openwrt/staging/blogic.git
5 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Fri, 19 Jul 2019 16:45:58 +0000 (09:45 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge yet more updates from Andrew Morton:
 "The rest of MM and a kernel-wide procfs cleanup.

  Summary of the more significant patches:

   - Patch series "mm/memory_hotplug: Factor out memory block
     devicehandling", v3. David Hildenbrand.

     Some spring-cleaning of the memory hotplug code, notably in
     drivers/base/memory.c

   - "mm: thp: fix false negative of shmem vma's THP eligibility". Yang
     Shi.

     Fix /proc/pid/smaps output for THP pages used in shmem.

   - "resource: fix locking in find_next_iomem_res()" + 1. Nadav Amit.

     Bugfix and speedup for kernel/resource.c

   - Patch series "mm: Further memory block device cleanups", David
     Hildenbrand.

     More spring-cleaning of the memory hotplug code.

   - Patch series "mm: Sub-section memory hotplug support". Dan
     Williams.

     Generalise the memory hotplug code so that pmem can use it more
     completely. Then remove the hacks from the libnvdimm code which
     were there to work around the memory-hotplug code's constraints.

   - "proc/sysctl: add shared variables for range check", Matteo Croce.

     We have about 250 instances of

          int zero;
          ...
                  .extra1 = &zero,

     in the tree. This is a tree-wide sweep to make all those private
     "zero"s and "one"s use global variables.

     Alas, it isn't practical to make those two global integers const"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (38 commits)
  proc/sysctl: add shared variables for range check
  mm: migrate: remove unused mode argument
  mm/sparsemem: cleanup 'section number' data types
  libnvdimm/pfn: stop padding pmem namespaces to section alignment
  libnvdimm/pfn: fix fsdax-mode namespace info-block zero-fields
  mm/devm_memremap_pages: enable sub-section remap
  mm: document ZONE_DEVICE memory-model implications
  mm/sparsemem: support sub-section hotplug
  mm/sparsemem: prepare for sub-section ranges
  mm: kill is_dev_zone() helper
  mm/hotplug: kill is_dev_zone() usage in __remove_pages()
  mm/sparsemem: convert kmalloc_section_memmap() to populate_section_memmap()
  mm/hotplug: prepare shrink_{zone, pgdat}_span for sub-section removal
  mm/sparsemem: add helpers track active portions of a section at boot
  mm/sparsemem: introduce a SECTION_IS_EARLY flag
  mm/sparsemem: introduce struct mem_section_usage
  drivers/base/memory.c: get rid of find_memory_block_hinted()
  mm/memory_hotplug: move and simplify walk_memory_blocks()
  mm/memory_hotplug: rename walk_memory_range() and pass start+size instead of pfns
  mm: make register_mem_sect_under_node() static
  ...

5 years agoproc/sysctl: add shared variables for range check
Matteo Croce [Thu, 18 Jul 2019 22:58:50 +0000 (15:58 -0700)]
proc/sysctl: add shared variables for range check

In the sysctl code the proc_dointvec_minmax() function is often used to
validate the user supplied value between an allowed range.  This
function uses the extra1 and extra2 members from struct ctl_table as
minimum and maximum allowed value.

On sysctl handler declaration, in every source file there are some
readonly variables containing just an integer which address is assigned
to the extra1 and extra2 members, so the sysctl range is enforced.

The special values 0, 1 and INT_MAX are very often used as range
boundary, leading duplication of variables like zero=0, one=1,
int_max=INT_MAX in different source files:

    $ git grep -E '\.extra[12].*&(zero|one|int_max)' |wc -l
    248

Add a const int array containing the most commonly used values, some
macros to refer more easily to the correct array member, and use them
instead of creating a local one for every object file.

This is the bloat-o-meter output comparing the old and new binary
compiled with the default Fedora config:

    # scripts/bloat-o-meter -d vmlinux.o.old vmlinux.o
    add/remove: 2/2 grow/shrink: 0/2 up/down: 24/-188 (-164)
    Data                                         old     new   delta
    sysctl_vals                                    -      12     +12
    __kstrtab_sysctl_vals                          -      12     +12
    max                                           14      10      -4
    int_max                                       16       -     -16
    one                                           68       -     -68
    zero                                         128      28    -100
    Total: Before=20583249, After=20583085, chg -0.00%

[mcroce@redhat.com: tipc: remove two unused variables]
Link: http://lkml.kernel.org/r/20190530091952.4108-1-mcroce@redhat.com
[akpm@linux-foundation.org: fix net/ipv6/sysctl_net_ipv6.c]
[arnd@arndb.de: proc/sysctl: make firmware loader table conditional]
Link: http://lkml.kernel.org/r/20190617130014.1713870-1-arnd@arndb.de
[akpm@linux-foundation.org: fix fs/eventpoll.c]
Link: http://lkml.kernel.org/r/20190430180111.10688-1-mcroce@redhat.com
Signed-off-by: Matteo Croce <mcroce@redhat.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Aaron Tomlin <atomlin@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: migrate: remove unused mode argument
Keith Busch [Thu, 18 Jul 2019 22:58:46 +0000 (15:58 -0700)]
mm: migrate: remove unused mode argument

migrate_page_move_mapping() doesn't use the mode argument.  Remove it
and update callers accordingly.

Link: http://lkml.kernel.org/r/20190508210301.8472-1-keith.busch@intel.com
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: cleanup 'section number' data types
Dan Williams [Thu, 18 Jul 2019 22:58:43 +0000 (15:58 -0700)]
mm/sparsemem: cleanup 'section number' data types

David points out that there is a mixture of 'int' and 'unsigned long'
usage for section number data types.  Update the memory hotplug path to
use 'unsigned long' consistently for section numbers.

[akpm@linux-foundation.org: fix printk format]
Link: http://lkml.kernel.org/r/156107543656.1329419.11505835211949439815.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: David Hildenbrand <david@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agolibnvdimm/pfn: stop padding pmem namespaces to section alignment
Dan Williams [Thu, 18 Jul 2019 22:58:40 +0000 (15:58 -0700)]
libnvdimm/pfn: stop padding pmem namespaces to section alignment

Now that the mm core supports section-unaligned hotplug of ZONE_DEVICE
memory, we no longer need to add padding at pfn/dax device creation
time.  The kernel will still honor padding established by older kernels.

Link: http://lkml.kernel.org/r/156092356588.979959.6793371748950931916.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Jeff Moyer <jmoyer@redhat.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agolibnvdimm/pfn: fix fsdax-mode namespace info-block zero-fields
Dan Williams [Thu, 18 Jul 2019 22:58:36 +0000 (15:58 -0700)]
libnvdimm/pfn: fix fsdax-mode namespace info-block zero-fields

At namespace creation time there is the potential for the "expected to
be zero" fields of a 'pfn' info-block to be filled with indeterminate
data.  While the kernel buffer is zeroed on allocation it is immediately
overwritten by nd_pfn_validate() filling it with the current contents of
the on-media info-block location.  For fields like, 'flags' and the
'padding' it potentially means that future implementations can not rely on
those fields being zero.

In preparation to stop using the 'start_pad' and 'end_trunc' fields for
section alignment, arrange for fields that are not explicitly
initialized to be guaranteed zero.  Bump the minor version to indicate
it is safe to assume the 'padding' and 'flags' are zero.  Otherwise,
this corruption is expected to benign since all other critical fields
are explicitly initialized.

Note The cc: stable is about spreading this new policy to as many
kernels as possible not fixing an issue in those kernels.  It is not
until the change titled "libnvdimm/pfn: Stop padding pmem namespaces to
section alignment" where this improper initialization becomes a problem.
So if someone decides to backport "libnvdimm/pfn: Stop padding pmem
namespaces to section alignment" (which is not tagged for stable), make
sure this pre-requisite is flagged.

Link: http://lkml.kernel.org/r/156092356065.979959.6681003754765958296.stgit@dwillia2-desk3.amr.corp.intel.com
Fixes: 32ab0a3f5170 ("libnvdimm, pmem: 'struct page' for pmem")
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: <stable@vger.kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/devm_memremap_pages: enable sub-section remap
Dan Williams [Thu, 18 Jul 2019 22:58:33 +0000 (15:58 -0700)]
mm/devm_memremap_pages: enable sub-section remap

Teach devm_memremap_pages() about the new sub-section capabilities of
arch_{add,remove}_memory().  Effectively, just replace all usage of
align_start, align_end, and align_size with res->start, res->end, and
resource_size(res).  The existing sanity check will still make sure that
the two separate remap attempts do not collide within a sub-section (2MB
on x86).

Link: http://lkml.kernel.org/r/156092355542.979959.10060071713397030576.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Michal Hocko <mhocko@suse.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: document ZONE_DEVICE memory-model implications
Dan Williams [Thu, 18 Jul 2019 22:58:29 +0000 (15:58 -0700)]
mm: document ZONE_DEVICE memory-model implications

Explain the general mechanisms of 'ZONE_DEVICE' pages and list the users
of 'devm_memremap_pages()'.

[dan.j.williams@intel.com: update ZONE_DEVICE memory model documentation]
Link: http://lkml.kernel.org/r/156109575458.1409767.1885676287099277666.stgit@dwillia2-desk3.amr.corp.intel.com
Link: http://lkml.kernel.org/r/156092354985.979959.15763234410543451710.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Mike Rapoport <rppt@linux.ibm.com>
Reviewed-by: Mike Rapoport <rppt@linux.ibm.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: support sub-section hotplug
Dan Williams [Thu, 18 Jul 2019 22:58:26 +0000 (15:58 -0700)]
mm/sparsemem: support sub-section hotplug

The libnvdimm sub-system has suffered a series of hacks and broken
workarounds for the memory-hotplug implementation's awkward
section-aligned (128MB) granularity.

For example the following backtrace is emitted when attempting
arch_add_memory() with physical address ranges that intersect 'System
RAM' (RAM) with 'Persistent Memory' (PMEM) within a given section:

    # cat /proc/iomem | grep -A1 -B1 Persistent\ Memory
    100000000-1ffffffff : System RAM
    200000000-303ffffff : Persistent Memory (legacy)
    304000000-43fffffff : System RAM
    440000000-23ffffffff : Persistent Memory
    2400000000-43bfffffff : Persistent Memory
      2400000000-43bfffffff : namespace2.0

    WARNING: CPU: 38 PID: 928 at arch/x86/mm/init_64.c:850 add_pages+0x5c/0x60
    [..]
    RIP: 0010:add_pages+0x5c/0x60
    [..]
    Call Trace:
     devm_memremap_pages+0x460/0x6e0
     pmem_attach_disk+0x29e/0x680 [nd_pmem]
     ? nd_dax_probe+0xfc/0x120 [libnvdimm]
     nvdimm_bus_probe+0x66/0x160 [libnvdimm]

It was discovered that the problem goes beyond RAM vs PMEM collisions as
some platform produce PMEM vs PMEM collisions within a given section.
The libnvdimm workaround for that case revealed that the libnvdimm
section-alignment-padding implementation has been broken for a long
while.

A fix for that long-standing breakage introduces as many problems as it
solves as it would require a backward-incompatible change to the
namespace metadata interpretation.  Instead of that dubious route [1],
address the root problem in the memory-hotplug implementation.

Note that EEXIST is no longer treated as success as that is how
sparse_add_section() reports subsection collisions, it was also obviated
by recent changes to perform the request_region() for 'System RAM'
before arch_add_memory() in the add_memory() sequence.

[1] https://lore.kernel.org/r/155000671719.348031.2347363160141119237.stgit@dwillia2-desk3.amr.corp.intel.com

[osalvador@suse.de: fix deactivate_section for early sections]
Link: http://lkml.kernel.org/r/20190715081549.32577-2-osalvador@suse.de
Link: http://lkml.kernel.org/r/156092354368.979959.6232443923440952359.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: prepare for sub-section ranges
Dan Williams [Thu, 18 Jul 2019 22:58:22 +0000 (15:58 -0700)]
mm/sparsemem: prepare for sub-section ranges

Prepare the memory hot-{add,remove} paths for handling sub-section
ranges by plumbing the starting page frame and number of pages being
handled through arch_{add,remove}_memory() to
sparse_{add,remove}_one_section().

This is simply plumbing, small cleanups, and some identifier renames.
No intended functional changes.

Link: http://lkml.kernel.org/r/156092353780.979959.9713046515562743194.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: kill is_dev_zone() helper
Dan Williams [Thu, 18 Jul 2019 22:58:18 +0000 (15:58 -0700)]
mm: kill is_dev_zone() helper

Given there are no more usages of is_dev_zone() outside of 'ifdef
CONFIG_ZONE_DEVICE' protection, kill off the compilation helper.

Link: http://lkml.kernel.org/r/156092353211.979959.1489004866360828964.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Acked-by: David Hildenbrand <david@redhat.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Michal Hocko <mhocko@suse.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/hotplug: kill is_dev_zone() usage in __remove_pages()
Dan Williams [Thu, 18 Jul 2019 22:58:15 +0000 (15:58 -0700)]
mm/hotplug: kill is_dev_zone() usage in __remove_pages()

The zone type check was a leftover from the cleanup that plumbed altmap
through the memory hotplug path, i.e.  commit da024512a1fa "mm: pass the
vmem_altmap to arch_remove_memory and __remove_pages".

Link: http://lkml.kernel.org/r/156092352642.979959.6664333788149363039.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Michal Hocko <mhocko@suse.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: convert kmalloc_section_memmap() to populate_section_memmap()
Dan Williams [Thu, 18 Jul 2019 22:58:11 +0000 (15:58 -0700)]
mm/sparsemem: convert kmalloc_section_memmap() to populate_section_memmap()

Allow sub-section sized ranges to be added to the memmap.

populate_section_memmap() takes an explict pfn range rather than
assuming a full section, and those parameters are plumbed all the way
through to vmmemap_populate().  There should be no sub-section usage in
current deployments.  New warnings are added to clarify which memmap
allocation paths are sub-section capable.

