławiński implicates this commit leading to bad
page state issues after VM shutdown, likely due to unbalanced page
references.  The original commit was intended only as a performance
improvement, therefore revert for offline rework.

Fixes: 356e88ebe447 ("vfio/type1: Improve memory pinning process for raw PFN mapping")
Cc: Jason Cai (Xiang Feng) <>
Reported-by: Amadeusz Sławiński <>
Signed-off-by: Alex Williamson <>
4 years agoMerge git://
David S. Miller [Sat, 2 Jun 2018 12:07:52 +0000 (08:07 -0400)]
Merge git://git./pub/scm/linux/kernel/git/bpf/bpf

Daniel Borkmann says:

pull-request: bpf 2018-06-02

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

The main changes are:

1) BPF uapi fix in struct bpf_prog_info and struct bpf_map_info in
   order to fix offsets on 32 bit archs.

This will have a minor merge conflict with net-next which has the
__u32 gpl_compatible:1 bitfield in struct bpf_prog_info at this
location. Resolution is to use the gpl_compatible member.

Signed-off-by: David S. Miller <>
4 years agobpf: fix uapi hole for 32 bit compat applications
Daniel Borkmann [Sat, 2 Jun 2018 03:21:59 +0000 (05:21 +0200)]
bpf: fix uapi hole for 32 bit compat applications

In 64 bit, we have a 4 byte hole between ifindex and netns_dev in the
case of struct bpf_map_info but also struct bpf_prog_info. In net-next
commit b85fab0e67b ("bpf: Add gpl_compatible flag to struct bpf_prog_info")
added a bitfield into it to expose some flags related to programs. Thus,
add an unnamed __u32 bitfield for both so that alignment keeps the same
in both 32 and 64 bit cases, and can be naturally extended from there
as in b85fab0e67b.


  # file test.o
  test.o: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), not stripped
  # pahole test.o
  struct bpf_map_info {
__u32                      type;                 /*     0     4 */
__u32                      id;                   /*     4     4 */
__u32                      key_size;             /*     8     4 */
__u32                      value_size;           /*    12     4 */
__u32                      max_entries;          /*    16     4 */
__u32                      map_flags;            /*    20     4 */
char                       name[16];             /*    24    16 */
__u32                      ifindex;              /*    40     4 */
__u64                      netns_dev;            /*    44     8 */
__u64                      netns_ino;            /*    52     8 */

/* size: 64, cachelines: 1, members: 10 */
/* padding: 4 */

After (same as on 64 bit):

  # file test.o
  test.o: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), not stripped
  # pahole test.o
  struct bpf_map_info {
__u32                      type;                 /*     0     4 */
__u32                      id;                   /*     4     4 */
__u32                      key_size;             /*     8     4 */
__u32                      value_size;           /*    12     4 */
__u32                      max_entries;          /*    16     4 */
__u32                      map_flags;            /*    20     4 */
char                       name[16];             /*    24    16 */
__u32                      ifindex;              /*    40     4 */

/* XXX 4 bytes hole, try to pack */

__u64                      netns_dev;            /*    48     8 */
__u64                      netns_ino;            /*    56     8 */
/* --- cacheline 1 boundary (64 bytes) --- */

/* size: 64, cachelines: 1, members: 10 */
/* sum members: 60, holes: 1, sum holes: 4 */

Reported-by: Dmitry V. Levin <>
Reported-by: Eugene Syromiatnikov <>
Fixes: 52775b33bb507 ("bpf: offload: report device information about offloaded maps")
Fixes: 675fc275a3a2d ("bpf: offload: report device information for offloaded programs")
Signed-off-by: Daniel Borkmann <>
Acked-by: Alexei Starovoitov <>
Signed-off-by: Alexei Starovoitov <>
4 years agonet: usb: cdc_mbim: add flag FLAG_SEND_ZLP
Daniele Palmas [Thu, 31 May 2018 09:18:29 +0000 (11:18 +0200)]
net: usb: cdc_mbim: add flag FLAG_SEND_ZLP

Testing Telit LM940 with ICMP packets > 14552 bytes revealed that
the modem needs FLAG_SEND_ZLP to properly work, otherwise the cdc
mbim data interface won't be anymore responsive.

Signed-off-by: Daniele Palmas <>
Acked-by: Bjørn Mork <>
Signed-off-by: David S. Miller <>
4 years agoMerge branch 'tunnel-mtus'
David S. Miller [Fri, 1 Jun 2018 17:56:31 +0000 (13:56 -0400)]
Merge branch 'tunnel-mtus'

Nicolas Dichtel says:

ip[6] tunnels: fix mtu calculations

The first patch restores the possibility to bind an ip4 tunnel to an
interface whith a large mtu.
The second patch was spotted after the first fix. I also target it to net
because it fixes the max mtu value that can be used for ipv6 tunnels.

v2: remove the 0xfff8 in ip_tunnel_newlink()

Signed-off-by: David S. Miller <>
4 years agoip6_tunnel: remove magic mtu value 0xFFF8
Nicolas Dichtel [Thu, 31 May 2018 08:59:33 +0000 (10:59 +0200)]
ip6_tunnel: remove magic mtu value 0xFFF8

I don't know where this value comes from (probably a copy and paste and
paste and paste ...).
Let's use standard values which are a bit greater.

Signed-off-by: Nicolas Dichtel <>
Signed-off-by: David S. Miller <>
4 years agoip_tunnel: restore binding to ifaces with a large mtu
Nicolas Dichtel [Thu, 31 May 2018 08:59:32 +0000 (10:59 +0200)]
ip_tunnel: restore binding to ifaces with a large mtu

After commit f6cc9c054e77, the following conf is broken (note that the
default loopback mtu is 65536, ie IP_MAX_MTU + 1):

$ ip tunnel add gre1 mode gre local remote dev lo
add tunnel "gre0" failed: Invalid argument
$ ip l a type dummy
$ ip l s dummy1 up
$ ip l s dummy1 mtu 65535
$ ip tunnel add gre1 mode gre local remote dev dummy1
add tunnel "gre0" failed: Invalid argument

dev_set_mtu() doesn't allow to set a mtu which is too large.
First, let's cap the mtu returned by ip_tunnel_bind_dev(). Second, remove
the magic value 0xFFF8 and use IP_MAX_MTU instead.
0xFFF8 seems to be there for ages, I don't know why this value was used.

With a recent kernel, it's also possible to set a mtu > IP_MAX_MTU:
$ ip l s dummy1 mtu 66000
After that patch, it's also possible to bind an ip tunnel on that kind of

CC: Petr Machata <>
CC: Ido Schimmel <>
Fixes: f6cc9c054e77 ("ip_tunnel: Emit events for post-register MTU changes")
Signed-off-by: Nicolas Dichtel <>
Reviewed-by: Ido Schimmel <>
Signed-off-by: David S. Miller <>
4 years agoMerge branch 'master' of git://
David S. Miller [Fri, 1 Jun 2018 17:25:41 +0000 (13:25 -0400)]
Merge branch 'master' of git://git./linux/kernel/git/klassert/ipsec

Steffen Klassert says:

pull request (net): ipsec 2018-05-31

1) Avoid possible overflow of the offset variable
   in  _decode_session6(), this fixes an infinite
   lookp there. From Eric Dumazet.

2) We may use an error pointer in the error path of
   xfrm_bundle_create(). Fix this by returning this
   pointer directly to the caller.

Please pull or let me know if there are problems.

Signed-off-by: David S. Miller <>
4 years agonet: dsa: b53: Add BCM5389 support
Damien Thébault [Thu, 31 May 2018 07:04:01 +0000 (07:04 +0000)]
net: dsa: b53: Add BCM5389 support

This patch adds support for the BCM5389 switch connected through MDIO.

Signed-off-by: Damien Thébault <>
Reviewed-by: Florian Fainelli <>
Signed-off-by: David S. Miller <>
4 years agokcm: Fix use-after-free caused by clonned sockets
Kirill Tkhai [Fri, 1 Jun 2018 11:30:38 +0000 (14:30 +0300)]
kcm: Fix use-after-free caused by clonned sockets

(resend for properly queueing in patchwork)

kcm_clone() creates kernel socket, which does not take net counter.
Thus, the net may die before the socket is completely destructed,
i.e. kcm_exit_net() is executed before kcm_done().

Signed-off-by: Kirill Tkhai <>
Signed-off-by: David S. Miller <>
4 years agonet-sysfs: Fix memory leak in XPS configuration
Alexander Duyck [Thu, 31 May 2018 19:59:46 +0000 (15:59 -0400)]
net-sysfs: Fix memory leak in XPS configuration

This patch reorders the error cases in showing the XPS configuration so
that we hold off on memory allocation until after we have verified that we
can support XPS on a given ring.

Fixes: 184c449f91fe ("net: Add support for XPS with QoS via traffic classes")
Signed-off-by: Alexander Duyck <>
Signed-off-by: David S. Miller <>
4 years agoixgbe: fix parsing of TC actions for HW offload
Ondřej Hlavatý [Thu, 31 May 2018 21:21:04 +0000 (23:21 +0200)]
ixgbe: fix parsing of TC actions for HW offload

The previous code was optimistic, accepting the offload of whole action
chain when there was a single known action (drop/redirect). This results
in offloading a rule which should not be offloaded, because its behavior
cannot be reproduced in the hardware.

For example:

$ tc filter add dev eno1 parent ffff: protocol ip \
    u32 ht 800: order 1 match tcp src 42 FFFF \
    action mirred egress mirror dev enp1s16 pipe \

The controller is unable to mirror the packet to a VF, but still
offloads the rule by dropping the packet.

Change the approach of the function to a pessimistic one, rejecting the
chain when an unknown action is found. This is better suited for future

Note that both recognized actions always return TC_ACT_SHOT, therefore
it is safe to ignore actions behind them.

Signed-off-by: Ondřej Hlavatý <>
Signed-off-by: Jeff Kirsher <>
Signed-off-by: David S. Miller <>
4 years agoMerge tag 'xfs-4.17-fixes-3' of git://
Linus Torvalds [Thu, 31 May 2018 21:23:07 +0000 (16:23 -0500)]
Merge tag 'xfs-4.17-fixes-3' of git://git./fs/xfs/xfs-linux

Pull xfs fix from Darrick Wong:
 "Clear out i_mapping error state when we're reinitializing inodes.

  This last minute fix prevents writeback error state from persisting
  past the end of the in-core inode lifecycle and causing EIO errors to
  be reported to userspace when no error has occurred.

  This fix for the behavioral regression has been soaking in for-next
  for a while, but various fs developers persuaded me to try to get it
  upstream for 4.17 because the patch that broke things was introduced
  in 4.17-rc4"

* tag 'xfs-4.17-fixes-3' of git://
  fs: clear writeback errors in inode_init_always

4 years agonet: ethernet: davinci_emac: fix error handling in probe()
Dan Carpenter [Thu, 31 May 2018 06:44:49 +0000 (09:44 +0300)]
net: ethernet: davinci_emac: fix error handling in probe()

The current error handling code has an issue where it does:

if (priv->txchan)

The problem is that ->txchan is either valid or an error pointer (which
would lead to an Oops).  I've changed it to use multiple error labels so
that the test can be removed.

Also there were some missing calls to netif_napi_del().

Fixes: 3ef0fdb2342c ("net: davinci_emac: switch to new cpdma layer")
Signed-off-by: Dan Carpenter <>
Signed-off-by: David S. Miller <>
4 years agonet/ncsi: Fix array size in dumpit handler
Samuel Mendoza-Jonas [Thu, 31 May 2018 04:10:04 +0000 (14:10 +1000)]
net/ncsi: Fix array size in dumpit handler

With CONFIG_CC_STACKPROTECTOR enabled the kernel panics as below when
parsing a NCSI_CMD_PKG_INFO command:

[  150.149711] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: 805cff08
[  150.149711]
[  150.159919] CPU: 0 PID: 1301 Comm: ncsi-netlink Not tainted 4.13.16-468cbec6d2c91239332cb91b1f0a73aafcb6f0c6 #1
[  150.170004] Hardware name: Generic DT based system
[  150.174852] [<80109930>] (unwind_backtrace) from [<80106bc4>] (show_stack+0x20/0x24)
[  150.182641] [<80106bc4>] (show_stack) from [<805d36e4>] (dump_stack+0x20/0x28)
[  150.189888] [<805d36e4>] (dump_stack) from [<801163ac>] (panic+0xdc/0x278)
[  150.196780] [<801163ac>] (panic) from [<801162cc>] (__stack_chk_fail+0x20/0x24)
[  150.204111] [<801162cc>] (__stack_chk_fail) from [<805cff08>] (ncsi_pkg_info_all_nl+0x244/0x258)
[  150.212912] [<805cff08>] (ncsi_pkg_info_all_nl) from [<804f939c>] (genl_lock_dumpit+0x3c/0x54)
[  150.221535] [<804f939c>] (genl_lock_dumpit) from [<804f873c>] (netlink_dump+0xf8/0x284)
[  150.229550] [<804f873c>] (netlink_dump) from [<804f8d44>] (__netlink_dump_start+0x124/0x17c)
[  150.237992] [<804f8d44>] (__netlink_dump_start) from [<804f9880>] (genl_rcv_msg+0x1c8/0x3d4)
[  150.246440] [<804f9880>] (genl_rcv_msg) from [<804f9174>] (netlink_rcv_skb+0xd8/0x134)
[  150.254361] [<804f9174>] (netlink_rcv_skb) from [<804f96a4>] (genl_rcv+0x30/0x44)
[  150.261850] [<804f96a4>] (genl_rcv) from [<804f7790>] (netlink_unicast+0x198/0x234)
[  150.269511] [<804f7790>] (netlink_unicast) from [<804f7ffc>] (netlink_sendmsg+0x368/0x3b0)
[  150.277783] [<804f7ffc>] (netlink_sendmsg) from [<804abea4>] (sock_sendmsg+0x24/0x34)
[  150.285625] [<804abea4>] (sock_sendmsg) from [<804ac1dc>] (___sys_sendmsg+0x244/0x260)
[  150.293556] [<804ac1dc>] (___sys_sendmsg) from [<804ad98c>] (__sys_sendmsg+0x5c/0x9c)
[  150.301400] [<804ad98c>] (__sys_sendmsg) from [<804ad9e4>] (SyS_sendmsg+0x18/0x1c)
[  150.308984] [<804ad9e4>] (SyS_sendmsg) from [<80102640>] (ret_fast_syscall+0x0/0x3c)
[  150.316743] ---[ end Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: 805cff08

This turns out to be because the attrs array in ncsi_pkg_info_all_nl()
is initialised to a length of NCSI_ATTR_MAX which is the maximum
attribute number, not the number of attributes.

Fixes: 955dc68cb9b2 ("net/ncsi: Add generic netlink family")
Signed-off-by: Samuel Mendoza-Jonas <>
Signed-off-by: David S. Miller <>
4 years agoMerge tag 'wireless-drivers-for-davem-2018-05-30' of git://
David S. Miller [Thu, 31 May 2018 19:27:39 +0000 (15:27 -0400)]
Merge tag 'wireless-drivers-for-davem-2018-05-30' of git://git./linux/kernel/git/kvalo/wireless-drivers

Kalle Valo says:

wireless-drivers fixes for 4.17

Two last minute fixes, hopefully they make it to 4.17 still.


* revert a fix which caused even more problems


* fix a crash when there are 16 or more logical CPUs

Signed-off-by: David S. Miller <>
4 years agocls_flower: Fix incorrect idr release when failing to modify rule
Paul Blakey [Wed, 30 May 2018 08:29:15 +0000 (11:29 +0300)]
cls_flower: Fix incorrect idr release when failing to modify rule

When we fail to modify a rule, we incorrectly release the idr handle
of the unmodified old rule.

Fix that by checking if we need to release it.

Fixes: fe2502e49b58 ("net_sched: remove cls_flower idr on failure")
Reported-by: Vlad Buslov <>
Reviewed-by: Roi Dayan <>
Acked-by: Jiri Pirko <>
Signed-off-by: Paul Blakey <>
Signed-off-by: David S. Miller <>
4 years agonet/sonic: Use dma_mapping_error()
Finn Thain [Wed, 30 May 2018 03:03:51 +0000 (13:03 +1000)]
net/sonic: Use dma_mapping_error()

With CONFIG_DMA_API_DEBUG=y, calling sonic_open() produces the
message, "DMA-API: device driver failed to check map error".
Add the missing dma_mapping_error() call.

Cc: Thomas Bogendoerfer <>
Signed-off-by: Finn Thain <>
Acked-by: Thomas Bogendoerfer <>
Signed-off-by: David S. Miller <>
4 years agodrm/amd/display: Update color props when modeset is required
Leo (Sunpeng) Li [Thu, 31 May 2018 14:23:37 +0000 (10:23 -0400)]
drm/amd/display: Update color props when modeset is required

This fixes issues where color management properties don't persist
over DPMS on/off, or when the CRTC is moved across connectors.

Signed-off-by: Leo (Sunpeng) Li <>
Reviewed-by: Harry Wentland <>
Signed-off-by: Alex Deucher <>
4 years agodrm/amd/display: Make atomic-check validate underscan changes
David Francis [Thu, 31 May 2018 17:48:31 +0000 (13:48 -0400)]
drm/amd/display: Make atomic-check validate underscan changes

When the underscan state was changed, atomic-check was triggering a
validation but passing the old underscan values.  This change adds a
somewhat hacky check in dm_update_crtcs_state that will update the
stream if old and newunderscan values are different.
This was causing 4k on Fiji to allow underscan when it wasn't permitted.

Signed-off-by: David Francis <>
Reviewed-by: David Francis <>
Acked-by: Harry Wentland <>
Signed-off-by: Alex Deucher <>
4 years agoMerge tag 'platform-drivers-x86-v4.17-4' of git://
Linus Torvalds [Thu, 31 May 2018 14:39:57 +0000 (09:39 -0500)]
Merge tag 'platform-drivers-x86-v4.17-4' of git://

Pull x86 platform driver fix from Andy Shevchenko:
 "Fix NULL pointer dereference in asus-wmi on rfkill cleanup.

  The effective change is just one new condition - two lines of code.
  But it required moving one static helper function, which is why the
  diff looks a bit bigger"

* tag 'platform-drivers-x86-v4.17-4' of git://
  platform/x86: asus-wmi: Fix NULL pointer dereference

4 years agoplatform/x86: asus-wmi: Fix NULL pointer dereference
João Paulo Rechi Vita [Tue, 22 May 2018 21:30:15 +0000 (14:30 -0700)]
platform/x86: asus-wmi: Fix NULL pointer dereference

Do not perform the rfkill cleanup routine when
(asus->driver->wlan_ctrl_by_user && ashs_present()) is true, since
nothing is registered with the rfkill subsystem in that case. Doing so
leads to the following kernel NULL pointer dereference:

  BUG: unable to handle kernel NULL pointer dereference at           (null)
  IP: [<ffffffff816c7348>] __mutex_lock_slowpath+0x98/0x120
  PGD 1a3aa8067
  PUD 1a3b3d067
  PMD 0