Link: http://lkml.kernel.org/r/156092352058.979959.6551283472062305149.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/hotplug: prepare shrink_{zone, pgdat}_span for sub-section removal
Dan Williams [Thu, 18 Jul 2019 22:58:07 +0000 (15:58 -0700)]
mm/hotplug: prepare shrink_{zone, pgdat}_span for sub-section removal

Sub-section hotplug support reduces the unit of operation of hotplug
from section-sized-units (PAGES_PER_SECTION) to sub-section-sized units
(PAGES_PER_SUBSECTION).  Teach shrink_{zone,pgdat}_span() to consider
PAGES_PER_SUBSECTION boundaries as the points where pfn_valid(), not
valid_section(), can toggle.

[osalvador@suse.de: fix shrink_{zone,node}_span]
Link: http://lkml.kernel.org/r/20190717090725.23618-3-osalvador@suse.de
Link: http://lkml.kernel.org/r/156092351496.979959.12703722803097017492.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: add helpers track active portions of a section at boot
Dan Williams [Thu, 18 Jul 2019 22:58:04 +0000 (15:58 -0700)]
mm/sparsemem: add helpers track active portions of a section at boot

Prepare for hot{plug,remove} of sub-ranges of a section by tracking a
sub-section active bitmask, each bit representing a PMD_SIZE span of the
architecture's memory hotplug section size.

The implications of a partially populated section is that pfn_valid()
needs to go beyond a valid_section() check and either determine that the
section is an "early section", or read the sub-section active ranges
from the bitmask.  The expectation is that the bitmask (subsection_map)
fits in the same cacheline as the valid_section() / early_section()
data, so the incremental performance overhead to pfn_valid() should be
negligible.

The rationale for using early_section() to short-ciruit the
subsection_map check is that there are legacy code paths that use
pfn_valid() at section granularity before validating the pfn against
pgdat data.  So, the early_section() check allows those traditional
assumptions to persist while also permitting subsection_map to tell the
truth for purposes of populating the unused portions of early sections
with PMEM and other ZONE_DEVICE mappings.

Link: http://lkml.kernel.org/r/156092350874.979959.18185938451405518285.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Qian Cai <cai@lca.pw>
Tested-by: Jane Chu <jane.chu@oracle.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: introduce a SECTION_IS_EARLY flag
Dan Williams [Thu, 18 Jul 2019 22:58:00 +0000 (15:58 -0700)]
mm/sparsemem: introduce a SECTION_IS_EARLY flag

In preparation for sub-section hotplug, track whether a given section
was created during early memory initialization, or later via memory
hotplug.  This distinction is needed to maintain the coarse expectation
that pfn_valid() returns true for any pfn within a given section even if
that section has pages that are reserved from the page allocator.

For example one of the of goals of subsection hotplug is to support
cases where the system physical memory layout collides System RAM and
PMEM within a section.  Several pfn_valid() users expect to just check
if a section is valid, but they are not careful to check if the given
pfn is within a "System RAM" boundary and instead expect pgdat
information to further validate the pfn.

Rather than unwind those paths to make their pfn_valid() queries more
precise a follow on patch uses the SECTION_IS_EARLY flag to maintain the
traditional expectation that pfn_valid() returns true for all early
sections.

Link: https://lore.kernel.org/lkml/1560366952-10660-1-git-send-email-cai@lca.pw/
Link: http://lkml.kernel.org/r/156092350358.979959.5817209875548072819.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Qian Cai <cai@lca.pw>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparsemem: introduce struct mem_section_usage
Dan Williams [Thu, 18 Jul 2019 22:57:57 +0000 (15:57 -0700)]
mm/sparsemem: introduce struct mem_section_usage

Patch series "mm: Sub-section memory hotplug support", v10.

The memory hotplug section is an arbitrary / convenient unit for memory
hotplug.  'Section-size' units have bled into the user interface
('memblock' sysfs) and can not be changed without breaking existing
userspace.  The section-size constraint, while mostly benign for typical
memory hotplug, has and continues to wreak havoc with 'device-memory'
use cases, persistent memory (pmem) in particular.  Recall that pmem
uses devm_memremap_pages(), and subsequently arch_add_memory(), to
allocate a 'struct page' memmap for pmem.  However, it does not use the
'bottom half' of memory hotplug, i.e.  never marks pmem pages online and
never exposes the userspace memblock interface for pmem.  This leaves an
opening to redress the section-size constraint.

To date, the libnvdimm subsystem has attempted to inject padding to
satisfy the internal constraints of arch_add_memory().  Beyond
complicating the code, leading to bugs [2], wasting memory, and limiting
configuration flexibility, the padding hack is broken when the platform
changes this physical memory alignment of pmem from one boot to the
next.  Device failure (intermittent or permanent) and physical
reconfiguration are events that can cause the platform firmware to
change the physical placement of pmem on a subsequent boot, and device
failure is an everyday event in a data-center.

It turns out that sections are only a hard requirement of the
user-facing interface for memory hotplug and with a bit more
infrastructure sub-section arch_add_memory() support can be added for
kernel internal usages like devm_memremap_pages().  Here is an analysis
of the current design assumptions in the current code and how they are
addressed in the new implementation:

Current design assumptions:

 - Sections that describe boot memory (early sections) are never
   unplugged / removed.

 - pfn_valid(), in the CONFIG_SPARSEMEM_VMEMMAP=y, case devolves to a
   valid_section() check

 - __add_pages() and helper routines assume all operations occur in
   PAGES_PER_SECTION units.

 - The memblock sysfs interface only comprehends full sections

New design assumptions:

 - Sections are instrumented with a sub-section bitmask to track (on
   x86) individual 2MB sub-divisions of a 128MB section.

 - Partially populated early sections can be extended with additional
   sub-sections, and those sub-sections can be removed with
   arch_remove_memory(). With this in place we no longer lose usable
   memory capacity to padding.

 - pfn_valid() is updated to look deeper than valid_section() to also
   check the active-sub-section mask. This indication is in the same
   cacheline as the valid_section() so the performance impact is
   expected to be negligible. So far the lkp robot has not reported any
   regressions.

 - Outside of the core vmemmap population routines which are replaced,
   other helper routines like shrink_{zone,pgdat}_span() are updated to
   handle the smaller granularity. Core memory hotplug routines that
   deal with online memory are not touched.

 - The existing memblock sysfs user api guarantees / assumptions are not
   touched since this capability is limited to !online
   !memblock-sysfs-accessible sections.

Meanwhile the issue reports continue to roll in from users that do not
understand when and how the 128MB constraint will bite them.  The current
implementation relied on being able to support at least one misaligned
namespace, but that immediately falls over on any moderately complex
namespace creation attempt.  Beyond the initial problem of 'System RAM'
colliding with pmem, and the unsolvable problem of physical alignment
changes, Linux is now being exposed to platforms that collide pmem ranges
with other pmem ranges by default [3].  In short, devm_memremap_pages()
has pushed the venerable section-size constraint past the breaking point,
and the simplicity of section-aligned arch_add_memory() is no longer
tenable.

These patches are exposed to the kbuild robot on a subsection-v10 branch
[4], and a preview of the unit test for this functionality is available
on the 'subsection-pending' branch of ndctl [5].

[2]: https://lore.kernel.org/r/155000671719.348031.2347363160141119237.stgit@dwillia2-desk3.amr.corp.intel.com
[3]: https://github.com/pmem/ndctl/issues/76
[4]: https://git.kernel.org/pub/scm/linux/kernel/git/djbw/nvdimm.git/log/?h=subsection-v10
[5]: https://github.com/pmem/ndctl/commit/7c59b4867e1c

This patch (of 13):

Towards enabling memory hotplug to track partial population of a section,
introduce 'struct mem_section_usage'.

A pointer to a 'struct mem_section_usage' instance replaces the existing
pointer to a 'pageblock_flags' bitmap.  Effectively it adds one more
'unsigned long' beyond the 'pageblock_flags' (usemap) allocation to house
a new 'subsection_map' bitmap.  The new bitmap enables the memory
hot{plug,remove} implementation to act on incremental sub-divisions of a
section.

SUBSECTION_SHIFT is defined as global constant instead of per-architecture
value like SECTION_SIZE_BITS in order to allow cross-arch compatibility of
subsection users.  Specifically a common subsection size allows for the
possibility that persistent memory namespace configurations be made
compatible across architectures.

The primary motivation for this functionality is to support platforms that
mix "System RAM" and "Persistent Memory" within a single section, or
multiple PMEM ranges with different mapping lifetimes within a single
section.  The section restriction for hotplug has caused an ongoing saga
of hacks and bugs for devm_memremap_pages() users.

Beyond the fixups to teach existing paths how to retrieve the 'usemap'
from a section, and updates to usemap allocation path, there are no
expected behavior changes.

Link: http://lkml.kernel.org/r/156092349845.979959.73333291612799019.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> [ppc64]
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Qian Cai <cai@lca.pw>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agodrivers/base/memory.c: get rid of find_memory_block_hinted()
David Hildenbrand [Thu, 18 Jul 2019 22:57:53 +0000 (15:57 -0700)]
drivers/base/memory.c: get rid of find_memory_block_hinted()

No longer needed, let's remove it.  Also, drop the "hint" parameter
completely from "find_memory_block_by_id", as nobody needs it anymore.

[david@redhat.com: v3]
Link: http://lkml.kernel.org/r/20190620183139.4352-7-david@redhat.com
[david@redhat.com: handle zero-length walks]
Link: http://lkml.kernel.org/r/1c2edc22-afd7-2211-c4c7-40e54e5007e8@redhat.com
Link: http://lkml.kernel.org/r/20190614100114.311-7-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Qian Cai <cai@lca.pw>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Mike Travis <mike.travis@hpe.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: move and simplify walk_memory_blocks()
David Hildenbrand [Thu, 18 Jul 2019 22:57:50 +0000 (15:57 -0700)]
mm/memory_hotplug: move and simplify walk_memory_blocks()

Let's move walk_memory_blocks() to the place where memory block logic
resides and simplify it.  While at it, add a type for the callback
function.

Link: http://lkml.kernel.org/r/20190614100114.311-6-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Mike Travis <mike.travis@hpe.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: rename walk_memory_range() and pass start+size instead of pfns
David Hildenbrand [Thu, 18 Jul 2019 22:57:46 +0000 (15:57 -0700)]
mm/memory_hotplug: rename walk_memory_range() and pass start+size instead of pfns

walk_memory_range() was once used to iterate over sections.  Now, it
iterates over memory blocks.  Rename the function, fixup the
documentation.

Also, pass start+size instead of PFNs, which is what most callers
already have at hand.  (we'll rework link_mem_sections() most probably
soon)

Follow-up patches will rework, simplify, and move walk_memory_blocks()
to drivers/base/memory.c.

Note: walk_memory_blocks() only works correctly right now if the
start_pfn is aligned to a section start.  This is the case right now,
but we'll generalize the function in a follow up patch so the semantics
match the documentation.

[akpm@linux-foundation.org: remove unused variable]
Link: http://lkml.kernel.org/r/20190614100114.311-5-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <lenb@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Rashmica Gupta <rashmica.g@gmail.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Michael Neuling <mikey@neuling.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: make register_mem_sect_under_node() static
David Hildenbrand [Thu, 18 Jul 2019 22:57:43 +0000 (15:57 -0700)]
mm: make register_mem_sect_under_node() static

It is only used internally.

Link: http://lkml.kernel.org/r/20190614100114.311-4-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agodrivers/base/memory: use "unsigned long" for block ids
David Hildenbrand [Thu, 18 Jul 2019 22:57:40 +0000 (15:57 -0700)]
drivers/base/memory: use "unsigned long" for block ids

Block ids are just shifted section numbers, so let's also use "unsigned
long" for them, too.

Link: http://lkml.kernel.org/r/20190614100114.311-3-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: section numbers use the type "unsigned long"
David Hildenbrand [Thu, 18 Jul 2019 22:57:37 +0000 (15:57 -0700)]
mm: section numbers use the type "unsigned long"

Patch series "mm: Further memory block device cleanups", v1.

Some further cleanups around memory block devices.  Especially, clean up
and simplify walk_memory_range().  Including some other minor cleanups.

This patch (of 6):

We are using a mixture of "int" and "unsigned long".  Let's make this
consistent by using "unsigned long" everywhere.  We'll do the same with
memory block ids next.

While at it, turn the "unsigned long i" in removable_show() into an int
- sections_per_block is an int.

[akpm@linux-foundation.org: s/unsigned long i/unsigned long nr/]
[david@redhat.com: v3]
Link: http://lkml.kernel.org/r/20190620183139.4352-2-david@redhat.com
Link: http://lkml.kernel.org/r/20190614100114.311-2-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoresource: avoid unnecessary lookups in find_next_iomem_res()
Nadav Amit [Thu, 18 Jul 2019 22:57:34 +0000 (15:57 -0700)]
resource: avoid unnecessary lookups in find_next_iomem_res()

find_next_iomem_res() shows up to be a source for overhead in dax
benchmarks.

Improve performance by not considering children of the tree if the top
level does not match.  Since the range of the parents should include the
range of the children such check is redundant.

Running sysbench on dax (pmem emulation, with write_cache disabled):

  sysbench fileio --file-total-size=3G --file-test-mode=rndwr \
   --file-io-mode=mmap --threads=4 --file-fsync-mode=fdatasync run

Provides the following results:

events (avg/stddev)
-------------------
  5.2-rc3: 1247669.0000/16075.39
  w/patch: 1286320.5000/16402.72 (+3%)

Link: http://lkml.kernel.org/r/20190613045903.4922-3-namit@vmware.com
Signed-off-by: Nadav Amit <namit@vmware.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoresource: fix locking in find_next_iomem_res()
Nadav Amit [Thu, 18 Jul 2019 22:57:31 +0000 (15:57 -0700)]
resource: fix locking in find_next_iomem_res()

Since resources can be removed, locking should ensure that the resource
is not removed while accessing it.  However, find_next_iomem_res() does
not hold the lock while copying the data of the resource.

Keep holding the lock while the data is copied.  While at it, change the
return value to a more informative value.  It is disregarded by the
callers.