  Oops: 0002 [#1] PREEMPT SMP
  Modules linked in: bnep ccm binfmt_misc uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_core hid_a4tech videodev x86_pkg_temp_thermal intel_powerclamp coretemp ath3k btusb btrtl btintel bluetooth kvm_intel snd_hda_codec_hdmi kvm snd_hda_codec_realtek snd_hda_codec_generic irqbypass crc32c_intel arc4 i915 snd_hda_intel snd_hda_codec ath9k ath9k_common ath9k_hw ath i2c_algo_bit snd_hwdep mac80211 ghash_clmulni_intel snd_hda_core snd_pcm snd_timer cfg80211 ehci_pci xhci_pci drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm xhci_hcd ehci_hcd asus_nb_wmi(-) asus_wmi sparse_keymap r8169 rfkill mxm_wmi serio_raw snd mii mei_me lpc_ich i2c_i801 video soundcore mei i2c_smbus wmi i2c_core mfd_core
  CPU: 3 PID: 3275 Comm: modprobe Not tainted 4.9.34-gentoo #34
  Hardware name: ASUSTeK COMPUTER INC. K56CM/K56CM, BIOS K56CM.206 08/21/2012
  task: ffff8801a639ba00 task.stack: ffffc900014cc000
  RIP: 0010:[<ffffffff816c7348>]  [<ffffffff816c7348>] __mutex_lock_slowpath+0x98/0x120
  RSP: 0018:ffffc900014cfce0  EFLAGS: 00010282
  RAX: 0000000000000000 RBX: ffff8801a54315b0 RCX: 00000000c0000100
  RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff8801a54315b4
  RBP: ffffc900014cfd30 R08: 0000000000000000 R09: 0000000000000002
  R10: 0000000000000000 R11: 0000000000000000 R12: ffff8801a54315b4
  R13: ffff8801a639ba00 R14: 00000000ffffffff R15: ffff8801a54315b8
  FS:  00007faa254fb700(0000) GS:ffff8801aef80000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 0000000000000000 CR3: 00000001a3b1b000 CR4: 00000000001406e0
   ffff8801a54315b8 0000000000000000 ffffffff814733ae ffffc900014cfd28
   ffffffff8146a28c ffff8801a54315b0 0000000000000000 ffff8801a54315b0
   ffff8801a66f3820 0000000000000000 ffffc900014cfd48 ffffffff816c73e7
  Call Trace:
   [<ffffffff814733ae>] ? acpi_ut_release_mutex+0x5d/0x61
   [<ffffffff8146a28c>] ? acpi_ns_get_node+0x49/0x52
   [<ffffffff816c73e7>] mutex_lock+0x17/0x30
   [<ffffffffa00a3bb4>] asus_rfkill_hotplug+0x24/0x1a0 [asus_wmi]
   [<ffffffffa00a4421>] asus_wmi_rfkill_exit+0x61/0x150 [asus_wmi]
   [<ffffffffa00a49f1>] asus_wmi_remove+0x61/0xb0 [asus_wmi]
   [<ffffffff814a5128>] platform_drv_remove+0x28/0x40
   [<ffffffff814a2901>] __device_release_driver+0xa1/0x160
   [<ffffffff814a29e3>] device_release_driver+0x23/0x30
   [<ffffffff814a1ffd>] bus_remove_device+0xfd/0x170
   [<ffffffff8149e5a9>] device_del+0x139/0x270
   [<ffffffff814a5028>] platform_device_del+0x28/0x90
   [<ffffffff814a50a2>] platform_device_unregister+0x12/0x30
   [<ffffffffa00a4209>] asus_wmi_unregister_driver+0x19/0x30 [asus_wmi]
   [<ffffffffa00da0ea>] asus_nb_wmi_exit+0x10/0xf26 [asus_nb_wmi]
   [<ffffffff8110c692>] SyS_delete_module+0x192/0x270
   [<ffffffff810022b2>] ? exit_to_usermode_loop+0x92/0xa0
   [<ffffffff816ca560>] entry_SYSCALL_64_fastpath+0x13/0x94
  Code: e8 5e 30 00 00 8b 03 83 f8 01 0f 84 93 00 00 00 48 8b 43 10 4c 8d 7b 08 48 89 63 10 41 be ff ff ff ff 4c 89 3c 24 48 89 44 24 08 <48> 89 20 4c 89 6c 24 10 eb 1d 4c 89 e7 49 c7 45 08 02 00 00 00
  RIP  [<ffffffff816c7348>] __mutex_lock_slowpath+0x98/0x120
   RSP <ffffc900014cfce0>
  CR2: 0000000000000000
  ---[ end trace 8d484233fa7cb512 ]---
  note: modprobe[3275] exited with preempt_count 2

Signed-off-by: João Paulo Rechi Vita <>
Signed-off-by: Andy Shevchenko <>
4 years agoMerge tag 'perf-urgent-for-mingo-4.17-20180531' of git://
Ingo Molnar [Thu, 31 May 2018 10:37:07 +0000 (12:37 +0200)]
Merge tag 'perf-urgent-for-mingo-4.17-20180531' of git://git./linux/kernel/git/acme/linux into perf/urgent

Pull perf/urgent fixes from Arnaldo Carvalho de Melo:

- Fix 'perf test Session topology' segfault on s390 (Thomas Richter)

- Fix NULL return handling in bpf__prepare_load() (YueHaibing)

- Fix indexing on Coresight ETM packet queue decoder (Mathieu Poirier)

- Fix format description of NRCPUS header (Arnaldo Carvalho de Melo)

- Update documentation section on cpu topology

- Handle uncore event aliases in small groups properly (Kan Liang)

- Add missing perf_sample.addr into python sample dictionary (Leo Yan)

Signed-off-by: Arnaldo Carvalho de Melo <>
Signed-off-by: Ingo Molnar <>
4 years agosched/headers: Fix typo
Davidlohr Bueso [Wed, 30 May 2018 22:49:40 +0000 (15:49 -0700)]
sched/headers: Fix typo

I cannot spell 'throttling'.

Signed-off-by: Davidlohr Bueso <>
Signed-off-by: Peter Zijlstra (Intel) <>
Cc: Davidlohr Bueso <>
Cc: Linus Torvalds <>
Cc: Peter Zijlstra <>
Cc: Thomas Gleixner <>
Signed-off-by: Ingo Molnar <>
4 years agosched/deadline: Fix missing clock update
Juri Lelli [Wed, 30 May 2018 16:08:09 +0000 (18:08 +0200)]
sched/deadline: Fix missing clock update

A missing clock update is causing the following warning:

 rq->clock_update_flags < RQCF_ACT_SKIP
 WARNING: CPU: 10 PID: 0 at kernel/sched/sched.h:963 inactive_task_timer+0x5d6/0x720
 Call Trace:
 hardirqs last  enabled at (793919): [<ffffffffa27c5f6e>] cpuidle_enter_state+0x9e/0x360
 hardirqs last disabled at (793920): [<ffffffffa2a0096e>] interrupt_entry+0xce/0xe0
 softirqs last  enabled at (793922): [<ffffffffa20bef78>] irq_enter+0x68/0x70
 softirqs last disabled at (793921): [<ffffffffa20bef5d>] irq_enter+0x4d/0x70

This happens because inactive_task_timer() calls sub_running_bw() (if
TASK_DEAD and non_contending) that might trigger a schedutil update,
which might access the clock. Clock is however currently updated only
later in inactive_task_timer() function.

Fix the problem by updating the clock right after task_rq_lock().

Reported-by: kernel test robot <>
Signed-off-by: Juri Lelli <>
Signed-off-by: Peter Zijlstra (Intel) <>
Cc: Claudio Scordino <>
Cc: Linus Torvalds <>
Cc: Luca Abeni <>
Cc: Peter Zijlstra <>
Cc: Thomas Gleixner <>
Signed-off-by: Ingo Molnar <>
4 years agosched/core: Require cpu_active() in select_task_rq(), for user tasks
Paul Burton [Sat, 26 May 2018 15:46:47 +0000 (08:46 -0700)]
sched/core: Require cpu_active() in select_task_rq(), for user tasks

select_task_rq() is used in a few paths to select the CPU upon which a
thread should be run - for example it is used by try_to_wake_up() & by
fork or exec balancing. As-is it allows use of any online CPU that is
present in the task's cpus_allowed mask.

This presents a problem because there is a period whilst CPUs are
brought online where a CPU is marked online, but is not yet fully
initialized - ie. the period where CPUHP_AP_ONLINE_IDLE <= state <
CPUHP_ONLINE. Usually we don't run any user tasks during this window,
but there are corner cases where this can happen. An example observed

  - Some user task A, running on CPU X, forks to create task B.

  - sched_fork() calls __set_task_cpu() with cpu=X, setting task B's
    task_struct::cpu field to X.

  - CPU X is offlined.

  - Task A, currently somewhere between the __set_task_cpu() in
    copy_process() and the call to wake_up_new_task(), is migrated to
    CPU Y by migrate_tasks() when CPU X is offlined.

  - CPU X is onlined, but still in the CPUHP_AP_ONLINE_IDLE state. The
    scheduler is now active on CPU X, but there are no user tasks on
    the runqueue.

  - Task A runs on CPU Y & reaches wake_up_new_task(). This calls
    select_task_rq() with cpu=X, taken from task B's task_struct,
    and select_task_rq() allows CPU X to be returned.

  - Task A enqueues task B on CPU X's runqueue, via activate_task() &

  - CPU X now has a user task on its runqueue before it has reached the
    CPUHP_ONLINE state.

In most cases, the user tasks that schedule on the newly onlined CPU
have no idea that anything went wrong, but one case observed to be
problematic is if the task goes on to invoke the sched_setaffinity
syscall. The newly onlined CPU reaches the CPUHP_AP_ONLINE_IDLE state
before the CPU that brought it online calls stop_machine_unpark(). This
means that for a portion of the window of time between
CPUHP_AP_ONLINE_IDLE & CPUHP_ONLINE the newly onlined CPU's struct
cpu_stopper has its enabled field set to false. If a user thread is
executed on the CPU during this window and it invokes sched_setaffinity
with a CPU mask that does not include the CPU it's running on, then when
__set_cpus_allowed_ptr() calls stop_one_cpu() intending to invoke
migration_cpu_stop() and perform the actual migration away from the CPU
it will simply return -ENOENT rather than calling migration_cpu_stop().
We then return from the sched_setaffinity syscall back to the user task
that is now running on a CPU which it just asked not to run on, and
which is not present in its cpus_allowed mask.

This patch resolves the problem by having select_task_rq() enforce that
user tasks run on CPUs that are active - the same requirement that
select_fallback_rq() already enforces. This should ensure that newly
onlined CPUs reach the CPUHP_AP_ACTIVE state before being able to
schedule user tasks, and also implies that bringup_wait_for_ap() will
have called stop_machine_unpark() which resolves the sched_setaffinity
issue above.

I haven't yet investigated them, but it may be of interest to review
whether any of the actions performed by hotplug states between
CPUHP_AP_ONLINE_IDLE & CPUHP_AP_ACTIVE could have similar unintended
effects on user tasks that might schedule before they are reached, which
might widen the scope of the problem from just affecting the behaviour
of sched_setaffinity.

Signed-off-by: Paul Burton <>
Signed-off-by: Peter Zijlstra (Intel) <>
Cc: Linus Torvalds <>
Cc: Peter Zijlstra <>
Cc: Thomas Gleixner <>
Signed-off-by: Ingo Molnar <>
4 years agosched/core: Fix rules for running on online && !active CPUs
Peter Zijlstra [Tue, 25 Jul 2017 16:58:21 +0000 (18:58 +0200)]
sched/core: Fix rules for running on online && !active CPUs

As already enforced by the WARN() in __set_cpus_allowed_ptr(), the rules
for running on an online && !active CPU are stricter than just being a
kthread, you need to be a per-cpu kthread.

If you're not strictly per-CPU, you have better CPUs to run on and
don't need the partially booted one to get your work done.

The exception is to allow smpboot threads to bootstrap the CPU itself
and get kernel 'services' initialized before we allow userspace on it.

Signed-off-by: Peter Zijlstra (Intel) <>
Cc: Linus Torvalds <>
Cc: Paul E. McKenney <>
Cc: Peter Zijlstra <>
Cc: Steven Rostedt <>
Cc: Tejun Heo <>
Cc: Thomas Gleixner <>
Fixes: 955dbdf4ce87 ("sched: Allow migrating kthreads into online but inactive CPUs")
Signed-off-by: Ingo Molnar <>
4 years agoxfrm Fix potential error pointer dereference in xfrm_bundle_create.
Steffen Klassert [Thu, 31 May 2018 07:45:18 +0000 (09:45 +0200)]
xfrm Fix potential error pointer dereference in xfrm_bundle_create.

We may derference an invalid pointer in the error path of
xfrm_bundle_create(). Fix this by returning this error
pointer directly instead of assigning it to xdst0.

Fixes: 45b018beddb6 ("ipsec: Create and use new helpers for dst child access.")
Signed-off-by: Steffen Klassert <>
4 years agofs: clear writeback errors in inode_init_always
Darrick J. Wong [Thu, 31 May 2018 02:43:53 +0000 (19:43 -0700)]
fs: clear writeback errors in inode_init_always

In inode_init_always(), we clear the inode mapping flags, which clears
any retained error (AS_EIO, AS_ENOSPC) bits.  Unfortunately, we do not
also clear wb_err, which means that old mapping errors can leak through
to new inodes.

This is crucial for the XFS inode allocation path because we recycle old
in-core inodes and we do not want error state from an old file to leak
into the new file.  This bug was discovered by running generic/036 and
generic/047 in a loop and noticing that the EIOs generated by the
collision of direct and buffered writes in generic/036 would survive the
remount between 036 and 047, and get reported to the fsyncs (on
different files!) in generic/047.

Signed-off-by: Darrick J. Wong <>
Reviewed-by: Jeff Layton <>
Reviewed-by: Brian Foster <>
4 years agoMerge tag 'drm-misc-fixes-2018-05-30' of git://
Dave Airlie [Wed, 30 May 2018 22:35:47 +0000 (08:35 +1000)]
Merge tag 'drm-misc-fixes-2018-05-30' of git:// into drm-fixes

dw-hdmi: Fix Oops regression from rc1 (Neil)

Cc: Neil Armstrong <>
* tag 'drm-misc-fixes-2018-05-30' of git://
  drm/bridge/synopsys: dw-hdmi: fix dw_hdmi_setup_rx_sense

4 years agoMerge tag 'for-linus-20180530' of git://
Linus Torvalds [Wed, 30 May 2018 21:37:59 +0000 (16:37 -0500)]
Merge tag 'for-linus-20180530' of git://

Pull block fix from Jens Axboe:
 "Just a single fix that should make it into this release, fixing a
  regression with T10-DIF on NVMe"

* tag 'for-linus-20180530' of git://
  nvme: fix extended data LBA supported setting

4 years agoMerge tag 'selinux-pr-20180530' of git://
Linus Torvalds [Wed, 30 May 2018 21:35:07 +0000 (16:35 -0500)]
Merge tag 'selinux-pr-20180530' of git://git./linux/kernel/git/pcmoore/selinux

Pull SELinux fix from Paul Moore:
 "One more small fix for SELinux: a small string length fix found by

  I dislike sending patches this late in the release cycle, but this
  patch fixes a legitimate problem, is very small, limited in scope, and
  well understood.

  There are two threads with more information on the problem, the latest
  is linked below:

  Stephen points out in the thread linked above:

   'Such a setxattr() call can only be performed by a process with
    CAP_MAC_ADMIN that is also allowed mac_admin permission in SELinux
    policy. Consequently, this is never possible on Android (no process
    is allowed mac_admin permission, always enforcing) and is only
    possible in Fedora/RHEL for a few domains (if enforcing)'"

* tag 'selinux-pr-20180530' of git://
  selinux: KASAN: slab-out-of-bounds in xattr_getsecurity

4 years agoMerge branch 'linus' of git://
Linus Torvalds [Wed, 30 May 2018 21:33:22 +0000 (16:33 -0500)]
Merge branch 'linus' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fix from Herbert Xu:
 "This fixes a potential kernel panic in the inside-secure driver"

* 'linus' of git://
  crypto: inside-secure - do not use memset on MMIO

4 years agoperf tools: Fix format description of NRCPUS header
Arnaldo Carvalho de Melo [Fri, 25 May 2018 19:37:36 +0000 (16:37 -0300)]
perf tools: Fix format description of NRCPUS header

In the HEADER_CPUDESC feadure header we store first the number
of available CPUs in the system, then the number of CPUs at the time of
writing the header, not the other way around.

Reported-by: Thomas-Mich Richter <>
Acked-by: Andi Kleen <>
Cc: Adrian Hunter <>
Cc: David Ahern <>
Cc: He Kuang <>
Cc: Hendrik Brueckner <>
Cc: Jin Yao <>
Cc: Jiri Olsa <>
Cc: Kim Phillips <>
Cc: Lakshman Annadorai <>
Cc: Namhyung Kim <>
Cc: Simon Que <>
Cc: Stephane Eranian <>
Cc: Wang Nan <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoperf script python: Add addr into perf sample dict
Leo Yan [Mon, 28 May 2018 08:45:01 +0000 (16:45 +0800)]
perf script python: Add addr into perf sample dict

ARM CoreSight auxtrace uses 'sample->addr' to record the target address
for branch instructions, so the data of 'sample->addr' is required for
tracing data analysis.

This commit collects data of 'sample->addr' into perf sample dict,
finally can be used for python script for parsing event.

Signed-off-by: Leo Yan <>
Cc: Alexander Shishkin <>
Cc: Jiri Olsa <>
Cc: Jonathan Corbet <>
Cc: Mathieu Poirier <>
Cc: Mike Leach <>
Cc: Namhyung Kim <>
Cc: Peter Zijlstra <>
Cc: Robert Walker <>
Cc: Tor Jeremiassen <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoperf data: Update documentation section on cpu topology
Thomas Richter [Mon, 28 May 2018 07:44:33 +0000 (09:44 +0200)]
perf data: Update documentation section on cpu topology

Add an explanation of each cpu's core and socket identifier to the file format documentation.

Signed-off-by: Thomas Richter <>
Cc: Heiko Carstens <>
Cc: Hendrik Brueckner <>
Cc: Martin Schwidefsky <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoperf cs-etm: Fix indexing for decoder packet queue
Mathieu Poirier [Fri, 25 May 2018 23:10:54 +0000 (17:10 -0600)]
perf cs-etm: Fix indexing for decoder packet queue

The tail of a queue is supposed to be pointing to the next available
slot in a queue.  In this implementation the tail is incremented before
it is used and as such points to the last used element, something that
has the immense advantage of centralizing tail management at a single
location and eliminating a lot of redundant code.

But this needs to be taken into consideration on the dequeueing side
where the head also needs to be incremented before it is used, or the
first available element of the queue will be skipped.

Signed-off-by: Mathieu Poirier <>
Tested-by: Leo Yan <>
Cc: Alexander Shishkin <>
Cc: Jiri Olsa <>
Cc: Namhyung Kim <>
Cc: Peter Zijlstra <>
Cc: Robert Walker <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoperf bpf: Fix NULL return handling in bpf__prepare_load()
YueHaibing [Fri, 11 May 2018 11:21:42 +0000 (19:21 +0800)]
perf bpf: Fix NULL return handling in bpf__prepare_load()

bpf_object__open()/bpf_object__open_buffer can return error pointer or
NULL, check the return values with IS_ERR_OR_NULL() in bpf__prepare_load
and bpf__prepare_load_buffer

Signed-off-by: YueHaibing <>
Acked-by: Daniel Borkmann <>
Cc: Alexander Shishkin <>
Cc: Namhyung Kim <>
Cc: Peter Zijlstra <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agodrm/bridge/synopsys: dw-hdmi: fix dw_hdmi_setup_rx_sense
Neil Armstrong [Wed, 30 May 2018 09:43:58 +0000 (11:43 +0200)]
drm/bridge/synopsys: dw-hdmi: fix dw_hdmi_setup_rx_sense

The dw_hdmi_setup_rx_sense exported function should not use struct device
to recover the dw-hdmi context using drvdata, but take struct dw_hdmi
directly like other exported functions.

This caused a regression using Meson DRM on S905X since v4.17-rc1 :

Internal error: Oops: 96000007 [#1] PREEMPT SMP
CPU: 0 PID: 124 Comm: irq/32-dw_hdmi_ Not tainted 4.17.0-rc7 #2
Hardware name: Libre Technology CC (DT)
pc : osq_lock+0x54/0x188
lr : __mutex_lock.isra.0+0x74/0x530
Process irq/32-dw_hdmi_ (pid: 124, stack limit = 0x00000000adf418cb)
Call trace:
 Code: 34000964 d00050a2 51000484 9135c042 (f864d844)
 ---[ end trace 945641e1fbbc07da ]---
 note: irq/32-dw_hdmi_[124] exited with preempt_count 1
 genirq: exiting task "irq/32-dw_hdmi_" (124) is an active IRQ thread (irq 32)

Fixes: eea034af90c6 ("drm/bridge/synopsys: dw-hdmi: don't clobber drvdata")
Signed-off-by: Neil Armstrong <>
Tested-by: Koen Kooi <>
Signed-off-by: Sean Paul <>
4 years agovhost_net: flush batched heads before trying to busy polling
Jason Wang [Tue, 29 May 2018 06:18:19 +0000 (14:18 +0800)]
vhost_net: flush batched heads before trying to busy polling

After commit e2b3b35eb989 ("vhost_net: batch used ring update in rx"),
we tend to batch updating used heads. But it doesn't flush batched
heads before trying to do busy polling, this will cause vhost to wait
for guest TX which waits for the used RX. Fixing by flush batched
heads before busy loop.

1 byte TCP_RR performance recovers from 13107.83 to 50402.65.

Fixes: e2b3b35eb989 ("vhost_net: batch used ring update in rx")
Signed-off-by: Jason Wang <>
Acked-by: Michael S. Tsirkin <>
Signed-off-by: David S. Miller <>
4 years agoMerge branch 'for-linus' of git://
Linus Torvalds [Wed, 30 May 2018 15:30:30 +0000 (10:30 -0500)]
Merge branch 'for-linus' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Martin Schwidefsky:

 - a missing -msoft-float for the compile of the kexec purgatory

 - a fix for the dasd driver to avoid the double use of a field in the
   'struct request'

[ That latter one is being discussed, and Christoph asked for something
  cleaner, but for now it's a fix ]

* 'for-linus' of git://
  s390/dasd: use blk_mq_rq_from_pdu for per request data
  s390/purgatory: Fix endless interrupt loop

4 years agoperf test: "Session topology" dumps core on s390
Thomas Richter [Mon, 28 May 2018 07:36:57 +0000 (09:36 +0200)]
perf test: "Session topology" dumps core on s390

The "perf test Session topology" entry fails with core dump on s390. The root
cause is a NULL pointer dereference in function check_cpu_topology() line 76
(or line 82 without -v).