[akpm@linux-foundation.org: fix find_next_iomem_res() documentation]
Link: http://lkml.kernel.org/r/20190613045903.4922-2-namit@vmware.com
Fixes: ff3cc952d3f00 ("resource: Add remove_resource interface")
Signed-off-by: Nadav Amit <namit@vmware.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Toshi Kani <toshi.kani@hpe.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: thp: fix false negative of shmem vma's THP eligibility
Yang Shi [Thu, 18 Jul 2019 22:57:27 +0000 (15:57 -0700)]
mm: thp: fix false negative of shmem vma's THP eligibility

Commit 7635d9cbe832 ("mm, thp, proc: report THP eligibility for each
vma") introduced THPeligible bit for processes' smaps.  But, when
checking the eligibility for shmem vma, __transparent_hugepage_enabled()
is called to override the result from shmem_huge_enabled().  It may
result in the anonymous vma's THP flag override shmem's.  For example,
running a simple test which create THP for shmem, but with anonymous THP
disabled, when reading the process's smaps, it may show:

  7fc92ec00000-7fc92f000000 rw-s 00000000 00:14 27764 /dev/shm/test
  Size:               4096 kB
  ...
  [snip]
  ...
  ShmemPmdMapped:     4096 kB
  ...
  [snip]
  ...
  THPeligible:    0

And, /proc/meminfo does show THP allocated and PMD mapped too:

  ShmemHugePages:     4096 kB
  ShmemPmdMapped:     4096 kB

This doesn't make too much sense.  The shmem objects should be treated
separately from anonymous THP.  Calling shmem_huge_enabled() with
checking MMF_DISABLE_THP sounds good enough.  And, we could skip stack
and dax vma check since we already checked if the vma is shmem already.

Also check if vma is suitable for THP by calling
transhuge_vma_suitable().

And minor fix to smaps output format and documentation.

Link: http://lkml.kernel.org/r/1560401041-32207-3-git-send-email-yang.shi@linux.alibaba.com
Fixes: 7635d9cbe832 ("mm, thp, proc: report THP eligibility for each vma")
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: thp: make transhuge_vma_suitable available for anonymous THP
Yang Shi [Thu, 18 Jul 2019 22:57:24 +0000 (15:57 -0700)]
mm: thp: make transhuge_vma_suitable available for anonymous THP

transhuge_vma_suitable() was only available for shmem THP, but anonymous
THP has the same check except pgoff check.  And, it will be used for THP
eligible check in the later patch, so make it available for all kind of
THPs.  This also helps reduce code duplication slightly.

Since anonymous THP doesn't have to check pgoff, so make pgoff check
shmem vma only.

And regroup some functions in include/linux/mm.h to solve compile issue
since transhuge_vma_suitable() needs call vma_is_anonymous() which was
defined after huge_mm.h is included.

[akpm@linux-foundation.org: fix typo]
[yang.shi@linux.alibaba.com: v4]
Link: http://lkml.kernel.org/r/1563400758-124759-2-git-send-email-yang.shi@linux.alibaba.com
Link: http://lkml.kernel.org/r/1560401041-32207-2-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/sparse.c: set section nid for hot-add memory
Wei Yang [Thu, 18 Jul 2019 22:57:21 +0000 (15:57 -0700)]
mm/sparse.c: set section nid for hot-add memory

In case of NODE_NOT_IN_PAGE_FLAGS is set, we store section's node id in
section_to_node_table[].  While for hot-add memory, this is missed.
Without this information, page_to_nid() may not give the right node id.

BTW, current online_pages works because it leverages nid in
memory_block.  But the granularity of node id should be mem_section
wide.

Link: http://lkml.kernel.org/r/20190618005537.18878-1-richardw.yang@linux.intel.com
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: remove "zone" parameter from sparse_remove_one_section
David Hildenbrand [Thu, 18 Jul 2019 22:57:17 +0000 (15:57 -0700)]
mm/memory_hotplug: remove "zone" parameter from sparse_remove_one_section

The parameter is unused, so let's drop it.  Memory removal paths should
never care about zones.  This is the job of memory offlining and will
require more refactorings.

Link: http://lkml.kernel.org/r/20190527111152.16324-12-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: make unregister_memory_block_under_nodes() never fail
David Hildenbrand [Thu, 18 Jul 2019 22:57:12 +0000 (15:57 -0700)]
mm/memory_hotplug: make unregister_memory_block_under_nodes() never fail

We really don't want anything during memory hotunplug to fail.  We
always pass a valid memory block device, that check can go.  Avoid
allocating memory and eventually failing.  As we are always called under
lock, we can use a static piece of memory.  This avoids having to put
the structure onto the stack, having to guess about the stack size of
callers.

Patch inspired by a patch from Oscar Salvador.

In the future, there might be no need to iterate over nodes at all.
mem->nid should tell us exactly what to remove.  Memory block devices
with mixed nodes (added during boot) should properly fenced off and
never removed.

Link: http://lkml.kernel.org/r/20190527111152.16324-11-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: remove memory block devices before arch_remove_memory()
David Hildenbrand [Thu, 18 Jul 2019 22:57:06 +0000 (15:57 -0700)]
mm/memory_hotplug: remove memory block devices before arch_remove_memory()

Let's factor out removing of memory block devices, which is only
necessary for memory added via add_memory() and friends that created
memory block devices.  Remove the devices before calling
arch_remove_memory().

This finishes factoring out memory block device handling from
arch_add_memory() and arch_remove_memory().

Link: http://lkml.kernel.org/r/20190527111152.16324-10-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: drop MHP_MEMBLOCK_API
David Hildenbrand [Thu, 18 Jul 2019 22:57:01 +0000 (15:57 -0700)]
mm/memory_hotplug: drop MHP_MEMBLOCK_API

No longer needed, the callers of arch_add_memory() can handle this
manually.

Link: http://lkml.kernel.org/r/20190527111152.16324-9-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: create memory block devices after arch_add_memory()
David Hildenbrand [Thu, 18 Jul 2019 22:56:56 +0000 (15:56 -0700)]
mm/memory_hotplug: create memory block devices after arch_add_memory()

Only memory to be added to the buddy and to be onlined/offlined by user
space using /sys/devices/system/memory/...  needs (and should have!)
memory block devices.

Factor out creation of memory block devices.  Create all devices after
arch_add_memory() succeeded.  We can later drop the want_memblock
parameter, because it is now effectively stale.

Only after memory block devices have been added, memory can be onlined
by user space.  This implies, that memory is not visible to user space
at all before arch_add_memory() succeeded.

While at it
 - use WARN_ON_ONCE instead of BUG_ON in moved unregister_memory()
 - introduce find_memory_block_by_id() to search via block id
 - Use find_memory_block_by_id() in init_memory_block() to catch
   duplicates

Link: http://lkml.kernel.org/r/20190527111152.16324-8-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Qian Cai <cai@lca.pw>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: allow arch_remove_memory() without CONFIG_MEMORY_HOTREMOVE
David Hildenbrand [Thu, 18 Jul 2019 22:56:51 +0000 (15:56 -0700)]
mm/memory_hotplug: allow arch_remove_memory() without CONFIG_MEMORY_HOTREMOVE

We want to improve error handling while adding memory by allowing to use
arch_remove_memory() and __remove_pages() even if
CONFIG_MEMORY_HOTREMOVE is not set to e.g., implement something like:

arch_add_memory()
rc = do_something();
if (rc) {
arch_remove_memory();
}

We won't get rid of CONFIG_MEMORY_HOTREMOVE for now, as it will require
quite some dependencies for memory offlining.

Link: http://lkml.kernel.org/r/20190527111152.16324-7-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Rich Felker <dalias@libc.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Qian Cai <cai@lca.pw>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agodrivers/base/memory: pass a block_id to init_memory_block()
David Hildenbrand [Thu, 18 Jul 2019 22:56:46 +0000 (15:56 -0700)]
drivers/base/memory: pass a block_id to init_memory_block()

We'll rework hotplug_memory_register() shortly, so it no longer consumes
pass a section.

[cai@lca.pw: fix a compilation warning]
Link: http://lkml.kernel.org/r/1559320186-28337-1-git-send-email-cai@lca.pw
Link: http://lkml.kernel.org/r/20190527111152.16324-6-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Qian Cai <cai@lca.pw>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoarm64/mm: add temporary arch_remove_memory() implementation
David Hildenbrand [Thu, 18 Jul 2019 22:56:41 +0000 (15:56 -0700)]
arm64/mm: add temporary arch_remove_memory() implementation

A proper arch_remove_memory() implementation is on its way, which also
cleanly removes page tables in arch_add_memory() in case something goes
wrong.

As we want to use arch_remove_memory() in case something goes wrong
during memory hotplug after arch_add_memory() finished, let's add a
temporary hack that is sufficient enough until we get a proper
implementation that cleans up page table entries.

We will remove CONFIG_MEMORY_HOTREMOVE around this code in follow up
patches.

Link: http://lkml.kernel.org/r/20190527111152.16324-5-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agos390x/mm: implement arch_remove_memory()
David Hildenbrand [Thu, 18 Jul 2019 22:56:35 +0000 (15:56 -0700)]
s390x/mm: implement arch_remove_memory()

Will come in handy when wanting to handle errors after
arch_add_memory().

Link: http://lkml.kernel.org/r/20190527111152.16324-4-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agos390x/mm: fail when an altmap is used for arch_add_memory()
David Hildenbrand [Thu, 18 Jul 2019 22:56:30 +0000 (15:56 -0700)]
s390x/mm: fail when an altmap is used for arch_add_memory()

ZONE_DEVICE is not yet supported, fail if an altmap is passed, so we
don't forget arch_add_memory()/arch_remove_memory() when unlocking
support.

Link: http://lkml.kernel.org/r/20190527111152.16324-3-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Oscar Salvador <osalvador@suse.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: simplify and fix check_hotplug_memory_range()
David Hildenbrand [Thu, 18 Jul 2019 22:56:25 +0000 (15:56 -0700)]
mm/memory_hotplug: simplify and fix check_hotplug_memory_range()

Patch series "mm/memory_hotplug: Factor out memory block devicehandling", v3.

We only want memory block devices for memory to be onlined/offlined
(add/remove from the buddy).  This is required so user space can
online/offline memory and kdump gets notified about newly onlined
memory.

Let's factor out creation/removal of memory block devices.  This helps
to further cleanup arch_add_memory/arch_remove_memory() and to make
implementation of new features easier - especially sub-section memory
hot add from Dan.

Anshuman Khandual is currently working on arch_remove_memory().  I added
a temporary solution via "arm64/mm: Add temporary arch_remove_memory()
implementation", that is sufficient as a firsts tep in the context of
this series.  (we don't cleanup page tables in case anything goes wrong
already)

Did a quick sanity test with DIMM plug/unplug, making sure all devices
and sysfs links properly get added/removed.  Compile tested on s390x and
x86-64.

This patch (of 11):

By converting start and size to page granularity, we actually ignore
unaligned parts within a page instead of properly bailing out with an
error.

Link: http://lkml.kernel.org/r/20190527111152.16324-2-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Andrew Banman <andrew.banman@hpe.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chintan Pandya <cpandya@codeaurora.org>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "mike.travis@hpe.com" <mike.travis@hpe.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoMerge tag 'for-5.3/dm-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 18 Jul 2019 21:49:33 +0000 (14:49 -0700)]
Merge tag 'for-5.3/dm-changes-2' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull more device mapper updates from Mike Snitzer:

 - Fix zone state management race in DM zoned target by eliminating the
   unnecessary DMZ_ACTIVE state.

 - A couple fixes for issues the DM snapshot target's optional discard
   support added during first week of the 5.3 merge.

 - Increase default size of outstanding IO that is allowed for a each
   dm-kcopyd client and introduce tunable to allow user adjust.

 - Update DM core to use printk ratelimiting functions rather than
   duplicate them and in doing so fix an issue where DMDEBUG_LIMIT()
   rate limited KERN_DEBUG messages had excessive "callbacks suppressed"
   messages.

* tag 'for-5.3/dm-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm: use printk ratelimiting functions
  dm kcopyd: Increase default sub-job size to 512KB
  dm snapshot: fix oversights in optional discard support
  dm zoned: fix zone state management race

5 years agoMerge tag 'nfs-for-5.3-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Thu, 18 Jul 2019 21:32:33 +0000 (14:32 -0700)]
Merge tag 'nfs-for-5.3-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Stable fixes:

   - SUNRPC: Ensure bvecs are re-synced when we re-encode the RPC
     request

   - Fix an Oops in ff_layout_track_ds_error due to a PTR_ERR()
     dereference

   - Revert buggy NFS readdirplus optimisation

   - NFSv4: Handle the special Linux file open access mode

   - pnfs: Fix a problem where we gratuitously start doing I/O through
     the MDS

  Features:

   - Allow NFS client to set up multiple TCP connections to the server
     using a new 'nconnect=X' mount option. Queue length is used to
     balance load.

   - Enhance statistics reporting to report on all transports when using
     multiple connections.

   - Speed up SUNRPC by removing bh-safe spinlocks

   - Add a mechanism to allow NFSv4 to request that containers set a
     unique per-host identifier for when the hostname is not set.

   - Ensure NFSv4 updates the lease_time after a clientid update

  Bugfixes and cleanup:

   - Fix use-after-free in rpcrdma_post_recvs

   - Fix a memory leak when nfs_match_client() is interrupted

   - Fix buggy file access checking in NFSv4 open for execute

   - disable unsupported client side deduplication

   - Fix spurious client disconnections

   - Fix occasional RDMA transport deadlock

   - Various RDMA cleanups

   - Various tracepoint fixes

   - Fix the TCP callback channel to guarantee the server can actually
     send the number of callback requests that was negotiated at mount
     time"

* tag 'nfs-for-5.3-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (68 commits)
  pnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS
  pnfs: Fix a problem where we gratuitously start doing I/O through the MDS
  SUNRPC: Optimise transport balancing code
  SUNRPC: Ensure the bvecs are reset when we re-encode the RPC request
  pnfs/flexfiles: Fix PTR_ERR() dereferences in ff_layout_track_ds_error
  NFSv4: Don't use the zero stateid with layoutget
  SUNRPC: Fix up backchannel slot table accounting
  SUNRPC: Fix initialisation of struct rpc_xprt_switch
  SUNRPC: Skip zero-refcount transports
  SUNRPC: Replace division by multiplication in calculation of queue length
  NFSv4: Validate the stateid before applying it to state recovery
  nfs4.0: Refetch lease_time after clientid update
  nfs4: Rename nfs41_setup_state_renewal
  nfs4: Make nfs4_proc_get_lease_time available for nfs4.0
  nfs: Fix copy-and-paste error in debug message
  NFS: Replace 16 seq_printf() calls by seq_puts()
  NFS: Use seq_putc() in nfs_show_stats()
  Revert "NFS: readdirplus optimization by cache mechanism" (memleak)
  SUNRPC: Fix transport accounting when caller specifies an rpc_xprt
  NFS: Record task, client ID, and XID in xdr_status trace points
  ...

5 years agopnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS
Trond Myklebust [Thu, 18 Jul 2019 13:32:17 +0000 (09:32 -0400)]
pnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS

Add tracepoints to allow debugging of the event chain leading to
a pnfs fallback to doing I/O through the MDS.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
5 years agopnfs: Fix a problem where we gratuitously start doing I/O through the MDS
Trond Myklebust [Thu, 18 Jul 2019 19:33:42 +0000 (15:33 -0400)]
pnfs: Fix a problem where we gratuitously start doing I/O through the MDS

If the client has to stop in pnfs_update_layout() to wait for another
layoutget to complete, it currently exits and defaults to I/O through
the MDS if the layoutget was successful.

Fixes: d03360aaf5cc ("pNFS: Ensure we return the error if someone kills...")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: stable@vger.kernel.org # v4.20+
5 years agoMerge tag 'riscv/for-v5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv...
Linus Torvalds [Thu, 18 Jul 2019 19:26:59 +0000 (12:26 -0700)]
Merge tag 'riscv/for-v5.3-rc1' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V updates from Paul Walmsley:

 - Hugepage support

 - "Image" header support for RISC-V kernel binaries, compatible with
   the current ARM64 "Image" header

 - Initial page table setup now split into two stages

 - CONFIG_SOC support (starting with SiFive SoCs)

 - Avoid reserving memory between RAM start and the kernel in
   setup_bootmem()

 - Enable high-res timers and dynamic tick in the RV64 defconfig

 - Remove long-deprecated gate area stubs

 - MAINTAINERS updates to switch to the newly-created shared RISC-V git
   tree, and to fix a get_maintainers.pl issue for patches involving
   SiFive E-mail addresses

Also, one integration fix to resolve a build problem introduced during
in the v5.3-rc1 merge window:

 - Fix build break after macro-to-function conversion in
   asm-generic/cacheflush.h

* tag 'riscv/for-v5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: fix build break after macro-to-function conversion in generic cacheflush.h
  RISC-V: Add an Image header that boot loader can parse.
  RISC-V: Setup initial page tables in two stages
  riscv: remove free_initrd_mem
  riscv: ccache: Remove unused variable
  riscv: Introduce huge page support for 32/64bit kernel
  x86, arm64: Move ARCH_WANT_HUGE_PMD_SHARE config in arch/Kconfig
  RISC-V: Fix memory reservation in setup_bootmem()
  riscv: defconfig: enable SOC_SIFIVE
  riscv: select SiFive platform drivers with SOC_SIFIVE
  arch: riscv: add config option for building SiFive's SoC resource
  riscv: Remove gate area stubs
  MAINTAINERS: change the arch/riscv git tree to the new shared tree
  MAINTAINERS: don't automatically patches involving SiFive to the linux-riscv list
  RISC-V: defconfig: Enable NO_HZ_IDLE and HIGH_RES_TIMERS

5 years agoMerge branch 'parisc-5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller...
Linus Torvalds [Thu, 18 Jul 2019 19:23:45 +0000 (12:23 -0700)]
Merge branch 'parisc-5.3-2' of git://git./linux/kernel/git/deller/parisc-linux

Pull parisc fixes from Helge Deller:

 - Prevent kernel panics by adding proper checking of register values
   injected via the ptrace interface

 - Wire up the new clone3 syscall

* 'parisc-5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Wire up clone3 syscall
  parisc: Avoid kernel panic triggered by invalid kprobe
  parisc: Ensure userspace privilege for ptraced processes in regset functions
  parisc: Fix kernel panic due invalid values in IAOQ0 or IAOQ1

5 years agoMerge tag 'modules-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/jeyu...
Linus Torvalds [Thu, 18 Jul 2019 19:06:57 +0000 (12:06 -0700)]
Merge tag 'modules-for-v5.3' of git://git./linux/kernel/git/jeyu/linux

Pull module updates from Jessica Yu:
 "Summary of modules changes for the 5.3 merge window:

   - Code fixes and cleanups

   - Fix bug where set_memory_x() wasn't being called when rodata=n

   - Fix bug where -EEXIST was being returned for going modules

   - Allow arches to override module_exit_section()"

* tag 'modules-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/jeyu/linux:
  modules: fix compile error if don't have strict module rwx
  ARM: module: recognize unwind exit sections
  module: allow arch overrides for .exit section names
  modules: fix BUG when load module with rodata=n
  kernel/module: Fix mem leak in module_add_modinfo_attrs
  kernel: module: Use struct_size() helper
  kernel/module.c: Only return -EEXIST for modules that have finished loading

5 years agoMerge tag 'trace-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux...
Linus Torvalds [Thu, 18 Jul 2019 18:51:00 +0000 (11:51 -0700)]
Merge tag 'trace-v5.3' of git://git./linux/kernel/git/rostedt/linux-trace

Pull tracing updates from Steven Rostedt:
 "The main changes in this release include:

   - Add user space specific memory reading for kprobes

   - Allow kprobes to be executed earlier in boot

  The rest are mostly just various clean ups and small fixes"

* tag 'trace-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (33 commits)
  tracing: Make trace_get_fields() global
  tracing: Let filter_assign_type() detect FILTER_PTR_STRING
  tracing: Pass type into tracing_generic_entry_update()
  ftrace/selftest: Test if set_event/ftrace_pid exists before writing
  ftrace/selftests: Return the skip code when tracing directory not configured in kernel
  tracing/kprobe: Check registered state using kprobe
  tracing/probe: Add trace_event_call accesses APIs
  tracing/probe: Add probe event name and group name accesses APIs
  tracing/probe: Add trace flag access APIs for trace_probe
  tracing/probe: Add trace_event_file access APIs for trace_probe
  tracing/probe: Add trace_event_call register API for trace_probe
  tracing/probe: Add trace_probe init and free functions
  tracing/uprobe: Set print format when parsing command
  tracing/kprobe: Set print format right after parsed command
  kprobes: Fix to init kprobes in subsys_initcall
  tracepoint: Use struct_size() in kmalloc()
  ring-buffer: Remove HAVE_64BIT_ALIGNED_ACCESS
  ftrace: Enable trampoline when rec count returns back to one
  tracing/kprobe: Do not run kprobe boot tests if kprobe_event is on cmdline
  tracing: Make a separate config for trace event self tests
  ...

5 years agoMerge branch 'for-linus-5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad...
Linus Torvalds [Thu, 18 Jul 2019 18:48:05 +0000 (11:48 -0700)]
Merge branch 'for-linus-5.2' of git://git./linux/kernel/git/konrad/swiotlb

Pull swiotlb updates from Konrad Rzeszutek Wilk:
 "One compiler fix, and a bug-fix in swiotlb_nr_tbl() and
  swiotlb_max_segment() to check also for no_iotlb_memory"

* 'for-linus-5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/swiotlb:
  swiotlb: fix phys_addr_t overflow warning
  swiotlb: Return consistent SWIOTLB segments/nr_tbl
  swiotlb: Group identical cleanup in swiotlb_cleanup()

5 years agoSUNRPC: Optimise transport balancing code
Trond Myklebust [Tue, 16 Jul 2019 17:27:23 +0000 (13:27 -0400)]
SUNRPC: Optimise transport balancing code

Moves the balancing code to avoid doing cursor changes on every search
iteration.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
5 years agoSUNRPC: Ensure the bvecs are reset when we re-encode the RPC request
Trond Myklebust [Thu, 18 Jul 2019 01:22:38 +0000 (21:22 -0400)]
SUNRPC: Ensure the bvecs are reset when we re-encode the RPC request

The bvec tracks the list of pages, so if the number of pages changes
due to a re-encode, we need to reset the bvec as well.

Fixes: 277e4ab7d530 ("SUNRPC: Simplify TCP receive code by switching...")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: stable@vger.kernel.org # v4.20+
5 years agopnfs/flexfiles: Fix PTR_ERR() dereferences in ff_layout_track_ds_error
Trond Myklebust [Wed, 17 Jul 2019 17:57:44 +0000 (13:57 -0400)]
pnfs/flexfiles: Fix PTR_ERR() dereferences in ff_layout_track_ds_error

mirror->mirror_ds can be NULL if uninitialised, but can contain
a PTR_ERR() if call to GETDEVICEINFO failed.

Fixes: 65990d1afbd2 ("pNFS/flexfiles: Fix a deadlock on LAYOUTGET")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: stable@vger.kernel.org # 4.10+
5 years agoNFSv4: Don't use the zero stateid with layoutget
Trond Myklebust [Tue, 16 Jul 2019 19:38:28 +0000 (15:38 -0400)]
NFSv4: Don't use the zero stateid with layoutget

The NFSv4.1 protocol explicitly forbids us from using the zero stateid
together with layoutget, so when we see that nfs4_select_rw_stateid()
is unable to return a valid delegation, lock or open stateid, then
we should initiate recovery and retry.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
5 years agoMerge tag 'xfs-5.3-merge-13' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Thu, 18 Jul 2019 18:18:00 +0000 (11:18 -0700)]
Merge tag 'xfs-5.3-merge-13' of git://git./fs/xfs/xfs-linux

Pull xfs cleanups from Darrick Wong:
 "We had a few more lateish cleanup patches come in for 5.3 -- a couple
  of syncups with the userspace libxfs code and a conversion of the XFS
  administrator's guide to ReST format.

  Summary:

   - Bring fs/xfs/libxfs/xfs_trans_inode.c in sync with userspace
     libxfs.

   - Convert the xfs administrator guide to rst and move it into the
     official admin guide under Documentation"

* tag 'xfs-5.3-merge-13' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  Documentation: filesystem: Convert xfs.txt to ReST
  xfs: sync up xfs_trans_inode with userspace
  xfs: move xfs_trans_inode.c to libxfs/

5 years agoMerge tag '4.3-rc-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Thu, 18 Jul 2019 18:11:51 +0000 (11:11 -0700)]
Merge tag '4.3-rc-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs updates from Steve French:
 "Fixes (three for stable) and improvements including much faster
  encryption (SMB3.1.1 GCM)"

* tag '4.3-rc-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6: (27 commits)
  smb3: smbdirect no longer experimental
  cifs: fix crash in smb2_compound_op()/smb2_set_next_command()
  cifs: fix crash in cifs_dfs_do_automount
  cifs: fix parsing of symbolic link error response
  cifs: refactor and clean up arguments in the reparse point parsing
  SMB3: query inode number on open via create context
  smb3: Send netname context during negotiate protocol
  smb3: do not send compression info by default
  smb3: add new mount option to retrieve mode from special ACE
  smb3: Allow query of symlinks stored as reparse points
  cifs: Fix a race condition with cifs_echo_request
  cifs: always add credits back for unsolicited PDUs
  fs: cifs: cifsssmb: Change return type of convert_ace_to_cifs_ace
  add some missing definitions
  cifs: fix typo in debug message with struct field ia_valid
  smb3: minor cleanup of compound_send_recv
  CIFS: Fix module dependency
  cifs: simplify code by removing CONFIG_CIFS_ACL ifdef
  cifs: Fix check for matching with existing mount
  cifs: Properly handle auto disabling of serverino option
  ...

5 years agoMerge tag 'ceph-for-5.3-rc1' of git://github.com/ceph/ceph-client
Linus Torvalds [Thu, 18 Jul 2019 18:05:25 +0000 (11:05 -0700)]
Merge tag 'ceph-for-5.3-rc1' of git://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "Lots of exciting things this time!

   - support for rbd object-map and fast-diff features (myself). This
     will speed up reads, discards and things like snap diffs on sparse
     images.

   - ceph.snap.btime vxattr to expose snapshot creation time (David
     Disseldorp). This will be used to integrate with "Restore Previous
     Versions" feature added in Windows 7 for folks who reexport ceph
     through SMB.

   - security xattrs for ceph (Zheng Yan). Only selinux is supported for
     now due to the limitations of ->dentry_init_security().

   - support for MSG_ADDR2, FS_BTIME and FS_CHANGE_ATTR features (Jeff
     Layton). This is actually a single feature bit which was missing
     because of the filesystem pieces. With this in, the kernel client
     will finally be reported as "luminous" by "ceph features" -- it is
     still being reported as "jewel" even though all required Luminous
     features were implemented in 4.13.

   - stop NULL-terminating ceph vxattrs (Jeff Layton). The convention
     with xattrs is to not terminate and this was causing
     inconsistencies with ceph-fuse.

   - change filesystem time granularity from 1 us to 1 ns, again fixing
     an inconsistency with ceph-fuse (Luis Henriques).