The session->header.env.cpu variable is NULL because on s390 function
process_cpu_topology() returns with error:

    socket_id number is too big.
    You may need to upgrade the perf tool.

and releases the env.cpu variable via zfree() and sets it to NULL.

Here is the gdb output:
(gdb) n
76                      pr_debug("CPU %d, core %d, socket %d\n", i,
(gdb) n

Program received signal SIGSEGV, Segmentation fault.
0x00000000010f4d9e in check_cpu_topology (path=0x3ffffffd6c8
"/tmp/perf-test-J6CHMa", map=0x14a1740) at tests/topology.c:76
76  pr_debug("CPU %d, core %d, socket %d\n", i,

Make sure the env.cpu variable is not used when its NULL.
Test for NULL pointer and return TEST_SKIP if so.

Output before:

  [root@p23lp27 perf]# ./perf test -F 39
  39: Session topology  :Segmentation fault (core dumped)
  [root@p23lp27 perf]#

Output after:

  [root@p23lp27 perf]# ./perf test -vF 39
  39: Session topology                                      :
  --- start ---
  templ file: /tmp/perf-test-Ajx59D
  socket_id number is too big.You may need to upgrade the perf tool.
  ---- end ----
  Session topology: Skip
  [root@p23lp27 perf]#

Signed-off-by: Thomas Richter <>
Cc: Heiko Carstens <>
Cc: Hendrik Brueckner <>
Cc: Martin Schwidefsky <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoperf parse-events: Handle uncore event aliases in small groups properly
Kan Liang [Mon, 7 May 2018 21:13:43 +0000 (14:13 -0700)]
perf parse-events: Handle uncore event aliases in small groups properly

Perf stat doesn't count the uncore event aliases from the same uncore
block in a group, for example:

  perf stat -e '{unc_m_cas_count.all,unc_m_clockticks}' -a -I 1000
  #           time             counts unit events
       1.000447342      <not counted>      unc_m_cas_count.all
       1.000447342      <not counted>      unc_m_clockticks
       2.000740654      <not counted>      unc_m_cas_count.all
       2.000740654      <not counted>      unc_m_clockticks

The output is very misleading. It gives a wrong impression that the
uncore event doesn't work.

An uncore block could be composed by several PMUs. An uncore event alias
is a joint name which means the same event runs on all PMUs of a block.
Perf doesn't support mixed events from different PMUs in the same group.
It is wrong to put uncore event aliases in a big group.

The right way is to split the big group into multiple small groups which
only include the events from the same PMU.

Only uncore event aliases from the same uncore block should be specially
handled here. It doesn't make sense to mix the uncore events with other
uncore events from different blocks or even core events in a group.

With the patch:
  #           time             counts unit events
     1.001557653            140,833      unc_m_cas_count.all
     1.001557653      1,330,231,332      unc_m_clockticks
     2.002709483             85,007      unc_m_cas_count.all
     2.002709483      1,429,494,563      unc_m_clockticks

Reported-by: Andi Kleen <>
Signed-off-by: Kan Liang <>
Acked-by: Jiri Olsa <>
Cc: Agustin Vega-Frias <>
Cc: Ganapatrao Kulkarni <>
Cc: Jin Yao <>
Cc: Namhyung Kim <>
Cc: Peter Zijlstra <>
Cc: Shaokun Zhang <>
Cc: Will Deacon <>
Signed-off-by: Arnaldo Carvalho de Melo <>
4 years agoMerge branch 'for-linus' of git://
Linus Torvalds [Wed, 30 May 2018 03:22:15 +0000 (22:22 -0500)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input fixes from Dmitry Torokhov:
 "We are switching a bunch of Lenovo devices with Synaptics touchpads
  from PS/2 emulation over to native RMI/SMbus.

  Given that all commits are marked for stable there is no point
  delaying them till next release"

[ Also fix a too-small stack array for i2c communication in elan driver ]

* 'for-linus' of git://
  Input: elan_i2c_smbus - fix corrupted stack
  Input: synaptics - add Lenovo 80 series ids to SMBus
  Input: synaptics - add Intertouch support on X1 Carbon 6th and X280
  Input: synaptics - Lenovo Thinkpad X1 Carbon G5 (2017) with Elantech trackpoints should use RMI
  Input: synaptics - Lenovo Carbon X1 Gen5 (2017) devices should use RMI

4 years agoMerge branch 'drm-fixes-4.17' of git:// into drm...
Dave Airlie [Wed, 30 May 2018 00:58:31 +0000 (10:58 +1000)]
Merge branch 'drm-fixes-4.17' of git:// into drm-fixes

One last fix for 4.17.  Fix a suspend regression in DC.

* 'drm-fixes-4.17' of git://
  drm/amd/display: Fix BUG_ON during CRTC atomic check update

4 years agoMerge tag 'drm-misc-fixes-2018-05-29' of git://
Dave Airlie [Wed, 30 May 2018 00:57:57 +0000 (10:57 +1000)]
Merge tag 'drm-misc-fixes-2018-05-29' of git:// into drm-fixes

core: Add 220us psr setup time (Dhinakaran)
omap: Fix NULL deref (Tomi)

Cc: Dhinakaran Pandiyan <>
Cc: Tomi Valkeinen <>
* tag 'drm-misc-fixes-2018-05-29' of git://
  drm/omap: fix NULL deref crash with SDI displays
  drm/psr: Fix missed entry in PSR setup time table.

4 years agoselinux: KASAN: slab-out-of-bounds in xattr_getsecurity
Sachin Grover [Fri, 25 May 2018 08:31:39 +0000 (14:01 +0530)]
selinux: KASAN: slab-out-of-bounds in xattr_getsecurity

Call trace:
 [<ffffff9203a8d7a8>] dump_backtrace+0x0/0x428
 [<ffffff9203a8dbf8>] show_stack+0x28/0x38
 [<ffffff920409bfb8>] dump_stack+0xd4/0x124
 [<ffffff9203d187e8>] print_address_description+0x68/0x258
 [<ffffff9203d18c00>] kasan_report.part.2+0x228/0x2f0
 [<ffffff9203d1927c>] kasan_report+0x5c/0x70
 [<ffffff9203d1776c>] check_memory_region+0x12c/0x1c0
 [<ffffff9203d17cdc>] memcpy+0x34/0x68
 [<ffffff9203d75348>] xattr_getsecurity+0xe0/0x160
 [<ffffff9203d75490>] vfs_getxattr+0xc8/0x120
 [<ffffff9203d75d68>] getxattr+0x100/0x2c8
 [<ffffff9203d76fb4>] SyS_fgetxattr+0x64/0xa0
 [<ffffff9203a83f70>] el0_svc_naked+0x24/0x28

If user get root access and calls security.selinux setxattr() with an
embedded NUL on a file and then if some process performs a getxattr()
on that file with a length greater than the actual length of the string,
it would result in a panic.

To fix this, add the actual length of the string to the security context
instead of the length passed by the userspace process.

Signed-off-by: Sachin Grover <>
Signed-off-by: Paul Moore <>
4 years agoMerge tag 'drm-intel-fixes-2018-05-29' of git://
Dave Airlie [Wed, 30 May 2018 00:00:01 +0000 (10:00 +1000)]
Merge tag 'drm-intel-fixes-2018-05-29' of git:// into drm-fixes

- Fix for potential Spectre vector in the new query uAPI
- Fix NULL pointer deref (FDO #106559)
- DMI fix to hide LVDS for Radiant P845 (FDO #105468)

* tag 'drm-intel-fixes-2018-05-29' of git://
  drm/i915/query: nospec expects no more than an unsigned long
  drm/i915/query: Protect tainted function pointer lookup
  drm/i915/lvds: Move acpi lid notification registration to registration phase
  drm/i915: Disable LVDS on Radiant P845

4 years agoMerge tag 'afs-fixes-20180529' of git://
Linus Torvalds [Tue, 29 May 2018 20:30:16 +0000 (15:30 -0500)]
Merge tag 'afs-fixes-20180529' of git://git./linux/kernel/git/dhowells/linux-fs

Pull AFS fixes from David Howells:

 - fix a BUG triggerable from faccessat()

 - fix the mounting of backup volumes

* tag 'afs-fixes-20180529' of git://
  afs: Fix mounting of backup volumes
  afs: Fix directory permissions check

4 years agodrm/amd/display: Fix BUG_ON during CRTC atomic check update
Leo (Sunpeng) Li [Tue, 29 May 2018 13:51:51 +0000 (09:51 -0400)]
drm/amd/display: Fix BUG_ON during CRTC atomic check update

For cases where the CRTC is inactive (DPMS off), where a modeset is not
required, yet the CRTC is still in the atomic state, we should not
attempt to update anything on it.

Previously, we were relying on the modereset_required() helper to check
the above condition. However, the function returns false immediately if
a modeset is not required, ignoring the CRTC's enable/active state
flags. The correct way to filter is by looking at these flags instead.

Fixes: e277adc5a06c "drm/amd/display: Hookup color management functions"

Signed-off-by: Leo (Sunpeng) Li <>
Reviewed-by: Harry Wentland <>
Tested-by: Michel Dänzer <>
Signed-off-by: Alex Deucher <>
4 years agoMerge branch 'nvme-4.17' of git:// into for-linus
Jens Axboe [Tue, 29 May 2018 18:54:12 +0000 (12:54 -0600)]
Merge branch 'nvme-4.17' of git:// into for-linus

Pull NVMe fix from Christoph:

"Below is a one-liner fix from Max that unbreaks T10-DIF support, which
 got broken in 4.15."

* 'nvme-4.17' of git://
  nvme: fix extended data LBA supported setting

4 years agonvme: fix extended data LBA supported setting
Max Gurtovoy [Sun, 27 May 2018 15:50:10 +0000 (18:50 +0300)]
nvme: fix extended data LBA supported setting

This value depands on the metadata support value, so reorder the
initialization to fit.

Fixes: b5be3b392 ("nvme: always unregister the integrity profile in __nvme_revalidate_disk")
Signed-off-by: Max Gurtovoy <>
Signed-off-by: Christoph Hellwig <>
4 years agoi2c: ocores: update HDL sources URL
Federico Vaga [Mon, 28 May 2018 08:59:14 +0000 (10:59 +0200)]
i2c: ocores: update HDL sources URL

The URL is broken. This patch fixes it.

Signed-off-by: Federico Vaga <>
[wsa: shortened the URL a bit]
Signed-off-by: Wolfram Sang <>
4 years agoIB/core: Fix error code for invalid GID entry
Parav Pandit [Sun, 27 May 2018 11:49:16 +0000 (14:49 +0300)]
IB/core: Fix error code for invalid GID entry

When a GID entry is invalid EAGAIN is returned. This is an incorrect error
code, there is nothing that will make this GID entry valid again in
bounded time.

Some user space tools fail incorrectly if EAGAIN is returned here, and
this represents a small ABI change from earlier kernels.

The first patch in the Fixes list makes entries that were valid before
to become invalid, allowing this code to trigger, while the second patch
in the Fixes list introduced the wrong EAGAIN.

Therefore revert the return result to EINVAL which matches the historical
expectations of the ibv_query_gid_type() API of the libibverbs user space

Cc: <>
Fixes: 598ff6bae689 ("IB/core: Refactor GID modify code for RoCE")
Fixes: 03db3a2d81e6 ("IB/core: Add RoCE GID table management")
Reviewed-by: Daniel Jurgens <>
Signed-off-by: Parav Pandit <>
Signed-off-by: Leon Romanovsky <>
Signed-off-by: Jason Gunthorpe <>
4 years agotun: Fix NULL pointer dereference in XDP redirect
Toshiaki Makita [Mon, 28 May 2018 10:37:49 +0000 (19:37 +0900)]
tun: Fix NULL pointer dereference in XDP redirect

Calling XDP redirection requires bh disabled. Softirq can call another
XDP function and redirection functions, then the percpu static variable
ri->map can be overwritten to NULL.

This is a generic XDP case called from tun.

[ 3535.736058] BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
[ 3535.743974] PGD 0 P4D 0
[ 3535.746530] Oops: 0000 [#1] SMP PTI
[ 3535.750049] Modules linked in: vhost_net vhost tap tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter sunrpc vfat fat ext4 mbcache jbd2 intel_rapl skx_edac nfit libnvdimm x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm ipmi_ssif irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc ses aesni_intel crypto_simd cryptd enclosure hpwdt hpilo glue_helper ipmi_si pcspkr wmi mei_me ioatdma mei ipmi_devintf shpchp dca ipmi_msghandler lpc_ich acpi_power_meter sch_fq_codel ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm smartpqi i40e crc32c_intel scsi_transport_sas tg3 i2c_core ptp pps_core
[ 3535.813456] CPU: 5 PID: 1630 Comm: vhost-1614 Not tainted 4.17.0-rc4 #2
[ 3535.820127] Hardware name: HPE ProLiant DL360 Gen10/ProLiant DL360 Gen10, BIOS U32 11/14/2017
[ 3535.828732] RIP: 0010:__xdp_map_lookup_elem+0x5/0x30
[ 3535.833740] RSP: 0018:ffffb4bc47bf7c58 EFLAGS: 00010246
[ 3535.839009] RAX: ffff9fdfcfea1c40 RBX: 0000000000000000 RCX: ffff9fdf27fe3100
[ 3535.846205] RDX: ffff9fdfca769200 RSI: 0000000000000000 RDI: 0000000000000000
[ 3535.853402] RBP: ffffb4bc491d9000 R08: 00000000000045ad R09: 0000000000000ec0
[ 3535.860597] R10: 0000000000000001 R11: ffff9fdf26c3ce4e R12: ffff9fdf9e72c000
[ 3535.867794] R13: 0000000000000000 R14: fffffffffffffff2 R15: ffff9fdfc82cdd00
[ 3535.874990] FS:  0000000000000000(0000) GS:ffff9fdfcfe80000(0000) knlGS:0000000000000000
[ 3535.883152] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 3535.888948] CR2: 0000000000000018 CR3: 0000000bde724004 CR4: 00000000007626e0
[ 3535.896145] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 3535.903342] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 3535.910538] PKRU: 55555554
[ 3535.913267] Call Trace:
[ 3535.915736]  xdp_do_generic_redirect+0x7a/0x310
[ 3535.920310]  do_xdp_generic.part.117+0x285/0x370
[ 3535.924970]  tun_get_user+0x5b9/0x1260 [tun]
[ 3535.929279]  tun_sendmsg+0x52/0x70 [tun]
[ 3535.933237]  handle_tx+0x2ad/0x5f0 [vhost_net]
[ 3535.937721]  vhost_worker+0xa5/0x100 [vhost]
[ 3535.942030]  kthread+0xf5/0x130
[ 3535.945198]  ? vhost_dev_ioctl+0x3b0/0x3b0 [vhost]
[ 3535.950031]  ? kthread_bind+0x10/0x10
[ 3535.953727]  ret_from_fork+0x35/0x40
[ 3535.957334] Code: 0e 74 15 83 f8 10 75 05 e9 49 aa b3 ff f3 c3 0f 1f 80 00 00 00 00 f3 c3 e9 29 9d b3 ff 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 <8b> 47 18 83 f8 0e 74 0d 83 f8 10 75 05 e9 49 a9 b3 ff 31 c0 c3
[ 3535.976387] RIP: __xdp_map_lookup_elem+0x5/0x30 RSP: ffffb4bc47bf7c58
[ 3535.982883] CR2: 0000000000000018
[ 3535.987096] ---[ end trace 383b299dd1430240 ]---
[ 3536.131325] Kernel panic - not syncing: Fatal exception
[ 3536.137484] Kernel Offset: 0x26a00000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[ 3536.281406] ---[ end Kernel panic - not syncing: Fatal exception ]---

And a kernel with generic case fixed still panics in tun driver XDP
redirect, because it disabled only preemption, but not bh.

[ 2055.128746] BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
[ 2055.136662] PGD 0 P4D 0
[ 2055.139219] Oops: 0000 [#1] SMP PTI
[ 2055.142736] Modules linked in: vhost_net vhost tap tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter sunrpc vfat fat ext4 mbcache jbd2 intel_rapl skx_edac nfit libnvdimm x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc ses aesni_intel ipmi_ssif crypto_simd enclosure cryptd hpwdt glue_helper ioatdma hpilo wmi dca pcspkr ipmi_si acpi_power_meter ipmi_devintf shpchp mei_me ipmi_msghandler mei lpc_ich sch_fq_codel ip_tables xfs libcrc32c sd_mod mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm i40e smartpqi tg3 scsi_transport_sas crc32c_intel i2c_core ptp pps_core
[ 2055.206142] CPU: 6 PID: 1693 Comm: vhost-1683 Tainted: G        W         4.17.0-rc5-fix-tun+ #1
[ 2055.215011] Hardware name: HPE ProLiant DL360 Gen10/ProLiant DL360 Gen10, BIOS U32 11/14/2017
[ 2055.223617] RIP: 0010:__xdp_map_lookup_elem+0x5/0x30
[ 2055.228624] RSP: 0018:ffff998b07607cc0 EFLAGS: 00010246
[ 2055.233892] RAX: ffff8dbd8e235700 RBX: ffff8dbd8ff21c40 RCX: 0000000000000004
[ 2055.241089] RDX: ffff998b097a9000 RSI: 0000000000000000 RDI: 0000000000000000
[ 2055.248286] RBP: 0000000000000000 R08: 00000000000065a8 R09: 0000000000005d80
[ 2055.255483] R10: 0000000000000040 R11: ffff8dbcf0100000 R12: ffff998b097a9000
[ 2055.262681] R13: ffff8dbd8c98c000 R14: 0000000000000000 R15: ffff998b07607d78
[ 2055.269879] FS:  0000000000000000(0000) GS:ffff8dbd8ff00000(0000) knlGS:0000000000000000
[ 2055.278039] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 2055.283834] CR2: 0000000000000018 CR3: 0000000c0c8cc005 CR4: 00000000007626e0
[ 2055.291030] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 2055.298227] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 2055.305424] PKRU: 55555554
[ 2055.308153] Call Trace:
[ 2055.310624]  xdp_do_redirect+0x7b/0x380
[ 2055.314499]  tun_get_user+0x10fe/0x12a0 [tun]
[ 2055.318895]  tun_sendmsg+0x52/0x70 [tun]
[ 2055.322852]  handle_tx+0x2ad/0x5f0 [vhost_net]
[ 2055.327337]  vhost_worker+0xa5/0x100 [vhost]
[ 2055.331646]  kthread+0xf5/0x130
[ 2055.334813]  ? vhost_dev_ioctl+0x3b0/0x3b0 [vhost]
[ 2055.339646]  ? kthread_bind+0x10/0x10
[ 2055.343343]  ret_from_fork+0x35/0x40
[ 2055.346950] Code: 0e 74 15 83 f8 10 75 05 e9 e9 aa b3 ff f3 c3 0f 1f 80 00 00 00 00 f3 c3 e9 c9 9d b3 ff 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 <8b> 47 18 83 f8 0e 74 0d 83 f8 10 75 05 e9 e9 a9 b3 ff 31 c0 c3
[ 2055.366004] RIP: __xdp_map_lookup_elem+0x5/0x30 RSP: ffff998b07607cc0
[ 2055.372500] CR2: 0000000000000018
[ 2055.375856] ---[ end trace 2a2dcc5e9e174268 ]---
[ 2055.523626] Kernel panic - not syncing: Fatal exception
[ 2055.529796] Kernel Offset: 0x2e000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[ 2055.677539] ---[ end Kernel panic - not syncing: Fatal exception ]---

 - Removed preempt_disable/enable since local_bh_disable will prevent
   preemption as well, feedback from Jason Wang.