  On top of this there are some additional dentry name handling and cap
  flushing fixes from Zheng. Finally, Jeff is formally taking over for
  Zheng as the filesystem maintainer"

* tag 'ceph-for-5.3-rc1' of git://github.com/ceph/ceph-client: (71 commits)
  ceph: fix end offset in truncate_inode_pages_range call
  ceph: use generic_delete_inode() for ->drop_inode
  ceph: use ceph_evict_inode to cleanup inode's resource
  ceph: initialize superblock s_time_gran to 1
  MAINTAINERS: take over for Zheng as CephFS kernel client maintainer
  rbd: setallochint only if object doesn't exist
  rbd: support for object-map and fast-diff
  rbd: call rbd_dev_mapping_set() from rbd_dev_image_probe()
  libceph: export osd_req_op_data() macro
  libceph: change ceph_osdc_call() to take page vector for response
  libceph: bump CEPH_MSG_MAX_DATA_LEN (again)
  rbd: new exclusive lock wait/wake code
  rbd: quiescing lock should wait for image requests
  rbd: lock should be quiesced on reacquire
  rbd: introduce copyup state machine
  rbd: rename rbd_obj_setup_*() to rbd_obj_init_*()
  rbd: move OSD request allocation into object request state machines
  rbd: factor out __rbd_osd_setup_discard_ops()
  rbd: factor out rbd_osd_setup_copyup()
  rbd: introduce obj_req->osd_reqs list
  ...

5 years agoMerge tag 'dax-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm
Linus Torvalds [Thu, 18 Jul 2019 17:58:52 +0000 (10:58 -0700)]
Merge tag 'dax-for-5.3' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull dax updates from Dan Williams:
 "The fruits of a bug hunt in the fsdax implementation with Willy and a
  small feature update for device-dax:

   - Fix a hang condition that started triggering after the Xarray
     conversion of fsdax in the v4.20 kernel.

   - Add a 'resource' (root-only physical base address) sysfs attribute
     to device-dax instances to correlate memory-blocks onlined via the
     kmem driver with a given device instance"

* tag 'dax-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  dax: Fix missed wakeup with PMD faults
  device-dax: Add a 'resource' attribute

5 years agoMerge tag 'libnvdimm-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm...
Linus Torvalds [Thu, 18 Jul 2019 17:52:08 +0000 (10:52 -0700)]
Merge tag 'libnvdimm-for-5.3' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull libnvdimm updates from Dan Williams:
 "Primarily just the virtio_pmem driver:

   - virtio_pmem

     The new virtio_pmem facility introduces a paravirtualized
     persistent memory device that allows a guest VM to use DAX
     mechanisms to access a host-file with host-page-cache. It arranges
     for MAP_SYNC to be disabled and instead triggers a host fsync()
     when a 'write-cache flush' command is sent to the virtual disk
     device.

   - Miscellaneous small fixups"

* tag 'libnvdimm-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  virtio_pmem: fix sparse warning
  xfs: disable map_sync for async flush
  ext4: disable map_sync for async flush
  dax: check synchronous mapping is supported
  dm: enable synchronous dax
  libnvdimm: add dax_dev sync flag
  virtio-pmem: Add virtio pmem driver
  libnvdimm: nd_region flush callback support
  libnvdimm, namespace: Drop uuid_t implementation detail

5 years agoMerge tag 'linux-watchdog-5.3-rc1' of git://www.linux-watchdog.org/linux-watchdog
Linus Torvalds [Thu, 18 Jul 2019 17:47:59 +0000 (10:47 -0700)]
Merge tag 'linux-watchdog-5.3-rc1' of git://linux-watchdog.org/linux-watchdog

Pull watchdog updates from Wim Van Sebroeck:

 - add Allwinner H6 watchdog

 - drop warning after registering device patches

 - hpwdt improvements

 - gpio: add support for nowayout option

 - introduce CONFIG_WATCHDOG_OPEN_TIMEOUT

 - convert remaining drivers to use SPDX license identifier

 - Fixes and improvements on several watchdog device drivers

* tag 'linux-watchdog-5.3-rc1' of git://www.linux-watchdog.org/linux-watchdog: (74 commits)
  watchdog: digicolor_wdt: Remove unused variable in dc_wdt_probe
  watchdog: ie6xx_wdt: Use spinlock_t instead of struct spinlock
  watchdog: atmel: atmel-sama5d4-wdt: Disable watchdog on system suspend
  watchdog: convert remaining drivers to use SPDX license identifier
  dt-bindings: watchdog: Rename bindings documentation file
  watchdog: mei_wdt: no need to check return value of debugfs_create functions
  watchdog: bcm_kona_wdt: no need to check return value of debugfs_create functions
  docs: watchdog: Fix build error.
  docs: watchdog: convert docs to ReST and rename to *.rst
  watchdog: make the device time out at open_deadline when open_timeout is used
  watchdog: introduce CONFIG_WATCHDOG_OPEN_TIMEOUT
  watchdog: introduce watchdog.open_timeout commandline parameter
  dt-bindings: watchdog: move i.MX system controller watchdog binding to SCU
  watchdog: imx_sc: Add pretimeout support
  watchdog: renesas_wdt: Add a few cycles delay
  watchdog: gpio: add support for nowayout option
  watchdog: renesas_wdt: Use 'dev' instead of dereferencing it repeatedly
  dt-bindings: watchdog: add Allwinner H6 watchdog
  watchdog: jz4740: Avoid starting watchdog in set_timeout
  watchdog: jz4740: Use register names from <linux/mfd/ingenic-tcu.h>
  ...

5 years agoMerge tag 'sound-fix-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Thu, 18 Jul 2019 16:36:51 +0000 (09:36 -0700)]
Merge tag 'sound-fix-5.3-rc1' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of small fixes.

   - The optimization of PM resume with HD-audio HDMI codecs, which
     eventually work around weird issues

   - A correction of Intel Icelake HDMI audio code

   - Quirks for Dell machines with Realtek HD-audio codecs

   - The fix for too long sequencer write stall that was spotted by
     syzkaller

   - A few trivial cleanups reported by coccinelle"

* tag 'sound-fix-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Don't resume forcibly i915 HDMI/DP codec
  ALSA: hda/hdmi - Fix i915 reverse port/pin mapping
  ALSA: hda/hdmi - Remove duplicated define
  ALSA: seq: Break too long mutex context in the write loop
  ALSA: hda/realtek: apply ALC891 headset fixup to one Dell machine
  ALSA: rme9652: Unneeded variable: "result".
  ALSA: emu10k1: Remove unneeded variable "change"
  ALSA: au88x0: Remove unneeded variable: "changed"
  ALSA: hda/realtek - Fixed Headphone Mic can't record on Dell platform
  ALSA: ps3: Remove Unneeded variable: "ret"
  ALSA: lx6464es: Remove unneeded variable err

5 years agoMerge tag 'pm-5.3-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Thu, 18 Jul 2019 16:32:28 +0000 (09:32 -0700)]
Merge tag 'pm-5.3-rc1-2' of git://git./linux/kernel/git/rafael/linux-pm

Pull more power management updates from Rafael Wysocki:
 "These modify the Intel RAPL driver to allow it to use an MMIO
  interface to the hardware, make the int340X thermal driver provide
  such an interface for it, add Intel Ice Lake CPU IDs to the RAPL
  driver (these changes depend on the previously merged x86 arch
  changes), update cpufreq to use the PM QoS framework for managing the
  min and max frequency limits, and add update the imx-cpufreq-dt
  cpufreq driver to support i.MX8MN.

  Specifics:

   - Add MMIO interface support to the Intel RAPL power capping driver
     and update the int340X thermal driver to provide a RAPL MMIO
     interface (Zhang Rui, Stephen Rothwell).

   - Add Intel Ice Lake CPU IDs to the RAPL driver (Zhang Rui, Rajneesh
     Bhardwaj).

   - Make cpufreq use the PM QoS framework (instead of notifiers) for
     managing the min and max frequency constraints (Viresh Kumar).

   - Add i.MX8MN support to the imx-cpufreq-dt cpufreq driver (Anson
     Huang)"

* tag 'pm-5.3-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (27 commits)
  cpufreq: Make cpufreq_generic_init() return void
  intel_rapl: need linux/cpuhotplug.h for enum cpuhp_state
  powercap/rapl: Add Ice Lake NNPI support to RAPL driver
  powercap/intel_rapl: add support for ICX-D
  powercap/intel_rapl: add support for ICX
  powercap/intel_rapl: add support for IceLake desktop
  intel_rapl: Fix module autoloading issue
  int340X/processor_thermal_device: add support for MMIO RAPL
  intel_rapl: support two power limits for every RAPL domain
  intel_rapl: support 64 bit register
  intel_rapl: abstract RAPL common code
  intel_rapl: cleanup hardcoded MSR access
  intel_rapl: cleanup some functions
  intel_rapl: abstract register access operations
  intel_rapl: abstract register address
  intel_rapl: introduce struct rapl_if_private
  intel_rapl: introduce intel_rapl.h
  intel_rapl: remove hardcoded register index
  intel_rapl: use reg instead of msr
  cpufreq: imx-cpufreq-dt: Add i.MX8MN support
  ...

5 years agoMerge tag 'acpi-5.3-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Thu, 18 Jul 2019 16:12:34 +0000 (09:12 -0700)]
Merge tag 'acpi-5.3-rc1-3' of git://git./linux/kernel/git/rafael/linux-pm

Pull more ACPI updates from Rafael Wysocki:
 "These get rid of two clang warnings, add a new quirk mechanism to the
  ACPI backlight driver (and apply it to one machine) and update the
  table load object initialization in ACPICA (this is a replacement for
  a previously reverted ACPICA commit).

  Specifics:

   - Make ACPI table loading work more consistently regardless of the
     exact mechanism used for loading a table (Erik Schmauss).

   - Get rid of two clang warnings (Arnd Bergmann).

   - Add new quirk mechanism to the ACPI backlight driver and use it to
     add a quirk for PB Easynote MZ35 (Hans de Goede)"

* tag 'acpi-5.3-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: video: Add new hw_changes_brightness quirk, set it on PB Easynote MZ35
  ACPI: fix false-positive -Wuninitialized warning
  ACPI: blacklist: fix clang warning for unused DMI table
  ACPICA: Update table load object initialization

5 years agoMerge branch 'floppy'
Linus Torvalds [Thu, 18 Jul 2019 15:43:20 +0000 (08:43 -0700)]
Merge branch 'floppy'

Merge floppy ioctl verification fixes from Denis Efremov.

This also marks the floppy driver as orphaned - it turns out that Jiri
no longer has working hardware.

Actual working physical floppy hardware is getting hard to find, and
while Willy was able to test this, I think the driver can be considered
pretty much dead from an actual hardware standpoint.  The hardware that
is still sold seems to be mainly USB-based, which doesn't use this
legacy driver at all.

The old floppy disk controller is still emulated in various VM
environments, so the driver isn't going away, but let's see if anybody
is interested to step up to maintain it.

The lack of hardware also likely means that the ioctl range verification
fixes are probably mostly relevant to anybody using floppies in a
virtual environment.  Which is probably also going away in favor of USB
storage emulation, but who knows.

Will Decon reviewed the patches but I'm not rebasing them just for that,
so I'll add a

Reviewed-by: Will Deacon <will@kernel.org>
here instead.

* floppy:
  MAINTAINERS: mark floppy.c orphaned
  floppy: fix out-of-bounds read in copy_buffer
  floppy: fix invalid pointer dereference in drive_name
  floppy: fix out-of-bounds read in next_valid_format
  floppy: fix div-by-zero in setup_format_params

5 years agoMAINTAINERS: mark floppy.c orphaned
Jiri Kosina [Wed, 17 Jul 2019 22:03:51 +0000 (00:03 +0200)]
MAINTAINERS: mark floppy.c orphaned

I volunteered myself to maintain it quite some time ago back when I
fixed the concurrency issues which exhibited itself only with
VM-emulated devices, and at the same time I still had the physical 3.5"
reader to test all the changes.

The reader doesn't work any more though, so I guess it's time to step
down from this super-prestigious role :p and mark floppy.c as Orphaned.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoriscv: fix build break after macro-to-function conversion in generic cacheflush.h
Paul Walmsley [Wed, 17 Jul 2019 20:41:51 +0000 (13:41 -0700)]
riscv: fix build break after macro-to-function conversion in generic cacheflush.h

Commit c296d4dc13ae ("asm-generic: fix a compilation warning")
converted the various flush_*cache_* macros in
asm-generic/cacheflush.h to static inline functions.  This breaks
RISC-V builds, since RISC-V's cacheflush.h includes the generic
cacheflush.h and then undefines the macros to be overridden.

Fix by copying the subset of the no-op functions that are reused from
the generic cacheflush.h into the RISC-V cacheflush.h, and dropping
the include of the generic cacheflush.h.

Fixes: c296d4dc13ae ("asm-generic: fix a compilation warning")
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoMerge branches 'acpi-misc' and 'acpi-video'
Rafael J. Wysocki [Thu, 18 Jul 2019 08:22:20 +0000 (10:22 +0200)]
Merge branches 'acpi-misc' and 'acpi-video'

* acpi-misc:
  ACPI: fix false-positive -Wuninitialized warning
  ACPI: blacklist: fix clang warning for unused DMI table

* acpi-video:
  ACPI: video: Add new hw_changes_brightness quirk, set it on PB Easynote MZ35

5 years agoMerge branch 'pm-cpufreq'
Rafael J. Wysocki [Thu, 18 Jul 2019 07:49:30 +0000 (09:49 +0200)]
Merge branch 'pm-cpufreq'

* pm-cpufreq:
  cpufreq: Make cpufreq_generic_init() return void
  cpufreq: imx-cpufreq-dt: Add i.MX8MN support
  cpufreq: Add QoS requests for userspace constraints
  cpufreq: intel_pstate: Reuse refresh_frequency_limits()
  cpufreq: Register notifiers with the PM QoS framework
  PM / QoS: Add support for MIN/MAX frequency constraints
  PM / QOS: Pass request type to dev_pm_qos_read_value()
  PM / QOS: Rename __dev_pm_qos_read_value() and dev_pm_qos_raw_read_value()
  PM / QOS: Pass request type to dev_pm_qos_{add|remove}_notifier()

5 years agoSUNRPC: Fix up backchannel slot table accounting
Trond Myklebust [Tue, 16 Jul 2019 17:51:29 +0000 (13:51 -0400)]
SUNRPC: Fix up backchannel slot table accounting

Add a per-transport maximum limit in the socket case, and add
helpers to allow the NFSv4 code to discover that limit.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
5 years agoSUNRPC: Fix initialisation of struct rpc_xprt_switch
Trond Myklebust [Thu, 18 Jul 2019 05:10:51 +0000 (01:10 -0400)]
SUNRPC: Fix initialisation of struct rpc_xprt_switch

Ensure that we do initialise the fields xps_nactive, xps_queuelen
and xps_net.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
5 years agofloppy: fix out-of-bounds read in copy_buffer
Denis Efremov [Fri, 12 Jul 2019 18:55:23 +0000 (21:55 +0300)]
floppy: fix out-of-bounds read in copy_buffer

This fixes a global out-of-bounds read access in the copy_buffer
function of the floppy driver.

The FDDEFPRM ioctl allows one to set the geometry of a disk.  The sect
and head fields (unsigned int) of the floppy_drive structure are used to
compute the max_sector (int) in the make_raw_rw_request function.  It is
possible to overflow the max_sector.  Next, max_sector is passed to the
copy_buffer function and used in one of the memcpy calls.

An unprivileged user could trigger the bug if the device is accessible,
but requires a floppy disk to be inserted.

The patch adds the check for the .sect * .head multiplication for not
overflowing in the set_geometry function.

The bug was found by syzkaller.

Signed-off-by: Denis Efremov <efremov@ispras.ru>
Tested-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agofloppy: fix invalid pointer dereference in drive_name
Denis Efremov [Fri, 12 Jul 2019 18:55:22 +0000 (21:55 +0300)]
floppy: fix invalid pointer dereference in drive_name

This fixes the invalid pointer dereference in the drive_name function of
the floppy driver.

The native_format field of the struct floppy_drive_params is used as
floppy_type array index in the drive_name function.  Thus, the field
should be checked the same way as the autodetect field.

To trigger the bug, one could use a value out of range and set the drive
parameters with the FDSETDRVPRM ioctl.  Next, FDGETDRVTYP ioctl should
be used to call the drive_name.  A floppy disk is not required to be
inserted.

CAP_SYS_ADMIN is required to call FDSETDRVPRM.

The patch adds the check for a value of the native_format field to be in
the '0 <= x < ARRAY_SIZE(floppy_type)' range of the floppy_type array
indices.

The bug was found by syzkaller.

Signed-off-by: Denis Efremov <efremov@ispras.ru>
Tested-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agofloppy: fix out-of-bounds read in next_valid_format
Denis Efremov [Fri, 12 Jul 2019 18:55:21 +0000 (21:55 +0300)]
floppy: fix out-of-bounds read in next_valid_format

This fixes a global out-of-bounds read access in the next_valid_format
function of the floppy driver.

The values from autodetect field of the struct floppy_drive_params are
used as indices for the floppy_type array in the next_valid_format
function 'floppy_type[DP->autodetect[probed_format]].sect'.

To trigger the bug, one could use a value out of range and set the drive
parameters with the FDSETDRVPRM ioctl.  A floppy disk is not required to
be inserted.

CAP_SYS_ADMIN is required to call FDSETDRVPRM.

The patch adds the check for values of the autodetect field to be in the
'0 <= x < ARRAY_SIZE(floppy_type)' range of the floppy_type array indices.

The bug was found by syzkaller.

Signed-off-by: Denis Efremov <efremov@ispras.ru>
Tested-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agofloppy: fix div-by-zero in setup_format_params
Denis Efremov [Fri, 12 Jul 2019 18:55:20 +0000 (21:55 +0300)]
floppy: fix div-by-zero in setup_format_params

This fixes a divide by zero error in the setup_format_params function of
the floppy driver.

Two consecutive ioctls can trigger the bug: The first one should set the
drive geometry with such .sect and .rate values for the F_SECT_PER_TRACK
to become zero.  Next, the floppy format operation should be called.

A floppy disk is not required to be inserted.  An unprivileged user
could trigger the bug if the device is accessible.

The patch checks F_SECT_PER_TRACK for a non-zero value in the
set_geometry function.  The proper check should involve a reasonable
upper limit for the .sect and .rate fields, but it could change the
UAPI.

The patch also checks F_SECT_PER_TRACK in the setup_format_params, and
cancels the formatting operation in case of zero.

The bug was found by syzkaller.

Signed-off-by: Denis Efremov <efremov@ispras.ru>
Tested-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoparisc: Wire up clone3 syscall
Helge Deller [Mon, 15 Jul 2019 20:33:26 +0000 (22:33 +0200)]
parisc: Wire up clone3 syscall

Signed-off-by: Helge Deller <deller@gmx.de>
Tested-by: Sven Schnelle <svens@stackframe.org>
Acked-by: Christian Brauner <christian@brauner.io>
5 years agoparisc: Avoid kernel panic triggered by invalid kprobe
Helge Deller [Tue, 16 Jul 2019 19:16:26 +0000 (21:16 +0200)]
parisc: Avoid kernel panic triggered by invalid kprobe

When running gdb I was able to trigger this kernel panic:

 Kernel Fault: Code=26 (Data memory access rights trap) at addr 0000000000000060
 CPU: 0 PID: 1401 Comm: gdb-crash Not tainted 5.2.0-rc7-64bit+ #1053

      YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
 PSW: 00001000000001000000000000001111 Not tainted
 r00-03  000000000804000f 0000000040dee1a0 0000000040c78cf0 00000000b8d50160
 r04-07  0000000040d2b1a0 000000004360a098 00000000bbbe87b8 0000000000000003
 r08-11  00000000fac20a70 00000000fac24160 00000000fac1bbe0 0000000000000000
 r12-15  00000000fabfb79a 00000000fac244a4 0000000000010000 0000000000000001
 r16-19  00000000bbbe87b8 00000000f8f02910 0000000000010034 0000000000000000
 r20-23  00000000fac24630 00000000fac24630 000000006474e552 00000000fac1aa52
 r24-27  0000000000000028 00000000bbbe87b8 00000000bbbe87b8 0000000040d2b1a0
 r28-31  0000000000000000 00000000b8d501c0 00000000b8d501f0 0000000003424000
 sr00-03  0000000000423000 0000000000000000 0000000000000000 0000000000423000
 sr04-07  0000000000000000 0000000000000000 0000000000000000 0000000000000000

 IASQ: 0000000000000000 0000000000000000 IAOQ: 0000000040c78cf0 0000000040c78cf4
  IIR: 539f00c0    ISR: 0000000000000000  IOR: 0000000000000060
  CPU:        0   CR30: 00000000b8d50000 CR31: 00000000d22345e2
  ORIG_R28: 0000000040250798
  IAOQ[0]: parisc_kprobe_ss_handler+0x58/0x170
  IAOQ[1]: parisc_kprobe_ss_handler+0x5c/0x170
  RP(r2): parisc_kprobe_ss_handler+0x58/0x170
 Backtrace:
  [<0000000040206ff8>] handle_interruption+0x178/0xbb8
 Kernel panic - not syncing: Kernel Fault

Avoid this panic by checking the return value of kprobe_running() and
skip kprobe if none is currently active.

Cc: <stable@vger.kernel.org> # v5.2
Acked-by: Sven Schnelle <svens@stackframe.org>
Tested-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Helge Deller <deller@gmx.de>
5 years agoparisc: Ensure userspace privilege for ptraced processes in regset functions
Helge Deller [Thu, 4 Jul 2019 01:44:17 +0000 (03:44 +0200)]
parisc: Ensure userspace privilege for ptraced processes in regset functions

On parisc the privilege level of a process is stored in the lowest two bits of
the instruction pointers (IAOQ0 and IAOQ1). On Linux we use privilege level 0
for the kernel and privilege level 3 for user-space. So userspace should not be
allowed to modify IAOQ0 or IAOQ1 of a ptraced process to change it's privilege
level to e.g. 0 to try to gain kernel privileges.

This patch prevents such modifications in the regset support functions by
always setting the two lowest bits to one (which relates to privilege level 3
for user-space) if IAOQ0 or IAOQ1 are modified via ptrace regset calls.

Link: https://bugs.gentoo.org/481768
Cc: <stable@vger.kernel.org> # v4.7+
Tested-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Helge Deller <deller@gmx.de>
5 years agoparisc: Fix kernel panic due invalid values in IAOQ0 or IAOQ1
Helge Deller [Tue, 16 Jul 2019 19:43:11 +0000 (21:43 +0200)]
parisc: Fix kernel panic due invalid values in IAOQ0 or IAOQ1

On parisc the privilege level of a process is stored in the lowest two bits of
the instruction pointers (IAOQ0 and IAOQ1). On Linux we use privilege level 0
for the kernel and privilege level 3 for user-space. So userspace should not be
allowed to modify IAOQ0 or IAOQ1 of a ptraced process to change it's privilege
level to e.g. 0 to try to gain kernel privileges.

This patch prevents such modifications by always setting the two lowest bits to
one (which relates to privilege level 3 for user-space) if IAOQ0 or IAOQ1 are
modified via ptrace calls in the native and compat ptrace paths.

Link: https://bugs.gentoo.org/481768
Reported-by: Jeroen Roovers <jer@gentoo.org>
Cc: <stable@vger.kernel.org>
Tested-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Helge Deller <deller@gmx.de>
5 years agoMerge tag 'platform-drivers-x86-v5.3-2' of git://git.infradead.org/linux-platform...
Linus Torvalds [Wed, 17 Jul 2019 20:16:30 +0000 (13:16 -0700)]
Merge tag 'platform-drivers-x86-v5.3-2' of git://git.infradead.org/linux-platform-drivers-x86

Pull another x86 platform driver update from Andy Shevchenko:
 "Provide better naming for ABI, i.e. tell that we have fan boost mode.

  It won't break any ABI, but has to be done now to avoid confusion in
  the future"

* tag 'platform-drivers-x86-v5.3-2' of git://git.infradead.org/linux-platform-drivers-x86:
  platform/x86: asus: Rename "fan mode" to "fan boost mode"

5 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux
Linus Torvalds [Wed, 17 Jul 2019 20:13:41 +0000 (13:13 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/rzhang/linux

Pull thermal management updates from Zhang Rui:

 - Convert thermal documents to ReST (Mauro Carvalho Chehab)

 - Fix a cyclic depedency in between thermal core and governors (Daniel
   Lezcano)

 - Fix processor_thermal_device driver to re-evaluate power limits after
   resume (Srinivas Pandruvada, Zhang Rui)

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux:
  drivers: thermal: processor_thermal_device: Fix build warning
  docs: thermal: convert to ReST
  thermal/drivers/core: Use governor table to initialize
  thermal/drivers/core: Add init section table for self-encapsulation
  drivers: thermal: processor_thermal: Read PPCC on resume

5 years agoMerge tag 'gpio-v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux...
Linus Torvalds [Wed, 17 Jul 2019 20:05:21 +0000 (13:05 -0700)]
Merge tag 'gpio-v5.3-2' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fixes from Linus Walleij:

 - Revert a SPIO GPIO fix that didn't fix anything but instead created
   new problems.

 - Remove the EM GPIO irqdomain in a safe manner.

 - Fix a memory leak in the gpio quirks.

 - Make the DaVinci error path silent on probe deferral.

* tag 'gpio-v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  Revert "gpio/spi: Fix spi-gpio regression on active high CS"
  gpio: em: remove the gpiochip before removing the irq domain
  gpiolib: of: fix a memory leak in of_gpio_flags_quirks()
  gpio: davinci: silence error prints in case of EPROBE_DEFER

5 years agoMerge tag 'hwlock-v5.3' of git://github.com/andersson/remoteproc
Linus Torvalds [Wed, 17 Jul 2019 18:53:53 +0000 (11:53 -0700)]
Merge tag 'hwlock-v5.3' of git://github.com/andersson/remoteproc

Pull hwspinlock updates from Bjorn Andersson:
 "This contains support for hardware spinlock TI K3 AM65x and J721E
  family of SoCs, support for using hwspinlocks from atomic context and
  better error reporting when dealing with hardware disabled in
  DeviceTree"

* tag 'hwlock-v5.3' of git://github.com/andersson/remoteproc:
  hwspinlock: add the 'in_atomic' API
  hwspinlock: document the hwspinlock 'raw' API
  hwspinlock: stm32: implement the relax() ops
  hwspinlock: ignore disabled device
  hwspinlock/omap: Add a trace during probe
  hwspinlock/omap: Add support for TI K3 SoCs
  dt-bindings: hwlock: Update OMAP binding for TI K3 SoCs

5 years agoMerge tag 'rproc-v5.3' of git://github.com/andersson/remoteproc
Linus Torvalds [Wed, 17 Jul 2019 18:44:41 +0000 (11:44 -0700)]
Merge tag 'rproc-v5.3' of git://github.com/andersson/remoteproc

Pull remoteproc updates from Bjorn Andersson:
 "This adds support for the STM32 remoteproc, additional i.MX platforms
  with Cortex M4 remoteprocs and Qualcomm's QCS404 Compute DSP.