Fixes: 761876c857cb ("tap: XDP support")
Signed-off-by: Toshiaki Makita <>
Acked-by: Jason Wang <>
Signed-off-by: David S. Miller <>
4 years agobe2net: Fix error detection logic for BE3
Suresh Reddy [Mon, 28 May 2018 05:26:06 +0000 (01:26 -0400)]
be2net: Fix error detection logic for BE3

Check for 0xE00 (RECOVERABLE_ERR) along with ARMFW UE (0x0)
in be_detect_error() to know whether the error is valid error or not

Fixes: 673c96e5a ("be2net: Fix UE detection logic for BE3")
Signed-off-by: Suresh Reddy <>
Signed-off-by: David S. Miller <>
4 years agonet: qmi_wwan: Add Netgear Aircard 779S
Josh Hill [Mon, 28 May 2018 00:10:41 +0000 (20:10 -0400)]
net: qmi_wwan: Add Netgear Aircard 779S

Add support for Netgear Aircard 779S

Signed-off-by: Josh Hill <>
Acked-by: Bjørn Mork <>
Signed-off-by: David S. Miller <>
4 years agomlxsw: spectrum: Forbid creation of VLAN 1 over port/LAG
Petr Machata [Sun, 27 May 2018 06:48:41 +0000 (09:48 +0300)]
mlxsw: spectrum: Forbid creation of VLAN 1 over port/LAG

VLAN 1 is internally used for untagged traffic. Prevent creation of
explicit netdevice for that VLAN, because that currently isn't supported
and leads to the NULL pointer dereference cited below.

Fix by preventing creation of VLAN devices with VID of 1 over mlxsw
devices or LAG devices that involve mlxsw devices.

[  327.175816] ================================================================================
[  327.184544] UBSAN: Undefined behaviour in drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c:200:12
[  327.193667] member access within null pointer of type 'const struct mlxsw_sp_fid'
[  327.201226] CPU: 0 PID: 8983 Comm: ip Not tainted 4.17.0-rc4-petrm_net_ip6gre_headroom-custom-140 #11
[  327.210496] Hardware name: Mellanox Technologies Ltd. "MSN2410-CB2F"/"SA000874", BIOS 4.6.5 03/08/2016
[  327.219872] Call Trace:
[  327.222384]  dump_stack+0xc3/0x12b
[  327.234007]  ubsan_epilogue+0x9/0x49
[  327.237638]  ubsan_type_mismatch_common+0x1f9/0x2d0
[  327.255769]  __ubsan_handle_type_mismatch+0x90/0xa7
[  327.264716]  mlxsw_sp_fid_type+0x35/0x50 [mlxsw_spectrum]
[  327.270255]  mlxsw_sp_port_vlan_router_leave+0x46/0xc0 [mlxsw_spectrum]
[  327.277019]  mlxsw_sp_inetaddr_port_vlan_event+0xe1/0x340 [mlxsw_spectrum]
[  327.315031]  mlxsw_sp_netdevice_vrf_event+0xa8/0x100 [mlxsw_spectrum]
[  327.321626]  mlxsw_sp_netdevice_event+0x276/0x430 [mlxsw_spectrum]
[  327.367863]  notifier_call_chain+0x4c/0x150
[  327.372128]  __netdev_upper_dev_link+0x1b3/0x260
[  327.399450]  vrf_add_slave+0xce/0x170 [vrf]
[  327.403703]  do_setlink+0x658/0x1d70
[  327.508998]  rtnl_newlink+0x908/0xf20
[  327.559128]  rtnetlink_rcv_msg+0x50c/0x720
[  327.571720]  netlink_rcv_skb+0x16a/0x1f0
[  327.583450]  netlink_unicast+0x2ca/0x3e0
[  327.599305]  netlink_sendmsg+0x3e2/0x7f0
[  327.616655]  sock_sendmsg+0x76/0xc0
[  327.620207]  ___sys_sendmsg+0x494/0x5d0
[  327.666117]  __sys_sendmsg+0xc2/0x130
[  327.690953]  do_syscall_64+0x66/0x370
[  327.694677]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  327.699782] RIP: 0033:0x7f4c2f3f8037
[  327.703393] RSP: 002b:00007ffe8c389708 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[  327.711035] RAX: ffffffffffffffda RBX: 000000005b03f53e RCX: 00007f4c2f3f8037
[  327.718229] RDX: 0000000000000000 RSI: 00007ffe8c389760 RDI: 0000000000000003
[  327.725431] RBP: 00007ffe8c389760 R08: 0000000000000000 R09: 00007f4c2f443630
[  327.732632] R10: 00000000000005eb R11: 0000000000000246 R12: 0000000000000000
[  327.739833] R13: 00000000006774e0 R14: 00007ffe8c3897e8 R15: 0000000000000000
[  327.747096] ================================================================================

Fixes: 9589a7b5d7d9 ("mlxsw: spectrum: Handle VLAN devices linking / unlinking")
Suggested-by: Ido Schimmel <>
Signed-off-by: Petr Machata <>
Signed-off-by: Ido Schimmel <>
Acked-by: Jiri Pirko <>
Signed-off-by: David S. Miller <>
4 years agoatm: zatm: fix memcmp casting
Ivan Bornyakov [Fri, 25 May 2018 17:49:52 +0000 (20:49 +0300)]
atm: zatm: fix memcmp casting

memcmp() returns int, but eprom_try_esi() cast it to unsigned char. One
can lose significant bits and get 0 from non-0 value returned by the

Signed-off-by: Ivan Bornyakov <>
Signed-off-by: David S. Miller <>
4 years agoMerge tag 'trace-v4.17-rc4-3' of git://
Linus Torvalds [Tue, 29 May 2018 12:28:48 +0000 (07:28 -0500)]
Merge tag 'trace-v4.17-rc4-3' of git://git./linux/kernel/git/rostedt/linux-trace

Pull tracing fixes from Steven Rostedt:
 "While writing selftests for a new feature, I triggered two existing
  bugs that deal with triggers and instances.

   - a generic trigger bug where the triggers are not removed from a
     linked list properly when deleting an instance.

   - a bug specific to snapshots, where the snapshot is done in the top
     level buffer, when it is supposed to snapshot the buffer associated
     to the instance the snapshot trigger exists in"

* tag 'trace-v4.17-rc4-3' of git://
  tracing: Make the snapshot trigger work with instances
  tracing: Fix crash when freeing instances with event triggers

4 years agodrm/i915/query: nospec expects no more than an unsigned long
Chris Wilson [Tue, 22 May 2018 12:10:18 +0000 (13:10 +0100)]
drm/i915/query: nospec expects no more than an unsigned long

nospec quite reasonably asserts that it will never be used with an index
larger than unsigned long (that being the largest possibly index into an
C array). However, our ubi uses the convention of u64 for any large
integer, running afoul of the assertion on 32b. Reduce our index to an
unsigned long, checking for type overflow first.

  drivers/gpu/drm/i915/i915_query.c: In function 'i915_query_ioctl':
  include/linux/compiler.h:339:38: error: call to '__compiletime_assert_119' declared with attribute error: BUILD_BUG_ON failed: sizeof(_s) > sizeof(long)

Fixes: 84b510e22da7 ("drm/i915/query: Protect tainted function pointer lookup")
Signed-off-by: Chris Wilson <>
Cc: Lionel Landwerlin <>
Cc: Joonas Lahtinen <>
Cc: Tvrtko Ursulin <>
Reviewed-by: Lionel Landwerlin <>
(cherry picked from commit a33b1dc8a732144e11cb4bf067d24ba51e6b8ab0)
Signed-off-by: Joonas Lahtinen <>
4 years agoiwlwifi: pcie: compare with number of IRQs requested for, not number of CPUs
Hao Wei Tee [Tue, 29 May 2018 07:25:17 +0000 (10:25 +0300)]
iwlwifi: pcie: compare with number of IRQs requested for, not number of CPUs

When there are 16 or more logical CPUs, we request for
`IWL_MAX_RX_HW_QUEUES` (16) IRQs only as we limit to that number of
IRQs, but later on we compare the number of IRQs returned to
nr_online_cpus+2 instead of max_irqs, the latter being what we
actually asked for. This ends up setting num_rx_queues to 17 which
causes lots of out-of-bounds array accesses later on.

Compare to max_irqs instead, and also add an assertion in case
num_rx_queues > IWM_MAX_RX_HW_QUEUES.

This fixes

Fixes: 2e5d4a8f61dc ("iwlwifi: pcie: Add new configuration to enable MSIX")
Signed-off-by: Hao Wei Tee <>
Tested-by: Sara Sharon <>
Signed-off-by: Luca Coelho <>
Signed-off-by: Kalle Valo <>
4 years agoRevert "rt2800: use TXOP_BACKOFF for probe frames"
Stanislaw Gruszka [Mon, 28 May 2018 11:25:06 +0000 (13:25 +0200)]
Revert "rt2800: use TXOP_BACKOFF for probe frames"

This reverts commit fb47ada8dc3c30c8e7b415da155742b49536c61e.

In some situations when we set TXOP_BACKOFF, the probe frame is
not sent at all. What it worse then sending probe frame as part
of AMPDU and can degrade 11n performance to 11g rates.

Signed-off-by: Stanislaw Gruszka <>
Signed-off-by: Kalle Valo <>
4 years agonet: netsec: reduce DMA mask to 40 bits
Ard Biesheuvel [Fri, 25 May 2018 12:50:37 +0000 (14:50 +0200)]
net: netsec: reduce DMA mask to 40 bits

The netsec network controller IP can drive 64 address bits for DMA, and
the DMA mask is set accordingly in the driver. However, the SynQuacer
SoC, which is the only silicon incorporating this IP at the moment,
integrates this IP in a manner that leaves address bits [63:40]

Up until now, this has not resulted in any problems, given that the DDR
controller doesn't decode those bits to begin with. However, recent
firmware updates for platforms incorporating this SoC allow the IOMMU
to be enabled, which does decode address bits [47:40], and allocates
top down from the IOVA space, producing DMA addresses that have bits
set that have been left unconnected.

Both the DT and ACPI (IORT) descriptions of the platform take this into
account, and only describe a DMA address space of 40 bits (using either
dma-ranges DT properties, or DMA address limits in IORT named component
nodes). However, even though our IOMMU and bus layers may take such
limitations into account by setting a narrower DMA mask when creating
the platform device, the netsec probe() entrypoint follows the common
practice of setting the DMA mask uncondionally, according to the
capabilities of the IP block itself rather than to its integration into
the chip.

It is currently unclear what the correct fix is here. We could hack around
it by only setting the DMA mask if it deviates from its default value of
DMA_BIT_MASK(32). However, this makes it impossible for the bus layer to
use DMA_BIT_MASK(32) as the bus limit, and so it appears that a more
comprehensive approach is required to take DMA limits imposed by the
SoC as a whole into account.

In the mean time, let's limit the DMA mask to 40 bits. Given that there
is currently only one SoC that incorporates this IP, this is a reasonable
approach that can be backported to -stable and buys us some time to come
up with a proper fix going forward.