  Also initial support for vendor specific resource table entries and
  support for unprocessed Qualcomm firmware files"

* tag 'rproc-v5.3' of git://github.com/andersson/remoteproc:
  remoteproc: stm32: fix building without ARM SMCC
  remoteproc: qcom: q6v5-mss: Fix build error without QCOM_MDT_LOADER
  remoteproc: copy parent dma_pfn_offset for vdev
  remoteproc: qcom: q6v5-mss: Support loading non-split images
  soc: qcom: mdt_loader: Support loading non-split images
  remoteproc: stm32: add an ST stm32_rproc driver
  dt-bindings: remoteproc: add bindings for stm32 remote processor driver
  dt-bindings: stm32: add bindings for ML-AHB interconnect
  remoteproc: Use struct_size() helper
  remoteproc: add vendor resources handling
  remoteproc: imx: Fix typo in "failed"
  remoteproc: imx: Broaden the Kconfig selection logic
  remoteproc,rpmsg: add missing MAINTAINERS file entries
  remoteproc: qcom: qdsp6-adsp: Add support for QCS404 CDSP
  dt-bindings: remoteproc: Rename and amend Hexagon v56 binding

5 years agoMerge tag 'rpmsg-v5.3' of git://github.com/andersson/remoteproc
Linus Torvalds [Wed, 17 Jul 2019 18:31:11 +0000 (11:31 -0700)]
Merge tag 'rpmsg-v5.3' of git://github.com/andersson/remoteproc

Pull rpmsg updates from Bjorn Andersson:
 "This contains a DT binding update and a change to make the remote
  function of rpmsg_devices optional"

* tag 'rpmsg-v5.3' of git://github.com/andersson/remoteproc:
  rpmsg: core: Make remove handler for rpmsg driver optional.
  dt-bindings: soc: qcom: Add remote-pid binding for GLINK SMEM

5 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Wed, 17 Jul 2019 18:26:09 +0000 (11:26 -0700)]
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio, vhost updates from Michael Tsirkin:
 "Fixes, features, performance:

   - new iommu device

   - vhost guest memory access using vmap (just meta-data for now)

   - minor fixes"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  virtio-mmio: add error check for platform_get_irq
  scsi: virtio_scsi: Use struct_size() helper
  iommu/virtio: Add event queue
  iommu/virtio: Add probe request
  iommu: Add virtio-iommu driver
  PCI: OF: Initialize dev->fwnode appropriately
  of: Allow the iommu-map property to omit untranslated devices
  dt-bindings: virtio: Add virtio-pci-iommu node
  dt-bindings: virtio-mmio: Add IOMMU description
  vhost: fix clang build warning
  vhost: access vq metadata through kernel virtual address
  vhost: factor out setting vring addr and num
  vhost: introduce helpers to get the size of metadata area
  vhost: rename vq_iotlb_prefetch() to vq_meta_prefetch()
  vhost: fine grain userspace memory accessors
  vhost: generalize adding used elem

5 years agoMerge tag 'vfio-v5.3-rc1' of git://github.com/awilliam/linux-vfio
Linus Torvalds [Wed, 17 Jul 2019 18:23:13 +0000 (11:23 -0700)]
Merge tag 'vfio-v5.3-rc1' of git://github.com/awilliam/linux-vfio

Pull VFIO updates from Alex Williamson:

 - Static symbol cleanup in mdev samples (Kefeng Wang)

 - Use vma help in nvlink code (Peng Hao)

 - Remove unused code in mbochs sample (YueHaibing)

 - Send uevents around mdev registration (Alex Williamson)

* tag 'vfio-v5.3-rc1' of git://github.com/awilliam/linux-vfio:
  mdev: Send uevents around parent device registration
  sample/mdev/mbochs: remove set but not used variable 'mdev_state'
  vfio: vfio_pci_nvlink2: use a vma helper function
  vfio-mdev/samples: make some symbols static

5 years agodm: use printk ratelimiting functions
Mike Snitzer [Wed, 17 Jul 2019 16:57:06 +0000 (12:57 -0400)]
dm: use printk ratelimiting functions

DM provided its own ratelimiting printk wrapper but given printk
advances this is no longer needed.

Also, switching DMDEBUG_LIMIT to using pr_debug_ratelimited() fixes the
reported issue where DMDEBUG_LIMIT() still caused a flood of "callbacks
suppressed" messages.

Reported-by: Milan Broz <gmazyland@gmail.com>
Depends-on: 29fc2bc7539386 ("printk: pr_debug_ratelimited: check state first to reduce "callbacks suppressed" messages")
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
5 years agoMerge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Linus Torvalds [Wed, 17 Jul 2019 17:07:48 +0000 (10:07 -0700)]
Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull clk updates from Stephen Boyd:
 "This round of clk driver and framework updates is heavy on the driver
  update side. The two main highlights in the core framework are the
  addition of an bulk clk_get API that handles optional clks and an
  extra debugfs file that tells the developer about the current parent
  of a clk.

  The driver updates are dominated by i.MX in the diffstat, but that is
  mostly because that SoC has started converting to the clk_hw style of
  clk registration. The next big update is in the Amlogic meson clk
  driver that gained some support for audio, cpu, and temperature clks
  while fixing some PLL issues. Finally, the biggest thing that stands
  out is the conversion of a large part of the Allwinner sunxi-ng driver
  to the new clk parent scheme that uses less strings and more pointer
  comparisons to match clk parents and children up.

  In general, it looks like we have a lot of little fixes and tweaks
  here and there to clk data along with the normal addition of a handful
  of new drivers and a couple new core framework features.

  Core:
   - Add a 'clk_parent' file in clk debugfs
   - Add a clk_bulk_get_optional() API (with devm too)

  New Drivers:
   - Support gated clk controller on MIPS based BCM63XX SoCs
   - Support SiLabs Si5341 and Si5340 chips
   - Support for CPU clks on Raspberry Pi devices
   - Audsys clock driver for MediaTek MT8516 SoCs

  Updates:
   - Convert a large portion of the Allwinner sunxi-ng driver to new clk parent scheme
   - Small frequency support for SiLabs Si544 chips
   - Slow clk support for AT91 SAM9X60 SoCs
   - Remove dead code in various clk drivers (-Wunused)
   - Support for Marvell 98DX1135 SoCs
   - Get duty cycle of generic pwm clks
   - Improvement in mmc phase calculation and cleanup of some rate defintions
   - Switch i.MX6 and i.MX7 clock drivers to clk_hw based APIs
   - Add GPIO, SNVS and GIC clocks for i.MX8 drivers
   - Mark imx6sx/ul/ull/sll MMDC_P1_IPG and imx8mm DRAM_APB as critical clock
   - Correct imx7ulp nic1_bus_clk and imx8mm audio_pll2_clk clock setting
   - Add clks for new Exynos5422 Dynamic Memory Controller driver
   - Clock definition for Exynos4412 Mali
   - Add CMM (Color Management Module) clocks on Renesas R-Car H3, M3-N, E3, and D3
   - Add TPU (Timer Pulse Unit / PWM) clocks on Renesas RZ/G2M
   - Support for 32 bit clock IDs in TI's sci-clks for J721e SoCs
   - TI clock probing done from DT by default instead of firmware
   - Fix Amlogic Meson mpll fractional part and spread sprectrum issues
   - Add Amlogic meson8 audio clocks
   - Add Amlogic g12a temperature sensors clocks
   - Add Amlogic g12a and g12b cpu clocks
   - Add TPU (Timer Pulse Unit / PWM) clocks on Renesas R-Car H3, M3-W, and M3-N
   - Add CMM (Color Management Module) clocks on Renesas R-Car M3-W
   - Add Clock Domain support on Renesas RZ/N1"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (190 commits)
  clk: consoldiate the __clk_get_hw() declarations
  clk: sprd: Add check for return value of sprd_clk_regmap_init()
  clk: lochnagar: Update DT binding doc to include the primary SPDIF MCLK
  clk: Add Si5341/Si5340 driver
  dt-bindings: clock: Add silabs,si5341
  clk: clk-si544: Implement small frequency change support
  clk: add BCM63XX gated clock controller driver
  devicetree: document the BCM63XX gated clock bindings
  clk: at91: sckc: use dedicated functions to unregister clock
  clk: at91: sckc: improve error path for sama5d4 sck registration
  clk: at91: sckc: remove unnecessary line
  clk: at91: sckc: improve error path for sam9x5 sck register
  clk: at91: sckc: add support to free slow clock osclillator
  clk: at91: sckc: add support to free slow rc oscillator
  clk: at91: sckc: add support to free slow oscillator
  clk: rockchip: export HDMIPHY clock on rk3228
  clk: rockchip: add watchdog pclk on rk3328
  clk: rockchip: add clock id for hdmi_phy special clock on rk3228
  clk: rockchip: add clock id for watchdog pclk on rk3328
  clk: at91: sckc: add support for SAM9X60
  ...

5 years agoMerge tag 'rtc-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux
Linus Torvalds [Wed, 17 Jul 2019 17:03:50 +0000 (10:03 -0700)]
Merge tag 'rtc-5.3' of git://git./linux/kernel/git/abelloni/linux

Pull RTC updates from Alexandre Belloni:
 "A quiet cycle this time.

   - ds1307: properly handle oscillator failure flags

   - imx-sc: alarm support

   - pcf2123: alarm support, correct offset handling

   - sun6i: add R40 support

   - simplify getting the adapter of an i2c client"

* tag 'rtc-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux: (37 commits)
  rtc: wm831x: Add IRQF_ONESHOT flag
  rtc: stm32: remove one condition check in stm32_rtc_set_alarm()
  rtc: pcf2123: Fix build error
  rtc: interface: Change type of 'count' from int to u64
  rtc: pcf8563: Clear event flags and disable interrupts before requesting irq
  rtc: pcf8563: Fix interrupt trigger method
  rtc: pcf2123: fix negative offset rounding
  rtc: pcf2123: add alarm support
  rtc: pcf2123: use %ptR
  rtc: pcf2123: port to regmap
  rtc: pcf2123: remove sysfs register view
  rtc: rx8025: simplify getting the adapter of a client
  rtc: rx8010: simplify getting the adapter of a client
  rtc: rv8803: simplify getting the adapter of a client
  rtc: m41t80: simplify getting the adapter of a client
  rtc: fm3130: simplify getting the adapter of a client
  rtc: tegra: Drop MODULE_ALIAS
  rtc: sun6i: Add R40 compatible
  dt-bindings: rtc: sun6i: Add the R40 RTC compatible
  dt-bindings: rtc: Convert Allwinner A31 RTC to a schema
  ...

5 years agoMerge tag 'dmaengine-5.3-rc1' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Wed, 17 Jul 2019 16:55:43 +0000 (09:55 -0700)]
Merge tag 'dmaengine-5.3-rc1' of git://git.infradead.org/users/vkoul/slave-dma

Pull dmaengine updates from Vinod Koul:

 - Add support in dmaengine core to do device node checks for DT devices
   and update bunch of drivers to use that and remove open coding from
   drivers

 - New driver/driver support for new hardware, namely:
     - MediaTek UART APDMA
     - Freescale i.mx7ulp edma2
     - Synopsys eDMA IP core version 0
     - Allwinner H6 DMA

 - Updates to axi-dma and support for interleaved cyclic transfers

 - Greg's debugfs return value check removals on drivers

 - Updates to stm32-dma, hsu, dw, pl330, tegra drivers

* tag 'dmaengine-5.3-rc1' of git://git.infradead.org/users/vkoul/slave-dma: (68 commits)
  dmaengine: Revert "dmaengine: fsl-edma: add i.mx7ulp edma2 version support"
  dmaengine: at_xdmac: check for non-empty xfers_list before invoking callback
  Documentation: dmaengine: clean up description of dmatest usage
  dmaengine: tegra210-adma: remove PM_CLK dependency
  dmaengine: fsl-edma: add i.mx7ulp edma2 version support
  dt-bindings: dma: fsl-edma: add new i.mx7ulp-edma
  dmaengine: fsl-edma-common: version check for v2 instead
  dmaengine: fsl-edma-common: move dmamux register to another single function
  dmaengine: fsl-edma: add drvdata for fsl-edma
  dmaengine: Revert "dmaengine: fsl-edma: support little endian for edma driver"
  dmaengine: rcar-dmac: Reject zero-length slave DMA requests
  dmaengine: dw: Enable iDMA 32-bit on Intel Elkhart Lake
  dmaengine: dw-edma: fix semicolon.cocci warnings
  dmaengine: sh: usb-dmac: Use [] to denote a flexible array member
  dmaengine: dmatest: timeout value of -1 should specify infinite wait
  dmaengine: dw: Distinguish ->remove() between DW and iDMA 32-bit
  dmaengine: fsl-edma: support little endian for edma driver
  dmaengine: hsu: Revert "set HSU_CH_MTSR to memory width"
  dmagengine: pl330: add code to get reset property
  dt-bindings: pl330: document the optional resets property
  ...

5 years agoMerge tag 'mips_5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux
Linus Torvalds [Wed, 17 Jul 2019 16:42:03 +0000 (09:42 -0700)]
Merge tag 'mips_5.3' of git://git./linux/kernel/git/mips/linux

Pull MIPS updates from Paul Burton:
 "A light batch this time around but significant improvements for
  certain systems:

   - Removal of readq & writeq for MIPS32 kernels where they would
     simply BUG() anyway, allowing drivers or other code that #ifdefs on
     their presence to work properly.

   - Improvements for Ingenic JZ4740 systems, including support for the
     external memory controller & pinmuxing fixes for qi_lb60/NanoNote
     systems.

   - Improvements for Lantiq systems, in particular around SMP & IPIs.

   - DT updates for ralink/MediaTek MT7628a systems to probe & configure
     a bunch more devices.

   - Miscellaneous cleanups & build fixes"

* tag 'mips_5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux: (30 commits)
  MIPS: fix some more fall through errors in arch/mips
  MIPS: perf events: handle switch statement falling through warnings
  mips/kprobes: Export kprobe_fault_handler()
  MAINTAINERS: Add myself as Ingenic SoCs maintainer
  MIPS: ralink: mt7628a.dtsi: Add watchdog controller DT node
  MIPS: ralink: mt7628a.dtsi: Add SPI controller DT node
  MIPS: ralink: mt7628a.dtsi: Add GPIO controller DT node
  MIPS: ralink: mt7628a.dtsi: Add pinctrl DT properties to the UART nodes
  MIPS: ralink: mt7628a.dtsi: Add pinmux DT node
  MIPS: ralink: mt7628a.dtsi: Add SPDX GPL-2.0 license identifier
  MIPS: lantiq: Add SMP support for lantiq interrupt controller
  MIPS: lantiq: Shorten register names, remove unused macros
  MIPS: lantiq: Fix bitfield masking
  MIPS: lantiq: Remove unused macros
  MIPS: lantiq: Fix attributes of of_device_id structure
  MIPS: lantiq: Change variables to the same type as the source
  MIPS: lantiq: Move macro directly to iomem function
  mips: Remove q-accessors from non-64bit platforms
  FDDI: defza: Include linux/io-64-nonatomic-lo-hi.h
  MIPS: configs: Remove useless UEVENT_HELPER_PATH
  ...