Fixes: 533dd11a12f6 ("net: socionext: Add Synquacer NetSec driver")
Cc: Robin Murphy <>
Cc: Jassi Brar <>
Cc: Masahisa Kojima <>
Cc: Ilias Apalodimas <>
Signed-off-by: Ard Biesheuvel <>
Reviewed-by: Robin Murphy <>
Acked-by: Jassi Brar <>
Signed-off-by: David S. Miller <>
4 years agoipv6: sr: fix memory OOB access in seg6_do_srh_encap/inline
Mathieu Xhonneux [Fri, 25 May 2018 12:29:41 +0000 (13:29 +0100)]
ipv6: sr: fix memory OOB access in seg6_do_srh_encap/inline

seg6_do_srh_encap and seg6_do_srh_inline can possibly do an
out-of-bounds access when adding the SRH to the packet. This no longer
happen when expanding the skb not only by the size of the SRH (+
outer IPv6 header), but also by skb->mac_len.

[   53.793056] BUG: KASAN: use-after-free in seg6_do_srh_encap+0x284/0x620
[   53.794564] Write of size 14 at addr ffff88011975ecfa by task ping/674

[   53.796665] CPU: 0 PID: 674 Comm: ping Not tainted 4.17.0-rc3-ARCH+ #90
[   53.796670] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS 1.11.0-20171110_100015-anatol 04/01/2014
[   53.796673] Call Trace:
[   53.796679]  <IRQ>
[   53.796689]  dump_stack+0x71/0xab
[   53.796700]  print_address_description+0x6a/0x270
[   53.796707]  kasan_report+0x258/0x380
[   53.796715]  ? seg6_do_srh_encap+0x284/0x620
[   53.796722]  memmove+0x34/0x50
[   53.796730]  seg6_do_srh_encap+0x284/0x620
[   53.796741]  ? seg6_do_srh+0x29b/0x360
[   53.796747]  seg6_do_srh+0x29b/0x360
[   53.796756]  seg6_input+0x2e/0x2e0
[   53.796765]  lwtunnel_input+0x93/0xd0
[   53.796774]  ipv6_rcv+0x690/0x920
[   53.796783]  ? ip6_input+0x170/0x170
[   53.796791]  ? eth_gro_receive+0x2d0/0x2d0
[   53.796800]  ? ip6_input+0x170/0x170
[   53.796809]  __netif_receive_skb_core+0xcc0/0x13f0
[   53.796820]  ? netdev_info+0x110/0x110
[   53.796827]  ? napi_complete_done+0xb6/0x170
[   53.796834]  ? e1000_clean+0x6da/0xf70
[   53.796845]  ? process_backlog+0x129/0x2a0
[   53.796853]  process_backlog+0x129/0x2a0
[   53.796862]  net_rx_action+0x211/0x5c0
[   53.796870]  ? napi_complete_done+0x170/0x170
[   53.796887]  ? run_rebalance_domains+0x11f/0x150
[   53.796891]  __do_softirq+0x10e/0x39e
[   53.796894]  do_softirq_own_stack+0x2a/0x40
[   53.796895]  </IRQ>
[   53.796898]  do_softirq.part.16+0x54/0x60
[   53.796900]  __local_bh_enable_ip+0x5b/0x60
[   53.796903]  ip6_finish_output2+0x416/0x9f0
[   53.796906]  ? ip6_dst_lookup_flow+0x110/0x110
[   53.796909]  ? ip6_sk_dst_lookup_flow+0x390/0x390
[   53.796911]  ? __rcu_read_unlock+0x66/0x80
[   53.796913]  ? ip6_mtu+0x44/0xf0
[   53.796916]  ? ip6_output+0xfc/0x220
[   53.796918]  ip6_output+0xfc/0x220
[   53.796921]  ? ip6_finish_output+0x2b0/0x2b0
[   53.796923]  ? memcpy+0x34/0x50
[   53.796926]  ip6_send_skb+0x43/0xc0
[   53.796929]  rawv6_sendmsg+0x1216/0x1530
[   53.796932]  ? __orc_find+0x6b/0xc0
[   53.796934]  ? rawv6_rcv_skb+0x160/0x160
[   53.796937]  ? __rcu_read_unlock+0x66/0x80
[   53.796939]  ? __rcu_read_unlock+0x66/0x80
[   53.796942]  ? is_bpf_text_address+0x1e/0x30
[   53.796944]  ? kernel_text_address+0xec/0x100
[   53.796946]  ? __kernel_text_address+0xe/0x30
[   53.796948]  ? unwind_get_return_address+0x2f/0x50
[   53.796950]  ? __save_stack_trace+0x92/0x100
[   53.796954]  ? save_stack+0x89/0xb0
[   53.796956]  ? kasan_kmalloc+0xa0/0xd0
[   53.796958]  ? kmem_cache_alloc+0xd2/0x1f0
[   53.796961]  ? prepare_creds+0x23/0x160
[   53.796963]  ? __x64_sys_capset+0x252/0x3e0
[   53.796966]  ? do_syscall_64+0x69/0x160
[   53.796968]  ? entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   53.796971]  ? __alloc_pages_nodemask+0x170/0x380
[   53.796973]  ? __alloc_pages_slowpath+0x12c0/0x12c0
[   53.796977]  ? tty_vhangup+0x20/0x20
[   53.796979]  ? policy_nodemask+0x1a/0x90
[   53.796982]  ? __mod_node_page_state+0x8d/0xa0
[   53.796986]  ? __check_object_size+0xe7/0x240
[   53.796989]  ? __sys_sendto+0x229/0x290
[   53.796991]  ? rawv6_rcv_skb+0x160/0x160
[   53.796993]  __sys_sendto+0x229/0x290
[   53.796996]  ? __ia32_sys_getpeername+0x50/0x50
[   53.796999]  ? commit_creds+0x2de/0x520
[   53.797002]  ? security_capset+0x57/0x70
[   53.797004]  ? __x64_sys_capset+0x29f/0x3e0
[   53.797007]  ? __x64_sys_rt_sigsuspend+0xe0/0xe0
[   53.797011]  ? __do_page_fault+0x664/0x770
[   53.797014]  __x64_sys_sendto+0x74/0x90
[   53.797017]  do_syscall_64+0x69/0x160
[   53.797019]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   53.797022] RIP: 0033:0x7f43b7a6714a
[   53.797023] RSP: 002b:00007ffd891bd368 EFLAGS: 00000246 ORIG_RAX:
[   53.797026] RAX: ffffffffffffffda RBX: 00000000006129c0 RCX: 00007f43b7a6714a
[   53.797028] RDX: 0000000000000040 RSI: 00000000006129c0 RDI: 0000000000000004
[   53.797029] RBP: 00007ffd891be640 R08: 0000000000610940 R09: 000000000000001c
[   53.797030] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000040
[   53.797032] R13: 000000000060e6a0 R14: 0000000000008004 R15: 000000000040b661

[   53.797171] Allocated by task 642:
[   53.797460]  kasan_kmalloc+0xa0/0xd0
[   53.797463]  kmem_cache_alloc+0xd2/0x1f0
[   53.797465]  getname_flags+0x40/0x210
[   53.797467]  user_path_at_empty+0x1d/0x40
[   53.797469]  do_faccessat+0x12a/0x320
[   53.797471]  do_syscall_64+0x69/0x160
[   53.797473]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

[   53.797607] Freed by task 642:
[   53.797869]  __kasan_slab_free+0x130/0x180
[   53.797871]  kmem_cache_free+0xa8/0x230
[   53.797872]  filename_lookup+0x15b/0x230
[   53.797874]  do_faccessat+0x12a/0x320
[   53.797876]  do_syscall_64+0x69/0x160
[   53.797878]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

[   53.798014] The buggy address belongs to the object at ffff88011975e600
                which belongs to the cache names_cache of size 4096
[   53.799043] The buggy address is located 1786 bytes inside of
                4096-byte region [ffff88011975e600ffff88011975f600)
[   53.800013] The buggy address belongs to the page:
[   53.800414] page:ffffea000465d600 count:1 mapcount:0
mapping:0000000000000000 index:0x0 compound_mapcount: 0
[   53.801259] flags: 0x17fff0000008100(slab|head)
[   53.801640] raw: 017fff0000008100 0000000000000000 0000000000000000
[   53.803147] raw: dead000000000100 dead000000000200 ffff88011b185a40
[   53.803787] page dumped because: kasan: bad access detected

[   53.804384] Memory state around the buggy address:
[   53.804788]  ffff88011975eb80: fb fb fb fb fb fb fb fb fb fb fb fb
fb fb fb fb
[   53.805384]  ffff88011975ec00: fb fb fb fb fb fb fb fb fb fb fb fb
fb fb fb fb
[   53.805979] >ffff88011975ec80: fb fb fb fb fb fb fb fb fb fb fb fb
fb fb fb fb
[   53.806577]                                                                 ^
[   53.807165]  ffff88011975ed00: fb fb fb fb fb fb fb fb fb fb fb fb
fb fb fb fb
[   53.807762]  ffff88011975ed80: fb fb fb fb fb fb fb fb fb fb fb fb
fb fb fb fb
[   53.808356] ==================================================================
[   53.808949] Disabling lock debugging due to kernel taint

Fixes: 6c8702c60b88 ("ipv6: sr: add support for SRH encapsulation and injection with lwtunnels")
Signed-off-by: David Lebrun <>
Signed-off-by: Mathieu Xhonneux <>
Signed-off-by: David S. Miller <>
4 years agoMerge git://
David S. Miller [Tue, 29 May 2018 02:39:09 +0000 (22:39 -0400)]
Merge git://git./pub/scm/linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

Netfilter/IPVS fixes for net

The following patchset contains Netfilter/IPVS fixes for your net tree:

1) Null pointer dereference when dumping conntrack helper configuration,
   from Taehee Yoo.

2) Missing sanitization in ebtables extension name through compat,
   from Paolo Abeni.

3) Broken fetch of tracing value, from Taehee Yoo.

4) Incorrect arithmetics in packet ratelimiting.

5) Buffer overflow in IPVS sync daemon, from Julian Anastasov.

6) Wrong argument to nla_strlcpy() in nfnetlink_{acct,cthelper},
   from Eric Dumazet.

7) Fix splat in nft_update_chain_stats().

8) Null pointer dereference from object netlink dump path, from
   Taehee Yoo.

9) Missing static_branch_inc() when enabling counters in existing
   chain, from Taehee Yoo.

Signed-off-by: David S. Miller <>