5 years agoMerge tag 'h8300-for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux...
Linus Torvalds [Wed, 17 Jul 2019 16:36:38 +0000 (09:36 -0700)]
Merge tag 'h8300-for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux-h8/linux

Pull h8300 update from Yoshinori Sato:
 "Remove unused barrier defines"

* tag 'h8300-for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux-h8/linux:
  H8300: remove unused barrier defines

5 years agoMerge tag 'for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux-h8/linux
Linus Torvalds [Wed, 17 Jul 2019 16:34:10 +0000 (09:34 -0700)]
Merge tag 'for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux-h8/linux

Pull SH updates from Yoshinori Sato.

kprobe fix, defconfig updates and a SH Kconfig fix.

* tag 'for-linus-20190617' of git://git.sourceforge.jp/gitroot/uclinux-h8/linux:
  arch/sh: Check for kprobe trap number before trying to handle a kprobe trap
  sh: configs: Remove useless UEVENT_HELPER_PATH
  Fix allyesconfig output.

5 years agoplatform/x86: asus: Rename "fan mode" to "fan boost mode"
Daniel Drake [Wed, 17 Jul 2019 05:10:58 +0000 (13:10 +0800)]
platform/x86: asus: Rename "fan mode" to "fan boost mode"

The Asus WMI spec indicates that the function being controlled here
is called "Fan Boost Mode". The user-facing documentation also calls it
this.

The spec uses the term "fan mode" is used to refer to other things,
including functionality expected to appear on future products.
We missed this before as we are not dealing with the most readable of
specs, and didn't forsee any confusion around shortening the name.

Rename "fan mode" to "fan boost mode" to improve consistency with the
spec and to avoid a future naming conflict.

There is no interface breakage here since this has yet to be included
in an official kernel release. I also updated the kernel version listed
under ABI accordingly.

Signed-off-by: Daniel Drake <drake@endlessm.com>
Acked-by: Yurii Pavlovskyi <yurii.pavlovskyi@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
5 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Wed, 17 Jul 2019 15:58:04 +0000 (08:58 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge more updates from Andrew Morton:
 "VM:
   - z3fold fixes and enhancements by Henry Burns and Vitaly Wool

   - more accurate reclaimed slab caches calculations by Yafang Shao

   - fix MAP_UNINITIALIZED UAPI symbol to not depend on config, by
     Christoph Hellwig

   - !CONFIG_MMU fixes by Christoph Hellwig

   - new novmcoredd parameter to omit device dumps from vmcore, by
     Kairui Song

   - new test_meminit module for testing heap and pagealloc
     initialization, by Alexander Potapenko

   - ioremap improvements for huge mappings, by Anshuman Khandual

   - generalize kprobe page fault handling, by Anshuman Khandual

   - device-dax hotplug fixes and improvements, by Pavel Tatashin

   - enable synchronous DAX fault on powerpc, by Aneesh Kumar K.V

   - add pte_devmap() support for arm64, by Robin Murphy

   - unify locked_vm accounting with a helper, by Daniel Jordan

   - several misc fixes

  core/lib:
   - new typeof_member() macro including some users, by Alexey Dobriyan

   - make BIT() and GENMASK() available in asm, by Masahiro Yamada

   - changed LIST_POISON2 on x86_64 to 0xdead000000000122 for better
     code generation, by Alexey Dobriyan

   - rbtree code size optimizations, by Michel Lespinasse

   - convert struct pid count to refcount_t, by Joel Fernandes

  get_maintainer.pl:
   - add --no-moderated switch to skip moderated ML's, by Joe Perches

  misc:
   - ptrace PTRACE_GET_SYSCALL_INFO interface

   - coda updates

   - gdb scripts, various"

[ Using merge message suggestion from Vlastimil Babka, with some editing - Linus ]

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (100 commits)
  fs/select.c: use struct_size() in kmalloc()
  mm: add account_locked_vm utility function
  arm64: mm: implement pte_devmap support
  mm: introduce ARCH_HAS_PTE_DEVMAP
  mm: clean up is_device_*_page() definitions
  mm/mmap: move common defines to mman-common.h
  mm: move MAP_SYNC to asm-generic/mman-common.h
  device-dax: "Hotremove" persistent memory that is used like normal RAM
  mm/hotplug: make remove_memory() interface usable
  device-dax: fix memory and resource leak if hotplug fails
  include/linux/lz4.h: fix spelling and copy-paste errors in documentation
  ipc/mqueue.c: only perform resource calculation if user valid
  include/asm-generic/bug.h: fix "cut here" for WARN_ON for __WARN_TAINT architectures
  scripts/gdb: add helpers to find and list devices
  scripts/gdb: add lx-genpd-summary command
  drivers/pps/pps.c: clear offset flags in PPS_SETPARAMS ioctl
  kernel/pid.c: convert struct pid count to refcount_t
  drivers/rapidio/devices/rio_mport_cdev.c: NUL terminate some strings
  select: shift restore_saved_sigmask_unless() into poll_select_copy_remaining()
  select: change do_poll() to return -ERESTARTNOHAND rather than -EINTR
  ...

5 years agodm kcopyd: Increase default sub-job size to 512KB
Nikos Tsironis [Wed, 17 Jul 2019 11:24:10 +0000 (14:24 +0300)]
dm kcopyd: Increase default sub-job size to 512KB

Currently, kcopyd has a sub-job size of 64KB and a maximum number of 8
sub-jobs. As a result, for any kcopyd job, we have a maximum of 512KB of
I/O in flight.

This upper limit to the amount of in-flight I/O under-utilizes fast
devices and results in decreased throughput, e.g., when writing to a
snapshotted thin LV with I/O size less than the pool's block size (so
COW is performed using kcopyd).

Increase kcopyd's default sub-job size to 512KB, so we have a maximum of
4MB of I/O in flight for each kcopyd job. This results in an up to 96%
improvement of bandwidth when writing to a snapshotted thin LV, with I/O
sizes less than the pool's block size.

Also, add dm_mod.kcopyd_subjob_size_kb module parameter to allow users
to fine tune the sub-job size of kcopyd. The default value of this
parameter is 512KB and the maximum allowed value is 1024KB.

We evaluate the performance impact of the change by running the
snap_breaking_throughput benchmark, from the device mapper test suite
[1].

The benchmark:

  1. Creates a 1G thin LV
  2. Provisions the thin LV
  3. Takes a snapshot of the thin LV
  4. Writes to the thin LV with:

      dd if=/dev/zero of=/dev/vg/thin_lv oflag=direct bs=<I/O size>

Running this benchmark with various thin pool block sizes and dd I/O
sizes (all combinations triggering the use of kcopyd) we get the
following results:

+-----------------+-------------+------------------+-----------------+
| Pool block size | dd I/O size | BW before (MB/s) | BW after (MB/s) |
+-----------------+-------------+------------------+-----------------+
|       1 MB      |      256 KB |       242        |       280       |
|       1 MB      |      512 KB |       238        |       295       |
|                 |             |                  |                 |
|       2 MB      |      256 KB |       238        |       354       |
|       2 MB      |      512 KB |       241        |       380       |
|       2 MB      |        1 MB |       245        |       394       |
|                 |             |                  |                 |
|       4 MB      |      256 KB |       248        |       412       |
|       4 MB      |      512 KB |       234        |       432       |
|       4 MB      |        1 MB |       251        |       474       |
|       4 MB      |        2 MB |       257        |       504       |
|                 |             |                  |                 |
|       8 MB      |      256 KB |       239        |       420       |
|       8 MB      |      512 KB |       256        |       431       |
|       8 MB      |        1 MB |       264        |       467       |
|       8 MB      |        2 MB |       264        |       502       |
|       8 MB      |        4 MB |       281        |       537       |
+-----------------+-------------+------------------+-----------------+

[1] https://github.com/jthornber/device-mapper-test-suite

Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
5 years agodm snapshot: fix oversights in optional discard support
Mike Snitzer [Wed, 17 Jul 2019 15:12:30 +0000 (11:12 -0400)]
dm snapshot: fix oversights in optional discard support

__find_snapshots_sharing_cow() should always be used with _origins_lock
held so fix snapshot_io_hints() accordingly.  Also, once a snapshot is
being merged discards must not be allowed -- otherwise incorrect or
duplicate work will be performed.

Fixes: 2e6023850e177d ("dm snapshot: add optional discard support features")
Reported-by: Nikos Tsironis <ntsironis@arrikto.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
5 years agodm zoned: fix zone state management race
Damien Le Moal [Tue, 16 Jul 2019 05:39:34 +0000 (14:39 +0900)]
dm zoned: fix zone state management race

dm-zoned uses the zone flag DMZ_ACTIVE to indicate that a zone of the
backend device is being actively read or written and so cannot be
reclaimed. This flag is set as long as the zone atomic reference
counter is not 0. When this atomic is decremented and reaches 0 (e.g.
on BIO completion), the active flag is cleared and set again whenever
the zone is reused and BIO issued with the atomic counter incremented.
These 2 operations (atomic inc/dec and flag set/clear) are however not
always executed atomically under the target metadata mutex lock and
this causes the warning:

WARN_ON(!test_bit(DMZ_ACTIVE, &zone->flags));

in dmz_deactivate_zone() to be displayed. This problem is regularly
triggered with xfstests generic/209, generic/300, generic/451 and
xfs/077 with XFS being used as the file system on the dm-zoned target
device. Similarly, xfstests ext4/303, ext4/304, generic/209 and
generic/300 trigger the warning with ext4 use.

This problem can be easily fixed by simply removing the DMZ_ACTIVE flag
and managing the "ACTIVE" state by directly looking at the reference
counter value. To do so, the functions dmz_activate_zone() and
dmz_deactivate_zone() are changed to inline functions respectively
calling atomic_inc() and atomic_dec(), while the dmz_is_active() macro
is changed to an inline function calling atomic_read().

Fixes: 3b1a94c88b79 ("dm zoned: drive-managed zoned block device target")
Cc: stable@vger.kernel.org
Reported-by: Masato Suzuki <masato.suzuki@wdc.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
5 years agovirtio_pmem: fix sparse warning
Pankaj Gupta [Fri, 12 Jul 2019 05:16:10 +0000 (10:46 +0530)]
virtio_pmem: fix sparse warning

This patch fixes below sparse warning related to __virtio
type in virtio pmem driver. This is reported by Intel test
bot on linux-next tree.

nd_virtio.c:56:28: warning: incorrect type in assignment
                                (different base types)
nd_virtio.c:56:28:    expected unsigned int [unsigned] [usertype] type
nd_virtio.c:56:28:    got restricted __virtio32
nd_virtio.c:93:59: warning: incorrect type in argument 2
                                (different base types)
nd_virtio.c:93:59:    expected restricted __virtio32 [usertype] val
nd_virtio.c:93:59:    got unsigned int [unsigned] [usertype] ret

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Pankaj Gupta <pagupta@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
5 years agodax: Fix missed wakeup with PMD faults
Matthew Wilcox (Oracle) [Thu, 4 Jul 2019 03:21:25 +0000 (23:21 -0400)]
dax: Fix missed wakeup with PMD faults

RocksDB can hang indefinitely when using a DAX file.  This is due to
a bug in the XArray conversion when handling a PMD fault and finding a
PTE entry.  We use the wrong index in the hash and end up waiting on
the wrong waitqueue.

There's actually no need to wait; if we find a PTE entry while looking
for a PMD entry, we can return immediately as we know we should fall
back to a PTE fault (which may not conflict with the lock held).

We reuse the XA_RETRY_ENTRY to signal a conflicting entry was found.
This value can never be found in an XArray while holding its lock, so
it does not create an ambiguity.

Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/CAPcyv4hwHpX-MkUEqxwdTj7wCCZCN4RV-L4jsnuwLGyL_UEG4A@mail.gmail.com
Fixes: b15cd800682f ("dax: Convert page fault handlers to XArray")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Tested-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Robert Barror <robert.barror@intel.com>
Reported-by: Seema Pandit <seema.pandit@intel.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
5 years agofs/select.c: use struct_size() in kmalloc()
Gustavo A. R. Silva [Tue, 16 Jul 2019 23:30:58 +0000 (16:30 -0700)]
fs/select.c: use struct_size() in kmalloc()

One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array.  For example:

  struct foo {
       int stuff;
       struct boo entry[];
  };

  size = sizeof(struct foo) + count * sizeof(struct boo);
  instance = kmalloc(size, GFP_KERNEL);

Instead of leaving these open-coded and prone to type mistakes, we can now
use the new struct_size() helper:

  instance = kmalloc(struct_size(instance, entry, count), GFP_KERNEL);

Also, notice that variable size is unnecessary, hence it is removed.

This code was detected with the help of Coccinelle.

Link: http://lkml.kernel.org/r/20190604164226.GA13823@embeddedor
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm: add account_locked_vm utility function
Daniel Jordan [Tue, 16 Jul 2019 23:30:54 +0000 (16:30 -0700)]
mm: add account_locked_vm utility function

locked_vm accounting is done roughly the same way in five places, so
unify them in a helper.

Include the helper's caller in the debug print to distinguish between
callsites.

Error codes stay the same, so user-visible behavior does too.  The one
exception is that the -EPERM case in tce_account_locked_vm is removed
because Alexey has never seen it triggered.

[daniel.m.jordan@oracle.com: v3]
Link: http://lkml.kernel.org/r/20190529205019.20927-1-daniel.m.jordan@oracle.com
[sfr@canb.auug.org.au: fix mm/util.c]
Link: http://lkml.kernel.org/r/20190524175045.26897-1-daniel.m.jordan@oracle.com
Signed-off-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Cc: Alan Tull <atull@kernel.org>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Moritz Fischer <mdf@kernel.org>
Cc: Paul Mackerras <paulus@ozlabs.org>
Cc: Steve Sistare <steven.sistare@oracle.com>
Cc: Wu Hao <hao.wu@intel.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>